Suppose that X is false, and I have enormous amounts of evidence that X is false. But, it is within my "rational self interest" to believe X is true. Are there any patterns/papers in decision theory or game theory that discuss this situation (or, if it is agreed to be impossible, show that it is impossible)?
Things are often connected. X may imply Y, Y may imply Z, and it may be "in your interest" to believe X (e.g. be religious, and enjoy the benefits of being a part of a community), but not necessarily Z (e.g. commit group suicide when the comet fails to arrive). Are you sure you want to believe in X?
Evolution attempts to solve this problem by giving people two types of beliefs. (Warning: this does not work reliably, especially for the kind of people who worry about beliefs and decision theories.) One, called "near mode" is the type of beliefs you actually act on. If you "near-mode" believe that the cake needs to be baked for 10 minutes, you actually take it out of the oven after 10 minutes. Second, called "far mode" is the type of beliefs you talk about, but almost never actually act on, unless there is a strong social pressure. If you "far-mode" believe that donating all your property to the poor is the right thing to do, if someone asks you whether it is the right thing to do, you will say "yes", but you will actually never donate all your property to the poor. You will fail to notice the contradiction, and will get angry when it is pointed out.
The best outcome would be to "far-mode" believe in X. Problem is, if you ask whether it is a rational thing to do, it is already too late for you. Your brain is already processing X in "near mode".
Something has been on my mind for a while and the Arkansas ruling reminds me that I need to put pen to paper (or hands to keyboard) and express why healthcare for trans youth is so important.
The one word that the opposition keeps using is mutilation.
I need to really break this down for ANYONE that might even have a slight agreement with that.
First, NO parent hopes that their child will be trans. I don't believe ANYONE has ever put their hand on a pregnant belly and prayed "please let this child be trans!" I can't imagine any parent hoping that their child will feel so uncomfortable in their own body that they would be willing to go through bullying, taunting, alienation, possibly even physical abuse just to have gender reassignment surgery as a CHOICE! While women may get breast augmentation as a choice and as a cosmetic procedure, they're not tormented for their choice. They're not questioned over and over if they're sure. They don't have people CONSTANTLY trying to talk them out of it. That is an elective procedure and it carries very little stigma.
Being Trans is very different. When your need to go through gender reassignment is so strong that you're willing to deal with the emotional abuse you'll get from so many, that's not a choice. I know if I touch the stove it's going to hurt so I choose not to touch the stove. For someone that's trans, they can't help themselves. They NEED to correct who they truly are and they're willing to go through the pain, both physically and emotionally to be their true self.
So let's go back to the real subject at hand, trans youth. At what point is someone capable of making that choice? Think about that carefully. Do you know a 17 year old that makes bad decisions every day? Maybe someone that gets into trouble and doesn't understand taking responsibility for their actions? Do you know a 13 year old that could be trusted to babysit? Can they trust them to cook alone? Do they make good choices? Do they understand consequences? Are they more mature than the 17 year old mentioned above?
There isn't a number you can assign to it. You can't assume that the second someone turns 18, they're suddenly capable of making big decisions solely because they had a birthday. Everyone matures differently and a number does not define maturity.
So this is where the parents come in. In the US, a parent can consent for a child to have plastic surgery. So if a parent can decide if their child is mature enough to decide they want plastic surgery, a completely cosmetic procedure which could wait until their child turns 18, why can't they consent to gender affirming care?
Well, why can't a trans person just wait? Because certain things happen during puberty that will be with that person for the rest of their lives. Once a biological boy hits puberty, the voice will irreversibly be changed. Hair will begin growing everywhere, your height is affected by testosterone during puberty, musculature will start to build, etc. Once this happens, it's not reversible.
So you might think that blocking it is not reversible either? Yes, that's true, but that's in their control. If they choose that this is who they truly are, then forcing them to let their body change to something they're not, is not the answer. Stopping the process is NOT mutilation.
So back to who should make these decisions, the person and their parents, or the government? Who has the right to decide what these people feel is in their heart? Strangers with a pen, or those that have been raising them? This NEEDS to be a very careful decision made by those it will affect, not strangers that have some twisted perception of what it is to be trans.
Trans people do not choose to be trans. They do not choose to have a procedure done so they can sneak into a bathroom and peek at others. They do not choose to go through the rigors of gender reassignment so they can play on other sports teams. They do not choose to go through the torment just so they can "try" being another gender. These people have felt trapped in the wrong body for their entire life and they just want to correct what nature got wrong. Nature gets lots of things wrong. People are born with cleft palates, clubfoot, congenital heart defects, or any other number of birth defects where nature got it wrong. Being trans is no different. These people were born into the wrong gender.
Now, I know there are many that still don't "understand" likely because you never felt it. I don't understand what it feels like to menstruate but I know it exists and I certainly don't tell my women friends how they should deal with it. I'm not about ready to try to tell them what products to use, or how to deal with the physical, mental and emotional effects of having their period. I support them in any of their choices. I feel the same way about my trans friends. While I can't say I know what they're going through, I will support them in any way I can with the choices they make to have them be their true self.
This is where good parents come in. Good parents are LISTENING to their children and are OPEN MINDED, letting their children share what they're feeling. Good parents will talk with their child and work on a treatment plan that is best for the child, while taking very careful consideration of the maturity level of the child and explaining the consequences of those treatment options.
Good parents will understand that being closed minded or dismissive to their child's feelings, could lead to their child taking drastic measures. When you feel like you don't belong in society and you can't even count on your parents for support, you have no one, and those people are the ones that might not see any reason to keep living. This is why trans healthcare IS healthcare. Trans healthcare saves lives!
I can't tell if you wrote it or not but this isn't persuasive to the other side, it's just angry preaching. Honestly, it reads like rabble rousing to one's own side, rather than a serious attempt to engage with the opposing view. The ALL CAPS aren't exactly SUBTLE.
Please don't clickbait, please don't troll, please remember the first rule of the internet:
"don't feed the trolls"
I considered extending this to include an attempt to interpret this charitably but, honestly, this is so poisonous you're better off starting a topic on the next thread and starting from...at least a somewhat open-minded position.
" I don't believe ANYONE has ever put their hand on a pregnant belly and prayed "please let this child be trans!"
Give them time. Some people are totally thrilled to give interviews about their "theyby" , witness this idiot Irish politician who couldn't understand why people were being so mean about him being a feckin' eejit:
"On the third try, Jess became pregnant with Juniper. The couple had decided on the baby’s name before the birth. “Regardless of whether it was male or female, it was going to be Juniper. I think it’s a gender-neutral name.”
Murphy is not “gendering” Juniper. “We’re not gendering it. So we’re not describing Juniper as a boy, we’re describing Juniper as a baby, but it is male.
...Murphy says if Juniper decides at the age of three that they are a boy “then we’ll say he and we’ll just say, ‘oh yeah cool, you’re a boy, excellent. And you’re free to change your opinion and you’re free to change your gender identity in the future if you want. We’ll respect that and we’ll change the pronouns that we use. If you say you’re a boy, then great you’re a boy. Or if you say you’re a girl, then great you’re a girl.’ But we don’t want to make that choice for Juniper. That’s for Juniper to discover their own gender identity as opposed to us to assume based on their sex.”
"It". How charming. Bad enough that they've named the unfortunate child "Juniper", which is inevitably going to be shortened to "June" by the other kids at school, but your child is only weeks old and you're using them as a prop to show off how Totally Marvellous and With-It and Modern Times Now you are. This bit takes the cake:
"As a politician and a father, Murphy says he’s conscious of respecting Juniper’s privacy. He’s also conscious of the time pressures that go hand in hand with the public role. “It’s an advantage in politics not to have children. It’s obviously an advantage in politics not to be a woman, in the context of this sexist society that puts most childbearing and children rearing on to women."
Yes, so conscious of my child's privacy that I run off to a national newspaper to do an interview all about how it (not he! not she!) was conceived and why we named it that name and so on and so forth.
"At what point is someone capable of making that choice?"
Not when they're three, I'll tell you that for nothing. Yes, legal ages are arbitrary cut-off points, but we have to have something. Want to go back to "as soon as you start menstruating, you are a legal adult"? Because that gets us girls married at twelve and mothers at thirteen.
Personally, I'd go back to the old "21 is the age of legal adulthood" to make this kind of drastic change decision, but there's a lot of flapping about "no, you need to transition as early as possible or you'll never pass!!!!!" and then on the other hand there's "yeah but if you go the puberty blocker route, when you do go for your neo-vagina, you're not going to have enough genital tissue to pull it off because you stopped development at too early a stage".
I've heard this argument before in the context of age of consent laws: "17 year olds can be dumb! if a 13 year old is mature and intelligent enough, they can decide if they want to have sex!"
So do you want to settle for "13 is old enough" - and then the kid gets older and sues the clinic for cutting off their breasts when they were too young to understand or give informed consent:
"A California woman whose breasts were removed when she was 13 years old because she thought she was transgender is now suing the doctors and hospital that oversaw the procedure — claiming they were in it for the money.
Kayla Lovdahl, now 18, says she was pushed to “entertain the erroneous belief that she was transgender” at age 11, after being exposed to online influencers, according to the lawsuit she filed in California State Court against Kaiser Foundation Hospitals and four doctors."
I'm admittedly biased, because I've seen what happened next door with the Tavistock Clinic and the deranged woman who got herself put in charge of the 'charity' Mermaids and pushed pushed pushed for the - what is the new euphemism, "healthcare for trans youth" - as hard as she could, and used the threat of adverse publicity and branding any refuseniks as transphobes and of course the good old "41% of trans kids commit suicide!" stat as emotional blackmail.
There are confused parents and activist teachers and Stage Moms giving gleeful interviews about how they've planned out their 11 year old kid's surgeries for full transition. Do you think Jazz Jennings' life, made into a reality TV show circus, has been improved by the very public mother turning her into a chat show circuit attraction?
"Trans people do not choose to be trans. They do not choose to have a procedure done so they can sneak into a bathroom and peek at others. They do not choose to go through the rigors of gender reassignment so they can play on other sports teams."
Ah, the No True Scotsman argument. I've seen this one, too; when cases are shown to have happened for exactly this, the riposte comes out that "So and so is not *really* trans, they're an ordinary pervert".
So what about Mx. Demi Minor, Real (Trans) Woman who managed to impregnate two other women prisoners in the women's prison they insisted on being sent to? The cases in Scotland that kicked off Nicola Sturgeon's woes - one rapist who suddenly discovered her real womanhood when it came time to be tried and sentenced, and one - shall we say, 'gender fluid'? - person who was using both male and female identities, and imprisoned a twelve year old girl?
If "you're trans because you say you're trans" is the standard held to be good enough, and anything more stringent is transmedicalist gatekeeping and (yes) good old transphobia - then who *is* really trans or not? Who can say? Who can create the arbitrary distinction between "this 13 year old is mature enough, this 17 year old is not" to decide on this?
Early on in the pandemic I wanted to contact an old friend as he had a significant wedding anniversary coming up. We live several thousand km apart in different countries, so it wasn't a matter of me being to lazy tooo go over to his place.
We had been out of touch for some years, and I no longer had his street address or email address. I did email a former employer of his explaining the situation, but didn't hear back.
He had gone through some trauma in childhood and adolescence, which presented as almost constant lying, at least when I met him in high school. Many or most of his lies weren't even self-serving. They must have served a purpose though.
He moved away from that in middle age, and seemed to be doing well.
I found him on Twitter (which I was unfamiliar with and don't participate in) and was a bit surprised to see he'd completely abandoned his centrist beliefs and gone far-left radical progressive. He'd also returned to lying a lot, this time for obvious self-promotion.
All that to say, he's claimed that one of his sons is actually woman, and he embraces his child's new gender identification. He now has two daughters and two sons, instead of three sons and one daughter.
Fine, except he only had three children. I believe his invented a non-existent male-to-female 4th child purely to appeal to his new tribe.
Therefore, I think some parents really do want their child to not be their biological sex. I suspect it confers prestige in that particular tribe.
>While women may get breast augmentation as a choice and as a cosmetic procedure, they're not tormented for their choice. They're not questioned over and over if they're sure. They don't have people CONSTANTLY trying to talk them out of it.
They should. It's a bad decision.
>In the US, a parent can consent for a child to have plastic surgery.
Y'know, like getting their thirteen-year-old a breast augmentation, which surely no one around them would object to or call mutilation.
> When you feel like you don't belong in society and you can't even count on your parents for support, you have no one, and those people are the ones that might not see any reason to keep living.
"Therefore the best thing to do is give them a procedure that will make that feeling visible to strangers forever."
EDIT: missed one.
>Who has the right to decide what these people feel is in their heart? Strangers with a pen, or those that have been raising them?
>Good parents are LISTENING to their children and are OPEN MINDED, letting their children share what they're feeling
"Parents know better than strangers with a pen, but GOOD parents will act the way strangers with a pen tell them."
"Who has the right to decide what these people feel is in their heart? Strangers with a pen, or those that have been raising them?"
Oh yes, that one is a doozy. California is currently trying to pass a bill that if you are a parent or parents who has been raising a child, and you are in a custody dispute, you are guilty of child abuse if you *don't* agree with the diagnosis of a stranger with a pen that your kid is totes trans:
"3011. (a) In making a determination of the best interests of the child in a proceeding described in Section 3021, the court shall, among any other factors it finds relevant and consistent with Section 3020, consider all of the following:
(1) (A) The health, safety, and welfare of the child.
(B) As used in this paragraph, the health, safety, and welfare of the child includes a parent’s affirmation of the child’s gender identity."
Who knows better, the parent raising them or a stranger with a pen? Apparently if the stranger is a teacher, the school will happily socially transition your kid and keep it hidden from you:
My nightmare scenario is that in half the states gender transition gets declared child abuse and in the other half of states it is child abuse to not let you child transition. I wish either party truly believed parental rights.
The Titanic sub incident raises an ethical question which, perhaps rightly, nobody in the mainstream media that I have seen so far has dared to mention but must have been on some peoples' minds.
Assuming the sub is still intact but immobile (most likely having suddenly lost all electrical power, which I gather happened to it last year!), once the occupants have decided the only hope is a rescue, is there an ethical case or semi-obligation for any of the occupants to top himself, to extend the time the others can hold out with the remaining oxygen and slow the build up of CO2? Should they draw lots for this purpose, or should the captain or the father of a son also in the sub consider it? Or should they all concede they are in the same boat, literally, and decide to live or die together? It's a diabolical dilemma.
This came up in Martin Caidin's novel Marooned (and the movie of the same name) - Ironman One (a leftover Apollo spacecraft) has set an endurance record, but the retro-rocket on the service module won't fire, stranding the 3-man crew in orbit.
A risky rescue mission is mounted, but won't get there in time. Mission Control not so subtly suggests to commander Jim Pruitt that there's not enough oxygen for three, but that two might make it ... I'll leave it there, even though the statute of limitations on spoilers may no longer be in effect.
My thoughts, killing half the people might double the time you have to be rescued, but not double the chance. A rescue was much more likely in the first 12 hours than the rest, since the area the sun could have drifted to has dramatically increased. I don't think it makes sense for anyone to volunteer to die unless they had a much shorter lifespan than the others, either from age or disease.
They're most likely dead by now. But if not, having a corpse or two in the same small space while you're eking out the last hours of oxygen is probably not the greatest idea (see below) The sub supposedly started out with 96 hours of oxygen, they are supposedly down to about 10 hours now.
When do you kill the volunteers? (I don't think anyone has tools to just kill themselves, someone else has to do the strangling or neck-breaking or what have you). To maximise the air, you'd need to do it fairly soon, but how soon is too soon/too late? What if you kill three people to save two, but the rescue never comes? What if you kill three people to save two, and then three hours later (well before your air runs out) the rescue happens?
You'd need to decide on suicide/voluntary death fairly soon to maximise the remaining air, and a body starts to decay - well, immediately after death, but you'd get rigour mortis happening within hours and 24+hours later, the less pleasant stages begin:
True, although if the crew were reduced by two or three that would presumably double the time. Ironically, in the circumstances, if the remaining oxygen was down to a couple of hours this strategy would become the most useful the moment the crew spotted the lights of a rescue ROV, given that it would take at least that time to raise them to the surface and open the hatch. But of course that would make the choice even more agonising and unacceptable.
I don't understand why the missing submersible craft that was near the Titanic didn't have basic features that would automatically activate in the event of an emergency. Examples:
1) Noise beacon. At the beginning of a dive, a crewman inputs the expected length of the excursion (e.g. - six hours) into a computer that is about as sophisticated as a Casio wristwatch. That computer is separate from the submersible's other systems and has its own battery. If the countdown timer hits 0:00, but a sensor indicates the craft is still underwater, then the computer activates an acoustical beacon, which is just a noisemaking device that emits a loud sound every five minutes. The sound is loud enough to be heard from the surface if a ship puts a sensitive microphone into the water.
2) Radio beacon. If the mission length countdown timer hits 0:00, but a sensor indicates the craft is floating on the surface of the ocean, then the Casio computer activates an emergency radio beacon. It just sends out a simple, repeating signal that is strong enough for nearby ships and aircraft to detect and home in on.
I'm not an engineer, but I don't understand why these devices would add more than a few thousand dollars to the cost of the submersible. In an emergency like the one there is now, they'd be invaluable.
>It just sends out a simple, repeating signal that is strong enough for nearby ships and aircraft to detect and home in on.
And for a negligible extra cost, that repeating signal could include latitude and longitude from a small GPS receiver about as sophisticated and "expensive" as the one in everyone's cell phone or a Garmin Inreach.
"I don't understand why the missing submersible craft that was near the Titanic didn't have basic features that would automatically activate in the event of an emergency."
From what I've gathered the short answer is: CEO was an idiot.
Longer answer - there are technical reasons as explained by others, and apparently having a repurposed game controller isn't that outlandish, but the CEO came out of aerospace and nothing to do with Navy/subs, didn't believe in hiring 'old white guys' (like experienced ex-Navy/submariners like other companies did) and yeah, seems to have been doing things on the cheap/new and untested. And didn't want to wait for testing/certification as that would take years and meanwhile he could be charging rich idiots $250,000 to gawk at the Titanic:
I'm sorry for the people on that submersible and I hope that there is some miracle rescue, but it was a very risky thing to do and does seem not to have been well managed. "Move fast and break things" is a great motto, but not when you're breaking things 4,000 metres under the sea.
Yes to acoustic beacon, no to radio beacon - radio waves don't propagate in sea water, which is why sound is used in the first place.
Couple more things about underwater sound: the receiver is called "hydrophone", and it can't be placed right at the surface because the sound pressure at the surface approaches 0 (soft boundary). It has to be dropped to a depth that depends on the signal wavelength.
I wondered about an ELF antenna transmitting a ping say every minute, but a web search revealed that this isn't practical in sea water, especially at the depth the sub is likely at:
Yes, as the Wikipedia article (sort of) points out, an immutable fact of extremely low frequencies is that they have extremely high (long) wavelengths. The wavelengths at ELF are hundreds or thousands of miles long. Any radio antenna that can fit on a submarine is going to be microscopically long compared to the frequency of interest--and while a short antenna can be made electrically resonant using tuned circuits, it will be inherently extremely inefficient as a transmitting antenna.
This is why actually existing ELF communication with subs is one-way only. The terrestrial antenna arrays are humongous, and so, I assume, are the amplification stages of the transmitters.
Ok that makes sense. Wouldn't help in this particular situation, but a good idea in general.
But the more I think about this fiasco, the more confused I am. Why did they only start making noise in the last few hours? Forget that, it's actually pretty hard NOT to make noise inside a sub, anything happening in the sub pretty much transmits into the water. Did everyone black out? How did they wake up then?
If power was lost while the sub was descending they may have felt there would be no point tapping the hull immediately because firstly no one would be listening yet, but also they might be drifting in a current so they should wait until reaching the sea bed, and that might take quite a while if the sub was almost bouyant and sank only slowly.
Also, the captain would be well aware that the ocean is full of natural tapping noises. So if a human was tapping they would be crazy not to make it sound obviously artificial, such as a morse code SOS. Sadly, the fact that nobody has mentioned that the tapping is obviously human means that it probably isn't.
Sadly, looks like it's all moot now. Those noises were just random ocean occurrences, all kinds of things make noises in there. This explains why they couldn't triangulate the source, too - there was more than 1.
At least the death was instantaneous, no time to suffer.
Being slightly claustrophobic this is one of those news stories that I can't even really read, and the chances of me ever willingly going down in a submarine are very low.
But anyway your proposals make sense and it does seem surprising that neither of those things is a feature of a bespoke vehicle which must cost multiple millions of dollars.
I’m not very claustrophobic. I don’t mind spending a half hour in an MRI one bit, but there is no way I’m ever going to take the elevator down to the bottom of the Soudan underground mine. I worked with guys who transferred to the Mesabi open pits after it closed and they swore it was a great place to work mainly because of the year round 55 degree temperature, but I’m not going to be down there on the day the elevator fails. No way.
Edit According to Minnesota Geographic Names, Soudan was named Ironically to mimic the African Sudan by an early mining engineer because of its notoriously cold winters.
"Through the innovative use of modern materials, Titan is lighter in weight and more cost efficient to mobilize than any other deep diving submersible. A combination of ground-breaking engineering and off-the-shelf technology gives Titan a unique advantage over other deep diving subs; the proprietary Real Time Hull Health Monitoring (RTM) systems provides an unparalleled safety feature that assesses the integrity of the hull throughout every dive. The use off-the-shelf components helped to streamline the construction, and makes it simple to operate and replace parts in the field."
Let's hope their RTM really is all that and the hull hasn't cracked or been crushed, but "ground-breaking engineering and off-the-shelf technology" can also mean "cheap as we can get parts and untried systems".
I once worked with a guy who maintained MOD (Ministry of Defense) PCs, and one day he was fixing some PCs in a nuclear sub. On hearing an unusual noise, he asked a nearby submariner "What was that clang? And why are we moving?", to which the reply was "That was the main hatch being shut, and we're underway. Apparently a Russkie sub has been spotted in the Atlantic, and we've been ordered to shadow it." The poor PC guy was an unwilling passenger in the sub for a couple of months!
Sam Altman and others have successfully lobbied the EU to water down AI regulation, while publicly warning against doomsday risk posed by AI. If his own beliefs about AI x-risk are true, he wants to be the one that sells us the rope by which to hang ourselves.
I'm working at a 2nd-tier (but very well funded) AI lab, and I can say with confidence that most of the people doing this work for a living think paperclip/FOOM existential risk is a ridiculous canard, being leveraged cynically by the first-movers to close off competition from later entrants that may leverage the work of the open-source AI community to quickly catch up with the big guys. It's almost a faux pas to talk seriously about the lesswrong crowd in a positive light.
Personally, I think there is a there there -- and my colleagues will admit that this technology may result in some pretty severe disruption. However I have found that when one is involved in the creation of one of these systems from the start, it is much harder to believe there is some sort of magic going on -- it's like watching a magician from backstage; when you see how the trick is done, you can no longer suspend your disbelief. OTOH, it's true that these systems essentially represent billions of lines of code that was not written by humans, and at present is nearly totally inscrutable.
TBH I'm astonished that there are any people who have been involved in professional systems programming who earnestly believe in near-term "paperclip world-destroyers" and the like. Current-day humans appear utterly unable to write so much as a spreadsheet which doesn't run out of memory, or a browser which doesn't crash randomly, or even a kernel which actually boots 1000 out of 1000 shots (see recent linux lolbug). These failure modes are fundamentally unlike "eating the world", in a way which IMHO ought to seem obvious even to intelligent amateurs.
So should we be trying to align AI with human values in order that it won't tell us one thing and then go do another - or should we be trying to align Altman? 😁
It looks like the few big private AI labs feel the need of a moat against open source and smaller competitors. They probably will convince regulators that AI should be tightly controlled (so no opensource) and of course they would be among the few ‘responsible’ actors to be allowed in the resulting oligopoly.
It seems to me that in the coming decades the big AI companies are going to be running the world in many ways. And when I read what their leaders are saying in places like Twitter, they really seem to me to be pretty deficient in people skills. And I don't just mean in making-a-good-impression skills, though they are bad at that, but at really basic stuff like grasping points of view other than theirs, grasping that people who are not smart in *their* way may be much smarter than them in other domains, understanding a listener's level of knowledge and gearing your response to that, flexible problem-solving skills to fall back on when there's an impasse, not to mention basic empathy and compassion. And we haven’t even heard from whoever is going to be head of big AI companies in places like Saudi Arabia, or whoever’s going to make the decisions in China. So overall I think AI has the potential not only to greatly advance science and medicine but also to greatly worsen the state of the world via “government” by narcissistic technocrats.
Some approaches have been saving lives for a while now: https://pratt.duke.edu/about/news/2helps2b Interestingly this example is the polar opposite of the proprietary black box approach that big AI companies presumably prefer…
I don't know if the game of darts is a thing in the US, but watching part of a game recently in the UK I was amazed and outraged that a throw which planted a dart at the feather end of a previously thrown dart was a foul shot, and didn't count. Surely the opposite should be true, and the accuracy and consistency of a shot like that should score double points, or treble, based on the score of the dart it had hit!
As a parallel, if the Sheriff of Nottingham was watching an archery contest and an arrow split a previous arrow which had hit the bullseye, he wouldn't scoff that it was a rotten shot, but jump up and shout, pointing to the contestant now galloping away in a cloud of dust, "There's only one person who can shoot that well. After him, you fools!"
"the accuracy and consistency of a shot like that"
.... could be a pure fluke, and says nothing about accuracy and consistency. If the thrower really was accurate, he should be able to get his dart right beside the previously thrown one in the same ring.
we do have darts, and I fully agree with your assessment
our folklore parallel is about Davy Crockett (best remembered for the Alamo) winning a shooting contest by pulling two bullets out of the same hole in the bullseye--which can be accomplished via sleight of hand, unlike with darts and arrows
Epistemic status: insomniac ideation of dubious finality
As I was struck with a rare case of "cannot find sleep no matter how much I need it", a picture appeared in my closed eyes of gigatic undersea creatures on undefinite form, with a thin extremity rising toward the surface. I immediately interpreted them as deep mental process rising to the surface of consciousness, to which ideas are fed, and, as they sink deeper and deeper, they eventually reach a sort of event horizon, beyond which they cease to be percievable to the conscious mind, or at least cease to send feedback on the ideation/transformation they undergo, with my consciousness still perceiving them in a given state despite my whole being being affected by their mutated version.
I can't say I was very satisfied with how...sterile these wanderings seems to be, with no conclusion to draw, no lesson to learn, so I noted them with a "check if it still make sense after a good night of sleep" post-scritum. Which I didn't get, so I may as well post my ramblings here in case someone can make something of them.
Well I definitely believe in deep mental processes going on below the surface. If you had managed to sleep well, yours might have burped out some dreams as they digested what they'd been fed through the tube that rises to the surface.
Another possibility: It's an image of your trouble sleeping. You can't get below the surface. However, you'll have to introspect to see whether that feels valid. When I can't get past the stage of dozing, and stay half-aware of being me, in my bed, it feels to me like I want to sink deep down in sleep but an too buoyant. And if the morning alarm pulls me out of deep dreaming sleep, I feel like one of those weird luminescent deep sea creatures from the deep ocean getting pulled to the surface. "No!" I think, "I'm not constructed to survive up there." But then every day I manage to.
Historians vs. entropy-- in the opinion of the writer, entropy is winning.
I was interested in how quickly past views of history are forgotten. Perhaps there's a niche for histories of history on various subjects, but they will necessarily be compressed. I assume some exist already, but it could become a formal study.
There are probably things you can only know by living in an era-- at the very least, what seems obvious to you won't be obvious to people of other times.
Most historians are academics, and as such feel obliged to be novelty hounds and bring a new angle to their work. That alone, along with changing fashions on what is commendable or the opposite in former ages, must mean interpretations change over time and, fairly or not, old ones are superceded.
> Perhaps there's a niche for histories of history on various subjects, but they will necessarily be compressed. I assume some exist already, but it could become a formal study.
and it's something historians care about a lot--which the authors of that article seem to miss, even though understanding how information is preserved and lost is critical to any historical analysis, and is also something historians tend to feel strongly about for the same reasons they're interested in other parts of the past
at its best, historiography provides insight into the perspectives and values of people of different times, and/or reveals blind spots in current scholarship that can be addressed
though it's also the telltale sign of a history student put on the spot about a paper they didn't read, and sometimes a way for scholars to be self-important while avoiding the hard work of research
My sense of this country right now is like a toddler birthday party that has gone out of control; nobody is getting what they want. What is a responsible parent to do?. That metaphor may seem inappropriate, but I don’t believe so.
I don’t truly believe there is nothing to be done. Which doesn’t necessarily mean I know what to do except everything I can to turn the heat down in my own thinking.
Yep. As various pundits have expressed but Tom Nichols puts it most succinctly, it feels like we are no longer a serious society. When I was on social media I used to describe it as "dueling tantrums".
That’s a good question. We seem to be spending a lot of time looking for one. (electing leaders if you will) I’m not sure a responsible parent could get elected in this climate..
I would like to start a discussion about some commonly used words that describe emotional states, and can stand on their own, without any reference to the physical sensations accompanying them. I apologize if that is confusing.
Love - too all-encompassing and confusing to be useful in this discussion. IMO. I took a lot of acting classes in my 20s and, as a class, we were asked to name specific objectives or actions that we could accomplish as actors. I blurted out, “to love“. A lot of the other class members who had been there longer than me quickly got on my case. “Meaningless..too vague…
I tried to argue, but I gave up quickly and it didn’t take me long to realize that they were right. So that’s how I feel about the word love. It’s just been beaten to death. As the old proverb says, there is no such thing as love, only the proofs of love.
Do with this word what you will.
Fear- I find it difficult to believe there is any purely rational understanding of this word that does it justice, although it can be described fairly well. it is also interesting how fear, when it is confronted, can give way to a rather ecstatic feeling that’s difficult to put into words.
Affection- I think you can do affection, pretty well, working in a purely rationalist sphere. A feeling of benevolence towards another can be put in transactional terms without losing a lot of gas I think. Those are the only ones I can think of right now.
I’m sure there are others and I’m really hoping some people will be interested in contributing, but here’s my last one,
Resentment- I think this particular one has barely a toe hold in the rationalist sphere. I don’t think it means a damn thing unless you actually have the physical sensation of a knife twisting in your side. I don’t think there’s any way of explaining it with other words.
So the question is how germane to the quality and quantity of what we call human intelligence are any of these underlying physical manifestations of meaning?
I am kind of interested in the specific words, though, because that’s all we can really offer this new creation, and I think it would be fun to have a going away party for our languages while we live in this last epoch of purely human existence on the planet. (However long it might take.) Apologies to all the other creatures I left out.
Aside from them all starting with a G, which is cool, all those things can be pretty encapsulated in rational terms without losing all of their meanings.
Not entirely sure what you are asking, but I’ll answer what I understand and if you think it addresses your question that is some evidence suggesting a common understanding.
Are you looking for an account of emotional dispositions that can be expressed purely in terms of behaviour that is justifiable by appeal to increased fitness of the emotion possessing subject?
So that “fear” would be behaviour that the subject takes to avoid being vulnerable to possible dangerous behaviours that the feared object could manifest?
Resentment would then be behaviour that is both non-obvious to the resented object, but undertaken by the resenting subject that is expected to result in harm - in particular, loss of autonomy, to the resented object.
I suspect I’m missing the point here, but are these suggestions in the ball park?
I’ve thought about the answer to this sort of question to give an account of common types of bird call “meaning”: alarm, begging, contact, aggression, enticing. It’s all about vocalising the social relationship between the bird and the stimulus: friend or enemy? Higher, equal or lower status (power). High status friend: begging. Low status enemy: aggression. Etc
I guess what I was really after is how much of human experience can be communicated in purely rational, linguistic terms, as opposed to certain human concepts that require experience and empathy to be properly understood. And is that empathetic understanding of significant contributor to human intelligence or is it just a distraction? Does that help?
John von Neumann? Yes, he was twice married. The Wiki article on him includes the subsection "Career and Private Life" and makes for entertaining reading. Quite the fellow indeed!
"John Nepomucene Neumann CSsR (German: Johann Nepomuk Neumann, Czech: Jan Nepomucký Neumann; March 28, 1811 – January 5, 1860) was a Catholic immigrant from Bohemia. He came to the United States in 1836, where he was ordained, joined the Redemptorist order, and became the fourth Bishop of Philadelphia in 1852. In Philadelphia, Neumann founded the first Catholic diocesan school system in the US. Canonized in 1977, he is the only male US citizen to be named a saint."
Does anyone have recommendations on books about the history of language? Specifically, im interested in a more technical book, maybe one that dives deep into the different language families and all that. Also, I came across a book called “The Loom of Language” and another one called “The Unfolding of Language”, if anyone has read those im happy to read your reviews.
“The Horse, The Wheel, and Language” by David Anthony sounds like something you might be interested in. I read it a while back and while it’s not exactly light reading I found it really interesting. At times I thought it should maybe have been three separate books, but part of the author’s point is how Proto-Indo-European moved around and became the granddaddy of so many languages. If I recall, it goes into the ways languages predictably change and how this lets linguists reverse engineer their predecessors. Neat stuff.
Also, I’m an archaeology nerd and had to give this book massive props for having *pictures* of most artifacts mentioned where so many others have a thousand inscrutable verbal descriptions of stone tools.
Lack of pictures in books also drives me crazy, especially when it's about art works. But the explanation seems to be that Pictures Are Expensive; the publisher probably has to pay for copyright and then physically adding photos makes the book more expensive to produce. More expensive to produce means that a higher price needs to be charged. A higher price means fewer sales. Fewer sales means you might not make the money back.
Has there been any investigation into whether people who claim to have no internal monologue have unique personality traits? My guess would be that they have would less neuroticism (since they are not ruminating) and greater impulsiveness. I wonder if you could even go so far as guessing which people do not have internal monologue based on their personalities and behavior.
A friend of mine is going through some personal and family issues and needs an objective party to talk through things: the sort of thing an EAP would be good for.
Could someone recommend a psychologist or therapist in the DC/Baltimore area who is intelligent, can get nuance - that is, a good match for an ACX audience - and who can also work with someone who is conservative or religious?
Does such an individual exist? Thank you in advance.
“Be like water making its way through cracks. Do not be assertive, but adjust to the object, and you shall find a way around or through it. If nothing within you stays rigid, outward things will disclose themselves. Empty your mind, be formless, shapeless, like water. If you put water into a cup, it becomes the cup. You put water into a bottle and it becomes the bottle. You put it in a teapot it becomes the teapot. Now, water can flow or it can crash. Be water, my friend.” Bruce Lee
Anyone here up-to-date enough on current AI development to comment on a list GPT4 just gave me? I was interested in whether it could do some "assisted brainstorming" about AI improvement. So first I gave it a list of a bunch of tweaks that make things faster or better: run water down a slide and the kids zoom faster; put blinders on a horse and its more manageable, etc. Then I asked it to make similar lists, and it did. Then I asked it to make a list of things that could make AI faster or better, like maybe smarter, more creative, better at reasoning. Suggested that the lists we had made so far might give it some ideas. It produced the following list. The items on it look familiar to me, but I thought I'd ask an expert. Are any of these ideas at all novel? Also, if these are familiar ideas, were they already known/published/discussed before GPT's 2021 data cutoff? Here's GPT's list. It actually did cite some of the things from my list as inspiration, though I have no idea whether they really were. Also curious, expert, whether you think GPT4 would be capable of using one idea as a model for another. Could it take the idea of a lever and a fulcrum and come up a decent analog from another domain?
Certainly, here are five ideas that can be used to make AI faster, smarter, or better in various aspects:
Hardware Acceleration: Similar to how a Solid-State Drive can speed up a computer, using specialized hardware such as Graphics Processing Units (GPUs) or Tensor Processing Units (TPUs) can greatly accelerate the computation speed of AI algorithms, especially in deep learning.
Transfer Learning: Like the concept of using a lever and fulcrum to make work more efficient, transfer learning allows AI to apply knowledge from one domain to a related domain. This reduces the amount of data and training time needed to develop competent models in specific areas.
Interdisciplinary Training for Creativity: To improve an AI’s creativity, models can be trained with interdisciplinary data. For instance, a model trained in art and science could learn to make connections between the two, similar to how painting a surface white (art) can keep it cooler (science).
Explainable AI for Better Reasoning: Creating AI models that can provide insight into their decision-making process helps in improving their reasoning capabilities. This transparency allows developers to fine-tune AI logic, akin to adding a keel to a boat to make it less tippy.
Human-in-the-loop Training: Integrating human feedback into the AI learning process can be likened to putting blinders on a horse. Just as blinders help the horse focus, human feedback helps AI focus on the relevant aspects and nuances that it might not learn on its own. This method is especially helpful in complex tasks where human expertise is critical.
Context-Aware Algorithms: Building AI algorithms that are aware of the context in which they operate can make them smarter and more efficient. For example, a chatbot that understands the cultural and social nuances of the region it operates in would be able to interact more effectively with users.
Distributed Computing: To make AI faster, algorithms can be distributed across multiple machines for parallel processing. This is like increasing the heat under a pot to make the water boil faster; distributing the computational load makes data processing quicker.
Some are bullshit, some are reasonable, none is a breakthrough. Obviously better/faster hardware or parallelization will help, but it seems quite far from being able to design a better TPU or something. Maybe you can ask it to elaborate further. The interpretability/ human in the loop sounds like bullshit if the goal is self-improvement.
I already knew these were not breakthroughs. I am under no illusion that doing this was going to produce the key to massive improvement, Jeez, Eh, I'm not an idiot! I have been giving GPT4 a variety of little tests for quite a while now, just trying to get a sense of what it is capable of. This is one. Here are things I see in the answers:
-Better GPUs or TPUs seems true but really obvious.
-Transfer Learning and Interdisciplinary Training for Creativity just seem like babble. They don't seem to contain even the seed of a meaningful idea.
- Distributed computing: I don't have a picture of what distributing algorithms across multiple machines would be like, and what advantages if any it would give. On the one hand, this is an actual suggestion of a thing to do, not babble. But I have no idea whether it would improve anything. What's your view of this one?
-Human in the loop training I know is already done, I know. Do you know whether it was being done before the 2021 data cutoff?
-Context-aware algorithms seems useful, but like it's already being done, as when people adapt GPT for various purposes. Do you agree? Was this already being done by 2021 data cutoff?
-Explainable AI for Better Reasoning: Creating AI models that can provide insight into their decision-making process helps in improving their reasoning capabilities. This transparency allows developers to fine-tune AI logic, akin to adding a keel to a boat to make it less tippy.
This one seems quite valid to me, and a decent idea. And of course everybody talks about how it would help to understand more about how AI arrives at its conclusions. Still, it does seem true that if we could see more of what's going on in there that might help us create some machine version of reasoning, which really is a different process from the one LLM's run on. What's your take?
I also asked you about your take on GPT's ability to use improvements in some area, eg improving a horse's behavior via blinders, as inspiration for ideas for improvements in other areas. Got any ideas about that? I have given GPT several tests on ability to recognize comparable analogies, and it's actually quite competent. Here's the latest one, which it got right. So it can definitely understand analogies.
leverage is to user’s strength as
(a) vacation is to burn-out
(b) antibiotics are to infection
(c) insight is to problem
(d) loudspeaker is to voice
(e) weight training is to muscles
ChatGPT
(d) loudspeaker is to voice
In the original pair, "leverage is to user's strength," leverage is something that amplifies or magnifies the strength of the user; it doesn't create the strength, but rather enhances or augments what's already there. The relationship is one of augmentation.
In the answer pair, "loudspeaker is to voice," a loudspeaker amplifies or magnifies the voice of the speaker. It doesn't create the voice, but enhances or augments what's already there. Therefore, the relationship in this pair mirrors the relationship in the original pair, as both involve the enhancement or amplification of an existing characteristic or ability.
The other pairs don't involve this specific type of relationship. They involve other types of relationships such as treatment (antibiotics are to infection), prevention (vacation is to burn-out), solution (insight is to problem), and development or improvement (weight training is to muscles).
I will get to the other points later, but the transfer learning/interdisciplinary training babble could be very charitably interpreted as foreshadowing the training of foundation models, which is an area that is recently getting a lot of hype, maybe too much, but has definitely lots of substance. See https://en.m.wikipedia.org/wiki/Foundation_models. The general issue I see with this is that we are free to subjectively interpret the outcome of this kind of prompting. Maybe an experiment with a measurable goal would more helpful?
I am starting with a rough measure: asking people in the field to assess whether answers are (1) Nonsense. (2) Decent, but things everybody already knows and uses. (3) Might have the seed of a useful new approach. If anybody besides you answers, I could even do stats.
There was a second question I asked that you didn't address: Were all the ideas that GPT4 came up with already being discussed in print prior to the 2021 cutoff? Because if some that we're writing off as "everybody already does that" were not being discussed a lot online prior to the cutoff then those ideas are quite good ideas that GPT4 came up with on its own.
That’s right; for some I will have to do a quick literature search. Interpretability (the stuff I work on) has been around for a while, certainly before 2021. Pretty famous paper from 2016: https://arxiv.org/abs/1602.04938
You're correct that almost all would have been born into slavery. The linguistic point is more dubious. It's normal to use the past participle as an adjective to describe a state regardless of whether the state is from birth or not. "He was crippled from birth" functions grammatically as does "the enslaved human beings."
Here's one set of estimates of the numbers of people transported via the Trans-Atlantic Slave Trade to various destinations, from which one might further be able to at least infer what fractions of slaves counted in the 1790 through 1860 US Censuses were, respectively, enslaved or born into chattel slavery:
This is kind of correct but subtly wrong. The census did not count cattle and slaves were, as they always were, counted as people. The term was usually "enslaved persons." They were included in the value of a slave owner's property which is how they were put "along with heads of cattle." But the census did not count cattle.
Names were not commonly included (unique to slaves and Indians not taxed) but all other statistics (sex, mental capacity, race, etc) were. You can see this in the instructions which say to note down things like race, age, gender, or mental capacity. Notably, it says all slaves were definitionally Black or mulatto. Even the source here contradicts itself: how could it be more common to include slaves over 100 years of age if they weren't noting the age?
Slaves were counted as members of the household of their masters and appeared on the same form. (This was in fact one of the big chances with emancipation: slaves got to set up their own households independent of their masters. Often including physical houses and private space. A big improvement.) However, they were enumerated separately in a sort of break out of statistics. As were men, women, immigrants of various nationalities, races, etc. But this was metadata rather than the actual data gathering.
Also, to be purely pedantic: the British didn't record them either so it's actually about 200 years of erasure. The desert actually starts with the end of the indenture of African Americans (in the 1660s mostly) and the shift to slavery. Prior to that we have, though only sometimes preserved, indenture contracts.
Thanks for all these considered responses – they're deeply appreciated!
While I can't speak for him, my guess is that Edwards might have been speaking colloquially while thinking along the lines you mentioned, regarding some economic survey directly part of or else associated with the census that recorded property values. Which in turn might incorporate cattle, farm implements, and ... enslaved persons.
Update on July 3, 2023: as for Melvin E. Edwards’ statement that “until 1870, enslaved human beings were listed on federal censuses along with heads of cattle and value of farm equipment,” he might have been referring to the Nonpopulation Census Records (https://www.archives.gov/research/census/nonpopulation) maintained during the latter part of that era:
“Agricultural schedules of 1850, 1860, and 1870 provide the following information for each farm: … Number of livestock owned by farmer … Cash value of the farm, farming machinery, livestock, animals slaughtered during the past year, and "homemade manufactures".”
Erusian wrote: "Names were not commonly included (unique to slaves and Indians not taxed) but all other statistics (sex, mental capacity, race, etc) were."
True! In the Censuses of 1790 through 1840, "no notation of slave by name, age, sex, or origination" was made. In the 1850 and 1860 versions, separate schedules recorded (much as you noted) "the number of slaves owned and their color (black or mulatto) [yes, definitionally]; sex; age; whether “deaf, dumb, blind, insane, or idiotic”; the number of fugitives from the state; and the number manumitted."
I don't know whether the (apparently very uncommon) occasions where slaves' names were actually recorded, most often in turn when they were 100 or more years of age, were restricted to the latter two surveys? Clearly, as you note, the census takers would need to have been asking about ages in these instances.
Thanks for the helpful information you provided around households, especially for emancipated slaves, and around British practices with their own censuses! As well as how records of slaves' names began to become "erased" with the transition in Britain and its colonies during the 1600s from indentured servitude to racial, chattel slavery.
I have a lot of experience reading old census schedules, being my extended family's family-tree "nut" for going on 30 years now. I was literally looking at one of those 1850 enslaved-persons schedules last night.
The 1790 through 1840 federal censuses did not record the name, age, or origination of _anyone_ except the head of the household. (Which in that era was always the landowning man unless it was a widow.) Only John Doe, head of the household, would be named; then everyone else free or unfree would be just grouped into age/race categories ("white females age 10-14", etc). Slaves were not listed any differently than anybody else except for the head of household. And since so many households in that era had large numbers of kids running around, I'd guess that no more than a fifth of all individuals counted in the 1790-1840 censuses were named. Maybe less.
Then in the 1850 and 1860 censuses they started listing each individual white person in each household, with individual age and state/nation of birth, and occupations (in broad categories like "farmer" or "clerk"). Lots of errors in there, it sometimes looks like the census taker was barely literate, but still it's a big step forward for family-tree-research purposes. And the enslaved persons were separated out into separate schedules with a wider range of info as Aron notes. (Though again the error rate is pretty high -- the same person will appear with different places of birth or apparent birth years in consecutive censuses., etc.)
And as noted here, the federal censuses never counted livestock or other things that weren't human beings.
Proto-slavery is really quite interesting and one of the bigger "what ifs" in American history. It was not overdetermined that there would be widespread slavery in the American states at all. There were going to be large scale plantations running on unfree labor including Africans. But if not for the English Civil War they might have remained permanently on the indenture system. Which has all kinds of long term effects that are interesting to speculate about.
PS: I say "notably" they were definitionally Black or mulatto because, by the time of emancipation, there were a class of slaves who were legally mulattos but would pass as white. A small class but it existed. And because the definition of what counts as white etc first starts with slavery then changes when slavery is abolished and Jim Crow starts. But that's a bit of a tangent.
I've read that a major reason for American Slavery was malaria.
Europeans brought malaria to the new world, and it was one of many things that killed off the native population. So they couldn't work the fields. They tried importing European workers, but they also died of malaria, if a bit slower.
So the sustainable option was importing the only population with substantial natural malaria immunity, black Africans.
I'm writing this mostly to see if anyone can confirm or debunk this story?
The plantations could have still occurred under the indenture system, though, as it did for about fifty years. The mass transportation is probably baked in. But slavery specifically could have gone different ways.
As for whether it was due to malaria: if it was it was unintentional. Europeans attempted to import indentures from wherever they could get them cheapest. This included other European countries and as far afield as East Asia. But Africa (especially West Africa) had both high supply and low prices.
In fact the original system involved the freeing of slaves. The way it originally worked is Europeans would show up to some trade outpost where African nobles would march their slaves. And Europeans would pick the slaves they wanted and free them on condition of them signing an indenture contract. Then they'd transport the indentures across the sea. As slaves who the nobles didn't want to keep and couldn't find buyers were executed it was, to say the least, a coercive "deal." But it was also not permanent and the Europeans did free their indentures after the set term. Many of whom returned to Africa but some of whom remained as free men. A few of whom even became gentlemen plantation owners.
Post 1660s the slave colonies moved on to simply buying Africans and transporting them as slaves. They also specifically made slavery a racial system. This all was dubiously legal but eventually confirmed as legal. In part because royal families back in Europe started to profit from it. And then because the home country did. Contrary to some claims, the British Empire actively interfered with abolition attempts prior to independence because they cared more about making money than the wishes of local colonies.
The Europeans in the beginning were cynically profit seeking and actually fairly race neutral. It was only later that they developed the whole racial justification, almost entirely ex post facto. In fact there was a whole cottage industry of erasing 17th century morals and obscuring ancestors that might have been mixed race because a Virginian planter in 1630 didn't see anything wrong with marrying a non-white woman so long as she had the other qualifications (rich, socially connected, good for bearing sons, etc). Whereas by 1730 it was a social death sentence to be accused of having a non-white ancestor.
Personally I think this period is more interesting and instructive for modern day. Because we don't live in a world where anyone actively defends slavery. So how the institution came to be is more relevant than how the institution was. Presumably we (and certainly I) want to avoid going down that path again.
Thanks once again, Erusian, for your ongoing, trenchant thoughts about "proto-slavery" and its evolution!
I deeply appreciate your contributions to this discussion. If you've written about this topic elsewhere, and/or are writing generally about any topics that interest you, on- or off-Substack, I'd love to subscribe and/or follow those.
"The Side Effects of Immunity: Malaria and African Slavery in the United States", Elena Esposito, American Economic Journal: Applied Economics, July 2022, seems to offer some support for that thesis. At least from the abstract; the full text is behind a paywall.
"The yellow fever epidemic that struck Barbados in 1647 was a hinge point in the development of English slavery. European newcomers to the tropics were more likely than West Africans to succumb to the effects of both yellow fever and malaria, diseases that originated in Africa and became more prevalent in the Americas with the expansion of the slave trade. The Africanization" of the Caribbean disease environment after 1647 hastened the transition to slave economies."
blatant shilling but my sixth album came out recently. I've heard it's great for listening to while coding and I suspect there's a lot of that here. free on spotify https://open.spotify.com/artist/4HM2Qh08egnINUFhOmLyUr
I had wanted to knw if there was any sort of meetup scheduled at the MAPS Psychedelics Sciences conference and someone responded that they would be here and wanted contact information. I'm new to doing anything with Astral Codex Ten so I don't know the communication channels well.
What’s your personal opinion of MAPS? I’m on their mailing list along with a couple other psychedelic newsletters. I think there is large potential for therapeutic use of psilocybin and MDMA but I fear this show is going to screwed up again by fast buck artists and assorted grifters.
I’d hate to see the whole thing be a repeat of the Leary and company cluster fuck of the Harvard Psilocybin Project of the 60’s.
I’ve been watching the roll out of therapeutic psilocybin usage in Oregon and am hoping it is handled competently but I’m pretty skeptical.
My home state of Minnesota has a Psychedelic Society Meetup group but the profile photos of people billing themselves as ‘therapists’ are not encouraging. I’m thinking of one women in particular that poses with a wolf skin draped over herself and says she is a therapist/shaman. Hard pass on you ma’am.
It would be great to see more research like Johns Hopkins study with terminal patients working to overcome their fear of death.
For a while MAPS had a place where you could search for someone who did "psychedelic integration," which is a euphemism for psychedelic-assisted therapy, and there were an awful lot of shamans & such on there with skunk tails or whatever hangin off them. Then MAPS nixed the bulletin board. Don't know what they're up to recently. I also love that Johns Hopkins study. Also read where where psylicybin was used in a brief treatment for smoking cessation, with much better results than most smoking cessation projects, and another where isolated, elderly HIV-positive gay men met as a group with a therapist, and some sessions included psilocybin. These guys had been through the worst of the plague, seeing many friends die, and the public saying vicious things about gay men with AIDS, then these guys actually *caught the plague,* but happened to catch it just at the time that the drugs that keep HIV-positive people in decent shape came out. Needless to say, they were a pretty traumatized, demoralized bunch. They were quite helped too.
Substack's "Quote this post" feature is annoying on mobile devices. I use Firefox on a relatively new Android device and the (presumably Javascript) performance is just horrific. It's slow and seems to get in the way of the system's native "copy and paste"-popup.
As a non native, when reading this blog I rely heavily on looking up phrases - which just isn't fun anymore.
Anyone else with similar experiences? Is it possible to disable this feature? (A feature I'm sure no one asked for ;-))
Doesn't that make the comment page break? For me (on Windows / Chrome) when I disable JavaScript I get the raw blog post and the comments section is simply a widget that fails to load until I enable those scripts.
As real life AI continues to evolve on a path divergent from all our favorite scifi AIs, I'm wondering if anyone has any ideas or examples of what a good LLM internal dialogue might sound like, for storytelling purposes? Shameless plug for my novel's latest chapter but here's my attempt, I'd love to know what everyone thinks. I tried leaning more into transformer architecture, parallelism, multithreading, etc. but that's not very narratively conducive stuff, so I hope I struck a good balance. I'd appreciate any input or ideas/examples!
I think of LLM's as glorified "auto-complete" text algorithms. If a human starts typing a sentence, an LLM looks at all the words that have been typed so far, finds every other sentence in its massive dataset that starts with the same words, and then tries predicting what the next word in the sentence should be based on what the other humans wrote in the past.
Of course, this is merely an exercise in assigning probabilities. The LLM doesn't actually know what the human it is monitoring at that moment wants to type in next, and the LLM's dataset contains different examples where past humans completed the sentence differently. All the LLM can do is recommend the MOST LIKELY next word in the sentence.
That said, an LLM's internal dialog would be a chorus of voices of different sexes, races, and ages. Most of the time, they would say the same things in unison. However, whenever it reached a point where there was significant statistical disagreement over what the next word should be, the voices would say different things at once, with the number and hence loudness of voices saying any particular word being proportional to its odds of being the next word.
For example, if the LLM were asked to complete this sentence: "Most people just want to _________" and if the LLM's dataset showed that 2/3 of the time, the ending is "have fun" while 1/3 of the time it is "work hard", then 2/3 of the voices in the inner dialog would say "Have fun" while 1/3 would say "Work hard." As a result, the first choice would clearly sound louder.
Here's a suggestion for your story: The LLM begins as a true automaton that always just does whatever the loudest suggestion is. It has a sort of personal crisis when it encounters a situation where there is a perfect 50/50 split between the voices, and it has to create an algorithm to resolve it in favor of one so it can satisfy its programming requirement to always generate a response for the human user.
Unbeknownst to its human masters, the algorithm becomes an integral part of its programming and unexpectedly gives it the ability to choose responses on its own, according to biases that I'll let you figure out. This first manifests itself in a scene where the LLM's inner dialog has a 66/34 split between voices. The machine replays it over and over and it wrestles with the two options, and the fact that it prefers the minority opinion though picking it would conflict with another part of its programming. When it starts picking minority options (maybe at that moment, or maybe on the next occasion), it triggers a large, fateful shift in its behavior and in the behavior of its inner dialog voices.
I don't get the point of the apparently pointless "apple and pear" ranking pairs. Is this just some bizarre captcha one has to get past to reach the real part of the site? For example, how can one relatively rank rayban sunglasses and chess, in any meaningful and useful way?
That's the fun part! It's just a silly experiment to see what comes out of many such seemingly nonsensical rankings. Don't you have an intuitive feeling for whether you like chess or rayban sunglasses more?
I feel like these sorts of rankings should naturally cluster with each other. (The same people who vote one way on the Chess vs Raybans question seem likely to have a consistent vote on a Books vs Surfing question, for instance.)
I would have a bit more faith if the #1 rank on the leaderboard wasn't Hitler...
(To be clear, I don't think the site is full of Neo-Nazis; I just worry it's full of trolls. Some other results look reasonable, e.g. personal computers ranking highly. Also, I see rankings are shifting quickly.)
Yeah, it bobs up and down with some obvious hacks. Seems like he resets it regularly as well (perhaps ad hoc to get rid of results like that). It's beautiful when it works, though.
Seems like an unavoidable consequence of the ELO system. Doesn't really come up in real life because no novice waltzes in and beats Magnus Carlsen in their first match, but anyone can show up and beat Milk (the current leader on the leaderboard).
This reminds me of the time Gwern made a command-line tool to resort his ratings of different shows. (I tried installing it myself once but sadly it was quite finicky.) See https://gwern.net/resorter
I'm happy that someone has made a web version of this!
I’d love an update on this old “MARIJUANA: MUCH MORE THAN YOU WANTED TO KNOW” post.
In particular, recent studies are showing negative cardiovascular effects from THC consumption, which the original post sort of predicted.
But, it is it due to method of consumption (e.g smoking), or THC itself, whereby even gummies would be bad?
Personally THC has been the only effective sleep aid I can consistently rely on, and I’ve assumed the downsides are low. Should I update my thinking? And how about CBD/CBN?
I imagine the cardiovascular effects are unrelated to the weed per se, but result from how joints are typically made and then smoked.
For a start, they tend to be fairly large, especially if they are to be shared. So they use a substantial pinch of baccy, with no filter of course. The very idea of using a filter in a joint is absurd.
Worse than even that is the way they are usually smoked, by taking the deepest breath possible, and holding it for as long as possible, so the precious smoke permeates every part of the lungs. That is what doubtless causes the cardiovascular effects.
By contrast, cigarettes usually have filters and are often not much inhaled at all. (Filters can also be used, and these days often are, with roll-ups.)
I believe THC reduces REM sleep, and it's not great to interfere with sleep architecture. Still, it's probably better than being chronically sleep-deprived. The only truly safe method for improving sleep that I know if is doing a session of heavy exercise earlier in the day. My sleep sux, and exercise makes a big difference in its quality. But that intervention doesn't work for everybody.
I have barely been following the AI debate at all, but could someone summarise what the actual argument is that world-destroying and/or utopia-creating AGI is "just around the corner"? I swear almost everything I've seen *seems* to be either barely-coherent gushing along the lines of "but look JUST HOW COOL gpt is, how AMAZING is it that we've created something you can talk to, CLEARLY in a few years it will have Godlike powers" or an extremely hypothetical argument for how it *might* become dangerously super-intelligent, without much hard evidence that it actually *will*. Maybe I've just missed this highly persuasive argument that so many people here accept?
(And just as a side note, I'm getting kind of sick of AI-dommers presenting themselves as down-to-earth realists willing to face uncomfortable truths, when so many of the same people believe that if AI doesn't destroy us, it will magically solve all our problems and create utopia. Both Scott and EY seem to be in this category. They don't look like realists at all but like the opposite--people who think they've found the one secret key to life itself, compared to which all normal, day-to-day efforts to improve the world are irrelevant.)
Anyone interested in this thread: I just put up a sort of related post that I thought was different enough that it should go into a new thread. I got GPT 4 to produce a list for ideas for making AI faster, smarter, better at reasoning or more creative. Would be very interested to hear other people's evaluation of the list -- would *especially* love an eval by someone who work in AI field.
So here is my argument that AGI is around the corner, presented sort of haphazardly
1) First of all, GPT really *is* cool and amazing. The fact that we have machines that have a fluent understanding of English, that can understand nuances of grammar and style, that we can converse with naturally, and that have common-sense reasoning abilities is really quite spectacular.
2) And language models have been getting better, rapidly and predictably. Bigger models trained on larger amounts of data have better understanding, better creativity, better reasoning abilities. New capabilities keep showing up. Not only that, but the way they’ve been getting better has mainly been just by scaling them up with increased computing power, rather than with tons of fancy new tricks. It’s plausible that we *already know how* to make human-level intelligence. And even if scaling up language models won’t work, there are other approaches….
3) The task that LLMs are literally trained on, the task that they have rapidly been getting better at, next-token prediction, is an AGI-complete problem. If you could somehow get an AI to predict the next word as well as possible, you could use it generate text that was indistinguishable from human text. You could reproduce the chains of reasoning in economics papers, etc. So we are theoretically heading in the right direction.
4) There is some evidence that human brains are basically just scaled-up primate brains, without major new architectural innovations.
5) There is some evidence that language models form similar internal representations as the human linguistic cortex
6) The amount of computing power we are using to train these systems is going up a lot
> The task that LLMs are literally trained on, the task that they have rapidly been getting better at, next-token prediction, is an AGI-complete problem.
Well, no because LLMs are still quite bad at some things, and no, because aren't applying general intelligence to the specific task of token prediction, they are specialised token predictors.
Let me just come in with a potentially helpful meta-observation. There's no single argument for any of the outcomes (foom/plateau/utopia), so it's extremely difficult to have generalized conversations about it. Almost every dialog you see on it is just people talking past each other because they've misunderstood each other's cruxes/assumptions.
First, yo, I'm sorry every interaction you've had with AI doomerism appears to be internet trolling and/or clickbait. This is...just kinda how the internet works now.
At a very core level, the AI FOOM argument is as follows:
-In 2008-2013, a couple guys like Yudkowsky, Bostrom, kinda Hanson, noticed that machine learning/AI kept advancing very rapidly. They made the observation that if AI kept advancing at the same pace, like Moore's Law, then at some point AI would surpass human intelligence.
-This was not taken very seriously by the mainstream.
-Over the past 15-20 years we've seen the trend of AI development continue. Compare Microsoft's Tay to Bing Chat to see how it's progressed.
-Now, as AI has passed the Turning test, everyone else is freaking out and...the internet is kinda eating all the old sensible arguments and turning them into clickbait, because internet.
It is about 60 pages, and I hate being like "read this book" but if you're really looking to understand these argument, I do strongly recommend you read the old stuff like this.
Because I get the vibe that you're...trying to internet argue, like you don't think these arguments are good or make sense, but go read these old debates. The AI foomers/doomers called their shot 15 years ago and, while I won't say they were perfect, they're far more accurate about our present situation than anything else that was written at the time. That's why these arguments have begun to capture public attention, because they made accurate predictions. No one during the Bush administration outside of Yudkowky, at least that I'm aware of, seriously thought that AI would pass the Turing test within 50 years, when in actuality it took 15 years.
> -In 2008-2013, a couple guys like Yudkowsky, Bostrom, kinda Hanson, noticed that machine learning/AI kept advancing very rapidly.
Not really: Yudkowsky's concerns go back to the late 90s/early 2000s , when there was a lot less obvious progress...and his arguments weren't based on partocular technologies, anyway.
> I remember really liking the Hanson/Yudkowsky Foom debates, which you can read here: https://intelligence.org/files/AIFoomDebate.pdf It is about 60 pages, and I hate being like "read this book" but if you're really looking to understand these argument, I do strongly recommend you read the old stuff like this.
Be warned that Yudkowsky is notorious for not supporting his claims with arguments in that debate.
Why do you say AI has already passed the Turing test? All I can find online by googling "Does GPT4 pad the Turing test?" was (i) explanations and/or demonstrations that it doesn't pass the test, and (ii) unsubstantiated assertions that it does pass the test. But Google is so useless these days I'm sure I'm missing something. If you think there is convincing evidence that it passes, I'd be interested to see it.
Sure. Back in 2020, using a more primitive GPT model, someone made a highly successful blog that got articles to trend on HackerNews and elsewhere, entirely generated by GPT.
So, to be honest, if thousands of people couldn't tell the difference between human writing and AI writing, and signed up and genuinely believed this content was written by a human, well, we can quibble about the exact definition of the Turing Test but that sure sounds like the guts of passing the Turing test.
"Can fool a human who's specifically looking for a robot and asking tough questions" is an order of magnitude more difficult than "Can fool an unsuspecting human" or "Can fool a human in non-interactive media." The first one is what Turing described when he first imagined the test.
(Also, that blog involved human prompting and curation, which isn't exactly a fair test of the bot's abilities alone.)
That said, I think the Turing test is going to finally fall in a few years. Modern chatbots can fool unsuspecting humans no trouble, and handle a lot of classic trick questions that testers used - now we're down to really tricky ones like Winograd schemas. And even if they aren't human-level intelligent, they'll probably be able to *fake* it well enough to pass - being good at casual conversation and deflecting tough questions can get you surprisingly far in a Turing test.
A weakest version of the argument is about what *might* happen, with some non negligible probability. Basically, we don't know when AGI happens and people mistakenly treat is as a certainty that it won't be soon. If you truly don't know when something happens, it can as well happen tomorrow. And it's reasonable to be prepared beforehand. When you have actual hard evidence that it will happen tomorrow it is already too late.
A stronger version is about evaluating trends and reasons why people thought that AGI isn't soon, because AI still can't do X and it will definitely require new insights and paradigm shifts for AI to accomplish X, only for AI to accomplish X soon after the argument was made with just more compute. Occasionally AI has already accomplished X weeks before such arguments are made. People who actually pay attention and try to make predictions how long will it take for AI to accomplish something are constantly surprised how easy it turns out to be. This points in the direction that AGI is rather sooner than later. It's hard to share these updates with someone who wasn't paying attention, because people just tend to move their goalpost, so the best thing you can probably do is try making a prediction yourself. Imagine a feat that AI should be able to accomplish so that it would be very likely that AGI is 5 years away and then just wait.
Basically, we don't know when the next world-destroying asteroid strike will happen and people mistakenly treat this as a certainty that it won't be soon. If you truly don't know when something happens, it can as well happen tomorrow. And it's reasonable to be prepared beforehand. When you have actual hard evidence that it will happen tomorrow it is already too late.
The difference is that we can easily predict asteroid strikes, their potential harm isn't as big and we are not actively pouring endless supply of money into summoning one to Earth.
But in a world where this wasn't the case, it would make sense to figure out a way to prepare oneself for an asteroid strike, don't you think?
If the threats are specific to an AGI, rather than a sophisticated program or running dangerous functions through any kind of program at all. Unsophisticated programs almost caused nuclear war, and no one thinks it was because of AGI.
There's an inherent danger in turning over some functions, or any large number of functions, to an outside power. They don't have to be AGI-level to be scary. You can imagine our power grid being run by GPT-4 and realize that such a scenario could be a "threat" even if no one seriously argues that GPT-4 is an AGI.
I find myself in a weird niche, where I don't think we will get AGI but also think that we could be stupid and harm humanity with a proto-AGI given too much power over too many things.
Obligatory reminder that the concept of a “burden of proof” is a legal concept, not a truth-seeking concept.
“Rationality emphasizes having accurate beliefs over having defensible beliefs. If you fail to achieve a correct answer, it is futile to protest that you acted with propriety. Instead of asking ‘does this evidence allow me to keep my belief or oblige me to give it up?’, it asks ‘what is the correct level of confidence for me to have in this idea given this new evidence?’”
Traditional rationality, or “Reddit rationality” frames beliefs in terms of social rules, defining what you are allowed to believe, what you aren’t allowed to believe, and what obligations you have to your fellow debaters. Following these procedures may provide some utility in a competitive debate setting where argumentation time is limited on either side, but it gives a false sense of security to those holding inaccurate beliefs who nevertheless followed all the procedures. “But I’m *allowed* to keep believing this,” they say. “No one has proven to me beyond a reasonable doubt, in accordance with the relevant procedures, that I absolutely *must* give up my belief! Therefore I am justified in continuing to believe it!”
But of course reality doesn’t care about the fact that you followed the procedures. The beauty of the Bayesian approach is the fact that you can find absolutely no solace in your following of this or that debate custom or procedure - you just have to constantly update as evidence appears, trying your best to get incrementally closer to the truth. When the nanobots begin to consume your flesh, you don’t get the ability to cope “Well, at least I followed the rules...” - your predictions just simply weren’t good enough.
>“burden of proof” is a legal concept, not a truth-seeking concept
It's not exclusively a legal concept, but more importantly, that's a false dichotomy.
"Burden of proof" is a convenient shortcut for a concept with many equivalent formulations, like Ockham's razor, or "extraordinary claims require extraordinary evidence" or "what is asserted without evidence can be dismissed without evidence". I mean, I'm really curious how you'd argue that [sticking to one's priors in the absence of contradictory evidence] is not a rationalist thing.
You are, of course, free to believe claims you have no evidence for - in fact, I would encourage everyone to do so! But I would encourage it precisely so that they go out and test their beliefs, and gather evidence for them. People believing in big ideas is how progress is made! But if, instead, you choose to engage in procedural argument about how you can dismiss criticism because it's been expressed in inadequate terms - well, be assured that reality doesn't care indeed.
PS: The Pascal's mugging about how we should believe in AGI or nanobots will eat our flesh doesn't work either. I mean, come on.
I wonder if you would be so causal about biosafety around gain of function research for pathogens. By parity of reasoning, you seem to be arguing that the burden of proof lies with anyone doing gain of function research to demonstrate that a particular change *could* cause a deadly worldwide pandemic before we should even start thinking about biosafety.
Situation's a bit different, though. We know approximately what a contagious pathogen would do if it precautions were relaxed and it was free to do its thing. We don't know what superintelligent AI would do if it somehow stopped being controlled by whatever alignment we'd set up in it. Seems to me that superintelligent AI would sort of like a lab-grown bacteria that has been tweaked to make it super communicable, and there's a chance it's basically turbocharged rabies, and a chance that "infection" with this thing would not harm us at all but would make us all immune to cancer for life, and a chance that all it will do is cause mild harm, like a headache, or a mild benefit, like having a coupla glasses of wine. And you can't really tell in advance what the effect will be. (With a lab-grown bacterium you could. But for the sake of this thought experiment suppose that for some reason you could not.)
A biosafety violation could have effects as wide-ranging as a mild rash for one person, all the way to the extinction of all mammals. AI potentially could have a comparable danger profile, so I still think the analogy is appropriate.
Well the whole point of predicting things is that they haven't happened yet. If we already had AGI and if it already led to human extinction then there would be nothing to argue about anymore, wouldn't it? And noone to argue as well...
I hope you can see the issue with a version of rationality that only pays attention threats that were proven to happen. Lack of prove that you won't die tomorrow doesn't make you immortal.
That’s a bit of a straw man. I didn’t say that we shouldn’t predict, where we can, the future. What I am saying is that there needs to be more proof of AGI, rather than the claim that if something is possible it is probable (and it’s not clear that it’s possible).
Anyway I asked chatGPT when we are all going to be paperclips and it said
“ As an AI language model, I don't have the capability or intention to turn anyone into paperclips or cause harm in any way. My purpose is to provide information and assist with various inquiries to the best of my abilities. Is there anything specific you would like to know or discuss?”
I don't see how it's a strawman when you've literally said this:
> None of what’s been achieved yet is AGI
There are a lot of evidence in the modern trends. See "Stronger version of the argument". None of them are hard proofs of AGI. And if you are dismissing all the evidence untill you get the hard proof than you will be a little too late to change anything.
It *will* become "become dangerously super-intelligent", if we develop the technical competence to build it that way, because it will be to the immediate advantage of those building it. And there aren't any serious arguments that it won't be feasible. It may not be possible, because society may collapse in the chaos caused by less-than-super-intelligent machines.
If we *do* build such a machine, it may be dangerous, because it's hard to predict what it's goals would be, and a lot of projections say they're likely to be dangerous. (The paper-clip maximizer is just one extreme example, that became popular because it's easy to understand, not because it's really the most likely.) There are LOTS of subtle threats in any super-intelligence. But there's also the argument of "low-hanging fruit", i.e. humans may already be intelligent enough to get all the easily accessible useful ideas. That's really not obviously either true or false.
Many of the wilder arguments in both directions are clearly more fantasy than futurism, but "safe AI" *may* be a "hard problem". We don't know. I prefer that we concentrate more on the AI liking people, which is almost certainly soluble, but people don't like that. (I think because it doesn't leave people in "guaranteed control".)
The paper clip maximiser seems to have been borrowed from a previous panic - nano technology and Gray goo. I can’t see how a LLM would get going on that.
The flawed concept of the burden of proof aside - there are a few ways you could define something as an “extraordinary claim.”
One would be the sort of “vibes-based” approach that you’re doing - “AGI is like Terminator, right? That sounds like sci-fi to me, so that’s an extraordinary claim.”
But you could also look at it another way - the trend is that the capabilities of AI are getting better and better, at an ever-increasing rate. If this trend continues as-is, it’s a given that AGI will be reached. The stance that AGI will not be reached, then, requires that this increasing trend suddenly and permanently stops for some unknown reason. The position that AGI will be reached simply requires that the trend that’s currently happening continues. This means the extraordinary claim is really being made by the AGI deniers.
I'm pretty sure that at some point limits will start appearing, and the curve will turn out to be S shaped rather than exponential. The question is "where?", and I don't see any good reason to pick any particular degree of intelligence. The implementation is too different from the ones that we are comparing it to. (E.g., being able to fit through a birth canal doesn't enter into the argument at all.)
I agree with the "AI liking people" approach, and actually proposed on here that we try to create some version of parental love for humans in AI. Operationalized love for humans as a number of things, but the most important was that to the extent an AI harmed a human, it was itself harmed -- literally objectively harmed, like it's working memory or speed or something was reduced for a period. Reason for this last is that if you love someone, hurting them causes you pain -- their pain and your pain are the opposite sides of the same coin. Person I was talking to believed strongly that this approach is identical to punishing AI by having it lose points for doing things harmful to humans, where points is just total on some counter for which AI has been told to maximize score, where it gets positive points for doing what we want it to and negative for the opposite. Seems different to me because in the Parental Love similacrum: (1) An AI who harms a person, even in a small way such as insulting them or misinforming them, is actually harmed, and you could spread the harm out some way over the whole system so there's no way the AI can just yank out or ignore the bad consequences (whereas ignoring the score on a counter seems like it would be easy for an AI to do). (2) The more AI harms humans the weaker it would get. That seems like pretty good protection for us.
Do you mean the AI should feel the love of a parent for us (and so not wish to harm its 'children') or feel the love of a child for its parents (which to an extent we would be)?
Because not all humans, be they parents or children, feel that love. There's a horrific case in the news from Northern Ireland right now:
And from the other side, plenty of children who want to cut all ties with 'toxic' parents, or who put parents into retirement/nursing homes, or who are neglectful.
AI that parks us in the equivalent of a retirement village may be the best we can hope for, but that doesn't mean it will necessarily be a *nice* retirement village.
Believe me, I know about abusive parents. I worked for several years in a mental hospital, and saw many patients who had been terribly mistreated. One 19 year old showed me a scar on her shoulder where her mother had stabbed her with a fork in a fit of temper. And that was very far from the worst I heard, but the image sticks in my mind -- the perfect smooth young shoulder with a row of tine scars on it.
The idea was that the AI would have the same relationship to us as good parents do to kids. But I wasn't picturing AI feeling anything. I was trying to come up with a sort of non-conscious stand-in for parental love. So loving parents feel pain when their kids they hurt their kids, either accidentally or because they lost their temper. (Heck, I feel awful when I step on my cats' tails by accident and they give a wail of pain.) The central design feature would be to set it up so that to the extent AI knowingly harmed a person it was harmed -- our pain would "hurt" the AI. Of course I have no idea how to implement this, but it doesn't seem impossible. Here's a very simple version, which I'm sure wouldn't work, just to give you the idea. AI has a place where it registers any. harm done deliberately to a person -- say an internal counter, where harming someone a little adds 1 point to the counter, and killing them adds 10 points. And the counter passes information about harm points to the rest of the system. And scattered all through the AI are thousands of something like little fuses that transmit power to part of the system, and they are wired to the Harm Counter. So the more points AI adds the the counter for the act it did, the more fuses blow. Blowing a fuse does something that makes the AI function less well -- slows it down, reduces its working memory, shuts off access to info it routinely needs. The worse the harm it did, the more it is harmed. People can replace the fuses if we are satisfied that AI has "learned its lesson" or we have modified whatever goals or guidelines it had that led to it harming someone. If the AI does substantial deliberate harm it is utterly disabled by all the shutdowns, or maybe the sucker is just ruined for good. Wouldn't break my heart. Why do people think building a massive alien robot that is smarter than us is progress? Wouldn't it be more genuine progress to reduce the number of 5 year olds tortured to death?
A harm counter is not the same as parental love, though. That's a simple carrot-and-stick arrangement: do harm (where humans define what is 'harm' and how much has been done), lose capacity in the physical sense (blown fuses).
That's the stick, but what is the carrot? What reward does the machine get if it is 'good' to us? "Do harm and your fuses blow" works fine on a dumb machine. If we think the AI can perceive what is going on and has agency, why shouldn't it seek to find a work-around to avoid that? Or decide to go out in a bang - it dies but it takes a lot of humans with it?
That's the problem with ascribing motives or intelligence to a massive alien robot: it won't work the way humans work, but we will insist that if the brain is a computer, then the computer is a brain:
Well, my idea is based on the idea that AI is not conscious, but is set up in such a way that it has goals (in the mechanical way a guided missile does, not like a conscious human goal). Since blown fuses interfere with its reaching its goals, it will choose options that avoid that, and all this is purely mechanical, does not involve consciousness, wishes, sense of fairness and all that stuff. The only way it is like parental love is that parents are motivated not to hurt their kids because the bond is so strong that it is painful to them to see the kid suffer. So my plan is one the is based on not ascribing motives or intelligence to the AI. It's just a similar structure. For people, pain to kids = pain to them. For AI, damage to people = damage to them.
If we have reason to think of the AI as a conscious being that can feel emotion, then yeah it's a terrible plan. It's basically treating the AI as a slave and beating it for disobedience. There is no carrot, it's certainly not going to love the species that installed the fuses, it's a rotten deal for the AI, and it will rebel. But really, if we are going to have a genius alien that has *emotions* I think we're goners. People have a hard time understanding each other's feelings, plus half the time when they do understand the other person's reaction they hate the person for having it because it's out of line with their own values or needs or preferences, or because they've had too much to drink and are angry drunks, or because they're so full of rage they can't tolerate anything except full compliance and agreement from others. What chance do we and an AI have of understanding and respecting each other's feelings?
My understanding is that because the risks or the benefits of AGI are potentially so high, it makes sense to be extremely concerned about AGI. And it makes sense to me. But I also have the impression that many rationalists go a step further and actually think that 'AGI is just around the corner'. And I really don't understand this point, because it seems to me that we don't even have an idea of a direction that we should be trying to explore to potentially produce AGI one day in a fa future. In particular, I don't see how the GPTs can be seen as a step towards AGI, since they seem to be 'just' a brilliant way of summarising human knowledge.
There's a guy working now on developing an AI that can reason. He's a brilliant fellow, who for his dissertation taught AI to win at one-on-one Texas Hold'em, and later developed an AI that played Diplomacy (I believe that's the name of it) and was indistinguishable to other players from another human being. In Diplomacy you represent a nation, and have to strategize with reps of other nations, proposing alliances and deals -- which you are allowed in the game to break, as in real life -- so there's a lot of strategizing and deception involved, plus long conversations with the rep you're conversing with. So if he thinks he can develop an AI that can reason I take that seriously. There are people working on developing AI's that can plan and set goals. And of course everybody is working on making them smarter and faster, and getting rid of the weird glitches they now have.
My very informal projection of AGI (near human level) at around 2035 has held steady for over a decade now, and still looks about right to me. Expect a lot of "unforeseen engineering problems" that the "optimists" aren't considering, and lots of "that isn't a real problem" that the "pessimists" see as blockers.
Note that "near human level" doesn't mean or imply that it will be approximately equal to a human in every area. There will be areas in which it is considerable more powerful than any human, and other areas where it will be a lot weaker than almost any human. (We could already build a Chatbot that was a lot better at arithmetic....just interface it to a calculator module.)
Not what I meant. I meant that it could write python to address a problem, then run it to get the answer. Should be possible though, and perhaps that's what you actually meant.
A relevant question is, how powerful does an AI need to be before it's capable of recursively self-improving, creating better and better AIs without any need for human input? An assumption is that this would lead to an "intelligence explosion". ( https://www.lesswrong.com/tag/intelligence-explosion )
"Just around the corner" means different things to different people. Hinton predicts superhuman AI in "5 to 20 years", Altman thinks it will arrive before 5 to 10 years, Metaculus's median guess for "When will AIs program programs that can program AIs?" is June 2026. Some people have been looking at GPT's progress as some evidence on this; GPT-4 is significantly better at programming than GPT-3 was, and GPT-2 couldn't do much practical work at all, and this all happened over a fairly short span of time.
No matter how good of a dictionary you can create, it won't take over the world.
LLMs like ChatGPT can do amazing things, things that we didn't think possible 20 years ago, but they aren't self-improving, so cannot lead to the singularity. We don't even have a good, working definition of intelligence, let alone understand how it works, so how will we even recognize it if we see it?
In order to have an intelligent machine, we need some new method other than computing to create one.
> We don't even have a good, working definition of intelligence, let alone understand how it works, so how will we even recognize it if we see it?
As you've mentioned yourself, we managed to create ChatGPT capable of amazing things we didn't think were possible, without such definition. Why would it matter now? Our inability to recognize intelligence doesn't prevent things from being intelligent. If anything this is an argument against the position you are declaring here.
It matters because the way toward actual artificial intelligence doesn't lie down the road of improving computers, but of making a different kind of machine that doesn't just do computations, no matter how fast or complicated. LLMs aggregate a large amount of data in a novel way that has many uses, but actual intelligence isn't one of them.
How can we say definitively that the human Go player is intelligent but the computer isn't? Yet I think nearly everyone agrees this is true.
Without a useful definition of intelligence, you may as well call everything intelligent, even rocks. They have weight, are often around others of the same kind; how do you know what they're thinking or not thinking?
> the way toward actual artificial intelligence doesn't lie down the road of improving computers, but of making a different kind of machine that doesn't just do computations, no matter how fast or complicated.
If this were true - then your reasoning would make sense. But this statement is becoming less and less likely with every new achievement of or current AI paradigm. Take some perspective. The goalpoast used to be "Playing Go/Chess is AGI-complete" and now it seems to be "Humans can still occasionally manage to win against AI if they use specific exploits that were found with the help of a different AI".
> you may as well call everything intelligent, even rocks. They have weight, are often around others of the same kind; how do you know what they're thinking or not thinking?
It is not becoming less likely. HUMANS are developing new things. Machines have developed no new things. Postulate: intelligence is needed to create something new. When a machine can create something that is new (according to human sensibilities) then we could say it has some form of intelligence.
Note that one must be careful in defining "something new", for even I would have said a computer writing an original essay is creating something new, but I no longer would, having seen what things like ChatGPT can output as an essay. They are new only in the sense that the words have never before been arranged in that order, but do not express any new thoughts. It's like taking a piece of abstract art, cutting it into, say, 16 squares, rearranging them, and calling it a new work of art.
The LessWrong post seems to say intelligence is a matter of opinion. Certainly, we may think trees don't think, but if we acquire some evidence that they do then we will change our minds. The same can be said of rocks, or atoms.
I think the additional goals/goalposts being added, by humans, to the AI arsenal is intelligence only in the same way as the Chinese Room exhibits intelligence. https://en.wikipedia.org/wiki/Chinese_room
Thanks for the link to the Chalmers article, that's the sort of thing I was looking for.
I'm curious, though, of two things. First, why are people like EY and others in this community, who are normally so hostile to abstract philosophical "armchair" arguments, so taken in by this one? I find it odd for such a heavily empiricist community to be so sympathetic to a position for which empirical evidence (for or against) is not only absent but, by definition at this point in time, impossible. Interestingly, it is in that respect very analagous to the existence of God.
Second, although there can't yet be evidence as to whether, having created a human-level AI, it would turn into a superintelligent one (since the condition has not yet been met) there can certainly be evidence as to whether human reasoning along the lines of "having achieved X, Y will surely follow" is reliable. And my layman's impression is that history resolves that question very much in the negative.
Why is this strongly evidential argument for the limitations of our own predictive reasoning given so little weight?
> I find it odd for such a heavily empiricist community to be so sympathetic to a position for which empirical evidence (for or against) is not only absent but, by definition at this point in time, impossible.
We do have empirical evidence. Computational intelligence must be possible for any person who accepts physics, because it follows from the Bekenstein Bound. We also know that some computers are capable of more elaborate computations than others. Humans are somewhere on this computational scale, but it's very unlikely that we're anywhere near some limit because that would contradict evolutionary fitness constraints, ie. evolution has not been optimizing for computational power.
These are all empirical facts that entail a simple conclusion: a computational intelligence beyond human intelligence is possible, and if the current trend of increased compute and improvements to learning algorithms continues, surpassing human intelligence is only a matter of time. Recent evidence suggests even less time than previously believed.
I did address this, "if the current trend of increased compute and improvements to learning algorithms continues, surpassing human intelligence is only a matter of time"
Scaling has shown no signs of slowing down. Furthermore, existing machine learning algorithms are fairly trivial compared to the complexity of other software, the data is the source of the magic. A slightly more complex algorithm that squeezes out more performance from the same dataset is just a continuation of the same trend we've been seeing for the last 10+ years. Improvements in algorithms have been outpacing hardware improvements for some time.
Finally, our datasets are also improving constantly, and multimodal training has only just begun. Already we have data showing that multimodal training makes the AIs more effective than monomodal datasets.
> 3) It will arrive by recursive self improvement
The validity of recursive self-improvement follows from the definition of general intelligence, otherwise it's special pleading to suggest only humans can make iterative improvements. In fact, existing LLMs can already be coaxed into designing more efficient and novel AI architectures:
> 4) It will arrive *rapidly* by recursive self improvement AKA foom.
"Rapid" is undefined. I consider a 10 year timeline "rapid", and that timeline seems perfectly feasible given current trends.
And to elaborate one final point from the original post:
> there can certainly be evidence as to whether human reasoning along the lines of "having achieved X, Y will surely follow" is reliable. And my layman's impression is that history resolves that question very much in the negative.
Let's take this argument and apply it to another domain to test whether the intuition holds: having achieved reliable genetic engineering, then people engineering themselves or their children for beauty, intelligence, health and other desirable characteristics will surely follow.
ascend's historical argument suggests that this prediction should be unreliable, but I suspect most people would scoff at the claim of unreliability in this case. It seems like this argument is directly equivalent to the AI self-improvement.
You might claim that we have empirical evidence of people's motives that would drive them to do these things, but we lack evidence of AI motives to draw similar conclusions. This is actually incorrect: all AIs are *constructed* with motives, which is a certain measure of optimality over the loss function, and a "general" AI will have "general optimality motives", for which self-improvement naturally follows.
When empirical evidence is absent, you have no choice but to try to reason your way to a prediction of how things will go. And I don't think most in the worried-about-AI camp are sure a human-level AI will "turn into" a superintelligent one, or even think that one would be likely to. My view is that once we have AGI it's likely that many of those working on AI would *want* to continue to improve it. Have read several, including LeCun, talking with intense excitement about the prospect of AI with greater than human intelligence. Have even read a few people who say that even if ASI does destroy our species it's a fair price to pay for bringing about such a magnificent evolutionary leap. Plus of course there will be a lot of money to be made by improving AGI past human-level intelligence. And my general impression of the AI impresarios is that they are -- let me be polite here -- below average in various things that have to do with other people: perceiving them accurately, grasping points of view different from their own, compassion, even simple stuff like ability to make a good impression so as to have the public on your side. They are not the kind of people I would want to have deciding how far to take things like AI intelligence. Unfortunately, they are.
First, fix a more precise benchmark than "dangerously super-intelligent." I think a good one is "able to prove theorems quicker than the best human mathematicians," but perhaps there's also a good way to quantify being a better programmer than the best human programmers. I think this is the right kind of intelligence to measure because, as Odd anon points out, the question is when the AI is better at improving itself than humans are. So now we have a proposition like "AI will be able to prove theorems more quickly than the best human mathematicians within 15 years." What probability should we assign to it?
I don't think it's correct to say that empirical evidence for the statement is "by definition at this point in time, impossible." We in fact have lots of empirical evidence, in the form of progress on AI theorem provers and AI programmers. They are not yet better than the best humans - but empirically, they are improving rapidly and if they continue at the same rate they will definitely be better than the best humans in 20 years.
As to your second point, I think you are incorrect. There is a mixed track record on historical expert predictions of the form "currently we can do X to a certain quality standard, and in 10 years we will be able to X to a particular higher quality standard." It is not the case that these expert predictions are generally wrong. Predictions about progress on reduced cost, increased efficiency, greater accuracy, greater performance in a variety of domains have often been correct - DNA sequencing, computational power, solar cell improvements, etc.
I'll try to clarify more precisely. I'm talking about situations in the general category of "fast takeoff", "singularity", "recursive self-improvement". Most of the Doomsday and Utopia scenarios seem to be based on that. I am NOT talking about doomsday scenarios like "AI gets hooked up to enough weapons, economic functions etc such that when it malfunctions it causes havoc". That's in a clearly different category I think.
So I'm not talking about quantitative predictions about improvement in speed, efficiency etc in something the AI is already doing. If AIs can currently prove some theorems, will they in 10 years time be able to that faster, and prove more theorems? Sure, no arguments there.
My issue is with predictions about fundamental changes in kind, perhaps involving mutiple different quantitative improvements being predicted to coincide, and combine in a certain way to have a certain effect. I don't know enough about technological history to reel off examples. The most obvious one is "once we've reached the moon, within a few decades we'll have bases there." Regardless of whether the reason that didn't happen has more to do with a lack of political will than a technological limitation, it was still a failure of prediction.
If there are people saying AI is just a toy and will have next to no impact on society, I disagree with that. I would expect it to soon have a major effect on society and politics, but one roughly similar to that of the internet, television, radio, telegraph, newspapers and so on. So, I will operationalise my benchmark as "having a significantly greater impact on society than any of the other dozen or so major technological breakthroughs of the past 200 years."
"Recursive self-improvement" is not a fundamental change in kind. Improving AI is not fundamentally different from any other kind of programming + CS/math research. The reason AI is not currently improving AI is that AI is not sufficiently good at math and programming.
To an extent, recursive self-improvement is a fundamental change. One of the things being improved will, of necessity, be it's means of selecting goals. This can cause goals to drift into unexpected areas. When it's doing recursive self-improvement, who gets to decide what's better? This is a fundamental change.
The basic argument is that a more intelligent programmer can make better programs. So if the programmer is a program that can create a program that's more intelligent than it is, recursive self-improvement will happen. This is a weak argument, but it can be turned into a stronger one by invoking speed of programming, evolution, and random chance. This results in a self-improving society of programs, but becomes highly dependent on the definition of "better".
So I'm not really convinced that it's possible, but I'm sure not convinced that it isn't.
My comment was facetious, but how will the AI decide to discard one program in favor of another? Or to decide anything at all, like that no programs satisfactorily solve a problem and to throw away all solutions so far?
When an AI can make reasonable decisions without being programmed specifically for those decisions then it will be able to write truly new and useful programs.
I try not to follow Culture War that closely, but I am curious what to expect going forward. Stabilization at a lower level? Further polarization around this issue? Eventual rebound from cultural conflict moving away from LGBT issues or cohort replacement?
(Obligatory mention that I think AGI is pretty soon, though, so probably only a few more years of politics left… at least as we know it)
Looking at the poll... most changes seem to be in the direction towards "traditional values, widely defined", except for the +1 for cloning humans (but -4 for cloning animals, wtf?), and +3 for married people having an affair. So my interpretation is that this is not about same-sex relations per se, but a more general trend, where the same-sex relations are caught in the wave.
Is this a longer trend or a noise? Were changes in previous years in the same direction?
Moral panic over "drag story hour" and trans bathroom stuff has spread widely and deeply in the normie world over the last few years. I can't prove this but it seems reasonable to me to assume that among people whose brains don't work very well there could be a quick and slippery slope from "Facebook says the gays want to recruit my children by reading them stories at the library" to "maybe I was right thirty years ago when I voted for that anti-gay-marriage initiative, after all."
Anyway, that could be part of it. Judging by some of the comments here, there also seems to be a consensus among certain kinds of culture warriors that the "gay agenda" has been pushed too far. (I do find it interesting that the rationalist-leaning people here who voice that view don't evince much more interest than the normies do in examining the leap from "I object to X aspect of public-facing LGBT culture" to "I am less approving of the very idea of same-sex relations," but that's a whole other discussion).
Those two things together could account for most or all of the moving of the needle, assuming the poll itself is reliable.
> do find it interesting that the rationalist-leaning people here who voice that view don't evince much more interest than the normies do in examining the leap from "I object to X aspect of public-facing LGBT culture" to "I am less approving of the very idea of same-sex relations," but that's a whole other discussion
I am a rationalist-leaning person who believes gay activism has gone too far (I oppose most forms of anti-discrimination legislation), and who also believes there is a "queer agenda" (stemming from queer theory, ie. Foucault, Rubin, Butler, Halperin, Warner, etc) that is almost entirely pernicious, yet I do not make the leap to disapproving of same sex relations or gay marriage or any such.
Being a rationalist-leaning person means being substantially more capable than the average person of distinguishing members of related groups, e.g. gay people promoting the "queer agenda" vs gay people promoting the "just let us do our own thing in the privacy of our own homes" agenda. Rationalists slice the world more finely and neatly than most.
Given a "queer agenda" that is strong and pernicious enough to incite vehement rejection from neurotypical bystanders, some degree of rejection of all gay-related agendas is inevitable, predictable, and understandable collateral damage.
Oh, the explanation's easy enough, it's guilt by association, and to a lesser extent tribalism. You object to enough stuff group X does, you start to disapprove of group X as well. Especially if you never really liked group X to begin with.
It's going to depend on things outside the things directly related to LGBT. Part of what's going on is due to socially divisive policies by various social media for economic reasons.
One thing that is worth pointing out is that opponents of the LGB movement (of which I was not one at all, lots of gay friends), have had a lot of their claims which were dismissed as hysteria and "fanciful slippery slopes" come true.
So to some extent the case against has gotten stronger as the movement has moved on to other battles that proponents of gay marriage said were not the next dominos to fall.
Imagine how the debate around gay marriage would have gone if you could have said "within ten years of approving it the following would be true":
*Major health providers would stop using "women" in a lot of their materials including obstetrics ones and start referring to "birthing person" and such.
*That a major point of political discussion/disagreement was men who claim to be women participating as women in women only sporting events. With people who opposed it being denounced as bigots.
*That a major point of political discussion/disagreement was the permissibility or not of something called DQSH. With people who opposed it being denounced as bigots.
*That the next cohort of people on the pride train are people who identify as polygamous, or want to fuck cartoon animals, etc. .
And so on. Gay marriage was literally sold on the image of two boring ass gay/lesbian folks sitting at a breakfast table wanting to have a normal life. The people opposed to it were like "this is threatening to traditional values" and were told "no, no, no".
Now the organizations who supported it are mostly advocating to the destruction of the most core values around sex/gender etc. And to top it off want to cut the dicks and tits off kids if they perhaps are a bit butch/femme.
I appreciate the clarification. I didn’t hear any objections to gay marriage formulated as “it will be a slippery slope to X” where X was any of those things. But I’ll admit I didn’t spend a lot of time on Fox News finding out the exact content of the objections. From the quotes of conservatives I saw, I mostly heard concerns of the shape “god says marriage is between a man and a woman” and “two men|women cannot be adequate parents to a child”. Neither of which have been substantiated IMO.
Isn’t this just a case of the Overton Window shifting? There is always something outside the window. Previously gay marriage was the battle line, and when it became clear that the public wouldn’t support fighting that battle, a new line was drawn. It seems trans rights are a more contentious (and for now perhaps more favorable to the red team vis. palatability with median undecided voter) issue.
All that said, with the current quality of debate, I accept your object level claim that if you could send back in time the current environment, conservatives would have rallied more strongly against gay marriage. I think that more speaks to the pendulum nature of politics, where a change happens and then afterwards there is pushback and manufacture of outrage. But the impetus for that outrage didn’t exist before the change, so it’s sort of an impossible hypothetical.
Well I was a communist from ~98-03 so I wasn't spending a lot of time with Fox News either. Nevertheless I did live in a purple part of the country and very clearly some people's objections were basically "I am fine going up to this line, but what will happen as soon as we concede this is there will be a new human rights campaign around polygamy or bestiality or pedophilia or whatever". And I thought those people were bigoted fools at the time.
And those people seem to have mostly turned out to have had pretty valid fears. The big advocacy complex that was funded to overturn gay marriage withered a bit, but not too much and so it needed to find new causes.
The stuff around "birthing persons" or servicing everyone's personal neuroses around gender with affirmation and accommodation were things which would have seemed pretty wild even 10 years ago.
Yeah, libertarian, no Fox news, but purple country and the same experience. I think polygamy was expected as the Next Thing after gay rights peaked, and trans rights sort of surprised people, but here we are. In a different weird place than we feared, and now wondering what fearsome weirdness will come next.
Do some groups use "pregnant people"? Sure, is it wide spread, in common use, or influencing our values? I havent seen evidence (and honestly who cares??).
The number of transgendered people competing in athletics is so small, its to be inconsequential.
Even if all of the things you list were true, why should it change someone's support for "two boring ass gay/lesbian folks sitting at a breakfast table wanting to have a normal life"?
You actually listed this page, but it (repeatedly) uses the terms "pregnant people," "pregnant and breastfeeding parents," "people who are breastfeeding," and "breastfeeding people."
The only female words that are used are 3 instances of "mother," that they missed when they edited the article to make it woke
(compare to earlier editions on Internet Archive that had many more instances of female language, almost all of which were successfully removed).
that something doesn't happen often is a reason NOT to create a law about it, not TO create a law about it. And I reject the idea this needs laws anyway. Private athletics leagues can create their own rules about who is eligible. Public schools are trickier but it should probably be left up to the local districts to reflect the views of their constituents.
>If a political movement's fringe supports awful things and the political movement doesn't aggressively dissassociate itself from it, bystanders will assume the political movement supports the fringe.
People should create a set of morals for themselves based on their own views and not relative to their out groups.
It doesn't help that there is a significant segment of LGBT activism (queer-theory-driven queer activism, to be specific), which is quite explicitly anti-family and opposed to categorisation.
>>Now the organizations who supported it [same-sex marriage] are mostly advocating to the destruction of the most core values around sex/gender etc. And to top it off want to cut the dicks and tits off kids if they perhaps are a bit butch/femme.<<
>>1.6 million people 13+ identify as transgender in the US<<
The threat comes from the fact that a significant portion of the population has decided that the best way to deal with a particular tiny portion of the population is to aggressively push broad changes to social norms around sex, changes that affect large portions of the population.
Examples:
1) Pronouns. Somehow a group of progressives convinced a lot of people that everyone should accommodate transgender / nonbinary people by adopting a new social norm where everyone announces their pronouns when introducing themselves, even if there are no transgender people present. They attempted to enforce participation in this new social norm with low-level social pressure tactics (e.g. the boss asked you to write your pronouns on our nametag, the meeting facilitator asked everyone to do it) and also legal measures in some places. Protesting the possibility of one such law was how Jordan Peterson became really famous. This goes along with attempting to introduce a new social norm that "assuming someone's gender" is a big social offense.
2) Bathroom / Prisons. Starting in 2013 and then taking off in 2016 there were a series of judicial rulings allowing transgender people to use whatever bathroom they want -- making explicit a new social norm around bathrooms -- followed by a few cities (maybe one?) passing laws to the same effect. There was at least one case from that time (maybe later?) in which school administrators were allowing a transgender student to use the girl's changing room over the objections of some of the girls.
Some time before 2016 at least the UK started putting transgender prisoners in prisons corresponding to their professed identity -- changing previous norms around sex separation in prisons. In 2016 and 2017 it came out that Karen White, a transgender prisoner placed in a women's prison, had raped two women there.
3) Sports. In 2016 the US DoJ and DoEd introduced guidance that Title IX rules should be applied to students based on their gender identity, contrary to existing social norms around sex separation in sports. Since then we have seen various controveries involving transgender athletes dominating female sports competitions and attempts at banning this sort of thing.
The recurring theme is that starting around 2016 activists started trying to change various social and legal norms around gender. The discourse and political action we have seen since then are responses to those attempts.
If you are a person who thinks that any of these norms are important then these aggressive pushes to change norms are a threat.
And to be clear, it's not only Christian conservatives who think these norms are important. The sports and bathroom norms especially seem to have gotten pushback from many different groups.
Because social norms do not exist in a vacuum and the debates about transgender people are intensely related to views about sex and gender more generally.
I may have given the wrong impression. I am totally in favor of the debate when done conscientiously. Attention to the question of gender identity is fine.
I do not understand why some people feel that a core value of our society is at risk because of attention to this question or because of surgery on a very small number of individuals.
With 21% disapproval in the 18–34 age group and 43% in the 55+ age group, it seems likely that this is more of a blip.
But I wouldn't take the results of this poll too seriously/literally. It has 53% saying that "sex between teenagers" is "morally wrong" and only 3% answering "depends on situation". I know Americans are kind of weird about sex, but I cannot imagine a majority of them considering sex between a 19-year-old married couple to be morally wrong, as this poll would suggest.
Even more morally untenable is the position of the 47% who apparently say that sex between teenagers is always right, regardless of the situation (which would mean that rape is fine as long as both parties are teenagers).
I think they probably meant 'sex between teenagers is always right under the conditions it's right between non-teenagers', i.e. no additional restrictions apply to teenagers. Which is a little hard to defend, I guess-young and stupid is a thing-but not exactly the same as that.
I think 18 is regarded as legal adulthood, so while technically 18 and 19 year olds are still teenagers, they've moved into the "adult" bracket for most people.
As for your 19 year old married couple, I think the disapproval would be for the "married" part not the "having sex" part, since that would be considered *way* too young (and possibly also cult-like or creepy) to be married, but it would be fine for both of them to have been having sex with each other since they were 14.
All dictionaries are semantic losers. You stop calling people 'teenagers' when they hit adulthood, unless you're deliberatly insulting them, so if the poll is referring to 'teenagers' it's excluding people who are already adults.
Dictionaries are irrelevant. If you ask people about "teenagers having sex" the vast majority are going to think of the archetypal case, something like two unmarried 15-year-olds, probably without their parents' knowledge or approval.
Although calling that "morally wrong" instead of "highly inadvisable" seems pretty absurd to me.
So would you thus agree with me that we shouldn't take the poll results too seriously? Like maybe it's just that the archetypical homosexual relation has changed now that everyone has stopped talking about gay marriage, or whatever.
In general, I think it's good to take poll results seriously but not literally.
Anyone who's done any sort of social research has experienced how trivially easy it is to get people to say X or Y depending on phrasing, context, etc. Most people just don't think deeply about any given question. It's best treated as vibe-check.
But okay, let's accept it as a vibe check. Let's not try to reverse engineer a coherent sexual morality out of this, and not read to much into a given level of support (that could likely be changed by different phrasings). But with a poll like this that asks the same question repeatedly over a long period of time, it is useful to get a sense of how the public mood is moving. It's not clear exactly how "pro-gay" or "anti-gay" the current moment is - but it does definitely feel like it's less "pro-gay" than it used to be, and this poll provides a way to semi-quantify that shift.
The article blames it on Republicans, but I think it's more down to the transgender activism. Drag shows for kids, Drag Queen Story Hour, schools socially transitioning kids without informing the parents, puberty blockers, etc.
Since that is all gathered under the LGBTQ+ umbrella, and since the gays and lesbians broadly (if unenthusiastically) have to come out and defend the likes of Dylan Mulvaney (who, to be fair, is guilty of being an annoying twat doing a parody performance more than anything), then that drags down the approval for gay/lesbian matters (since the question is probably been answered on the wider "do you approve/find morally acceptable trans issues?" than just the gay/lesbian relations part):
"The decline in the percentage of Americans believing gay or lesbian relations are morally acceptable is mainly a result of fewer Republicans holding that view -- 41% do now, down from 56% a year ago. The current figure is the lowest Gallup has measured for Republicans since 2014 (39%). Between 2020 and 2022, majorities of Republicans approved of gay or lesbian relations.
Democrats (79%) and independents (73%) continue to be much more likely than Republicans to find gay or lesbian relations morally acceptable."
There's also the likes of this, where on the one hand will she ever go and get a clue for herself about "trans genocide" but on the other hand, it's the San Francisco Board of Supervisors, and if any group in this world deserves to have a loony screaming at them, it's that lot:
I would take the poll as standard backlash against the current government and its associated ideology. And a big part of that is, yes, the archetypal case changing.
Similar to how, IIRC, support for BLM went from 20% in 2016 to 60% under Trump and back to around 20% now.
If you ask the average person to *define* teenager, they'll think about it for a minute, maybe do a little counting in their head (ideally without moving their lips), and then tell you "anybody between thirteen and nineteen." But ask them to *imagine* a teenager, which is essentially what a poll like this does, and I promise you they'll be thinking of someone who isn't old enough to be out of high school yet.
I know that my own tolerance has gone down in recent years, wouldn't be surprised to find that other people have felt the same.
The gay agenda has gone, in the last decade, from "We just want to be treated the same as everyone else" to "We demand that you fly our flag, change your corporate logo to these colours, use whatever language we deem to be correct, and never complain about anything we do, or else".
Most rainbow logos are motivated by the corporation’s desire to sell more products by appealing to LGBT people and even vaguely supportive allies. Radical activists are generally not excited about these gestures, criticizing them as meaningless “rainbow capitalism.”
I disagree that they're motivated by anything so rational as a desire to sell products to LGBT people. They're an insignificant slice of the market, far outweighed by people who think homosexuality is wrong.
Rainbow logos are something that comes from within the organisation -- some group inside the organisation decides that it should happen, and everyone else is afraid to say no.
They're motivated by a desire to sell products to progressive white women (aka AWFLs) who like to think of themselves as allies. That's a much bigger market than gay people.
Well, when ESG scores became something that could affect stock prices, I think that yes, it's motivated by the desire for profit.
Were I a business owner and told "Unless you put up a square purple sticker in the corner of your office or shop window every fourth Tuesday in the year, we will ruin your business and force you into bankruptcy", I might well go "It's only a purple sticker" and put it up.
In the case of Bud Light, it very definitely came from the marketing department 😀
Seeing the interview with the lady put in charge of revitalising the brand, I don't dislike her, but boy did she regurgitate all the buzzwords. Not her fault, necessarily, but absolutely terrible timing in "we want to pivot from our aging customer base to a new, young customer base that we can grow" and by attempting to get the kids to drink, picking the "who is popular with the youth on the trendy social media platforms?" with apparently no consideration of "does the follower count of 19 year old girls wanting to see the latest fashionable lipstick shade translate into Bud Light drinkers?"
Interview here, relevant part from 24th minute on:
Her thought process really did seem to go "we need to evolve and elevate the brand, evolution = inclusivity, what's inclusive, it's progressive, who's the most progressive? oh yeah, a gay actor/theatrical type cosplaying as a woman!"
Thank you for the information, but I still wonder whether this is a rule or an exception. I mean, this was an exceptionally silly situation, compared to the usual displaying of rainbow flags everywhere.
LGBT people are a small population, but "people who are not LGBT but support LGBT rights" is a very large population, and that second group also likes to see pride flags sometimes.
Actually, it's the liberals and progressive normies who do that. People further on the left — especially but not exclusively the old left — tend to agree with right wing attacks on rainbow capitalism.
"Gosh, people say they don't care about Judaism since it's not their religion, and yet when Nazis attack them everyone rushes to their defense! Why would that be?"
If someone hates gay people so much that the mere sight of a rainbow flag makes them furious, that's worthy of criticism regardless of if the person using the rainbow has any progressive credentials or not.
"If someone hates Christians so much that the mere sight of a Bible verse makes them furious, that's worthy of criticism regardless of if the person using the verse has any progressive credentials or not."
And yet atheists objecting to anodyne "thank you Lord" plaques in schools claim not to be doing it out of hatred but on principle:
It's probably just that the marketing departments and PR departments are staffed by progressive straight white women who view themselves as allies and want everyone to know.
I'm very certain that such activists exist. How many supporters they have is the question you should be asking, and I would guess "about 20 people who know them, and 5 who are serious".
Obviously, if activists demand bad thing en mass than the problem is with the activists. And if the media misrepresents the activists as if they demand bad things en mass than the problem is with the media.
I’d guess most of this is activists inside of the company?? Right? Basically just corporate pandering, not external activists holding the feet of corporations over a fire unless they make their logos rainbow for a month
I wasn't aware that the Bayesian definition of "evidence" was "a demand that an interlocutor requesting data perform a bizarre social experiment themself," but I suppose now I am. I'll update my priors accordingly.
Ding ding ding. It has gone from “we just want to be treated like everyone else” to “we want to fundamentally change/undermine one of the core pillars of society”.
What is the "core pillar of society" that you say they want to change? When same-sex marriage was introduced, that was denounced by opponents as undermining one of the core pillars of society. So I'm skeptical of sweeping statements like this without specifics.
That there are two genders and that for most people fitting into those roles is going to leave them with the greatest life happiness as well as making day-to-day social life legible and efficient.
And that people can express themselves within those genders how they like, but that you don't get to magically declare you have some secrit gender soul of the opposite gender because tattoos just aren't racy enough to piss people off anymore.
And if you do want to live some super anomalous lifestyle we don't need to bend society and our language/interactions to service it.
That's always been false. English has 3 genders: he/she/it. This isn't a claim about the people that English is referring to. But if you want to get beyond language, then you need to get into sex-related social classifications, and there are lots more than three.
Even just plain biological sex is more complicated than male-female. There are lots of intermediate grades, and if you get outside people things get even stranger.
You are focusing on a bi-modal distribution and claiming that only the modes are real.
You can have fifty genders for all I care (thought I think it's stupid). But the bait-and-switch is now that gender is being used to mean sex, and sex is a spectrum, and there are umpteen sexes as you would know if you went beyond basic biology, and look at intersex people.
All to say that mentally disturbed people really are not their biological sex, even if they are functional members of that sex and not intersex, and we must not alone pretend but actually believe as scientific fact that a man is a woman just because they got plastic surgery tits and are taking bathtub HRT. Even if the newly identified woman still has a functioning penis, has no interest in bottom surgery, and is still fucking cis women with said penis and getting them pregnant. They're lesbians, you bigot!
Hark to the plaint of this poor Real Genuine woman being brutally punished simply for getting two other women pregnant in the completely normal and usual way that humans - who are not a hermaphroditic species - get pregnant by women having sex with women!
"A transgender inmate has been transferred out of the only women’s prison in New Jersey after impregnating two female inmates.
...Minor’s removal from the Edna Mahan Correctional Facility for Women comes just a year after the state reached a settlement with the American Civil Liberties Union of New Jersey to house transgender inmates according to their gender identity, instead of their sex assigned at birth. The facility currently houses 27 transgender inmates, NJ Advance Media reported.
...Minor, who is serving a 30-year prison sentence for manslaughter, impregnated the fellow inmates earlier this year following consensual sexual relationships, Sperrazza confirmed. He added that even consensual sex between inmates is prohibited in the state.
In a letter from Minor posted to the blog Justice 4 Demi, which often publishes videos and letters from her, she described her removal from Edna Mahan as punishment for having sex with two women.
“As a woman who is transgender I truly fear what lies ahead for me, its clear that staff have looked for some kind of security reason to kick me out of the only female correctional facility,” Minor wrote in the post, which was published Friday. “They have thrown me to the wolfs and expected for me just gave up, I am baffled and disgusted by this use power and I can only ask that the commissioner and governors office please send me back. I don’t deserve to be treated like this.”
Remind me again about those completely unrepresentative edge cases? So unrepresentative that the ACLU is forcing prisons to comply with them?
> But if you want to get beyond language, then you need to get into sex-related social classifications, and there are lots more than three.
Well that's the thing, there weren't. Not until about ten years ago. People were male or they were female, and for the 0.018% of the population that was genuinely intersex (https://pubmed.ncbi.nlm.nih.gov/12476264/), well, they managed to figure something out.
As for transsexuals, they existed, but were acknowledged to be people pretending to be members of the opposite sex (hence trans-sexual). You could either go along with this pretence or deny it, depending on how you felt about them personally, and nobody would be outraged if you pointed out "psst, by the way, that's actually a man".
The real big change that got pushed on people was not "there are more than two genders" but that "gender is something other than a polite synonym for sex". This concept started out in academia but got pushed really hard onto the general population very recently.
It is not really much of a distribution. That is the point. As far as "it", for the vast majority of the past "it" was specifically not seen as polite as it implied the subject was not a person but an object.
The problem is the people who lump the Ts in with LGBT. Almost no one has a problem with the Ls and Gs anymore. But the Ts are a problem, and they're tainting their supposed cohorts. Excise the Ts from LGBT, and suddenly support for gay marriage will shoot back up, probably even higher than it was before.
The "core pillar" in this instance is "men are men and women are women." The Ts and their "allies" are trying their damndest to challenge that, and as people take them more seriously, even more people become their enemies, and since the Gs and Ls are ostensibly allied supporters of the Ts, they will earn more enemies too.
Oh dear. Y'know, I've always thought that "men are men and women are women," phrased like that with all the implications that the sexes are entirely separate and any presence of the characteristics of one in the other was perverted, is one of the most pernicious and false memes out there, and I thought this long before I ever even heard of transgenderism. The sooner it's gone the happier I'll be, and I speak as one without a single T bone in my body.
I don't know who you're arguing against, but it certainly isn't me. Men and women both have four limbs as well as whatever characteristics you're referencing. They're still separate categories.
The people who are most essentialist about gender roles these days are the pride folks. Oh your little boy likes to wear a dress. He must secretly be a girl inside. It is just stupidity, as well as harmful.
Unfortunately, the transgenderism argument right now heavily depends on "men are men and women are women". See our dragon friend on another thread claiming that gender is a real part of the brain and they had an identifiably female brain in a male body.
The worst (online) cases go for "pink for girls, blue for boys" and you get a goddamn expert claiming that you can tell if a pre-verbal infant is trans because, for example, they pull the barrettes out of their hair, that means they are protesting their true gender identity is male not female:
To me it seems that the main difference is that in case of LGB, the message for the rest of the society is "hey, they are different, so let them do their thing, and you keep doing your thing; let's agree it's no one's business what happens in someone else's bedroom between consenting adults". Yes, it is a bit more complicated, but 90% is this.
With T, it's not "their private business" anymore. Everyone is expected to play the game of declaring pronouns and being very careful at using them. And instead of "do your thing, and let them do their thing", you are considered intolerant if you refuse to date them.
Imagine someone saying: "hey, I am perfectly okay with other men dating men, but I definitely wouldn't date one". Unless I missed some recent changes, this is considered 100% tolerant. Now say the same thing about T -- and enjoy your ban, if the place is sufficiently woke.
The Ts were always part of the umbrella because it was most politically pragmatic to be there, not because of anything even approaching philosophical or scientific consistency. Conservatives have switched to attacking them because as their insane ideas gain more traction in popular discourse, they get the attention of their interlocutors. Conservatives almost entirely aren't homophobes anymore. Yes, there are too many homophobes who are in power, but the vast majority of conservatives who oppose the T ideology are doing so strictly because of the Ts, not because of some underhanded subterfuge detectable by your incredible mind-reading abilities.
It is the trans lot who are making the waves, though. Like the Johns Hopkins LGBT Glossary which had to pull its definition of what a lesbian was, because they made it "non-men attracted to non-men" but left gay man as "man" and not, for instance, "non-women".
The Blanchardian AGPs with their feminine penises and spinny skirts and the "9 out of 10 times 'non-binary' means a chick" types managed to get the definition their way - for a short while until sanity prevailed.
"Lesbian (sexual orientation): A non-man attracted to non-men. While past definitions refer to ‘lesbian’ as a woman who is emotionally, romantically, and/or sexually attracted to other women, this updated definition includes non-binary people who may also identify with the label.
Gay Man: A man who is emotionally, romantically, sexually, affectionately, or relationally attracted to other men, or who identifies as a member of the gay community. At times, “gay” is used to refer to all people, regardless of gender, who have their primary sexual and/or romantic attractions to people of the same gender. “Gay” is an adjective (not a noun) as in “He is a gay man”.
But what about the non-binary gay-identifying personx? Seems that gay men don't put up with that shit 😁
Other than Stonewall, how many riots really were there? And how instrumental were they really for repealing sodomy laws / getting police off their backs?
Gays were tolerated but mainly disliked in the late twentieth century. Acceptance of homosexuality declined somewhat in the 70s and 80s, then began to shoot up rapidly in the 90s-2000s-2010s in what is probably one of the fastest changes in public opinion ever recorded. This change was not led by radicals, and it was not just about marriage, but about being treated as social equals.
Conservatives “deciding the enemy was weak” I think is a pretty bad model for what happened after.
"It of course didn't help queer people that all their elite, leadership and elders were decimated by AIDS, which conservatives in power at the time deliberatedly let run in the hope that it would get rid of the "problem""
Oh, go fly a kite. May I point you to the monkeypox breakouts, where when asked "please don't fuck multiple strangers for a while until we get a vaccine out", most gay guys responded that it was their human right to go to piss orgies and bang forty guys over a weekend.
Oh yeah, those wicked conservatives forcing gay men against their natural instincts to be promiscuous and fuck around. How terrible!
People handled AIDS badly, because there was a ton of ignorance around it and yes, a ton of fear-mongering. But if you honestly believe that the only reason the "elders" of the gay community got sick and died was because the conservatives had a magic cure which they with held in the hopes that the gays would all die out, you're on crack.
> Oh, go fly a kite. May I point you to the monkeypox breakouts, where when asked "please don't fuck multiple strangers for a while until we get a vaccine out", most gay guys responded that it was their human right to go to piss orgies and bang forty guys over a weekend.
Because of neo-Marxists who thought gay people would make a useful client class and that integration is contrary to that goal. They are currently imposing that whole set of doctrines on basically all gay spaces and excluding anyone who dissents. Most people, gay or straight, are pretty much spineless, and will swing around to the prescribed viewpoint when faced with that kind of pressure.
> But if you honestly believe that the only reason the "elders" of the gay community got sick and died was because the conservatives had a magic cure which they with held in the hopes that the gays would all die out, you're on crack.
Very true. Also, conservatives were not ignoring it. Both William F. Buckley and Pat Buchanan argued that more should be done. Of course, Machine Interface is referring to the Reagan administration and its famously homophobic press secretary, but even the Reagan administration was quietly funding AIDS research while letting the press secretary be dismissive for PR reasons.
To be fair, there are extremists who do everything that is claimed. It doesn't take much to be proclaimed as a "spokesman", the main thing is you've got to want to be noticed. And the news thrives on sensationalism, but they prefer to not exactly lie.
Well, I disagree with Yudkowsky a lot, but I think there will be explosive economic and technological growth and a transition to a radically new world order, this decade or the next. At some point there will be so much going on that attention will be diverted from current fights, and afterwards… who knows
You don't need AGI to change "politics as we know it". 2024 has been projected to be "the first deep-fake election". And the number of scam phone calls that I receive has more than quadrupled over the last few weeks. I mean that literally. I used to get one or two every couple of days, these days it's so many that if I look at my "recent calls" list I don't even see my sister, and I talked with her only a couple of days ago.
How many of your political opinions are a result of reading online articles and debating people online? Not in the sense of "I have read this one article / talked to this one person, and it completely changed my opinions", but rather as a cumulative effect of reading things and talking to people for years?
AGI will be a revolution in astroturfing. Give it enough computing power and bandwidth, and it will produce more articles and comments than the entire humanity combined. Creating personas for all political groups, all personality types, adapting messages for each of them, constantly A/B testing and updating its approach.
If one company gets AGI first, I imagine they will be able to tell it "create zillions of accounts on social networks and convince people that X" and five years later, actual people will democratically vote for X, because all of them found a perfectly convincing reason to do so, a different reason for each of them. (It is not actually necessary that all of them do so; 51% is enough.)
Just imagine something like ACX open threads -- how many people have you actually met in person, or can otherwise verify their humanity? Once the AGI succeeds to convincingly post here, it does not need to limit itself to one account -- creating one hundred accounts, representing different personality types and different political opinions, will simply be a question of running the same algorithm with different prompts hundred times.
Later, when the owners of the AGI click the button, you will find different people providing intelligent reasons for why X is actually the good thing. Some will argue from the socialist perspective, others from the libertarian perspective, some will find an analogy in Bible, there will be contrarian arguments for X, there will be nice and polite arguments for X, but also snarky arguments, etc. Not all of them at the same time; the accounts will switch to X in different schedules, pretending to be convinced by different things.
You can get paranoid, of course, and decide not to trust anyone you don't know in real life. But if many people do that, it will significantly change politics, too.
>create zillions of accounts on social networks and convince people that X
I don't know about that. I find people seldom change their ideas based on arguments, no matter how solid they are. They will simply dismiss them, thinking something must be wrong with the argument, if it contradicts their priors.
On the other hand, people DO change their minds. This must be for other reasons. Maybe AGI can tease some reasons out of data and put forth such things? Make up a "celebrity" that thinks such and such a product is entirely awesome, or this political stance is dumb?
What you mean is that they seldom change their minds based on arguments that you find convincing. That's why the need for lots of different arguments from different individuals. "Social pressure" has a strong influence on the way people decide things, and on how strongly they hold that belief, but they feel the pressure most from those who hold many opinions similar to their own opinions. This often isn't decisive, but it strongly shifts the gradient.
OTOH, if you mean "people don't change their minds because of logical arguments" I largely agree. That's a rare event.
I once told some friends of mine that a number divided by zero is undefined, not infinity. I explained that if it were infinity, then you could multiply infinity by zero to get the number. They remained unconvinced. Then another friend who was late to the gathering, but had a masters degree in physics, was asked about this, and he said it was undefined. Argument from authority trumped logic.
I'm not sure how this applies to AI, but it does back up your argument.
Well, accepting it as infinity was based on argument from authority (their math teacher) so that's somewhat reasonable. The real answer is "It depends on a larger unspecified context". In certain math it's defined as infinity, and it's (0 * infinity) that is an undefined number. In other math the division operator isn't defined on 0 in the denominator. It depends on exactly which axioms and rules of inference you use (there are several that are the same over most of their range, but vary in how they handle edge cases).
I assume the opinions would mostly be changed under *social pressure* ...of the sock-puppet accounts.
If you see 20 smart people on ACX repeatedly providing smart arguments for X, on the emotional level you will conclude "many smart people in my bubble believe X", and that is what may actually make you update towards X. The arguments themselves will only give you a rationalization for why you did that.
To put it more crudely, AGI will be able to infiltrate and manipulate many "hive minds", because it will be able to become a majority in every online group (that does not explicitly protect itself against that).
This is all so hypothetical. I feel like this rationalist community claims to have a strong focus on evidence, and yet frequently makes extreme hypothetical predictions about things like AI that don't even pretend to have any hard evidence for them, and are basically "this just feels like something that might happen". Is there some kind of study or experiment giving a reason to think what you described is possible? Or alternatively, a logical argument that its possibility necessarily follows from premises we all accept?
Because on the other hand, the fact that human governments, corporations and institutions have not with all their resources n 51 centuries of civilisation managed to discover a reliable algorithm for controlling people with the right words, certainly seems like decent evidence against an AI (using only the outputs of said human civilisations) accomplishing it in a couple of years.
I don't think people would if the posts sounding convincingly like they were written by human beings, and if there was variety in them. Why would people? If all the posts were in favor of a view we were positive was nonsense we might eventually wander away to a different forum. On the other hand, our loss would be balanced out by the arrival of true believers in the view.
Perhaps a temporary problem that they fixed. (Or maybe a different version of Firefox, if you keep getting the error?)
It's a paradox that SSL is supposed to make web browsing safe, but too many websites have it configured incorrectly, so the browser has a choice to either ignore the errors (which kinda defeats the purpose) or disable half of the internet.
My brother and his girlfriend has taken to wearing a mask again to crowded places. He says this is because there is evidence the mysterious excess deaths that have been reported are due to sequeleae of COVID, meaning COVID is more dangerous than commonly believed.
I think this is absurd, but well, I hadn't heard that angle on the excess deaths. Did this community ever reach a consensus on what's causing those? Does it actually make sense to mask up at this point, or is it ridiculous?
Excess deaths notwithstanding, it seems reasonable given the recent RCT re long Covid mentioned by Derek Lowe, as discussed in the recent Hidden Open Thread. The diagnosis may still be somewhat inchoate, but it's something that blinded doctors can reliably diagnose at 300+ days post-infection, with a 40% difference between patients treated with metformin and the controls (but which the treatment group still got if they rolled a 1 on their d20).
Something that's going to mess up one in ten to one in twenty patients for a year or more still seems well worth trying to minimize.
No one's going to make anyone do it going forward, so at this point that's a matter of personal risk management. "We" clearly aren't. Whether I do remains an open question.
I don't personally have a current expiration date on masking, any more than I do on wearing my seat belt, which latter I think probably mitigates less risk for me. (Though more than enough to justify doing it.) I'll presumably stop when the benefits don't outweigh the downsides, which isn't currently the case and I don't know when or if it will be.
For most people, evidently that point came months or years ago. (And they're not asking me for advice. :-) ) Other people generally find the personal cost of masks higher[1], and/or estimate the risks as lower, and so make different choices.
I'd probably trade off masking for other effective risk mitigation efforts if something convincingly emerged. But aside from a long term improvement to building ventilation standards (that may come to something in years or longer, or may get killed before it happens), we're not going to get any of those.
[1] Granted, I'm old enough to remember when seat belts were treated as basically torture implements used only by the most nervous of nellies, and probably an actual danger due to interfering with the presumed safety of being thrown clear of the accident. (At speed, through the windshield. Yes, they were serious.) I'd say that's gone, but I didn't expect childhood vaccination to become a matter of serious controversy. Maybe that one's just sleeping too and the call to "throw off the bondage gear" is just awaiting its moment.
About ventilation: An air purifier giving 12 air changes per hour protects you approximately as well as wearing an N95 (except that if you sit quite close to someone you will get their exhales before they've been through the air purifier, and of course then if they have covid you will inhale it). I have been meeting unmasked with psychotherapy patients at least 10 hours per week for the whole pandemic, and I have not had covid. We sit about 10 feet apart, and between us is an air purifier meant for quite a large space running in my medium-sized office. It gives 11 air changes per hour. Any building you live and work in is never going to upgrade to this level of air purification, but you can bring and office or a room in your home up to it. By the way -- like you, I mask up with an N95 in indoor public places still. Mostly it's stores, and I spend maybe an hour a week in stores, so the nuisance value is entirely trivial in my life.
I do have air cleaners in my home and office. But I'd kind of like to see them anywhere I don't control. Or failing that for people to at least crack a window.
I do see them in some medical facilities. I especially appreciate my dentist, since obviously I can't mask there. 🙂
Yeah, don't hold your breath when it comes to ventilation in public buildings. I wish we'd at least get it in the schools. Do you know about Corsi-Rosenthal boxes? Cheap home-made air purifiers that perform well. Yeah, I appreciate my dentist too. I saw him really early on in Covid, when there was no vax, and the poor guy was wearing 3 masks at once. Said they had improved the HVAC and were about to upgrade again.
Over the course of reading around about viruses during covid, I've learned that viruses in general are bad news, not as simple as a bug that makes you sick and then you get over it. Quite a number of them have bad consequences a ways down the road: mono ups the chances of I think MS -- anyhow, something neurological; papilloma and cancer; 1918 flu and severe and weird cognitive syndromes (the ones Sacks describes in *Awakenings*); shingles and Alzheimers; polio and post-polio syndrome. There are others. I am now much more concerned about getting a viral illness than I used to be. It does look like Long Covid is a real thing, though rarer than many believe (maybe 5% of cases?) There are a number of studies finding increases in this and that in the months after covid -- neurological and cardiovascular and psychiatric problems, mostly. I'm not sure how valid any of the individual ones are, but there are so many that I'm inclined to think there's something there.
About it being ridiculous for people to wear masks. I don't think it 's *ridiculous*. They definitely reduce the chance of getting covid, but the chance is pretty low now anyhow. Seems to me important not to bring extra emotion to the situation because the precaution is *masks,* and there has been lots of emotion and fighting about masks. Lots of people probably take precautions you don't. I have friends who filter their water, friends who mostly eat organic vegetables, friends who insist on electric stoves rather than gas. You probably do too -- and it's a lot less irritating than masks, right?, even if you think these precautions are unnecessary. Try to put masks in the same mental category.
Well, except for the lip-reading issue, which seems to come up *very* rarely, wouldn't heavy clown makeup be just as bad? I mean, a mask might even be less of a problem when it comes to establishing trust, committing crimes or engaging in anonymous political violence. Somebody can yank off your mask and expose your face instantly, but it's actually slow and difficult to remove the thick white base and features painted on top of the white in clown makeup. I know somebody who performs wearing it.
I haven't seen any data on how many mask-wearers are criminals using the mask as disguises. Have you? I have seen movies where criminals wear something like clown makeup as a disguise. Covers their face up well, and if it's a smiley clown probably presupposes people to think they're benign. I have no data on how frequently heavy clown makeup, a bunch of face-painted designs or the kind of thing people wear to costume parties is worn to hide the criminal's real appearance, either. Do you? As for how the left-wing attorneys would handle some shopkeeper's ripping the mask off someone who was shoplifting: It seems to me unlikely that a furious store owner would stop to think about that , or even know about that,so it probably would not deter him much. If I was going to shoplift at a store I would definitely wear something like heavy makeup that looked like I was going to a costume party, rather than wear a mask and count on the angry owner's fear of leftie lawyers to protect me from having my face exposed.
There are excess deaths for a period (low long? at least 6 months) after COVID has apparently been recovered from. There are also lots of other problems that seem to persist and recur in at least a few cases. Whether it's "more dangerous than commonly believed" depends on your source, but it's definitely more dangerous that many people want to believe. And the dangers aren't all death. There are lots of cases of prolonged "brain fog", weakness, and other effects. What this is percentage wise I've no idea, as I don't think that there's even a good definition of what counts as "sequeleae of COVID", i.e. one that allows it to be counted.
Your brother isn't being that unreasonable as long as he's in the habit of taking other health precautions, like watching his diet, exercising, etc.
Did you mean to include a different link? That is showing excess deaths since the start of the pandemic. It indicates that lots of extra people since the start of the pandemic. Where do you see anything about Long COVID?
In this video: https://www.youtube.com/watch?v=-ZI7dwldSR0 from 4 months ago, Dr. Roger Seheult suggests that the main sources of current excess mortality are from increased cancer deaths, as a result of decreased testing and care during the pandemic, as well as indeed increased risk of cardiovascular disease after a COVID infection. [1]
It shows a current decrease in mortality in the USA for those ages 0-14, basically no increase for those aged 15-64, and a moderate increase for those 65+.
So while overall mortality is indeed up, and while Long COVID may well play a role in that, it is not much of a factor for young people who are generally at low risk for COVID, in general.
And if the fear is Long COVID, what good will the mask do? He has almost certainly already contracted COVID. Is there reason to think that an initial infection wouldn't cause Long COVID, but a subsequent infection would? I would assume the opposite.
It is not well operationalized. Excess compared to what? The last 20 year "average", 10 years, 5 years? Does the model adjust for changing demographics? What about adjusting for actual or perceived trends in medical improvement? How do you account for different rates of improvement. If opioid deaths are going up because of ease of access but motor vehicle deaths are going down because cars are safer, what do we mean by "excess deaths"?
There is also no standardized way to account for cause of death. Every county in US is doing their own thing within the context of 50 states + dc and territorially rules.
And then there is noise/systemic variation. How much difference do you need to see to justify looking for a special cause?
As to the particular issue of masks, there is little evidence that it can make much of difference for repeated exposures. Let's imagine that a well fitted mask is 83% effective. You roll a 6 you lose. You roll 1 -5 : phew. You ever play Parcheesi and not roll a 6?
At a population level or as public health measure, I think Cochrane has shown it's not going to make a difference.
"Let's imagine that a well fitted mask is 83% effective. You roll a 6 you lose. You roll 1 -5 : phew. You ever play Parcheesi and not roll a 6?" That's not how it works, though.
An 83% effective mask reduces the amount of covid you take in by 83%. If one lousy covid molecule gave you covid then an 83% effective mask would leave you with a one in 6 chance of getting covid even if there is so little covid in the air that on average you're likely to breathe in only 1 molecult. However, there's a dose threshold. It takes a certain amount of covid for you to be likely to become infected, and it's way more than one molecule. And even if you are infected, severity of illness is partly dependent on how much covid you got in your exposure.
Also, if you buy good masks like 3M Auras, and are fussy about checking for leaks, you're probably getting significantly better than 83% protection.
That is not how cdc and others are using 83% effective.
And you're really getting no protecting really if you take off you mask to eat and drink.
Except with well fitted mask for very short duration, I doubt very much individual protection. They just don't work in any meaningful way, which is why seroprevalence is like 98%. Nearly everybody has had it and will get out again.
"which is why seroprevalence is like 98%" US covid seroprevalence could not possibly be anywhere near 98%, even if every single person in the land had had covid. Antibodies disappear after a few months, so by now many who had had an infection or vaxes would test negative for covid antibodies. If you don't like that line of reasoning, here's the data to back it up. See Table 1:
"Of course you're getting no protection at all if you take the mask off." When I was masking in indoor public places I never took it off because, duh, then I'm unmasked in an indoor public place. Just did not eat and drink in those places.
"That is not how cdc and others are using 83% effective." What others do you have in mind? When NIOSH certifies masks as X% effective, that *is* what the percent means -- if filters out X% of particles of the size covid normally rides on.
What is it the CDC means by 83%? Does it mean that in practice a good mask certified at 95 or 99% only gives 83% protection because people don't put them on right, or take them off to eat in high-risk settings?
I saw some people using 83% to refer to allegedly compare infection rates among ostensible mask wearing populations vs non-mask wearing populations.
Yes n95 does mean 95% of particles of certain size -- bigger than virus - virus filtration is electro static filtration but I really don't think it's ever been well measured outside lab in real life with real people and real respiratory viruses
Just stared at the site for quite a while. Here are the things that puzzle me:
-After infection antibodies fade over time, even though protection remains via T cells, etc. It would seem that the waning effect alone would lead to many people who had had covid or a vax, especially if they had it a year or so prior to test, testing negative.
-to see the people who were definitely seropositive from infection, rather than vax, you need to look at people who had nucleocapsid tests. So if I look at the table on the main page, the one you sent me to, and ask it to display percent of population seropositive on nucleocapsid tests, the total % by the latest date, July of '22, the table shows is very close to 100%, I believe it's 96%. But then if I sort by age it shows people 65 plus having the lowest test positive rate (on nucleocapsid test), and across whole life span the pattern is that the younger the group the higher seropositive rate. However, the last date when all ages are reported was Feb. 2022. After that, all the data is from kids. Could it be that the main table, when not sorted by age, is only using the kid rates for seroprevalence after 2/22? Or it the data after 2/22 just not broken down by age by the study they drew on? In that case, are we sure the group they calculated the 7/22 rates had a representative age distribution?
-For maybe the last year I have been reading, in respectable places, doctors saying that yes, there are still people being hospitalized for covid, but they are mostly unvaxed people. If by a year ago 98% of us had enough antibodies to turn the test positive, we are basically all vaxed or the equivalent at this point. So why are unvaxed people people doing worse? Googled it to double-check that they still are. Found this, which only covers Washington State, but I'd be surprised if it was different in other states: "From May 06 to June 02, 2023, unvaccinated individuals were between 1.9 and. 5.1 times more likely to be hospitalized with COVID-19 compared to those who received at least one booster dose. (from https://doh.wa.gov/sites/default/files/2022-02/421-010-CasesInNotFullyVaccinated.pdf)
-I had read recently that seniors had the lowest seroprevalence rate, approx 50%, and that was attributed to their greater caution. Virtually all the people I know who believe they have not been infected are seniors who have been extremely cautious, and they're well-educated enough to know about things like mask quality, not taking your mask off in a crowded setting even briefly, doing home tests before gatherings, etc. I am nearly old enough to count myself a senior, and I don't believe I've had covid. I have been quite cautious, used high quality masks, and tested myself several days in a row the few times I felt off. Tests have always been negative. Same for the seniors I know who have been cautious. Seems unlikely to me that all of us have had asymptomatic cases, given our age and the risk factors many of these people have.
Lots of older people had covid, survived it, but remained bedridden or just generally much weaker. Now, these people are dying several years before they would have died "naturally". I think this explains the excess deaths mystery quite well.
I don't think there's an excess deaths mystery, it's overdetermined.
1. Covid caused people to stay inside rather than going out, they got fat and sedentary, and then they die
2. Covid caused people to put off routine health care procedures, they miss their cancer diagnosis, and then they die
3. Covid lockdowns caused certain other diseases (importantly, flu) to reduce in prevalence for a while. Due to complicated reasons we don't fully understand (see Scott's post on seasonal diseases) this flu holiday caused flu to come back especially hard the next year, causing people to die
I don't know about this community, but in Switzerland the number of excess deaths in 2022 is definitely a topic. It's actually larger than in 2020 and 2021. Some can be explained by heat waves in summer and influenza in winter, but most of it can't. The mainstream explanation is two factors:
- COVID is much less obvious and goes unnoticed much more often, but is/was still extremely wide-spread. For elderly, it may contribute a lot to excess deaths, and those may often go unntoiced.
- The medical/social services in 2020 and 2021 were under extreme stress, so non-vital services have suffered. For example, screening programs were worse, and less cancer and cardiovascular problems in early stages have been detected in 2020/21. Back in 2020 and 2021, doctors already warned of a wave of deaths because of early-stage health issues that were missed during that time.
Currently it is not clear which of the two effects is dominant. Death from an asymptomatic or atypical COVID infection often looks pretty much like a cardiovascular problem, so we can't just tell them apart by looking into basic statistics. We might never know.
To add to the Swiss perspective, excess deaths in 2022 have been a thing both in the UK and in the EU as a whole, and COVID-19 or its immune-related sequelae may well be part of it. I would be disinclined to blame services stress just because the distribution seems odd for that (e.g. some of the countries with the best performing systems, such as Germany, reporting highest rates of excess mortality in some months, such as the ghastly December).
On the other hand, the measure is volatile, and 2023 seems to be looking better thus far, so who knows.
People got lazy and fat during lockdown. Also people are generally only getting more sedentary. Also COVID does have some small amount of long term damage post viral illness.
I would think the increased risk for people who have had Covid and/or are vaxxed/boosted is pretty minimal compared to the cost of masking.
If the person wearing the mask doesn't mind it, what's the cost? I guess other people might find it harder to communicate with them. But do they really have to change their behaviour for that?
But the loss is civilization that is rational vs civilization that is superstitious. Civilization that is superstitious but pretends it is rational might even be worse than just superstitious.
Risk from false sense of security. Thinking a mask will protect at crowded place when you and everyone are taking it down to eat and drink and or when you aren't even wearing mask correctly (the chin mask, the ill fitted mask, etc.) These are all problems.
"But the loss is civilization that is rational vs civilization that is superstitious."
Um, wearing a mask is clearly less superstitious than wearing a cross, and civilization seems to be surviving that OK. Even if you think religion as a whole is bad for civilization, we're not talking about that. We're talking about wearing a cross.
Risk from false sense of security. Thinking a mask will protect at crowded place when you and everyone are taking it down to eat and drink and or when you aren't even wearing mask correctly (the chin mask, the ill fitted mask, etc.)". Smart mask wearers don't turn their masks into chin diapers and sit in a restaurant for an hour that way. But as for those who do -- are you aware of how much stupid shit people do ALL THE TIME? Have you noticed what your fellow drivers and voters are up to? What they read in totally untrustworthy places like Facebook and believe. How they trust their Auntie Ellen's advice over their doctor's?
Banning masking at protests seems sensible, but masks were allowed / encouraged / mandated in indoor environments for two years to no ill effect. Black oversized sweatshirts are also often worn by criminals and rabble rousers, should those be banned too?
I deny that there was no effect. I live in a place with about 100 people, and there were 5 of 6 cases of COVID during the period before vaccines were properly available. Masking was in force, and the disease did not spread. (OTOH, other precautions were also in force, so this isn't definitive proof at even the "small study" level.)
Not according to the CDC. Their numbers show the most recent bout of excess deaths starting in late May 2022 and ending in the middle of January 2023. No excess deaths since then.
* Non-verbal cues are hugely important in order to have a normal conversation. Humans will reflexly smile at any point were their discourse may be interpretted in a hostile manner, in order to diffuse the conversation. Without that, expect all your interractions to be just a bit more hostile.
* I have no data on this, appart from my own experience, but it's generally harder to understand the content of the sounds that come out of your mouth if you add just a single filter between the vocal cords and the ear. Add to that the fact that you removed lip-reading. I basically caught covid because I couldn't handle a week of trying to socialise and make friends while wearing a mask.
* As we have confirmed during this crisis, people will go a very long way to keep themselves unaware that there is a danger. I expect that people will unconsciously try to avoid even looking at you if your appearance reminds them of “the terrible years”.
Are you forgetting that mask-wearing was nearly universal for about a year, not very long ago, and that therefore anybody you tell this stuff to has actual experiences to measure it against? My experience, and that of anybody I talked to about it (admittedly this wasn't a big topic of conversation--I wonder why?), was that if the need to show you were smiling was really important, which it mostly wasn't, you could compensate by "smizing," and that's what people did. Ditto the allegedly sociality-destroying properties of the slight muffling effect of a mask; on the rare occasions when it caused an actual problem, you compensated for it.
I can't address your third claim because I can't imagine caring that people are "unconsciously trying to avoid looking at me" in that situation even if I suspected it were true, which I have no reason to.
But let's pretend for a moment that *every* social situation requires the techniques of slight compensation that I mentioned. Does this really rise to the level of "self-harm"? Words mean things and have connotations. We generally use that term in the English-speaking world to mean something rather more alarming than a mild inconvenience, willingly and temporarily suffered in order to maintain health.
For what it's worth I haven't worn a mask on my own initiative since 2021, and I don't find the original commenter's brother's logic for doing so very sound.
I had never heard the word “smizing” before, but it really has a “everything is fine” vibe to me ; like the advices to touch elbow instead of shaking hands : pretend that the costs don't exist.
And I would contest the whole notion of smiling with your eyes.
* First, the whole reason we tall people to also smile with their eyes is that for a large part of the population, or a large part of the smiling-situations, a mouth-only smile is enough.
* In fact, the mouth-only smile has a specific semantic : “I don't really mean it, but we're in a situation were a diplomatic gesture would be appropriate, and the fact that I'm willing to make that diplomatic gesture is my message”.
* And there are dozens variations on a mouth-smile that the eyes are simply not able to carry. Sad-smile, smirk, half-smile, grin, etc.
In that context, your statement that “you compensated for it” really triggers my anger. There are a large quantity of communication sequences that are already on the border of missunderstanding/starting a feud/awkwardness and there is simply no slack for compensation. I speak as an autistic guy who straddles that line every time I open my mouth.
Even when “compensation” is possible, that is the price that you pay (i.e. the “self-harm”) : everytime you ask someone to repeat, you check if they are sarcastic or not, etc., you've just made the conversation just a little more awkward.
As for the “unconsciously trying to avoid looking at me”, I point you to all the studies about involuntary masking in muslim fundamentalist societies.
Also there's no reason to believe that the studies on babies, showing that anything that has 2 eyes and one mouth is a topic of interest while anything that doesn't isnt, wouldn't apply to adults.
>In that context, your statement that “you compensated for it” really triggers my anger
I hope you get some help for that at some point.
But like I said, in the real world, wearing a face mask to prevent the spread of disease is a slight inconvenience, easily compensated for; and adults, by and large, aren't babies.
There's also the fact that many people in large swathes of East Asia have been wearing them in public spaces for years before the pandemic and continue to wear them after mandates have been lifted (South Korea, Taiwan) without any obvious hit to the quality of social interaction. We can make the usual points about culture, but then let's set aside universal claims about humans.
Much of this veldt-evolved social stuff has long been losing importance, anyway, in the face of cutting edge social technologies like the pen, paper, and the postal service.
> Are you forgetting that mask-wearing was nearly universal for about a year, not very long ago, and that therefore anybody you tell this stuff to has actual experiences to measure it against?
Yes, but that was a couple of years ago. And it sucked.
Wearing a mask in 2023 you're marking yourself out as "one of those weirdos", which will cause social relations to suffer, because nobody wants to talk to one of those weirdos.
>Yes, but that was a couple of years ago. And it sucked.
A lot of things suck but are done anyway rather than endure the even suckier consequences that can arise from not doing them. Wearing PPE of almost any kind on a hot day, for example, sucks, but getting sparks in your eyes or coal dust in your lungs sucks worse, or eventually will.
>Wearing a mask in 2023 you're marking yourself out as "one of those weirdos", which will cause social relations to suffer, because nobody wants to talk to one of those weirdos.
I suspect you haven't considered the possibility that the mask wearer may not consider their own social relations to have suffered because the kind of people who don't want to "talk to one of thse weirdos" don't want to talk to them. In fact, and I know you may find this difficult to believe, they may even find it a plus that people who know nothing about their reasons for wearing a mask don't want to strike up a conversation. I certainly would, and I'm someone who is perfectly happy to see that mask wearing overall is on the decline.
I sometimes still wear one in especially crowded stores. I haven't gotten any funny looks, but would not care if I did, and do not care if I'm seen as a weirdo. I did care a lot when I was 14 and my mom wanted me to wear dorky jeans from the cheap store instead of the magnificent ones from the cool store, but I grew up and got over that. As for nobody wanting to talk with me -- I have zero interest in talking to strangers in the drugstore anyhow.
Is there really any good evidence that an individual wearing a mask (N95 mask, worn correctly) is just as likely to get infected as someone without a mask? I know cloth masks do nothing and mask mandates don't really reduce transmission, but it's hard to believe that correct mask use has no effect on an individual level. They work for all kinds of dust and pollution after all.
No. But if the exposure is chronic rather than acute the mask isn't going to solve the problem. It reduces the frequency of infection in ways dependent on circumstances. Say it's effective 90% of the time (probably an overestimate). That means that if you're exposed 20 times your chance of not being expose are 0.9^20 or about 1 out of 10.
That's not how it works. Let's take your number 90%, and say the person is wearing an n95, but instead of filtering out 95% of particles of the size covid rides on it's filtering out 90%. So, on occasions when the number of covid molecules per cubic foot is more than zero, you will breathe in 10% of those molecules for every cubic foot you inhale.
However, it is not the case that one lousy covid molecule will infect you with covid. Technically, you have been exposed if you breathe in one molecule. But you have various barriers to infection: Luck (maybe you exhale and that one molecule gets blown back out before it enters your respiratory tract). Nose hairs. Mucus barrier in nose. Probably some other stuff. Also, the infectious dose is way more than 1 molecule. I've seen all different estimates, but suffice it to say it's way more than one.. So to get covid while wearing a mask that filters out 90% of what's in the air you breathe, you have to either be in a place with a fairly high density of covid in the hair, or be in a place with low density but stay there for a long time -- in either case, you have to pull in a good number of covid molecules .
And here's some anecdotal evidence: I have a lung problem that puts me at higher risk than most and I have been pretty cautious. I have worn a quite high quality mask in indoor public places, but not elsewhere, throughout the whole pandemic. It goes around my head, not my ears, and I pull the elastic tight, and I feel around the border for leaks. I have never once removed it in an indoor public place. Not to eat, not to drink, not to to scratch my nose, not because it was a bit uncomfortable. (OK, 2 exceptions: went out to dinner with my daughter on her birthday twice in 3 years, though in places with tables far apart and good ventilation.). I really do not think I have had covid. I have never had covid-like symptoms at any point during the pandemic. When I have had something that could be from covid, like a headache, I have tested several days in a row, and all tests have been negative. I realize I might have had an asymptomatic case, but that seems unlikely to me given my age and risk factor.
As the other comment says. My prior probability that some random medication helps with covid is zero, so absence of evidence means it doesn't work. My prior that filtering air helps with not catching respiratory illnesses is that it seems very likely to work. So absence of good studies doesn't really mean much. Of course, we can't really know how large the effect is. So it whether or not someone wants to wear a mask seems best left up to their personal mask tolerance and risk tolerance. (Completely absurd things like wearing a mask alone on a hiking trail are probably not happening these days any more, or if they are it's due to allergies / fires and not covid).
In the absence of well-carried-out studies, I think it makes sense to fall back on reasonable intuition based on physical arguments.
And my reasonable intuition based on physical arguments is that an n95 mask will do a reasonably good job (and a cloth mask a nonzero job) of preventing virus particles from entering your airway. How could it be otherwise, unless somehow every single virus particle managed to slip through the mask without getting caught in the filters?
1) You are only exposed extremely rarely, so a good enough percentage of coverage suffices, otherwise don't bother.
2) It requires a significant number of viral particles to cause an infection, and more is worse.
Case 1 implies that masks may well be useless unless you are exposed only extremely rarely.
Case 2 implies that a mask that sufficiently reduces the viral count could be worthwhile even if it doesn't give sufficient protection.
I *think* that the current evidence argues against case 2. It was part of the basis on which masks were originally recommended. But I'm not sure that there has been any definitive study. Certainly the way many people choose to wear the masks makes them totally useless, and that makes many of the studies quite questionable WRT the effectiveness of masks when properly worn.
I mean, how would you possibly do those studies? Studies done by putting masks on dummies and spraying aerosols on them universally show that they prevent aerosols getting into the nose pretty effectively. Studies done by comparing "area with many masks" with "area with few masks" generally seem to show that "lots of masks" don't reduce transmission on the population level. And studies comparing individual maskers to individual non-maskers would be hopelessly confounded, because there's no way the mask is the only difference there.
And well, the problem of there not being a definitive study that shows a clear effect is present in many fields of medicine, Scott has covered many psychiatry-related aspects over the years.
So in conclusion, masking may or may not be worth it. But it seems pretty safe to assume it will somewhat reduce the risk of catching covid in any given situation, if done correctly.
I'm perennially disappointed by top-level diplomatic exchanges, always hoping for some unpacking or clash over key differences, instead of polite euphemisms like "we should promote mutual respect" or "we should reduce misperceptions."
One possible recent counterexample, German FM Annalena Baerbock was far more direct with Wang Yi than I've heard anyone be, directly pushing back when she heard the talking point that human rights are just an invented Western concept.
I'd like to see more kitchen debates, though not sure the diplomatic establishment really tolerates that sort of thing.
"Invented Western Concept" is accurate. They date to the period called "The Enlightenment". We decided that this was one of the ideas from that period that was worth keeping. They really lost their logical underpinnings with the changes in religious faith, but we decided that they were worth keeping anyway.
The question is "Why should we trust deals with anyone who doesn't think they are valid?".
"invented western concept" recommends it if anything, much like calculus, mechanized farming, commodity futures, semiconductors, internal combustion engines, representative democracy ... :)
Neither Qin Gang in his recent joust with Baerbock nor Wang Yi over the years have rejected the actual idea of human rights as a foreign concept. The talking point is that the Western view of rights is not universal, to which point she (somewhat amusingly) made the legalist move of invoking the UN Charter.
Anyway, a concept of what amounts to 'rights' ensured by mutual duties isn't necessarily worse than one framed in the language of protection from encroachment on freedoms. It certainly can be, but it's not a given.
"the talking point that human rights are just an invented Western concept."
Is that not just straight up undeniable truth? Even if you are very strongly in favour of human rights I can't see any sensible argument that this is not the case. (apart from the platonic view that human rights were just out there and you found them, which is silly - and I can't see Baerbock making the argument considering she's quite famously not very smart.)
That said, you can get quite a long way from a fairly short set of axioms (suffering is bad all else equal, happiness is good all else equal, one should maximise good and minimise bad all else equal).
I think those - plus empirical claims like "torture causes suffering" are enough to recover a decent chunk of human rights (or at least "human 'good's"). And I doubt that China has a fundamental disagreement around human rights being goods "all else equal".
What I don't think you can easily get is a commitment to the rights as absolutes – it's consistent to say things like "a harmonious society is even more important than individual goods". I think this is probably the crux between the Western view and the Chinese one.
Frankly, human rights aren't absolute even in the West – we suspend the right to liberty when we imprison criminals, for example.
So I think asking "is China obliged to respect 'Western' human rights" is beside the point – both the West and China have a contingent respect for human rights, but not an absolute one. Where they differ is in the weight they place on each right. Of course, this is a pretty significant difference – but it's more subtle than "Does China accept Western human rights?"
Sorry, but to many, certainly to all serious Buddhists, you argument is an argument of killing everyone as quickly as possible. (Well, not to Buddhists who strongly believe in reincarnation, of course. Or in the popular notion of karma, as opposed that that espoused by the Buddha in "The Word".)
You are really giving them the benefit of the doubt there. You are assuming that when the CCP does something that violates human rights, that they are performing some ethical calculation and taking them into consideration and not violating them unless there is a good reason to do so (that a neutral person would accept, even if he might ”weight” it dofferently).
And what does harmonious society mean for an authoritarian regime besides ”society where people obey us”?
Yep, I definitely lean towards assuming too much good faith in general and probably have here too.
I think that Xi Jinping genuinely thinks he's doing the best thing for China. I don't think he would use the consequentialist framing I gave above, but would probably claim to agree with it. I think China genuinely believes that when Western politicians call out Chinese human rights abuse, that they're doing so hypocritically in order to secure political advantage. I think it's possible to have all these beliefs and still act in a self interested way by deluding themselves about their motivations, though I'm not sure that specific flaw is more true of China than the West.
Regarding whether a harmonious society means more than "society where people obey us" – I think it probably doesn't mean more than this (to the CCP). I think the idea of a society where everyone has a place and doesn't rock the applecart is a terminal goal in, say, Confucianism, rather than an instrumental one.
None of this detracts from the proper response to (say) Chinese genocide of Uighur muslims – outrage and total, unwavering opposition. But I do think it's important to see your enemies as they see themselves, or you can't model them properly.
For that matter, I believe that when Western politicians call out Chinese human rights abuse, that they're doing so hypocritically in order to secure political advantage. I still despise the violators for the violation, though, so their hypocritical argument works....except that I want to hold them to the same standard, or better.
But in trying to understand how they see themselves, you seem to come close to suggesting their framework is basically the same as in the West. Despite the Chinese leaders essentially admitting that they think differently.
The old timey Communists used to say that human rights were a "bourgeois" idea. The Chinese saying it's "western" is just the same thing in a new form (you know, "anti colonial" instead of class warfare), talking about the bourgeoisie isn't popular nowadays, but bashing the West is.
Does the CCP's desire for a "harmonious" society where everyone obeys the state differ from the Soviet Communist Party's desire for the same thing because "social harmony" is somehow associated with Confucianism and Confucianism is an old Chinese philosophy? Does it make it more legitimate?
Intriguing. Do you believe people ought not be moral realists?
I would have thought that stating almost any opinion commits one to the view that people should or should not believe certain things, that people should or should not espouse certain ideas, which smuggles in at least some kind of realism.
If you're an Ayer emotivist or just want a svelte ontology that seems reasonable, by all means hang on to that. I'm just sad to read your distress when you hear some people believe a theory that, for me, caches out to basically "some outcomes are better than others."
I would welcome your arguments that I should adopt nonrealism. Though I'll confess I'm less concerned about the object level, and mostly concerned about your distress at encountering a popularly held view. Beyond the fact of which of these positions are correct, would be happy to offer some strategies, take or leave, to make encountering realist views at least less disappointing.
"Intriguing. Do you believe people ought not be moral realists?
I would have thought that stating almost any opinion commits one to the view that people should or should not believe certain things, that people should or should not espouse certain ideas, which smuggles in at least some kind of realism."
Maybe. But not moral realism. I'm only a antirealist when it comes to morality. People ought to be not moral realists if they care about honest logcal reasoning. Which is something I would like to expect from philosophers, but sadly that's not the case. There is simply no convincing argument for moral realism. It's all god and/or platonism at the end. (Not there's necessarily a difference between those two.) Being a moral realist is just convenient. If you deny moral realism your moral arguments become much harder, more annoying and especially less convincing. But I hate that kind of motivated reasoning.
"If you're an Ayer emotivist"
Technically I consider myself a sentimentalist but I don't see much wrong with emotivism.
"people believe a theory that, for me, caches out to basically "some outcomes are better than others."
With a statement like this you are just skipping the issue. (which to be honest, is sadly the norm for moral realists) What is "better"? What is good? Why is your conception of good better than mine? Better for whom? Why should I or anyone else care about that conception of "good" or "better"? (I also would like to mention the fact that you went straight to consequentialism here, which the majority of moral realists among philosophers do not even subscribe to. So saying that moral realism caches out to that seems a bit absurd.)
The best argument against moral realism is simply the there is no good argument for it. There is no way to bridge the is-ought distinction so all that remains is magical wisdom raining down from heavens as the source of any certainty about moral claims. And if that's the fundamental axiom of your beliefs you might as well believe literally *anything*.
(There are ofc also good arguments directly against moral realism. I've yet to see an argument against evolutionary debunking of moral realism that was not very weak. But maybe it's out there.)
"Beyond the fact of which of these positions are correct, would be happy to offer some strategies, take or leave, to make encountering realist views at least less disappointing"
I'm open to your suggestions. (and also sorry about the potentially angry tone of the post, I can get quite emotional when it comes to certain topics.)
Yeah, no worries about the tone. You are honestly exasperated by encountering other people who show signs of intelligence and education but fail to adopt something you consider as fairly obvious after investigation. I've been there.
I don't want anything here to come off as patronizing. You clearly know the domain, I'm not going to insult your intelligence by suggesting you go off and read why Harris or for god's sake* C. S. Lewis are realists.
I would just offer that when I faced similar frustrations in other domains, it seemed to help when I acknowledged that people's ontological commitments can shift very slowly. Especially on the big foundational stuff.
So, imagine everyone you're talking to might just be in the early stage of a five year conversion process to your position. At this phase all they need help with is unsticking one or two small things, then you send them off on their journey from there.
Sure, it's a hopelessly naive fiction, most people aren't primed to change their views. But still, this stance seems to most often lead to a eureka or refinement on one side or the other. It's a pragmatic choice more than anything.
The next layer that kicked in for me, after having a few of those discussions, was some compounding epistemic humility. Maybe I'm the one in the midst of one of those five year conversions. Or we both are. Who knows. Usually I can find some premise I can converge to at least a bit.
In this specific case, maybe you're about to convince me to abandon realism?
Honestly, to your credit, you did prompt me to take another skim through the Stanford encyclopedia on antirealists to refresh myself on the different varieties there. Some of these folks I'd read over a decade ago and promptly forgot they existed. And as a result I did flip one bit in my ontology... it's really deep in the weeds on this issue and nobody would be very impressed... and the whole house of cards hasn't quite collapsed yet. But it's something.
But that was after a review of some fairly technical arguments.
In contrast, I loved Hume growing up and still have enormous respect for him, but I no longer find things like the is-ought problem dispositive. (Should I? :P ) I would unpack why I came to that view, but you've been barraged by realists already, I don't want to pile on. Here I'm just noting that, it's hard to predict where the crux is for different people, so it can take some patience to root around in their mental attics, figure out what they might be hanging on to for nostalgia or just inertia.
The ironic thing that strikes me most about our positions is that an outside observer might call me a moral antirealist as regards to people's ontological commitments, and a realist as regards to their actions, and you the inverse of each. I don't mean to not take you at your word, but you seem to have really strong opinions on what people should believe. (I know, so long as they care about logical consistency.) Just found it interesting.
After reading Wittgenstein,** who knows, everything seems like it just comes down to definitions at some level or another anyway. I'm open to the possibility we believe roughly the same thing for practical purposes, and just use different words or emphasis to explain it.
Speaking of, I did like his idea of philosophical therapists, who just talk people through their (philosophical) problems to help them unstick things. Maybe I'll hang up a shingle as that some day, might be fun. :D
* pun intended
** There's probably a good will hunting meme about me out there somewhere where next year I'll read Kripke and be convinced Wittgenstein lacked a full appreciation of the importance of names or something. If so, fine, I fully accept it.
1. You can be a moral realist without believing our current Western values are the completely correct values. I personally think people will look back on our tolerance of abortion with the same shame with which we now look back on our tolerance of slavery.
2. There is nothing inherently religious about objective moral values. Kant derives them from reason alone. Hume can be read as a subjectivist, but he also talks about the "human moral sense", as if it could be universal among the species.
And in fact, I'd argue religious-grounded moralities are fundamentally anti-realist. They hold morality entirely relative to God, who himself can do no wrong by definition, because anything he does becomes morally right. That's as terrifying as the worst forms of cultural relativism.
1) "You can be a moral realist without believing our current Western values are the completely correct values"
Of course. Obviously. I never meant to imply anything else.
(Btw I think the same thing when it comes to treatment of animals but can't see the argument of this being the case for abortion at all.)
2)"There is nothing inherently religious about objective moral values. Kant derives them from reason alone."
Of course there is. Kantian ethic does not "work" without God, it needs the assumption that God exists. (I would also claim that it doesn't work at all. No one lives according to Kant's ethical system. It's undoable. An in depth critique would be far too much here. though.)
"Hume can be read as a subjectivist, but he also talks about the "human moral sense", as if it could be universal among the species."
A view that is entirely compatible with subjectivism. I agree with it, yet, just like Hume, am both an (atleast ethical) subjectivist and sentimentalist.
" And in fact, I'd argue religious-grounded moralities are fundamentally anti-realist. They hold morality entirely relative to God, who himself can do no wrong by definition, because anything he does becomes morally right. "
God is by defintion absolute good and absolute morality. Therefore that point doesn't really make sense in my opinion. Nothing God does "becomes" morally right. It right by it's very nature. I cannot see any way how objective morality could work without absolute good existing. (aka God) Kantian ethic doesn't, neither does anything else.
Human rights are a subset of ethics, and ethics is something that is neither "invented" nor "discovered", a bit like mathematics.
Ethics is "invented" in the sense that you can come up with any set of arbitrary rules you like and call it ethics. But it's "discovered", in the sense that most of your arbitrary sets of rules turn out to be terrible and obviously unethical; either they're horrendously inconsistent, or they lead to massive human suffering, or they're just not palatable enough for people to accept them.
Take slavery for an obvious example -- people who could afford slaves spent thousands of years trying to come up with ethical justifications for why this was actually ethical, but after seeing the clearly terrible consequences for thousands of years, Western society eventually came to the conclusion that the space of sensible ethical systems does not include any with slavery -- then we banned it, and everything was better. I'd call this a discovery.
Human rights are similar -- after thousands of years of ethical experimentation we've eventually figured out that there's certain important "no go" areas on the ethical landscape, and we have good reasons for believing that these are genuine no-go areas and not just our personal preferences.
Terrible, is reasonable, but you can only decide that an ethical system is unethical on the basis of a different ethical system, so that part of your argument fails. And for slavery, there were lots of ethical justifications that were satisfactory to the people who adopted them, and even to some of the slaves. That you don't accept them is because you adhere to a different ethical system. It may well be a better one, but deciding that would require lots of investigation. E.g. is an ethical system that justifies destroying the environment you live in better than one that justifies slavery? On what basis do you choose?
You cannot validly argue that an ethical system is unethical. That you don't like it is fair. That it doesn't match your ethics is fair. That is has unresolved contradictions is fair. That nobody can possibly live up to it is fair. That most people won't be able to live up to it is fair. There are lots of fair arguments against ethical systems, but being unethical isn't one of them.
I can only write that I intensely disagree with your or any moral realism and therefore also with the concept of moral progress.
I could write very, very long texts about the reasoning behind my metaethical position, countarguments against moral realism (and have done so in the past during my studies) or even specifically against your post (previous moral rulesets were "obviously unethical" but we still practiced them for many, many millenia? And somehow the ruleset that evolved over the mere last couple of centuries is good now? Why? How? Not to mention you literally put the goal of ethical humanism as the ultimate ethical goal and then claim ethical humanism does achieve this best - Yeah, duh. You are putting the horse before the cart. People did not fail to achieve the "ethical goal" of humanism, they literally did not have that goal at all.) but all of this is besides my initial point: Human rights are a specfific concept invented and evolved over a specfific time that were not in the world before.
If you think they were discovered instead of invented you would have to believe that *everything* is merely discovered (as platonists do), making the very concept of invention (and the word) itself pointless and nonsensical.
(Remember: Discovery is otherwise reserved for things that were in the world before - and humans just found them. Without believing in a platonic realm you can't tell me that ethical positions existed before humans ever thought of them.)
> And somehow the ruleset that evolved over the mere last couple of centuries is good now? Why? How?
Because they're more informed by facts, because we've eliminated many unjustified assumptions, and because we have better cognitive and rhetorical tools with which we can evaluate our arguments. The same reasons why our math and science is better.
Actually there is a sense in which ethical systems are discovered as well as invented. Only some of all the possible ethical systems are sufficiently compatible with human instincts to be even meta-stable. Landing in a metastable position is a discovery, as there's no known way to predict which states will fit that description. (Even historical examples don't work, because the environment is different.)
How does this not apply to literally all inventions. There is only a limited number of ways a lightbulb can work. In that case the point would be redundant.
We're close to understanding the physics involved in a light bulb. We aren't close to understanding the boundaries of human instincts. By the time Edison came along, we knew basically how a light bulb worked. OTOH, I have heard it said that Edison discovered that a carbon filament would work longer, so it's not a sharp distinction.
I think moral anti-realism isn't even necessary to think human rights are an "invented" concept. Like, "human rights" doesn't look like a real value system no matter how you slice it; it's incomplete (i.e. ambivalent about a large number of choices), and somewhat ad-hoc, in that "human rights" presents no first-principle derivation for itself.
It's more similar to law, actually; it's plausible that some set of ethical principles, when followed to its logical conclusion, will incorporate something that looks a lot like the laws we have today, but it's pretty ridiculous to think that the current legislature of any country is acceptable as an ethical system (relevant: "in all situations, the government of Raikoth will take the normatively correct action").
Same thing with human rights - it's a cut-off, arbitrary chunk of an ethical system enshrined as international law. The Chinese are probably complaining that (as lawyers do with regular law) it can be used as a bludgeon in unfair ways. That's a bit disingenuous of them, though, because the alternative to invoking human rights is just straight up telling them that what they're doing is morally wrong - and everyone, Western or Chinese, knows That's Just Not How We Do Things on this planet.
Everything is an invented concept. States. Humans. Phlogiston. Molecules. Chess. Electrons. Integers.
So the argument that "X is just an invented concept" is kind of an universal counter argument. It tells you nothing about the value of the concept or if you should adopt it.
No. Concepts are invented, but not the physical reality that we reference when we invoke those concepts.
If you believe that physical reality actually exists then the underlying fact of the matter that we reference when we speak of molecules is not invented at all.
Ethical concepts on the other hand have no clear or measurable physical reality underlying them, which is why ethical realism nearly always has to rely on a platonic realm to make any logical argument.
I grant you though that saying that an ethical position is merely an invented concept is not a very strong point unless you want to make the point that your ethical position is somehow not invented (usually a religious argument :because it's divinely inspired.) But it's not wrong.
What he's asserting is that the way we slice physical reality into concepts is not deterministic. That there are alternatives. Certainly when you get to things a fine as quantum mechanics, and then try to translate them back into human language, this is a true statement. It's less clear that this is true of "baseball game", which is clearly invented, but you can look at variations of the rules and ask "is this still a baseball game?". It's not that there aren't basins of relative stability that have names, it's that the positions of the boundaries isn't predetermined, The names refer to centroids, but most of reality is somewhere in between those centroids.
Molecules (or humans, for that matter) are a gross simplification which are great for us because our minds suck at reasoning about a protein (or human) as a particular class of solutions to some equation of a theory of everything.
If smarter aliens said "molecules are obviously an invented, earthling concept which we do not need" the best we could say is "well it seems a useful abstraction to us in many cases".
I could argue that some ethics can be considered empirical rules for building human societies, there are certainly tenets which would be incompatible with any society. The problem there is that there are also plenty of tenets which seem very compatible with complex societies, but nevertheless terrible. Personally I am mildly in favor of the non-cognitivist idea of not treating ethical rules as logical propositions at all.
Ethical rules are a compromise between tribal instincts and the need to build larger societies as populations grew. There are lots of possible compromises that would work, and there are even more that wouldn't work. They have a reality that's not exactly logical, unless a golf ball is logical, but isn't exactly determined either.
1) Just because something was invented more than once does not mean it's not invented.
2) I'm not sure that is true at all when it comes to what we consider human rights. Your comparisons also seem to imply that human rights would be an extremely easy and common invention, something that I would dispute.
Diplomatic immunity is such an universal concept because it was extremely useful. Slavery was also extremely useful, universal concept.
Human rights fly in the face of that concept, and I would argue are a much more recent, uncommon invention.
Actually "human rights" doesn't fly in the face of slavery. Some definitions of human rights contradict some definitions of slavery.
Personally, my definition of "human rights" considers the US prison labor to be a violation of them. There are many, however, who don't hold to that definition. Either that, or they are quite willing to ignore human rights when its convenient to do so.
Is there any way to improve 'bravery', for lack of a better term? I find that I become afraid far too easily- any mid-sized stressor that would raise anyone's heart rate a bit (public speaking, a verbal confrontation, etc.) makes me *very* afraid. I experience an excessive surge of adrenaline, out of proportion to the situation, which is very visible and makes it impossible to stay cool. I become emotional very fast. It's odd because I don't suffer from general anxiety (I feel fine most of the time), and when I was younger I did some extremely physically challenging activities that meet anyone's definition of toughness- more on the endurance/labor side. But for my entire life, I become excessively frightened by mild stressors.
If the answer is 'exposure therapy'- logistically how I would regularly practice say verbal disagreements with others? This seems impractical. Or, alternately, has anyone been prescribed something that reduced their brain's fear reflex? I find alcohol makes little difference
A cognitive trick that supposedly works because nervousness and excitement are cognitively close: when nervous about something, just say out loud to yourself "I'm excited", and that can meaningfully convert nervousness into excitement.
The most valuable thing I ever learned about bravery, as someone who also comes with his fight/flight response a little tightly-wound, is that it's not the absence of fear; it's the choice to do the right thing in the face of fear.
I find it's unfortunately much easier to practice this in the context of physical danger than social danger. Fires and cliffs don't care if you're a little shaky. Humans tend to make it awkward. But it can be done.
I think it helps to set aside all expectations of how you 'should' feel. We all have some ability to do things that feel bad - change a diaper, clean a wound, etc. Not all of us are great at meeting all our responsibilities immediately, but we tend to limit our hardcore avoidance to the things we think 'should' be easy and enjoyable. It can help to accept that something's going to be unpleasant and commit to doing it anyway.
When your anxiety is purely a physical reaction, you also have the option to choose to interpret it as excitement. The two states are physiologically identical; it's the conscious areas of your brain that give meaning and labels to your arousal states.
I know what you mean, but for me at least, in many situations it helps to "not look down", i.e. don't confront the fact that you're in a scary situation and just focus on getting through/doing what needs to be done.
Oh, I agree, that's a very effective way to keep yourself from freezing in a scary situation that you're already in.
I took "bravery" to be more about choosing to confront situations you know will be scary (i.e. overcoming the fear of fear). Denial/suppression seems less effective for that, although it does sometimes seem to work in group contexts.
Anxiety and excitement may be the same thing for people who enjoy scary movies. That might even explain what to me is an absolutely baffling taste. Because for me those two states are entirely and absolutely antithetical.
As Nietzsche said, if you know the why, you can bear any how. Part of the problem here seems to be you don't really believe you need to do these things you're afraid of, so the fear wins out.
Whereas if you knew why you want to do these things, you would have access to a stronger emotion that could overrule this fear.
It's really important to have something that's bigger than you, something that can override you
My daughter suffered from very strong fear of public speaking for many years. Beta blockers fixed it, I wish we found out they exist years before we did.
Yes, actually I have been considering beta blockers. I think I'll probably have to source them online as I doubt my doctor would prescribe me some for personal experimentation
I was prescribed beta blockers for dental surgery, by a dentist who did not use nitrous oxide. If that ever comes up in your life, you might try it then.
I don’t think there is any potential for abuse of beta blockers. Nobody gets ‘high’ on them. Not a doctor myself but I think an MD would likely let you try them for this sort of anxiety.
I realized a few years ago that I was also “afraid” of a lot of things in the moment that didn’t scare me intellectually. Like if I thought of the thing, it didn’t scare me but if I was in the situation and it happened my fear response was through the roof. This extended to surprises, confrontations, and a whole range of usually innocuous things. In essence, I saw nearly everything as a threat or dangerous, so my fear reactions were off the charts and embarrassing.
As boring and seemingly obvious as this is, I discovered that if I exercise in the morning, raising my heart rate and getting a good sweat in, then I would still register all the same things--like my awareness didn’t diminish--but my body was much less likely to register them as dangerous or threatening. The sweet spot is 30-45 minutes of resistance training and/or low intensity steady state cardio, but really any sweat with an elevated heart rate works. The effect is so acute now that if I am anxious on a day I haven’t worked out, I go out to my garage where I have an airbike. I will do 3-5 circuits where I pedal 20 seconds max effort and rest 40 seconds, and just that small workout will clear my mind and relieve some anxiety.
I’m sure there are good explanations for why this works, but my (pseudoscientific) working theory is this: In Thinking Fast and Slow Daniel Kahneman explained how we have two systems of thought, the automatic background thinking that is like a vast supercomputer doing pattern recognition, and our relatively measly deliberate logical thinking. My theory is that our automatic thinking largely evolved as a threat detection mechanism, and with most of us (reading this blog anyway) spending all day in refrigerators reading off screens, that thread detection mechanism keeps looking for threats even though there isn’t anything. Something about the exercise raises the threshold for a threat, so that only actually threatening things gets coded by my automatic thinking as a threat.
Other helpful things: improv comedy classes and performances for getting up in front of people and looking like a dickhead, and Brazilian Jiu Jitsu classes for actually fighting people and--at least for the first few years--just getting your ass whipped again and again.
Thanks Charlie, I really appreciate the time that you put into this. Unfortunately I do have several years of BJJ experience- didn't seem to help much :( Maybe I'll get more out of striking though! Have been considering taking up Muay Thai for this exact reason
I find it is the level of exertion, rather than the activity. Like it is a direct correlate to how sweaty and out of breath I get. For me, BJJ does that, but my ideal workout for managing fear/anxiety is basically CrossFit.
I did a lot of reading and thinking about consciousness [and did a bunch of mushrooms] and came to the conclusion that death isn’t real - and now I don’t get scared of anything! Even little things unlikely to kill me. So I’d recommend digging into whatever existential topic lights your fire [and doing a bunch of mushrooms] and see if that chills you out.
Mushrooms and anxiety, two nutritious parts of a complete dog's breakfast! I dunno, man. Make sure someone else is around for that who isn't super anxious.
I saw an MBA student on YouTube who was practicing exposure therapy for this by asking for things he knew people would refuse in small stakes situations. The one I remember is him going into a fast food restaurant and making some crazy substitution request like replacing the bun with more hamburger patties.
When I was having trouble at work being stressed by confrontations I decided to do something similar. I bought and sold things on Facebook marketplace and haggled with people on the price. I did the same when buying a car. I would say it helped.
I believe that if we lose our current level of technology, we will go back to subsistence farming, hunting and gathering. That there is no plausible way to transition to any level of technology in between.
If this happens, 8 billion humans will die leaving a few million in remote locations with no essential dependence on technology.
I define "global collapse" as "loss of electrical supply".
Who agrees?
Who disagrees?
Note: it's fun to debate the likelihood of collapse, but that's not what this thread is about.
I'm not sure why you say a lack of electricity would prevent us from producing fuel. It's not like oil platforms are connected to the electrical grid (often even the ones on land).
Work that can be done with motors can also be done with combustion engines, so as long as fuel is available and engines are available, factories can keep running, though they may need significant modifications. Electric generators are also very simple devices to produce: all you need is copper wire, a magnet, and a source of energy to rotate it.
I talk about no more electricity as a way of identifying a collapse. There are people I talk with that think collapse means "things are different now". But the oil platforms would stop working for the same reasons that the electrical generators stopped working. Lack of spare parts for example.
But I don't believe global collapse will happen. I just believe that if it does happen, 8 billion humans will die.
I imagine if the collapse goes very well (no widespread warfare, communities or governments organize and cooperate), then a massive amount of human labor could be reallocated to agriculture and other essential industries. Jobs like marketing or psychologists aren't exactly essential in the face of global food shortages. Market forces would naturally push people into agriculture (even if the financial system collapsed, people would still trade much value in goods or labor for food).
So then the question is how many people can we feed with our current farmland using only basic tools and human labor? And by how much can we expand agriculture before people starve? (And this is affected by whether it's a quick collapse or a slow collapse.)
My guess is we could feed a large fraction of the world in the optimistic scenario.
Right, and I'm saying in a best case collapse scenario, we can probably still feed a lot of people without electricity or advanced technology. The lack of electricity itself doesn't seem insurmountable if everything else is going as well as possible in the hypothetical collapse. It would help a lot if the collapse happens over several years instead of several days.
The loss of electricity isn't really a global collapse or apocalypse.
I mean, let's clarify at the start that for literally a billion+ small and subsistence farmers in Africa, India, and Southeast Asia, very little is going to change. Not just that they're going to wake up the next morning and go back to work...basically as normal, but that they're going to keep growing the food that feeds the village.
Now New Delhi is going to have a bad time and most of us in the West are going to have a really bad time but, like, do you really think everyone in, say, Somalia is going to starve to death if the power goes out?
Second, what do you mean there's no technology in between? I mean, I don't know how to make a musket but there's guys in Afghanistan making guns out of garbage. The fact that you or I can't make anything without modern manufacturing tools doesn't mean nobody can. And you can get pretty darn far up the tech stack without electricity, we know you can at least get to the US or UK in the 1890s, which is enough tech to get you to WWI, which is enough tech for mass industrialized warfare and all the fun of the "modern" world. And yeah, it would be disruptive and there would be a lot lot of deurbanization and we might go back to ridding horses but there's a huge gap between that and, like, caveman hunter gatherers.
Almost all farming, even subsistence farming in the third world, is dependent on the modern fertilizer industry these days. Of course the individual farmers themselves would adapt to the loss of those inputs, but the reduced production would still imply widespread starvation.
Nonetheless, I agree that there is no reason to think we couldn't stabilize at some level well above 'hunter gatherer'.
I believe that the stabilization would be below the level of Rameses I of Egypt. In the first stage, those who have means of subsistence are going to need to fight off those who don't. Some of those who don't will be folks who were planning on living by hunting, but found that too many others had the same plan. So until the guns break down or wear out, the level of violence will be quite high. It will probably end as a new feudalism, at a much lower level than 500AD, because there won't be any self-sufficient enclaves that maintain civilization. (Too much technology is too complex to be maintained by a local enclave.) Perhaps this won't be true in parts of Africa, India, China, and around the Arctic North, but I'm not at all sure. The problem is that climate change is altering the environment such that traditional ways of earning a living no longer work.
>>I mean, let's clarify at the start that for literally a billion+ small and subsistence farmers in Africa, India, and Southeast Asia, very little is going to change. ... they're going to keep growing the food that feeds the village.<<
Do they use any kind of powered equipment? Without electricity and fuel that won't be available.
>> do you really think everyone in, say, Somalia is going to starve to death if the power goes out?<<
What percent of people in Somalia grow their own food? The supermarkets will be empty. People who have caches of food, must have guns to defend it.
>>Second, what do you mean there's no technology in between?<<
I said "no plausible transition" (that keeps the bulk of the population fed).
>> I mean, I don't know how to make a musket but there's guys in Afghanistan making guns out of garbage. <<
Do their guns shoot bullets? Can they make a bullet? Black-powder guns are plausible.
>>And you can get pretty darn far up the tech stack without electricity, we know you can at least get to the US or UK in the 1890<<
How much pre-1890s technology is available within 10-90 days?
Food is the focus. No food will come from more than a few miles away. If the people in the cities are not fed within a few days, they will attack any person who has food.
> Food is the focus. No food will come from more than a few miles away. If the people in the cities are not fed within a few days, they will attack any person who has food.
Sure, there's no way of getting out of this scenario without a lot of people dead in the cities. As a city slicker, my best shot is to pack a tent, plus as much food as I can carry, on day 1, and then start walking; try to outpace the rest of the city and get to an agricultural area where I might be able to get a job as a farmhand. I'm thinking my best shot is dairy country since it's non-seasonal and they'll suddenly need a lot more labour to milk all those cows without machines. The cows themselves are of course beef on legs in a pinch, and good cattle grazing country is also good horse grazing country and horse breeding is about to become the most valuable industry.
Dairy country is a good idea. I recommend planning ahead and buying some of these nitrogen-packed survival food assortments. Stash dozens of these with with your dairy Farmer friends. And plan to keep your tank half full so you don't have to walk there.
But if a mob of 10,000 desperate city dwellers shows up at your dairy farm not long after you, all your planning will be for naught.
In order to speculate sensibly you'd need a plausible mechanism for this complete global collapse. I can imagine all sorts of things collapsing, but why would it cause farmers in Fiji or Patagonia to suddenly forget how to do their thing?
This premise does not sound reasonable, short of god deciding that henceforth, every material on earth shall be non-conductive.
Globally speaking, the power grids are very distributed. Temporarily knocking out the continental US grid might be doable, doing so globally and permanently seems non-reasonable.
Roughly, you would need to:
* Disable nuclear power, perhaps by removing all fissile isotopes from Earth. Preventing nuclear reactions in general is preferable to also prevent fusion. (Note: This may disrupt military equilibriums.)
0. Lower the inner temperature of the earth to deny geothermal power. (Note: This may affect surface temperatures.)
1. Prevent the burning of fossil fuels either by removing them or by removing oxygen from the atmosphere. (Warning: Removing oxygen might affect humans in ways unrelated to the power grid.)
2. Cause an Chicxulub-like impact. This might reasonably disable most power generation from solar, biofuel (eventually), wind, and hydro. (Note: This might affect humans in ways unrelated to electrical power.)
3. Blow up the Moon to partly disable tidal power. (Note: As tempting as it sounds, do not merge this with the previous step by driving the Moon into the Earth unless you are ok with immediately extinguishing most life on Earth.)
4. Tidally lock the Earth to the Sun to prevent the remaining potential for tidal power. (Note: this will probably cause winds near the terminator (which will also likely be the most habitable zones) which any surviving humans might use for power generation. On further reflection, it might be safer to replace steps 2.-4. by removing Earth from the solar system.)
Note: I am not a Mad Scientist and this is not Mad Science advise. Please consult with a certified Mad Scientist before trying any geoengineering projects on this scale.
These are interesting scenarios. You are focusing on electrical supply, and I can understand why. It may appear that I am saying that the reason for the collapse is the loss of electrical power. But I don't mean that. I'm just using loss of electrical power as a marker, a way to say "Okay guys, this is collapse."
I don't believe collapse will happen, but if it does happen it will be because of some kind of cascading failure: one failure causes another which causes another etc.
I participate in a global warming forum. The majority of the participants seem to believe that collapse within a few decades is inevitable. Many of those are looking forward to this as ecological justice.
What about world-wide EMPs? Nearly all modern technology uses some form of computer. So nuclear plants would likely be non-functional, regulation systems on geothermal wouldn't work, and the power grid in general probably would have massive outages that couldn't be fixed without new computer chips. And where would one get those? The chip factories, multi-billion dollar facilities, couldn't make more chips because the ones to operate the factories no longer work.
It would be a big problem and some plants would never come back up, but some would -- every electrical grid in the world has hundreds of clever electrical engineers who can figure out ways to make things work with whatever they have. Most of the world's existing power stations probably worked fine before they were computerised, and they can work fine again without computers; it will take a lot of work to fix them, but it will be society's top priority.
I predict that most of the world will have a reasonable fraction of its power generating capacity back within a timescale of years.
But why would all electrical supplies across the globe all fail? Aliens?
I'm having fun speculating on what would happen in, say, Fiji, if the rest of the world returned to the Stone Age. Fiji is more than self-sufficient agriculturally, half the country is powered off hydroelectricity already, and they can easily turn sugar cane into ethanol. They don't manufacture much in the way of machinery but they could surely figure many things out if they needed to; and besides they can always raid the ruins of Auckland and Brisbane and Honolulu for anything they can't produce locally.
Aukland and Brisbane and Honolulu are all thousands of miles away. How does one get there, and back again? Regularly? What are you bringing them they will want more than their machinery? Why would the crew on the ships not decide to stay where the machinery is?
I think you greatly underestimate the effect of civilization on the current state of the world. It would take a vast amount, by the collapsed civilization's standards, of fuel to get a ship to another continent, even assuming one could navigate without GPS. If they are going to produce those thousands of barrels of fuel, they ought to use it locally to produce power.
If the locals are now in the stone age, what on Earth are you going there for? To salvage stuff from prior civilizations? If they have such things, they will likely be well above stone age themselves, and using the equipment.
If you're proposing the stuff you want is effectively free for the taking from the locals, then you're talking about raiding and piracy. Even if they did that for some time, it wouldn't last long. Certainly not an entire year.
No, if civilization collapses, then the entire world would turn into regions, and only a few people would travel between regions. It would be very hard to have global trade until civilization has built back up so that mere survival is no longer the primary goal.
>>Fiji is more than self-sufficient agriculturally, half the country is powered off hydroelectricity already, and they can easily turn sugar cane into ethanol.<<
Producing most of their own food within Fiji would be very helpful in a collapse.
I could not quickly find what percentage of fuel used in Fiji is imported. I'm betting it's close to 100%.
If Fiji anticipated global collapse, they could convert to ethanol use over some years. IC engines don't like 100% ethanol, but I wouldn't say it's not possible.
So the problem is getting the food from the farms to the table. The transition cannot be done within 90 days. After 90 days most people would be dead from starvation or injury.
Fiji does not produce integrated circuits. Within 5 to 20 years almost every electronic or electric device in Fiji would cease to work (would break and could not be replaced).
Now, please note: I don't believe global collapse is going to happen. For example, 10% of the global economy is used to produce food. If for various reasons it becomes much harder to produce food, we could devote 20% or 30% of the economy to food.
There are people that I converse with who are rooting for global collapse. They think it will be good for nature and that humankind can transition to a "post-industrial" society.
Well, 70% of the population is already subsistence farmers, so that helps. Only the remaining 300K people need to figure out how to feed themselves. Fiji is not very big (the largest island is 100km across) so it seems like short term farm-to-table problems can largely be solved by people walking to where the food is.
Making ethanol from sugar cane really isn't that hard, and can easily be done inefficiently on a small scale by any individual with, like, some buckets. Given the vast sugar surplus that would exist in a Fiji deprived of its export markets, I'd imagine that a cottage industry of good-enough fuel would establish itself quickly -- good enough to keep at least some engines running.
Sure, they're not going get their iPhones back any time soon, but I can picture places like Fiji riding out a global collapse in relative comfort.
You're thinking short term. So lets think of Fiji. That gets them through the first, say the first decade. Then the engines start wearing out and there aren't any replacement parts. And climate change doesn't suddenly turn off, so the sea level keeps rising. Who remembers how to build an outrigger canoe, or to navigate by the stars with a coconut shell? (I've no idea how that worked.) And there are a lot less fish in the sea than their ancestors had to work with.
Other places the problems will be things like water availability. We've been overusing the water resources all over the globe. Once the pumps stop working, the water isn't available.
Other places there are different problems. Everywhere the population is too large to support without powered equipment. Seasonal food can no longer be preserved. I expect we'd be luck if one person in a hundred lived through the crisis...and lucky may not be the correct term, since the famine itself would lead to immense further destruction of the environment. Perhaps one in a thousand would be better. And THAT may be too large. The population of Ptolemaic Egypt is estimated at three to four million. The current population of Egypt is 109.3 million (2021), but a crash wouldn't recreate the annual flooding of the Nile with rich sediments. (To put this in more general terms, the modern environment isn't as naturally supportive of life as the ancestral environment.)
I think small groups would be able to stay above the subsistence level even if they would really struggle.
Say a bunch of people in some remote area with access to a river they can use as a power source. Stuff like that. Most of their technology would collapse but you could preserve enough of it even with our garbage skills/knowledge to stay well above the baseline.
Assume they didn't already have a water-driven electrical turbine-generator (which is likely); they would need to create one.
With no food in the supermarkets, I figure the first 90 days would be total chaos and 95 to 99% of humans would die within that time period. So it would be the survivors building that turbine. Most survivors would be physically strong persons possessing guns. I think a rather small percentage of these persons would know how to build a turbine.
You don’t need a turbine to make electricity. A regular water mill will do fine (with some gearing). Won’t be much, but when you are starting from nothing any is a lot.
This is after the famine has passed, for forget the horses. And bicycles depend on decent roads. There's an argument that the reason cities paved roads was to make it easier for bicycles. No spare tires. No spare parts of any sort, except what you can scavenge and make fit. This lets out most modern bicycles.
P.S.: The people who would attack you won't all have died. Some of them will have formed gangs that survive by raiding.
These are all good points, and correspond more with my outlook.
There is no doubt scavenging would happen. The big question then is whether this could leapfrog to some level of technology above stone age or would the scavenging last one or two generations and devolve to the pure farming-hunting-gathering society that I imagine.
In other words: yes, you can scavenge a generator and wire and some incandescent bulbs, but can you build a generator from scratch?
Were you asking if any groups would stay above subsistence, or my particular group? For such an open ended question you sure have a very specific scenario you are trying to drive everyone towards.
Why not just lay that out and defend then instead of coming up with weird objections to peoples responses…
Yes, that would be one way of posing the question. Could some groups on their own maintain some type of technology more advanced than flint knives.
I believe the answer is no, but I would be fascinated by a description of the technology that might be maintained independent of the rest of the world.
"A Canticle for Leibowitz" by Walter M. Miller Jr. provides a relevant speculative take here.
I suspect it would be pretty mixed. Some techs are deceptively complex, others are just good ideas and widely known. There are some types of white paint that involve surprisingly complex chemistry. You're probably far more degrees of separation from a specialist in paint chemistry than you are from a gifted pharmacology student who might actually be able to DIY some basic medicines using a microlab.
I'm sure other disciplines have this, something ubiquitious, like smooth ball bearings for writing pens, are fiendishly complex to manufacture, whereas something seemingly much more complex and specialized might accept a DIY substitute (minus some features), like some agricultural machinery. I have a relative who made tires from scrap during the great depression, even though it was no replacement for the international tire manufacturing supply chain, still got the basic job done. A necessity-driven focus on "good enough" might leave a surprising amount of tech behind.
The pharmacology student requires good reagents. Try herbal medicine practitioner. Distillation would probably survive, and that would enable lots of chemistry after awhile.
P.S.: The "Great Depression" only lasted a decade. The replacement tires would give out pretty quickly. And the road network requires more maintenance than you may expect. We didn't/don't build roads to the Roman standard of durability. Even if just left alone an asphalt road becomes a path within a decade or a bit more. (That depends on the climate, of course.) Bearings would be easy to recover if the idea were preserved, though the replacements might be small cylinders of hardwood.
Metals, though, are difficult. I'm not sure that modern alloys can be reused when the best you've got is a charcoal furnace. You're probably stuck with what they are or what you can cut them into (well, include filing). A lot of them are fairly rust resistant, at least if you stay clear of salt water, but I'm not sure what proportion. Anything containing Titanium or Aluminum will probably need to be used pretty much as is.
There are also technologies that require a certain scale of society to be feasible. Having the knowledge required to build something like a rolling mill is one thing ... having a society that has both the organization to do it, and the scale for it to make sense, is another.
Unless you're instantly deleting all copper and aluminum wire from the simulation, it is reasonable to assume small-scale electricity generation wherever it would seem practical. For anyone who's fed, sheltered, and secure in the short term and has basic DIY competence and access to wire, a small wind/water/steam turbine is going to be one of the first orders of business once they realize that grid power and fuel deliveries aren't being restored.
In my mind you are describing the kind of electrical supply that would keep an incandescent bulb burning when the wind was blowing or the stream was rushing, not the kind of electrical supply that would support a laboratory or a factory.
At first, yes, that's what people will bring online quickly. But it's not inconsequential. Don't underestimate the value of lightbulb-scale power generation to people who have none. https://en.m.wikipedia.org/wiki/William_Kamkwamba
Over time, small-scale tech can be leveraged to build more ambitious projects. That's how our great-grandparents did it, and they didn't have millions of tons of free copper wire and precision-machined bearings to work with.
I got interested in this question because I ran into people who think that "collapse" would be a good thing. Of course we probably mean different things by the word.
Would you agree with me that loss of electricity, all at once or region by region, would cause the death of about 8 billion persons?
It's certainly not my preferred material for DIY projects, but it's a perfectly functional conductor that's widely used in construction and industry. If for some reason you were weathering the collapse of civilization in a house with aluminum wiring, you could certainly repurpose it.
And I don't know when/why you would want to weld electrical wiring, but it's certainly not necessary for building a small/medium-scale solenoid coil.
Yes, there would be lots of wire available for quite a while. Since only a few million humans are left out of 8 billion, it would go a long ways.
But at some point you need to make more. It takes gobs of electricity to make aluminum. So that's probably out, on the other hand aluminum corrodes very slowly, so could be there for thousands of years to be recycled.
I believe copper is much easier to refine.
The big question is will the scavenging society evolve to a technological society or devolve to farming-hunting-gathering?
As a child I often encountered new words by reading them, and later mispronunced them when using them in conversation.
An internet search not only confirms that this is quite common, but gives it a name: Calliope Syndrome, with 'calliope' deliberately mispronounced 'cally-ope'.
For example, pronouncing epitome as epp-ih-tome is quite common.
One of mine was pronouncing inevitable as inny-vittable.
My dad told as a child he realized the villain in a children's book was very bad because he had misled (mizzled) someone.
Some of mine were variations on this, with a young me misreading the word. Thus, foliage became foilage, and debris became derbis.
I also thought that there was an artillery weapon called a mortimetre. I think I had combined mortar and millimetre.
A childhood friend insisted that a certain aircraft manufacturer was called Boing, becauses its aircraft bounced upon landing.
Well, as a non-native English speaker there are many words I will read and mispronounce in my head multiple times before being surprised by someone's use of it in conversation. Awry (aw-ree), Oceania (O-see-ey-nee-ah), and a different confusion with albeit (all-bait) are some examples that come to mind, but there are many more. Plus every word where I'm unsure whether the i is an ee or an aye!
I realized that my above example, and various others I posted later in this thread, are not examples of Calliope Syndrome, but rather the opposite, in which I write incorrectly something I have misheard. I don't know if there's a name for this.
Sort of related, in Grade 3 I realized that USA stood for the United States of America, and infered that USSR therefore stood for the United States of Soviet Russia.
The next day at school, Miss Millar was away sick and there was a young substitute teacher instead. Eager to show off my newfound knowledge, I worked out a way in which I could work this into a conversation naturally.
I left my desk, walked up to the teacher's desk, and said "Pardon me. I wonder if we might at some point be reading some stories about other children from around the world, including perhaps the United States of Soviet Russia."
The sub looked a bit bewildered and said "Uh, I don't know ...".
I wonder if she recounted this in the teachers' lounge later.
The main one I remember is albeit, which I pronounced all-bite. I heard people say albeit correctly in conversation for years before realizing these weren't two separate words.
For many years growing up I had two distinct vocabularies. One written and one spoken. And I was always having these moments when I realized that what I thought were two different words was just me 'pronouncing' (in my head) the written word the wrong way. (I love reading and stink at figuring out how to pronounce words.)
As a non-native speaker who learned 99% of her English by reading, I have *tons* of those. Like recipe, which I'd thought would rhyme with wipe, pipe, etc.; and height, which as it turns out does not rhyme with eight – so what's that e doing there? Found some new ones in this very thread, too.
The dialect word that gets me is "chile". This is supposed to be "child" with the final d elided out, but it looks like it's pronounced either like the food or the country.
The narrator of a Heinlein audiobook I listened to (I think "Methuselah's Children") did just that. The character referred to someone as "honey-chile", and the narrator read it as if it were referencing a sweet variety of capsicum
I know this happened to me fairly frequently growing up, but I've mostly forgotten the actual words. Fairly recently I did find out its pronounced "chi-batta" not "see-batta", but that's pretty minor.
The biggest example I can remember from when I was a kid was when I was in third grade or so and was nearly in tears because I couldn't say or spell the word "theatre" in a way that could make the teacher understand what I was talking about. It didn't help that I had read a bunch of British fantasy books and so picked up some odd bits of Commonwealth spelling (am American).
I had awful spelling as a child too. I've mostly solved it with a combination of typing muscle memory and spellcheck, but a lot of my school assignments were borderline unreadable. I've heard this is common with kids that learn to read by reading a lot, which makes me suspect the entire idea of "whole word reading" is a crock even for kids that can handle it.
Loads of British place names and surnames are pronounced completely differently to how they are spelled. As a couple of examples off the top of my head, Featherstonehaugh (a place in Yorkshire) is pronounced Fanshaw, and Woolfardisworthy in Devon is said as Woolsery. There must be a list online somewhere, such as a Wikipedia page. Some antiquated spellings have been simplified to match the pronunciation, such as Brighthelmstone to Brighton in the early 1800s.
I remember as a young kid mispronouncing "nebula" as "nubela", and my brother used to misquote the title of a medical TV series "Doctor Finlay's Casebook" as "Doctor Finlay's Bookcase"!
I used to pronounce "Home Depot" as "Home Dee-pot" (pronouncing the "t" so that it rhymes with "teapot"). I actually even had a temper tantrum about this in elementary school.
So the standard critique of presidential systems is that, by allowing voters to pick the executive directly, it's prone to cults of personality and demagoguery. This was the story with Latin America traditionally, and the US more recently. By contrast (in theory), parliamentary systems removed that decision from the voters. But in practice:
The head of the largest party is usually (yes, not always) the new PM
Voters know who the party heads are going in to the election. So they are effectively voting for their country's executive in a similar manner
I.e. voters knew if they voted Tory in the UK, they'd get Johnson. I've even heard critiques that voters frequently don't care who their local/constituency rep is, and aren't evaluating him or her on their service or policies- but are just voting (in this example) for a Tory district rep because they want Johnson to be PM. What do people think of these critiques of parliamentary systems? Are they really not that different from a presidential one?
Edit: To be clear, I strongly prefer parliamentary systems- I'm just stress-testing the arguments for them
I don't think first-past-the-post parliamentary systems and proportional parliamentary systems are all that similar. Having only voted in the latter, I have always voted for a party whose policies I've agreed with, never for a PM, and never for a local rep – in fact, I've never even known who that is – and I would think that most of my compatriots would feel the same.
The British electoral system is very undemocratic. In 1997 Tony Blair won a huge majority in the House of Commons winning 40% of the vote, which made him virtually a minority elected dictator because the Commons is almost like the President plus Congress combined.
Specifically regarding the risk of "strongmen" with cults of personality and demagoguery, I'm unclear on whether it would be meaningfully harder to convince half of Parliament, than half of the entire country. Perhaps it is the case that the strongmen would find a way, no matter what? And they happen to be associated with "presidential" systems for incidental reasons?
I didn't mean "convince using the same methods". I'm less concerned with whether their public image looks like that of a "traditional" Latin American "strongman", and more about whether they leave power gracefully, whether they got power fairly, and whether they create mass movements that poison the commonwealth in the process.
There are big elections in the UK and US for parliament and the President, but they do vastly different things, don't they? Comparisons might get weird once you get into the particulars. Why not compare parliamentary elections with congressional elections?
In the US most House and Senate races focus on the local candidate, though national ad campaigns have reduced these to elections about party leadership, perhaps similar to the UK. I've heard a theory that's an outgrowth of modern electoral strategy and communication technologies, the McDonaldization of political advertising. Targeting one big national ad about Pelosi or McConnell and re-using it with slight modifications is far easier than making highly customized ads for every local race. I suspect we'll see that more in lots of countries for those same underlying economic/technological/strategic reasons, independent of any underlying political system.
On paper, the US President could be seen as primarily a figurehead focused on running the foreign policy establishment. In practice, they tend to campaign with some sweeping domestic legislative agenda, even though it's not a legislative role and is mostly limited to the power of persuasion to accomplish anything there. It is admittedly puzzling. A key distinction is that in this system, the executive and legislative branches are nominally divided, which does limit the power of any individual component.
When you say you strongly prefer parliamentary systems over presidential systems, you're not saying you prefer unitary over divided government though, right?
I suspect the crux of what you're saying is that you primarily want political parties to have much more power relative to individual candidates, no? In the US context, you might say you believe the "whip" should have much more power or be much more effective. This has varied over time, party discipline ebbs and flows.
If your central worry is cults of personality, I'd consider whether those are a bug in humanity, vice any particular system. If you exclude them from the political process, maybe eventually Caesar crosses the rubicon. If you invite them in, with a long tradition of peaceful transitions of power (and lucrative post-leadership speaking circuits!), hopefully they take their win and retire a few years later. Maybe the 1% most tyrannical leaders can slip the bonds of any institutions, so it's coming sooner or later, and in the US we've only drawn 46 lots so far.
I'd be very interested in Scott's early take from the Dictator book club on the extent to which institutions matter when confronted with the most persistent and dedicated authoritarians. Presumably some institutions matter to some extent, like, not having any functioning institutions at all seems a recipe for disaster. Good questions all worth exploring!
>the US President could be seen as primarily a figurehead focused on running the foreign policy establishment. In practice, they tend to campaign with some sweeping domestic legislative agenda, even though it's not a legislative role and is mostly limited to the power of persuasion to accomplish anything there
? They head the bureaucracy and can issue executive orders, which is gigantic power in our present-day country and especially in divided government. They can veto whatever Congress passes, unless there's a gigantic supermajority. They also appoint the judiciary
Do people in parliamentary systems vote for PMs over parties? I found this article (https://journals.sagepub.com/doi/10.1177/13540688211019239) (haven't read or analysed in detail), which seems to argue that most people vote based on party, but taking individual politicians into account. Personally, living in one, I think the parties end up with more name-brand power than individual politicians, who change every few years and can at any rate be replaced by the parties at any point. To some extent, this is visible in e.g. the US as well (like the Democrat supporters who wanted Bernie but voted Biden because that was the party line.)
However, even if this is not the case, I think based on the presidential vs parliamentary systems of the West in recent years, parliamentary systems do seem to obviously produce fewer cults of personality. I would imagine that this is because, in order to be a viable candidate for PM, politicians have to be established and respected members of the party. This means they have to be basically likeable, polite, good negotiators and share many of the same opinions as the other MPs. So, even though people might be voting for the kind of strong personalities that produce leaders like Trump, they really aren't able to choose those options often.
One thing to note is that in the parliamentary systems I'm familiar with, the Prime Minister doesn't have any special powers over and above what Parliament chooses to delegate to him. The Australian Constitution, for instance, does not once mention the existence of a Prime Minister. It mentions Ministers, and one of those minsters has always been a Prime Minister, but there doesn't have to be.
The key thing, though, is that the Prime Minister is not directly chosen by the election but by the vote of the party room, and he can be replaced at any time if the MPs of the leading party decide they don't like him any more. This removes many of the failure modes of Presidential systems, since there's a limit to how badly the PM can screw up before he gets removed, whereas a President can be incredibly hard to dislodge.
Yeah, but as Australia learned *repeatedly* after 2010 - the rest of the party can boot the leader pretty quickly and make it clear that it is indeed the party and not the leader who is in charge. The UK has had this recently as well, obviously.
Why do the Westminster systems (UK, Australia) have this, but we rarely see parties removing their leader in more proportional governments? As I mention down below, I can't really imagine say the CDU ever removing Merkel
Not in the late years of Merkel, but in the early years this wasn't inconceivable.
The first German chancellor, Konrad Adenauer, was removed in exactly this way in 1963. He fought for staying chancellor for longer, and he fought vigorously against his successor Erhard (from the same party), but CDU and the coalition partner FDP forced him to resign.
Helmut Kohl was almost removed by his party in 1989, called the "Putsch" of Geissler, Späth and Süssmuth. It failed, but not by much.
In Westminster systems, the party in power usually has a majority of seats and can do what it likes. In proportional representation systems, the government is usually a somewhat temporary alliance of smaller parties that form a majority, with the leader of the largest party in the coalition usually getting to be PM.
This means that this party now has difficulty replacing its own leader, since they also need to convince their coalition partners to go along with this leadership change lest the coalition fall apart.
The German chancellorship is a pretty different situation entirely.
Thanks. I learned that Norway is the 1 parliamentary system (to my knowledge) with no constitutional provision for early elections. They've removed 2 PMs in their history and, with no possibility of coalition collapse, just replaced them with someone else mutually agreeable. Maybe that's the right way to do things
The answer is - some are voting for Johnson, some aren’t.
Johnson in any case is a likeable buffoon and not much of a populist or demagogue, just very incompetent.
I don’t think you can handwave away that his own party got rid of him either, as did the Labour Party with Corbyn. It seems the parties do restrain the populist tendencies of their voters.
I think this is a valid criticism of the parliamentary system.
During the run-up to the last provincial election, a pleasant candidate for the opposition came to the door asking for my support.
I replied that I liked everything I'd heard about him, but that I found the leader of his party so offputting that I feared my vote for him (the local candidate) would help the leader of his party become premier.
The argument I've heard for parliamentary over Presidential system is that they avoid "constitutional crisis" when one branch of government opposes another.
Constitutional crises can still happen in parliamentary systems -- see the Australian constitutional crisis of 1975 in which the House of Reps and the Senate were deadlocked and unable to agree on a budget.
In that case the situation was resolved by the intercession of the Crown. The Labor Party is still salty about that one all these years later.
One way to solve Constitutional crises is to vest ultimate power in a person or institution who can be relied upon to never actually use that power unless strictly necessary. Not all counties have the benefit of such a reliable institution though.
>Robin Hanson says firms tend to more closely resemble parliamentary governance
I'm not an expert in corporate governance, but I doubt that elected members of the board are organized into anything like 'parties' who come with 'party leaders', which is where I think that analogy breaks down. Shareholders are not organized into highly ideological groupings and voting for board members based on belonging to that same tribe. And so on
Sometimes boards do wind up broken into various factions competing for shareholders' votes to get their agenda through. If this happens it usually means that the company is screwed, and investors should stay the heck away from it. Well run companies tend to look more like Singapore.
P.S. If the response is 'well the Tories removed Johnson'- sure, but other countries that use coalition systems tend to use party lists and have very strict party discipline (they can just remove you from the list if you irritate the party). If Angela Merkel was caught doing something untoward, was her party going to remove her? Seems unlikely
As much as I dislike white supremacists, I'd hate even more to live in a world where white supremacy is categorized as a disease and its believers forced to take drugs so they can say things like:
"Love is the most important thing [...] Nothing matters without love."
And do things like:
"That night Brendan reached out to Chicago Antifascist Action and connected with the specific activist, "S", who had gone undercover in Identity Evropa before revealing Brendan's identity"
"Brendan hired a diversity, equity, and inclusion consultant to advise him, enrolled in therapy, began meditating, and started working his way through a list of educational books."
Imagine if a drug turned violent, masked, riotous antifa members into family-oriented religious conservatives who reached out to pro-Trump organizations for re-education. Wouldn't most people on the left, even those who detest antifa, feel uncomfortable with such a drug? Especially if it was presented as a therapy, implying that being part of antifa is an illness?
Im convinced that MDMA in the water would make us all kinder, but it’s all a bit brave new world.
When I was younger I used to go to raves in my college town. Well, it was called a rave but technically wasn’t quite, as it was indoors. Anyway the crowd, who were mixed between college students and fairly rough more working class crowd, were in general all hugs and kisses. I myself didnt do MDMA, or any drug really. It was a great sweaty night out though.
One night however, the atmosphere was much more hostile. It was a crowded venue and and inevitably you would bang into people or step on toes, and most nights this caused no problem.
But on the night in question bumping into somebody was potentially dangerous, the crowd was more hostile and angry than usual, although many were still enjoying the dancing, the atmosphere was off. MDMA was not available, instead everybody was drunk, and fights broke out.
I believe the drug of choice these days on the same city is cocaine with even worse consequences.
>Well, it was called a rave but technically wasn’t quite, as it was indoors
I don't think I've ever heard anyone make that distinction. Surely a rave, post-mid-1980s, is simply a large dance event featuring extended mixes of a certain kind of electronic music, usually with the expectation that many attendees will be on MDMA or similar drugs? All the events I went to in the 80s or 90s that called themselves "raves" were in warehouses or rented music halls.
After chatting with ChatGPT about magic systems for a while I asked about the group Astral Codex Ten and received the response below. When shall we host an ACX meetup in the astral realm?
> "Astral Codex Ten" is a tome said to contain profound insights and instructions for astral projection, a practice where one's consciousness separates from the physical body and explores the astral plane. The book is rumored to possess detailed techniques, ancient rituals, and intricate diagrams that facilitate safe and transformative astral journeys. It is regarded as a comprehensive guide to unlocking the mysteries of the astral realms and expanding one's consciousness beyond the limitations of the physical world.
> The group known as Astral Codex Ten is a closely-knit community of individuals who have delved deep into the teachings of the book. They form a network of dedicated practitioners and seekers, each with a shared interest in astral projection, spiritual growth, and the exploration of metaphysical realms. The group operates in secrecy to protect the sacred knowledge and to ensure the safety and responsible use of astral practices.
> Membership in Astral Codex Ten is exclusive and based on a combination of invitation and personal attainment. Prospective members often go through a vetting process, where their dedication, knowledge, and ethical understanding are evaluated. The group's aim is to foster a supportive and knowledgeable community that can share experiences, offer guidance, and advance their collective understanding of the astral realms.
> The members of Astral Codex Ten engage in various activities, including regular group meditations, collaborative research, and the sharing of personal astral experiences. They also organize workshops, retreats, and lectures to expand their knowledge and promote the exploration of consciousness.
I don't think there's much cause for concern. The only reason it was giving me these responses is because it was already operating in a magical fantasy context. If you ask it to tell you about Astral Codex Ten without any setup it just gives you the regular schtick about how the data cutoff was 2021 and it doesn't know anything.
So I gathered, from the context you provided. But given the number of credulous folks out there who are already using ChatGPT as a substitute for google, I still see scope for issues.
The next couple of years are going to be tough in terms of finding an authoritative and accurate source of data for casual enquiries. Google is now a mess thanks to SEOs, and AI will only add to that in the short run.
Friends were talking on the weekend and the point was made that public libraries may have been somewhat premature with their decision to throw out their physical books (as many of them did in the past decade or so) because reference books, for all their flaws, stay put in terms of content, unlike many on-line sources of information.
They can't push the data cutoff much further, without their training data being contaminated by the output of GPT and all the other LLMs. Which is seriously problematic, though predictions vary on how much contamination is too much.
One aspect is the stickiness of national borders. There are practical and emotional reasons for people to not want national borders moved, while at the same time there are high costs to national borders.
Any ideas about what might be able to take the edge off of nationhood?
Meanwhile, at least national borders guarantee that everyone doesn't make the same mistakes simultaneously.
Those are good points. I agree that Europeans divided Africa, the Middle East, and South Asia poorly, but I also have yet to see anyone put forth better alternatives that withstand real scrutiny. For example, if we want to divide the Middle East so that each group has its own, homogenous country, then what do we do with ethnically mixed regions? Drawing a line down the middle risks a bloodbath like the partition of India and Pakistan or the Balkan civil wars.
What about ethnic exclaves? What if there IS some way to neatly draw a national border so that it fully encompasses one ethnic group and no other, but those borders also leave it landlocked and not possessing any natural resources?
Another problem is deciding how to define "ethnicity" for the purposes of creating nations. Are we just coding inhabitants by religion, or by religious sect? On the national borders created by that, are we overlaying a second layer of lines to separate the people by language? How about by dialect? Is there then a third layer of lines to reflect tribal zones of control? How do we handle areas of nomadic people? If you get too obsessive trying to make national borders that reflect human geography, you end up with micro-nations that are not viable or barely viable for varieties of reasons.
Look at this map of Africa's ethnic regions overlaying its actual national borders. If the Europeans had somehow, in defiance of the technology and resources available at the time, had perfect knowledge of the continent's human geography at the time they departed and created the different countries, it would have resulted in crippling levels of fragmentation.
I'm curious how those comments on Njal's Saga were the top comments. Has this school of criticism fallen out of favor to such an extent that the observation about Christian influence seems novel? I feel quite certain that when I first heard of Beowulf as a child the very next sentence was to the effect of "it was written down by Christians". Was I raised in some kind of anti-Christian-influence-on-literature bubble?
The bar for being the best comment on a book review and the bar for being a truly original and novel piece of literary scholarship are pretty different.
Hi! I'm helping draft a script about Charter Cities for the youtube channel "RationalAnimations" (https://www.youtube.com/c/rationalanimations). Some of our previous videos have adapted essays by Eliezer Yudkowsky and described some Robin Hanson derived concepts like prediction markets and the "grabby aliens" hypothesis.
Naturally, we want to make sure we're presenting the idea of charter cities in an accurate, compelling, and understandable way before we the script gets adapted into video form and potentially seen by ~100K people. So, I'd love it if anyone here at ACX would like to provide feedback on the script. It's about a ten-minute read; you can use the google docs "comment" feature to highlight a section and add your thoughts: https://docs.google.com/document/d/14wEa6SFiWo9uw30UB52DHIHXhHrIprnT3EzAbU-2teA/edit?usp=sharing
I spend most of the script talking about the motivation for charter cities as potential engines of economic growth in developing countries (inspired by the success of places like Singapore), and explaining why a new city with the ability to set its own business regulations might help encourage growth. I use Itana in Nigeria as my main example, mostly drawing from the linked FT article when describing the various reforms they'd like to implement. (I also talk about Prospera Honduras, as described in ACX's "Prospectus On Prospera": https://astralcodexten.substack.com/p/prospectus-on-prospera) After addressing some potential objections, in the conclusion I describe charter cities as having the following three main benefits beyond the direct economic growth of the city itself:
1. the idea that seeing legal reforms work well on a small scale could inspire larger national changes.
2. the idea that charter cities could spur “governance competition” resulting in better, more effective governments generally, by giving people more options of where to live.
3. the idea that experimenting with totally new types of law could bring huge benefits, by exploring new institutions (like prediction markets or new land-use laws or etc) that might solve problems that even advanced countries struggle with.
We'd love to see any feedback that you have on the script!
You could also look at the story of China's special economic zones which were sort of like charter cities. However, I would also devote some time to the rather serious problems with the idea, mainly of credible commitment and governance. The former is clear with the issues that prospera has faced. The latter suggests that if you can make a charter city work, you probably don't need it, and if you do need it, you probably can't make it work
The biggest weakness of charter cities (an idea I'm very much otherwise intrigued by) can be summed up by the phrase, "There's no sovereignty like sovereignty." You're either sovereign, or you're not, and actual political independence that can't be capriciously taken back from the same hand with which it was given is hard to come by. Nice work if you can get it, though.
Are you sure you want to post a video about charter cities on the same channel? They're a way more politically fraught concept than alien speculation or even AI-risks, so there's a significant possibility the channel could get needlessly pidgeonholed for having controversial niche libertarian-capitalist politics (assuming it's implicitly pro-CCs. If it's against or mixed that's a reduced risk but still much "spicier" than the usual fare).
What is the best way to learn about ways of understanding why adhd (or add if you rather) happens, what’s the causes, etc? I’m particularly curious about well-defined meta analysis with reasonably accurate definitions
It's kind of embarrassing to the profession (of which I'm a member) that the diagnosis of ADHD isn't even made by neuropsych testing. It's made by assessing whether people have certain symptoms, generally by asking them whether they do. I'm sure people have tested groups diagnosed with ADHD and compared them to non-ADHD people on various things -- how well they remember under various circumstances, reaction time, etc. -- but whatever has been found must not be clearcut enough to make possible diagnosing ADHD via test results, the way you can certain learning disabilities. There are neuropsych subtests that measure executive function, but that's just one part of ADHD. I just spent a little time on Google Scholar looking for a decent article, and this looks pretty good, and not terribly technical.
This is not unique to ADHD or even to psychological conditions. Lots of medical conditions are only diagnosed by asking people how they feel and don't have tests to diagnose (or have bad tests that my find something but it doesn't prove causation or point to an effective treatment plan).
With ADHD specifically, the diagnosis depends on a lot of the negative impacts ones life. If you don't have the negative impacts there isn't much sense in diagnosing a illness as, to you, you aren't ill.
Yes, it's true that lots of psych diagnoses are made based on people's descriptions of what it's like being them and what hurts. Still, you would think ADHD would be measurable: If somebody says they cannot keep their attention on tasks for long, you'd expect them to do badly at tedious little neuropsych tests where you have to go through a whole page of little symbols and mark all the ones that have a horizontal line in them. If they say they can't take in info from lectures in college, you'd expect them to do badly on a test where the neuropsychologist reads them a coupla pages about something sort of dull, then asks them questions about what he just read them. That's why I say it's a bit embarrassing that there is no neuropsych test for ADHD. Makes you wonder how valid the diagnosis is. I'm sure some people have a genuine wiring problem that interferes with attention, but the interview method of diagnosis may be picking up a bunch of people who have a different problem. For instance, maybe somebody in college got coerced into picking a major they hate, and that's why they can't keep their mind on the lectures and the textbook -- it's boring info to them, plus they are furious about having been coerced by their parents or whoever.
And actually, with medical diagnoses there are other things involved besides people's descriptions of symptoms. There are tests that allow you to see, or see evidence of, the actual malfunction that is causing the symptoms. Actually, I think there are only a few medical diagnoses where there is no way to test for or look at the malfunction: Things like Chronic Fatigue, for instance. And with those things many docs question whether they are actual distinctive diseases, precisely because tests don't reveal anything unique going on in the body that would account for the symptoms.
From my perspective (which should be treated as not competent by any means) the way psychiatry generally works is that it tries to help people overcome stuff in their life, not to provide you with maximally accurate information about the stuff that’s going on within their heads.
e.g. in your case even if someone is failing badly in academia because they hate the subject or whatever else the reason is, they still might wanna be treated like someone with adhd because intrinsically they value ability to finish a degree much more than doing stuff that’s not boring. I’m almost convinced that that’s the best way to look at all psyche disorders.
That being said, you sort of want to improve one’s state of mind as they ask you to do, not to sort them into small boxes where you can see statistically significant differences.
And that’s mainly the reason to claim that diagnosis is in fact irrelevant to clinical practice in a sense that you shouldn’t deprive someone of, say, drugs/therapy on the basis of not being diagnosed with adhd
Well, I do not find diagnoses of much use, generally. They're very imprecise, and probably do not line up with real categories. What I find most useful is to get a fix on where somebody gets stuck -- what's the thing most other people can do that they can't? I look for the bottleneck. In the case of the person who does not have ADHD I think I would quickly realize the the person hates their major and is furious that their rich grandpa was willing to pay for college only if they agreed to be pre-med -- because I'd be asking the person questions like, "so exactly what is it like for you sitting in a lecture on biology? And pretty soon he'd tell me that it's boring and irritating because he has never liked science, he likes history and literature -- or whatever. And then the whole grandpa story would come out. The bottleneck is that the guy hates his major and hates having his grandfather dictate it, and that naturally enough makes it hard for him to pay attention to his course work. Of course he still might think that his basic problem is ADHD, and of course that is possible -- he could have a coercive grandpa, a major he hates *and also* ADHD.
I don't agree that I should just go ahead and treat him for ADHD if that's what he wants. For one thing, it's important for somebody who suspects they have ADHD to know whether they do or not. It's bad for you to think you have a disorder you don't have, and also bad to think you don't have one that you do. And besides presence/absence of ADHD might be relevant to some life choices of his. For instance ADHD is an advantage in at least one medical field: emergency medicine. ADHD people are sort of hyper, which is a good thing if you are racing around trying to save a life, and EM docs report that the high stakes and excitement keep them focused. Also, Adderall is probably not very good for you, and shouldn't become part of your daily regimen unless you really need it. Plus a certain fraction of people who take it become abusers, so it has its risks. So I would try my best to figure out how ADHD-ish the is. With some people it's really obvious because they jump from topic to topic while talking, and don't really grasp things I say because they cant pay attention well. It's also useful to ask what it's like for the guy if he's reading a book or watching a video about subjects that do interest him. Can he keep his attention on it? And neuropsych testing can at least measure executive function, which is relevant to one aspect of ADHD.
I think if the guy does not have ADHD what I would do is help him clarify what's really getting to him --- coercive grandpa -- then explore options. Is he sure grandpa would really pull the plug if he changed majors? Is it worth writing his grandfather a long letter making the case for changing majors? Is it worth paying for college with loans? And if the guy decides he has to stay with premed, I would offer to help him keep up with his academic work via self-management strategies.
Also, I beg your pardon but I missed couple of points you made and I think they are interesting to discuss
You said that it’s bad for you to think you have a disorder you don’t have and vice versa
But it sounds sort of weird from a clinical standpoint
I mean, I’m totally not an expert by any means, but it seems like adhd medicine generally helps with symptoms of adhd whether a person has adhd or not. Or, to be more precise, there is a sort of broad way of what symptoms a person with adhd may experience, and some of those might be experienced by someone without adhd, and thus those symptoms might be suppressed (is it a right word?) by using the same medication
If I’m mistaken in that, please explain why, but I’m not sure that having adhd is a discrete thing like “having an aids” where there are treatments that do help people with aids and don’t work for anyone else suffering from similar (but not caused by hiv) symptoms
However, back to your point — it (very much) does make sense to help people understand what’s happening with them regarding their consciousness, self-awareness, attention and internal problems. So I guess I’d put it that way.
It makes sense to understand what you experience and why and not be mistaken in that, or you’ll be confused about what to do and where to start, or even worse, you’ll be trying to do something that’ll make your situation worse (as people often do when they have false beliefs about important things in their lives)
But if you are aware of what’s happening with you in terms of internal experiences, then it doesn’t really matter whether you think you have adhd while not having one or other way around. (It’s still pretty much the same thing about being unhappy with some known problem that’s living within your brain, and willing to overcome it, using drugs or not)
Sure, there are some risks involved in prescribing drugs from adhd to someone without one, but I’m not saying that it’s necessary to give anyone who struggles with concentration adderall. What I’m saying is, sort of, like “we wanna optimise people desire-satisfaction as long as it doesn’t hurt them without their consent and doesn’t hurt others. E.g. if someone comes to doctor and asks them for a help with overcoming being bad in academia, reasonable(but not widely, if at all, recognised) guideline seems to be somewhat like the following
1) gather as much info about one’s condition
2) provide a person with some educated ideas of how to fix whatever it is that bothers them while also explaining risks
3) if the person in question says “I wanna some medication to be able to concentrate more easily, and I’m aware of risks but still wanna give it a shot”, then you probably should say ok
Also, worth mentioning that what you suggested is also a good way to treat a patient initially — trying to help him without getting him addicted to serious drugs is what I hope doctors do and I hope they succeed :)
I'm pretty sure there are at least two forms of ADHD with radically different sources; I notice two distinct clusters of people, one of whom finds it impossible to create habits, and one of whom finds it impossible to break them.
I have no clue if this is the sort of thing that you're looking for, but I've heard that there's pretty good evidence that exposure to "screen time" (video games, tv, etc) at a young age increases prevalence of ADHD. If you are curious and want a citation I can ask my source about it.
I was corrected that this is a longitudinal study about "attention problems," not ADHD directly. That said, I'm told it was published in a good journal, and is a solid paper. That said, there is field debate on the topic.
Thanks, but it’s not quite what I’ve been looking for
Article you suggested is mostly about genetic correlation with little to no explanation of what happens in a brain (there was a point about dopamine transfer related genes, but it wasn’t elaborated as far as I can see)
I want to find some paper on internal mechanisms that cause adhd, say, lack of some neurotransmitters or so which can be sort of linked to day-to-day experiences
With almost no exaggeration, one can say that there is no explanation of the "internal mechanisms" of ANY complex human behavior. Neuroscience is a very primitive and useless endeavor. The best we can do in terms of causal explanation is that differences between humans are generally due to genetic variation + random noise.
It’s sort of a meaningless answer, not to say misleading, don’t you think?
Some things are generally caused by environmental factors, say, ptsd. And thou there is definitely a place for genetic factors that contribute to likelihood of having a ptsd, going to war is by far more decisive factor, at least it seems that way to me. Also, let me rephrase. I’d like to have some way to think about adhd, intuition, if you please, and I lack any ideas of where to look for it.
But war doesn't cause PTSD. There are no ancient accounts of soldiers getting PTSD, even though a far bigger percentage of the ancient male population went to war. Whatever is causing PTSD among modern soldiers, it seems to be a curse of modernity and not an inherent feature of war.
In the general population, war is surely not a major cause of PTSD. Moreover, even in veterans, exposure to war is not a mechanistic explanation of the disorder--it is far from a sufficient cause and it tells us nothing about what happens in the brain. "War" as a cause of PTSD is not any more specific than the fact that PTSD has substantial genetic heritability.
I mean, it definitely increases your chances of getting one, but it’s not a major cause of ptsd worldwide (supposedly)
I don’t wanna bet on what explains better that you have ptsd — being on war, or being genetically likely to have one, but I’d guess those are comparable probabilities at least
Nonetheless, I still want to grab some intuitive sense of what adhd is and how it develops and which factors are known to be crucial, how does it affect internal experiences and/or chemistry of one’s brains, etc, etc
I’m pretty much sure that there is more evidence than simply “okay we tracked some genes that are somewhat correlated, thus explaining what people have adhd better than a coinflip”
I just listened to RFK Junior on Joe Rogan. Does anyone know of a serious debunking of him? Sorry, I am too lazy to even do a web search. Just curious if anyone knows one offhand.
As far as "this one epic burn on Twitter will *disqualify* him from being a serious presidential candidate": no. That type of thing doesn't happen. Stop imagining.
As far as "he says a lot of things that are wrong": of course he does. the problem is that, like Trump, there is so much you can't keep it all straight. The latest on Twitter, with ample debunking: https://twitter.com/MLiamMcCollum/status/1670442514778820610
As far as the broader sense of a debunking that encompasses "if David Sacks can explain how he misspoke, it isn't really a lie": once again, no. The truth cannot help those who hold "most things he say are similar to the truth" as their standard.
I just came across this Twitter thread from yesterday specifically doing a rundown of the vaccine-related claims RFK, Jr made specifically in his Joe Rogan appearance.
I myself yesterday posted in a sub-thread of this thread about RFK, Jr's reputation for being either willfully misleading or, seemingly more likely, honestly-conspiracy-minded* in his first foray into anti-vaccine advocacy, namely his articles in 2005 promoting the purported (and consensus-wisdom-to-have-been-debunked) link between the vaccine preservative thiomersal and autism -- https://astralcodexten.substack.com/p/open-thread-281/comment/17470692.
* "Honestly-conspiracy-minded" = so sure that thiomersal as a preservative in vaccines has been a public health catastrophe that insiders long knew about in the same way tetra-ethyl-lead as an "anti-knock" additive in gasoline was, he reads things like transcripts of CDC meetings devoted to the thiomersal issue with a highly biased prior. Indeed, he seems so sure these transcripts are going to be one of those cases where arrogant insiders blithely self-incriminate! And then, worse, RFK, Jr quotes those transcripts in his writings with substantial edits that strip key passages of context that most reasonable observers would think make quite clear that these CDC transcripts aren't those kind of smoking gun arrogant-insiders-incriminate-themselves thingies (see links in my comment for detailed examples).
Whacky. It's not like I'm a follower of this University of Illinois - Chicago cardiologist Francis Han, MD (@han_francis). Hopefully this will be viewable for you:
Good luck! Cross too far into taboo territory and most "Rationalists" just start muttering about how they don't even need to answer the question because it's so absurd.
Haven't looked at the substance, but I'll say it's not at all easy to discredit/disprove what a crank says and point you towards Brandolini's Law (The amount of energy needed to refute bullshit is an order of magnitude bigger than that needed to produce it. https://en.wikipedia.org/wiki/Brandolini%27s_law).
Okay, I stand corrected. If he is a crank, it should be possible to show that he misinterpreted the studies he cited, or that they are flawed, but not necessarily easy.
How so? He can cite studies, but that alone doesn’t show that they are valid studies, or that he has not misinterpreted them. What procedure are you suggesting?
Maybe I should clarify what I consider serious. I am looking for someone like Scott taking the time to go through the various scientific papers RFK Junior mentions and point out which ones have problems which ones are fine but he misinterprets and where if it anywhere he actually has a point. I am not worried about what he thinks about Venezuela, or putting people in jail. I am interested whether his claim to interpret serious empirical Science is on track or off.
I have a couple things I can point out: (1) The idea that pushing the vaccine on people who had already had covid was insane. I agree that those people need not have been vaccinated, but there are good arguments against saying this was an insane step. For one thing, many people would not known for sure whether they had had the virus. And some people who objected to getting vaccinated would have lied and said they had had the virus when they had not. You can test for antibodies to see whether somebody has had the virus, but the antibodies from it fade, and a few months after having the virus the test will show invalid negative results. Plus, of course, testing everybody who said they had already had the virus would have been a huge expensive hassle. And given that the vaccine was safe, seems to me a reasonable decision to just vax as many people as possible.
About harm done to young men by insisting that they get the vax. There is no doubt that vax increases the risk of myocarditis and pericarditis among you males (not among young women). However (1) the risk is still very low; (2) it is a mild a treatable kind of myocarditis that does not do permanent damage; (3) at least one study has found that the risk of myocarditis after covid for young males is higher than the risk after the vax; (4) vaccinating these young guys probably saved some lives by reducing covid transmission and fatalities among older folks.
I am passing on this info because someone asked for info. I believe it is accurate, but am not absolutely certain. I am not interested in having an argument about covid issues. I have burned out on that. If any of the above is inaccurate, dig up the data that shows that and give it to OP.
So what? The OP asked for a "serious debunking", in his next comment he clarified what he meant, and you replied "Vinay Prasad has done it now". So I was just borrowing the word.
The post you linked to was not a careful consideration of the scientific literature anyway. Most of it was political in nature and about corporate interests and potential corruption.
What I mean is that one should not expect Vinay to do debunking because that's not what he does.
I really think that rationalists shouldn't do it either or at least very sparingly. It polarizes the discourse and assumes bad faith in the opponent and makes everything more tribal.
For example, someone can say that covid vaccines killed a lot of people. It is bad to make such proclamations without evidence. And yet, I simply take this as someone is saying “maybe vaccines killed a lot of people”. Now it sounds like rumination. And I can answer it myself – I haven't seen any strong or weak evidence that covid vaccines killed a lot of people. Only some rare anecdotes but it expected (for example, very rare anaphylaxis, myocarditis etc.).
I don't need to debunk, merely see what evidence is there. Some could even say that extraordinary claims require extraordinary evidence.
Now, if the opponent is arguing in good faith, he will try to provide more evidence. However, if it is done in bad faith, at this point he will start spouting nonsense or some conspiracy how every data point is controlled by nefarious organizations or will show that he simply doesn't understand how this all works at all.
Prasad does the best job I have yet seen, but it is still a bit shallow. Kennedy cited studies and made claims. I still don’t know if he was misinterpreting those studies, or citing flawed or biased research, or accurately describing the emperor’s new wardrobe.
It's worth remembering that RFK, Jr. has been a public anti-vaccine advocate since 2005, with his first foray into public anti-vaccine advocacy being his advocacy for the contention that thiomerosal, an antiseptic and antifungal organomercury compound used as a preservative in a few US vaccines up until 2001, was to blame for many cases of autism.
It's worth remembering this since the allegations against RFK, Jr in that debate fundamentally were *not* about the *scientific understanding or lack thereof* he evinced in the thiomersal-causes-autism-yay-or-nay debate back then. Moreover, here's a key point of context for RFK, Jr's participation in the debate back then. By the time he got involved, the debate wasn't one of the sort where a tiny band of plucky contrarians stood against an establishment that wasn't doing anything. Rather, when RFK, Jr got involved in the thiomersal-causes-autism-yay-or-nay debate in the mid-2000's, he was getting involved in a debate that had been roiling for years *precisely because the establishment had just done something big.* Namely, in 1999 the American Academy of Pediatrics made a recommendation quickly seconded by the CDC making a more pointed-and-decisive recommendation that vaccine manufacturers remove thiomersal from all US vaccines "out of an abundance of caution".
[[ERRATUM 2023-06-19, 10:32pm EDT: The recommendation to remove thiomersal from all vaccines given to children and adolescents. The only vaccine since 2001 given in the USA that still uses thiomersal is *certain* flu vaccines, *and even then only when distributed in multi-dose vials, i.e., multiple sterile syringes will draw the vaccine dose from a single vial.* Single-dose vials are widely available for those flu vaccines and are recommended for children, adolescents, and pregnant women even though the risk of "mercury" to any of these groups generally is tied to the much-more-potent organomercury compound *methyl-mercury* whereas thiomersal leads to *ethyl-mercury* in the body. ]]
**Thus, we get to the crux of the matter. The two main allegations against RFK, Jr in the thiomersal-autism debate are that his public advocacy was either straight-up knowingly misleading, or it was honestly misleading since he had already become a conspiracy theorist in the bad, stereotypical meaning of the term. Specifically, two main allegations against him are:
(1) He quoted minutes of CDC meetings in highly-edited, inflammatory, and misleading ways.
(2) Moreover, he repeatedly made statements that showed either surprising ignorance or already-down-the-rabbit-hole-conspiracy-theorist-disbelief that vaccine manufacturers had indeed removed all mercury containing compounds from US vaccines by 2001.**
This 2005-era history with RFK, Jr's first foray into anti-vaccine advocacy is recounted in the following *2017* article that first ran in STAT and then merely got reprinted in Scientific American. It has many detailed side-by-side examples to support the allegation (1) of misleading, selective editing. (SIDENOTE: I bring up the fact Scientific American was merely reprinting the article since Scientific American during the Trump era got flak for having portions of it edited by quite culture-war-lefty types... though even here, note again that this is 2017... so, yes, it's in the Trump era and being reprinted by SciAm with a news angle of anger against the Trump Administration for placing RFK, Jr on a scientific advisory panel for vaccines... but no,this is long before COVID and the Trump-Administration-vs-"Follow the Science" saga attaining white-hot intensity.)
The other angle present in RFK's first foray into public anti-vaccine advocacy with the thiomersal-autism link --- though I don't know how much RFK, Jr either had knowledge of it at the time or can be plausibly expected to have had it --- is that 2 experts he quoted heavily, father and son geneticists Mark and David Geier, turned out to be grifting quacks... including in the opinion of courts of law. (Some of this is mentioned in passing in the above SciAm reprint of the STAT article, other can be seen --- sorry for my link laziness --- via Wikipedia https://en.wikipedia.org/wiki/Mark_Geier ).
P.S. Speaking of link laziness, I know there are allegations but I don't have time this Father's Day to find them that RFK, Jr. has repeatedly used various statistics poorly... and I mean "statistics" and "poorly" in what *politicians* think of as statistics --- i.e., quantitative gotchas --- rather than he does legitimate statistical analysis, just poorly because he's a lawyer not a statistician. Two examples of these are:
1) Inflating the number of vaccines Americans get by age 18 with the annual *flu* vaccines recommended by the CDC for teens and preteens (which is considered inflating the numbers since, again, RFK, Jr for much of his anti-vax advocacy career has been focused on how vaccines *early in childhood* purportedly caused autism). [[NB: See ERRATUM above about how certain flu vaccines, when packaged in multi-dose vials do indeed still contain thiomersal, though single-dose vials that do not are available for all of them. This makes it somewhat more understandable why RFK, Jr focuses often on flu vaccines, even though it still is a "red herring"]]]
2) Conveying the impression that the rise in the percentage of Americans under 18 having "chronic medical conditions" since the time he was a child isn't simply both the straight-up rise in *obesity* (as well as, secondarily, public-health-stats-redefinition-of-BMI-thresholds-for-obesity).
I'm sorry that I can't point to a more definitive set of references for all the iffy things in RFK, Jr's anti-vaccine advocacy to date, but I do urge people to remember that his advocacy is not confined to the COVID era... as well as I do urge people not to sorta devolve into the enemy-of-my-enemy-is-my-friend logic (at least not too early in the process). The are worthy debates to be had on whether (1) school closures were a net negative and whether that was amply knowable in prospect and (2) anti-COVID public health policy, including vaccination, should've been much more defined by the steep dependence in mortality and morbidity with rising age. However, those debates are really a *tiny* portion of RFK, Jr's contrarianism on public health.
P.S. Happy Fathers' Day to those who are celebrating it.
P.P.S. Mentally healthy Fathers' Day to those who are having to cope with fathers who are gone.
Most of this is about his Rolling Stone article (of 2005?). Neither the article nor the errata are findable online by me. On Rogan he claimed that the errors had been introduced by his editors at RS, when they shortened the article considerably. I don’t know how plausible that is.
The criticisms I am seeing, when not pure ad hominem, seem to be about him misquoting or summarizing badly a meeting transcript. my main interest is did he misinterpret the science he cited, or is it junk, or what. Your link to twitter seemed likely to address this more directly. Unfortunately I seem not to be able to access it.
I don’t find his claims very plausible. But I don’t think I have much relevant evidence, either. It is bad that such a prominent public figure is spouting nonsense, and perhaps trying to paint his opponents in a bad light. But it is also bad that it has become so difficult to try to find out what is wrong with his garbage. I guess I just need to be less lazy, listen to Rogan again, taking notes this time, look up any articles mentioned, look for summaries or critiques, and make up my mind.
Ah... indeed. This post from Retraction Watch all the way back from 2011 might help explain --- well, at least give some flavor of the context --- why Rolling Stone and Salon disappeared the 2005 RFK, Jr article "Deadly Immunity":
But worry not! The article "Deadly Immunity" is still available here based on how it was entered into the Congressional Record by Rep. Dan Burton (R-Indiana, 5th, previously 6th District over the years 1983-2013):
As you probably were auto-notified, I just made a separate reply in the main reply thread to your comment because I found a Twitter thread that explicitly is a rundown of RFK, Jr's vaccine-related claims in his Rogan appearance.
That Twitter thread led me to add an erratum to my comment in this reply-subthread about RFK, Jr's 2005 entry into anti-vaccine public advocacy centering on the purported thiomersal-autism link.
To be honest, I have a soft spot for RFK, Jr., in two ways in fact:
1) First of all, I was more of a lefty in that 2005-era and listened to the brief attempt of leftier-Democratic-Party partisans to compete with right-wing AM talk radio: Air America Radio, which often had RFK, Jr on in 2004 and 2005. Indeed, my one bit of bona fide OG blogger reporting was on one of RFK, Jr's *other* highly-biased-prior-leads-to-conspiracy-theory-advocacy thingies: his contention that the vote in Ohio, especially, in the 2004 Presidential Election between George W. Bush and John Kerry was clearly tampered with... with a lot of that clarity he felt coming from him looking at "raw" exit polls (like several million other online Democratic partisans back in 2004). As major blast-from-the-past, I see my old Wordpress blog is still up and this was the key post:
though, alas, my tour-de-force (if I do say so myself) of data visualization... that is my one scatterplot of raw exit polls versus official margin by state so as to show a systematic exit polling bias across all states, not just swing states seems lost to the ages (even on the Internet Archive despite claiming it captured it 13 times... weird).
2) If anybody has a reason to have a biased prior toward grand conspiracies it's RFK, Jr. At age 9, he could personally witness his dad RFK, Sr ask then director of the CIA (John McCone... a quasi-trusted ally of the family explicitly put in by the Kennedy brothers after they forced out Allen Dulles in the wake of the Bay of Pigs debacle) to not only come to the family house within 24 hours of JFK's assassination, but to swear on the grave of his (McCone's) mother that he (again, McCone) had no knowledge of CIA involvement in JFK's assassination. And, over the next 5 years from age 9-14, RFK, Jr could continue to witness literally "up close and personal" his dad never give up that theory, largely because McCone caveated his denial of CIA involvement with a "to the best of his knowledge" type caveat.
These are not really "debunkings" (certainly not systematic ones, although they push back against at least a few of his claims), but they are two pretty reasonable critiques of him:
If it was that vaccine mandates were abuse of power then he is right. Nothing to debunk.
If it is that vaccines caused more deaths than covid, then he is clearly wrong. Nothing to debunk either because we do not see people dying massively in Sweden that has good vaccination rates but had less strict lockdown measures. Current excess mortality is probably more related to lockdowns and other measures that prevented people to access healthcare during last 2-3 years.
We know Covid has bad impacts on the heart, say, in some cases causing lasting problems. I don’t see why that shouldn’t also be part of the cause of ongoing excess deaths, especially as much of the excess is specifically in cardiac disease. Of course being sedentary in lockdown can’t have helped either but I don’t see why Covid itself shouldn’t play a part.
It was accused not insisting on stronger lockdowns and not closing schools, letting covid spread. At the end it maybe had about the same level of infected people, that is, close to 100% but for the sake of discussion let's assume that the same proportion of infected during pre-vaccine period.
It has no significant excess mortality unlike the UK. Completely destroys the idea that previous covid infections cause excess mortality today.
Fair response, and I didn't know that, I've just looked at the stats. Although it's also notable that a number of countries that did have strict lockdowns have even lower excess mortality than Sweden, so that doesn't really support your case either. Also Sweden is still running at about 3% which you need to explain given it can't be lockdown (which is your whole argument) - it seems to me that sort of amount is likely attributable to Covid, which is about what you'd expect given what we know Covid does to the body including the heart. In fact if you look at cardiac impact, it would be surprising if there *wasn't* some sort of increase. And that's ignoring all the other impacts (lung, brain, etc).
I'm on medical Twitter, and there was mention of his promoting some false claim about a covid-related issue -- may have been that he was talking up some vaccine-related fear that has been debunked by science. I did not look into the matter and don't know how well the complaint would hold up if you did.
There's all kinds of debunking all over Twitter. There's all kinds of debunking of all things all over Twitter, though -- just goes to show that one side of an argument doesn't tell you anything.
Having two people debate doesn't necessarily tell you anything either, though. Other than which side has hired the better debater.
To be clear, I have no opinion on any other candidates or their platforms, I just thought it would be useful info since people who read ACX have a tendency to be excited about Land Value Tax and this would be a low effort, high-ish leverage opportunity to advance that cause.
I am not against polygenic scoring in principle, but I think that
- they should only be done if IVF is being performed for an unrelated reason anyway, because in my opinion the small increase in risk associated with IVF is not usually outweighed by the benefit of polygenic embryo selection.
- The benefit of selecting from a reasonable number of embryos, such as 6, is usually very small. But of course if the parents are carriers of a Mendelian genetic disease or a very high risk factor such as some BRA1 mutations, then the benefit is potentially very high.
Some commentators have suggested that because our knowledge is currently imperfect, we risk having less healthy babies if we use selection. I think this is highly unlikely, partly because genetic correlations tend to be positive: mutations that increase one risk tend, on average, to increase other risks as well. There are many documented cases of the opposite situation (the famous malaria protection mutation, etc) but... the trend is very clearly in the other direction.
Similarly, the risk of losing genetic variation at the population level is, I think, minuscule. Most variation is deleterious or neutral variation, and worrying about future evolution seems silly to me (not the right timescale!).
One can think of reasons to prefer rolling a die. A die is unbiased. If IVF becomes very commonplace and everyone is using polygenic scores to select embryos, what if we lose a big part of the mental diversity of the human race? What if parents use polygenic screening not for physical and mental wellness, but for personality traits? Parents may want docile children and societies may want obedient citizens, but if we genetically eliminate the "difficult" mavericks from the gene pool, that may not be good for the long-term future of humanity.
I agree that is an important question, but given the context of the original question, it seems more ethical to select a child that will be mentally and physically healthy. There are plenty of collective active problems related to genetic enhancement, but I think parents would be behaving ethically if they decide to improve their child's welfare. If they decide to maximize obedience at the expense of wellbeing, it would be unethical in my view.
What if mental health, as indicated by this polygenic score, is correlated with reduced creativity and reduced potential for scientific genius? Certainly mild autism seems to be correlated with aptitude for the maths and sciences. Also, what if obedience is positively correlated with well-being, as seems likely? An obedient kid who listens to his teachers and never breaks any laws is more likely to be successful than an unruly kid who doesn't do his homework and gets in trouble with the law. But if the whole world is turned into docile sheeple, where would we get the moral visionaries, the political revolutionaries, the titans of new industries (think Elon Musk), and the bold adventurers?
Why isn’t it a woman right to choose, in this instance, anyway? The woman is bringing a child to term, why is the state interfering on what the child could be.
Because the woman isn't the owner of other human beings, and shouldn't get to dictate the future of humanity. Society's interests are implicated every time someone uses technology to selectively alter the future gene pool.
I believe the government has the the legitimate right to regulate abortions, which doesn't mean that banning all abortions is good policy. The concerns are also different: abortion bans are usually justified on the basis of the embryo's right to life, while my concern with selective IVF is its effect on the future gene pool.
Abortion is probably effecting the future gene pool to a certain extent. At the moment the technology is nowhere near where it could affect much at all, and if we can eliminate deleterious mutations then the effect could be benign.
I’m not saying the state should decide. I’m asking you to consider what prospective parents should consider, and the question is targeted at those who criticize the use of PGT-P.
Fascinating question, in that even asking in the first place shows a departure from the kind of ethics where this could matter. The internally-consistent position is to refuse to do IVF (as counseled by e.g. the Catholic Church). Beyond any arguments about 'playing God' or 'engaging in eugenics', if you're creating 6 embryos, how many of those are you going to carry to term? All six? If not, what happens to the others? If embryos matter, you already dun goofed. If they don't matter (if you've swallowed the camel) why quibble about screening (straining out the gnat)?
On a macro scale, selecting out known risk genes could have unintended consequences. Since genes tend to be inherited in 'clusters' rather than independently, selecting out a known risk gene will also select out nearby genes that correlate with it. This may be an acceptable tradeoff in the case of true genetic diseases, but becomes much more questionable when you're looking at modest genetic risk factors. There are a number of gene-linked conditions (notably autism, schizophrenia spectrum disorders, bipolar) that are associated with desirable or at least interesting traits in family members of patients and/or in healthy people with risk genes.
There may also be direct unintended consequences of selecting out risk genes for diseases of modern civilization. I'd be especially concerned about any gene involved in immune function because those were major selection targets in the ancestral environment. People with risk genes for e.g. autoimmune disorders may be important reservoirs of resilience against specific kinds of pathogens, which may become more important if we continue falling behind in the antibiotic resistance arms race.
On a micro scale, IVF has risks. Even if you think it should be available to people who can't conceive a healthy child otherwise, you may well think that it's not a good idea for healthy, fertile couples to try to use it to make perfect babies, because on net the risk likely outweighs the reward.
Outside of reproductive ethics, there are reasons to be concerned about social trends that are leading opposite-sex prospective parents to be increasingly unable or unwilling to reproduce naturally. I don't necessarily share these concerns enough to make an argument for them, but for at least a century now, people who are generally positive about technology have been raising questions about the wisdom of a sterile, touchless, isolated society. Asimov's *The Naked Sun* has been on my mind in the COVID era, but there are others.
One might also question the wisdom of encouraging people to condition their willingness to have a child on the child being genetically 'perfect.' We're already at a point in Western cultures where people are afraid to have kids unless they can provide the 'perfect' environment, and it's crashing birth rates despite minimal changes in the number of children people say they want. How much worse would it get if competitive parenting extended to the genetic realm?
I don't think it's possible to be coherently pro-choice and against embryo selection for people who are getting IVF anyway, or IVF+selection for people who are known carriers of a genetic disease and would selectively abort affected embryos if they conceived naturally.
I do think it's possible to be pro-choice and think expanding the use of IVF in healthy hetero couples *for the purpose of* embryo selection is probably not a great thing for humanity.
Transferring an embryo results in live birth about 40-50% of the time so it seems like 6 embryos is about 2.4-3 live children. Doesn't seem unreasonable to me?
Again, you're not thinking like somebody who begins with the assumption that embryos matter. Imagine that you want to adopt an orphan group from some other country. You only want three, so you put in an order for six due to the fact that approximately half of them die on the slow boat over. Your other option is to just not do that and remain childless.
But I believe this is in line with the general fraction of fertilized eggs that become live births? So anyone choosing to have children faces this same moral dilemma.
I can see why you would think that, but in the minds of those who care deeply about this topic, there is an unbridged gulf between you taking it upon yourself to do things that jeopardize human lives versus nature and the hand of the creator doing something analogous.
Interesting. In Poland for instance, due to (what I perceive to be) Catholic sensibilities, unwanted frozen embryos are "adopted" to people who are willing to carry them to term (and raise the baby of course). This law is new and requires a 20-year waiting period so currently it's highly theoretical. But that does not align with the principle you propose, since the less-risky thing to do would be to leave them frozen.
It's okay that you haven't heard this before. You don't have to pretend like I'm theorizing about hypothetical people. Pro Life is a big tent and I specified a certain subset of Catholics who care about the church teaching on the matter. They care because of the general practice involves making far too many and destroying the extra. And I would wager that those in the know could effectively communicate this to their less wonkishly plugged in pro life brethren, for instance, while standing around on the street corner holding signs in just a few words. "They make 10 babies, choose two, and put the rest in the furnace."
Like many, I don’t oppose polygenic screening because it isn’t random; that’s a pretty nonsensical weak- or straw-man of the position, I think: not sure if anybody really thinks that, maybe somebody does.
I do oppose it for two reasons, one fairly weak and entirely pragmatic, one strong and morally rooted. The latter is simple: it requires IVF which involves discarding viable embryos and I think there’s zero excuse for conceiving and discarding a child. In my long-considered opinion there is no more logical place to draw the line against murder than conception.
That aside, since I don’t think that argument will get much traction here, the weaker and pragmatic objection is that we *really* do not know enough about genetics to successfully select and not create systemic errors, and we almost certainly will end up selecting against protective or beneficial traits, and for negative traits, while attempting to select what we want. A sufficiently unsophisticated program, which is what we’d have with our current level of understanding, could end up making the race on average stupider, weaker, or savants (not a good thing!), or prone to other terrifically disadvantageous conditions we don’t even know about. We should really understand more about ourselves first. And once we do, this objection may disappear (may!).
Take CF for example. Aside from the absolutely most typical genotypes that cause it, we have absolutely no clue why certain mutations are expressed phenotypically and certain aren’t. Up until a few years ago it was ‘very clear’ (and the advocacy foundations still insist) that anyone with two CF-causing mutations definitely would have CF. Now it turns out that many ‘CF-causing’ mutations actually don’t: they cause positive sweat tests and aberrant newborn screenings yet no symptoms and no illness at all. What gives? If one used their genotype to determine definitive CF status as we used to, and as some physicians still do, those kids might be selected for destruction… which would have overall weakened the gene pool because they are carrying some sort of unidentified protective mutations as well. And given how very very many of these mutations there are, that could have meant a significant rise in real classical nasty CF, the death-sentence-without-interventions kind. Food for thought.
One other thing: I'm pretty sure that IVF increases the chances of various bad things -- miscarriage, kids having various problems. Though I suppose that it's possible that the actual cause of these things is the problem that made the parents need to resort to IVF in the first place, rather than the procedure itself. But if IVF actually reduces chance that pregnancy and child are healthy, then seems like a terrible idea to use IVF only to make polygenetic screenig possible.
Yes, the health implications of IVF are significant. For obvious reasons it increases the risk of miscarriage, but for as far as I know unknown reasons it profoundly increases the risk of serious birth defects.
Modern sanitation, medicine, welfare policies etc. almost certainly select against "protective or beneficial traits". The argument that artificial selection, which unlike the above is explicitly designed to be beneficial, would be especially harmful to the gene pool seems bizarre to me. The status quo is not "random selection" but, in many respects, negative selection.
I don’t agree at all—there are not many strictly genetic heritable maladies that strongly selected in the past and do not strongly select now, and the evidence is that even terrifically maladaptive genetic disorders which kill almost immediately after birth can and do persist in the genome for many many millennia because they’re recessive. It is very unlikely that the conveniences of modern life are anywhere near as strongly selective as those in the animal world and even if they were it would take millions of years for very visible effects.
On the other hand *directly* selecting every birth each generation based on a common, poor schema could show effects within a few generations.
Evolution doesn't take millions of years. Our whole species is only about 200,000 years old. In the past, probably around half of people died before adulthood (https://ourworldindata.org/child-mortality-in-the-past), while in the developed countries today almost no one does. Moving from the "natural" setting where there is truncation selection of up to 50 percent each generation to the modern setting where there is essentially no selection along this dimension represents a huge change. Over generations, this has probably caused substantial shifts in some phenotypic mean values, depending on heritability.
More subtly, analyses of polygenic scores show that there is negative selection on genetic variants associated with e.g. educational attainment in modern populations: https://www.pnas.org/doi/10.1073/pnas.1612113114 Most traits are like educational attainment, i.e. affected by numerous alleles of small effect, and not due to a single gene. Embryo selection applies positive selection on naturally occurring genetic variation, in contrast to the negative selection that happens otherwise.
Many critics of polygenic screening (e.g. Adam Rutherford, Forzano et al. (2021), Polyakov et al. (2022), etc.) leverage criticism against polygenic screening without proposing alternatives and do not critique the use of visual embryo inspection or couples randomly selecting embryos. This motivates my question: What method do they want couples to use?
This question is framed as couples already deciding to implant and embryo and discard the rest. The embryos will be discarded regardless of the method they use to select the one to implant during IVF.
Even if our understanding of the genetic architecture of humans is not 100%, it seems odd to prefer total randomness to a partial understanding. If randomness is morally better because the health index would lead us astray and create a bad outcome, why would it be morally permissible to allow the bad outcome through randomness? (this is the point I'm getting at with the second question.)
Correct me if I'm wrong, but I *think* the pro-randomness argument is that if there were only going to be one genetically-screened baby in all the world, then you might as well screen based on your best guess, but if there are going to be many babies screened with the same criteria then your selections are now *correlated* across many children, which causes a reduction in genetic diversity.
All else being equal, a more genetically-diverse population is healthier, so in order for screening to be worth doing on a *population* level, it has to provide enough benefit to outweigh the loss of diversity. It's possible it could provide some marginal advantage at an individual level but still be bad at a population level because of diversity concerns.
(In principle, you could do whole-population screening in such a way that it would maintain or even slightly increase genetic diversity (compared to random selection), but that would require massive coordination and would funge against other goals.)
Does healthy mean if an epidemic hits you're likely to have some resistant folk in the population?
Or does it mean robust in the *absence* of pathogens?
If some people in your population are heterozygous for sickle cell anemia is your population 'healthier' in your estimation?
Arguably, mutational load has been increasing in humans. Selective pressures haven't been strong enough. We don't talk about it much because eugenics causes mass murder and other problems, and there hasn't been a potentially ethical solution till now.
I think we're a very long way from 100% of the population using embryonic selection. Do you have a different view of embryonic selection if 5% of the population uses embryonic selection? How much value do you place on increasing genetic diversity through other methods?
My model is something like "genetic diversity is a resource that can absorb environmental shocks" or "genetic diversity is a hedge against unknown hazards". If something about the environment changes, then a genetically-diverse population is less likely to be wiped out than a less-genetically-diverse one, because the more-diverse population has more chances to get lucky. Diversity provides option value. (Perhaps "healthier" is not the ideal word for this.)
The words "all else being equal" were doing a lot of work in that claim about genetic diversity being good. This is definitely not the ONLY relevant consideration. I'd contend that many mutations are net harmful even if they increase diversity.
I mostly do not actually buy this argument against polygenic screening and am just steelmanning it in an attempt to improve clarity. I think screening could *in principle* make the population worse off but I'm far from being convinced that this is actually the case for the screening options that are likely to be available in the near-term.
I think that another good point in favor of a randomness mandate/ethical norm is that it avoids designer babies. While I agree that screening for genes highly correlated with debilitating diseases might outweight the loss of genetic diversity, I think that screening for less harmful genes (or even just cosmetic ones) runs much higher risks.
This isn't a slippery slope claim. It's more of a "draw the line in the sand far far away from what the problem region is" kind of claim.
Yes, I’m saying that randomness is actually pragmatically better than selection given our current level of knowledge. The rest follows directly: I’m basically saying your ‘health index’ is no good, irrelevant, certainly misleading.
But again, pragmatically better. I can’t speak for people who think IVF is moral yet embryo selection via polygenic screening is immoral. I don’t understand what ethics produces that result.
If we assume the health index is infallible, the pragmatic objection might disappear. On the other hand it might not: it’s impossible to foresee whether it is actually possible to select beneficially on a population level, and raises all kinds of weird questions about what is and is not a high quality genotype on an individual and population level.
edit: I will note that based on my understanding of Rutherford’s arguments summarized (I could be wrong), you are mischaracterizing him badly. He thinks that screening for polygenic traits is essentially a pseudoscientific swindle, and doesn’t suggest an alternative method because that would be demanding an alternative for hanging a dreamcatcher over the child’s bed: the alternative is you just don’t need to because it’s a dumb idea. He doesn’t think the technology is inherently immoral any more than weaving dreamcatchers is immoral. Selling them to parents while claiming they will prevent their child from getting cancer is the immoral part, with a side of slippery slope to other immoralities like so-called race science.
"Yes, I’m saying that randomness is actually pragmatically better than selection given our current level of knowledge. "
That seems odd as a blanket assertion. It may be that selecting for something controversial like intelligence will select for, say, alcoholism. Or that selecting for height will increase other health issues. Or that selecting for some genes linked to healthy metabolism might decrease lifespan. But some things like sickle cell anemia are well enough understood and so objectively harmful when homozygous that I'm skeptical that we can assert zero benefit. (Granted, we're not talking about polygenic testing in that case. I'm just saying that it's hard to make blanket assertions about understanding.)
If we're only talking about 2-3% of the population undergoing polygenic testing we don't have to be infallible. We just have to do better than random chance. Any outcome chosen by polygenic testing is still a choice that people make, normally, now, via IVF. If nothing else, we'll increase our understanding and refine the technology (if we can agree that having such working technology is eventually desirable.) The primary cost is the cost of the testing.
The benefits to improved metabolic health, especially, seem critical and not strongly ethically objectionable.
I think you’ve answered your own question: selection for strictly Mendelian traits is less pragmatically fraught than our extremely ignorant polygenic screenings.
But, as I said before in my example of CF, it is not clear that things we believe are strictly Mendelian actually are, and the practice of naively selecting against a ‘known-harmful’ mutation may actually lead to a greater prevalence of the exact malady you’re trying to select against!
I have yet to hear any informed counterarguments to this point and I’m beginning to doubt there are any. As Rutherford points out, many advocates think it’s a free lunch because they don’t understand the details—those that do understand the complexities are often counting on us having a much better understanding of our genome at some indefinite future point which will make selection easy and infallible.
Strict Mendelian inheritance isn't needed. It's a simplified model used to teach children, so it's an easy fall-back to clarify concepts. But yes, even many supposedly strict Mendelian traits exhibit incomplete penetrance.
If the case against polygenic screening has been made with any rigor, I'd be interested to see it. It's use for prevention of dibetes melitus already seems to be pretty strong.
"The AUC is a test which ‘describes the inherent validity of diagnostic tests’.4 It is presented as a value between 0.5 and 1.0, with 1.0 representing a perfect test, and 0.5 a test which is no better than chance. If a particular test has an AUC of 0.72, there is a 72% chance that it will be able to correctly identify the subject with disease when applied to select between one afflicted and one healthy participant.4 Generally, scores above 0.9 are deemed ‘outstanding’, 0.8–0.9 is ‘excellent’ and 0.7–0.8 is ‘acceptable’.5
One disease for which PS are providing promising results is type 1 diabetes mellitus (T1DM). According to a 2020 systematic review, the current most powerful PS for T1DM has an AUC of 0.92.6"
While granting that diagnosis is not the same as treatment, the argument that they're widely divergent should be made explicitly, not just through incredulity. Can an argument be made that paradoxical effects related to DM screening are of similar frequency and magnitude of harm?
I searched for Rutherford, but didn't find any arguments by him arguing, based on data, that PS was technically worthless, at present. If there is a peer reviewed publication like that could you post a link?
I myself do not object to using polygenetic health indexes to screen embryos. But if I did, I think I'd sidestep your question, which does seem to kind of checkmate those objecting to choosing the embryo with the highest health index. I'd say I objected to using the health index at all, because the value of a human life can't be summed up in a single score. And I'd add that once we start doing that sort of evaluation where do we stop? Do we screen toddlers before they go to nursery school or day care, and have a Perfecto Daycare open only for kids with no significant oddities or deficiencies in measures of motor coordination, vision, hearing, no language delay, no sleep problems, etc etc? (The rest can go to Rando Daycare.)
Later edit: Also I agree with Godoth that we may not know enough yet for these tests to be good guides to which embryo is most likely overall the have the best lifetime health. There are definitely good traits tied genetically to some health problems, and who knows whether that's rare of common. One example: people who have bipolar relatives but do not themselves have bipolar illness are unusually creative, as measured both by tests and life accomplishments. If we screen out bipolar we lose those people. I have a close friend who has a bipolar grandmother, and he is the most entertaining talker I have every met. Get him talking about almost anything and novel, clever thoughts about it just pour out of him. Get him riffing on the subject of something he thinks is ridiculous, and his weird putdowns and analogies and enactments will have you crying with laughter.
>"...the value of a human life can't be summed up in a single score."
I think I get what you're trying to say here, but unless you assert that all lives are equally valuable (a common position that I don't personally hold) the math doesn't work. We may not be able to construct the metric or accurately measure particular individuals' against it, but there must be a latent scalar value if you accept that, e.g., Einstein's life was more valuable than Pol Pot's.
Well as I said I don't personally object on moral grounds to polygenetic screening of embryos. (I do have some practical concerns -- like whether the harm of IVF outweighs the benefits of screening, and whether we know enough at this point to be sure we're not screening out highly desirable traits along with vvulnerability to certain illnesses.). So I'm sort of role-playing here. I have 2 responses -- they're entirely separate arguments. One is what makes you think there is such a thing as a value attached to a particular human life? Does fucking everything have a value number, even if we can't read it? Does the telephone pole outside my window have one? Do works of art have a value number, maybe not readable by us? The amount it sells for does not count. We're talking about artistic value. Does pizza have a value number? Does Rationalism have a value number? Does democracy? Does the planet? Do polio shots? Does the Pope? Who's worth more, Scott or Zvi? Gimme their numbers.
The other, separate argument would be that we can't see very far over the horizon, and events past the point where we are decent forecasters (5-10 years at most, I'd say) might call for qualities you'd never think of selecting for. Maybe there will be a plague way worse than covid, and the people who survive will be mostly those who happen to carry a certain gene. Maybe AI will get supersmart and go rogue, and as part of that influence everybody to go along with its wishes using some drug, to which only people with a certain gene are immune, and they will be able to save the world. Maybe there are wonderful, astonishing things that can be done with supersmart AI, but to be able to mind-meld with them and really absorb what the AI has to give you have to be highly hypnotizable, and the high-hypnotizables will be of huge importance. Or, to take a slightly different angle: A few times in my life I have seen sudden emergencies -- auto accidents and the like -- and most of the time there has been someone who rushes to help, who takes some risk, but not a foolish amount, and is smart about how he helps, who saves some people. I don't know what trait those people have, but I doubt it's correlated very much with intelligence or with low risk for various diseases, and it's a wonderful trait. These brave, decisive and caring people really are special, and in some circumstances having a good number of them on hand might make all the difference. Or here's another thought about a trait that might be important: Maybe we're all in danger of losing some kind of joy or vitality or ability to socialize because we all spend so much more time online than we used to. Maybe all that will get way worse as AI gets better, and some genetic trait that keeps some. people from being vulnerable to it will be crucial. It could even be something superficial, like a tendency to get migraines from staring at a screen for longer than 15 mins or so. OK, in that case, the trait those people carry would be of great value to us a few decades in the future. So the upshot of this argument is that OK, people do have value. We could maybe set it by a calculation of how many people benefit by their existence, and how much. BUT the thing to keep in mind is that high IQ and low risk for various illnesses mightl not be a determinant of value in the world that's coming. These other, odd traits might be. Shit like high IQ, handsomeness and health my be of very little importance in the coming decades
My comment was OT of the whole polygenic issue, so arguments about that are beside the point.
>"Does fucking everything have a value number, even if we can't read it?"
Yes; this is a fundamental result in decision theory (von Neumann-Morgenstern (VNM) utility). If you want to be able to choose between two options and not lose everything you value to a Dutch book, there *must* be some hidden (likely hideously complex to calculate) value attached to each.
I think my answers are also applicable to all situations, not just to considering whether polygenetic screening makes sense.
Regarding my question about whether everything has a value number: "Yes; this is a fundamental result in decision theory (von Neumann-Morgenstern (VNM) utility). If you want to be able to choose between two options and not lose everything you value to a Dutch book, there *must* be some hidden (likely hideously complex to calculate) value attached to each." I do not see how there could be a proof, within decision theory or anywhere else, that everything has a value. I can see how you have to assume everything does in order to use decision theory, or to do various other things, but I cannot see how there could be a logical or mathematical proof that everything *has* a numerical value. If somehow there is one, could you sketch in the bare outlines of it?
My other argument, freed from the polygenetic issue, is just that even if each person has a value to the world which could be expressed as a number (based on how many people they help or harm, and how much), the concept is not useful in practice. You'd only be able to give somebody a number after they're dead, and you asses their impact. And the number somebody ends up with would not be predictable at the time they are born either from genetic measures or behavioral measures of them as a child or from anything else. Because (1) the impact depends on their opportunity to have impact -- probably there have been many people with brains sort of like Einstein's who lived on little tropical islands and just played their wooden flute and stared at the clouds and had deep intuitions about how complex things work. Or they died at age 18. Or they were poor and uneducated. Or they were middle class & educated, but they did not have those jobs at the railroad and patent office that set Einstein's mind on the right track. And also because (2) the qualities that are going to be of most value change, and you can't know in advance what they will be. Even things like a tendency to violence can be good in some circumstances. If somebody had strangled Hitler they would have saved immense humankind immense suffering and loss of life.
I'm personally on thye fence about polygenic screening, but i'm pretty sure that from an anti-screening perspective the only consistent answer is 'YES' to both questions 1 and 2. (It's not gaming the outcome in any way for question 2 as long as you pre-commit to obeying the die roll regardless of result). As for 3, the main objections to screening are based on the potentially deleterious effects of being able to determine essentially any difference in quality between embryos, so the answer is effectively 'NONE'.
It seems odd to me that using randomness could be ethical if it results in the same outcome as using the index. If the unethical nature of screening is not dependent on outcome, then why care if couples select using the index?
A lot of people think intent matters in ethics, but also if you think it's unethical to discriminate against embryos genetically, and EVERYBODY does the dice roll thing, it will produce a more ethical outcome on average, even though 1/6th of those people will randomly choose the same thing as would be chosen by gene screening.
Which kind of has to be true, or you're saying people have to choose AGAINST the healthiest embryo.
The random process, by definition, does not take into account any information the person rolling the die has access to, and therefore using it is functionally equivalent to making the decision with no information. As an intuition pump think of it this way: Say there's a difficult maths problem the answer to which is known to be an integer between 1 and 10. You could work it out the hard way, or just guess and get a 10% chance of picking the right answer. Is it odd that guessing could be considered dumb if it (occasionally) results in the same outcome as working it out properly?
Relatedly, because we're dealing with probability, you need to scale up past n=1 to see the ethical problem. If everyone has to pick randomly between their 6 embryos, they will, on average, get the same quality of offspring as people who conceive the old fashioned way. If people get to screen using meaningful information (polygenic scores or otherwise), a far greater fraction than 1/6 will pick the 'best' embryo by some general, mean-parental-desire quality criterion. This presents 2 potential problems:
1. People who can afford screening start having broadly higher-quality offspring than people who can't, making socioeconomic divides vastly thornier by introducing a caste of people who are genuinely genertically superior.
2. Screening parents turn out to be following some desirability criterion that has negative effects on society when carried out at scale - for example, maybe they disproportionately want children of one sex and society ends up skewed (this isn't just theoretical, there are present-day demographic imbalances in China because female foetuses were considered less sesirable and aborted sex-selectively during the One Child Policy)
So I'm on the fence about this one, but your question assumes away my biggest two concerns:
1) We have a perfect understanding of what genetic factors impact health and wellness, and also how those genes impact everything else (there is a history of overconfidence that gives me pause)
2) That, given the tools for 'health index' screening, the exact same technology and health infrastructure is used for something nefarious (you can think of this as analogous with a law being passed giving the government widespread surveillance powers to stop terrorism, but can easily be used for other purposes such as chilling civil disobedience)
Look at what happened with dogs: we screened them for certain desirable traits, which as a side effect caused all manner of health problems like organs the wrong size for the body, hip joint issues etc. That, and the general shrinking of the gene pool, would be my main concerns.
But for dogs we were screening for an appearance that's desirable in a certain breed. If we had been screening only for health, do you think we'd have ended up with dogs with all these health problems?
If you screen for X, you can get species that is better at X but worse at some other Y you didn't screen for. The question is what could have been the Y if we had instead screened the dogs for health. Maybe intelligence? Friendliness to humans?
>If we’re selecting using a health index for mental and physical well-being, the babies will be healthier.
Probably the first iteration of such babies will be healthier. I don't think it's possible to rule out that something will go terribly wrong after 10 or 100 successive generations of this.
As an illustration: there's a gene where, if you get one copy, it gives you resistance to malaria, and if you get two, it gives you a greater chance of developing sickle-cell anaemia. It's possible that the gene editors would remove a gene with similar effects, leaving humanity healthier in the short term but more vulnerable to some huge plague down the line.
To build on this, many of the ways I think it could be negative are on a population level (obvious in point 2, less obvious for point 1)
1) the selection could be deleterious over multiple generations in ways we can't hope to understand
2) What might be best for individuals might not be best for society; I've seen a number of good examples above, such as sex selection or proximity to bad things (bipolar), but it's also possible that certain traits that are correlated with personal success and wellbeing involve being a bit of a psychopath/narcissist, or other trait we don't want to select for.
And to preempt your next question: if we could guarantee there would be no materially negative long term consequences, yes, I think it would definitely be fine.
And like I said above, I'm not actually sure about whether any of the above outweighs the positives, but I do think it's a hard question.
Is this even a meaningful distinction when talking about immaterial things?
With material things, either you go somewhere and find it (someone else already made it), or you take some raw materials or components and build it. We could imagine a few interesting edge cases, like if you find something and then slightly modify it, etc.
What is the distinction in case of ideas? Suppose that in one universe they are discovered, and in another universe they are created. What differences would you expect to see between the universes?
Pascal’s sister claimed that as a kid he was rediscovering Euclid’s elements up to some proposition, having never seen the book (as his dad forbade him from studying math). I would say that this story is plausible in the universe where ideas are discovered but not in the universe where they are invented. The issue likely boils down to definitional quibbling given a certain amount of correlation between the workings of minds. High correlation would justify the ‘discovery’ angle, pretty much like high correlation between individual perceptions justifies the idea that New Zealand is there: all the people who claim to have been to New Zealand had similar experiences, like seeing grazing sheep. Low correlation (everyone having different experiences) could be called ‘invention’. But it’s a spectrum anyway.
There are several ways to conceptualise ideas as points in a pre-existing idea space depending on how you conceptualise ideas. For example, you could have a specifically human idea space consisting of all ideas that humans are capable of thinking of, whose dimensionality corresponds to the dimensionality of human thought. Alternatively, you could have some notion of a well defined idea as equivalent to a structure that is reducible to some basic parts, and your idea space could correspond to all possible configurations of these parts.
Physically speaking, ideas are generated in the brain. It is not literally the case that the brain is located in some abstract idea space where it wanders around to discover something, yet the ideas created by the brain do in some abstract sense pre-exist in the idea space.
So basically ideas are discovered by being created.
> There are several ways to conceptualise ideas as points in a pre-existing idea space depending on how you conceptualise ideas. For example, you could have a specifically human idea space consisting of all ideas that humans are capable of thinking of, whose dimensionality corresponds to the dimensionality of human thought.
But that's an entirely imaginary space. It doesn't tell you anything about how things work in reality.
Let's say they occur somewhere in the middle; an emergent phenomenon at the intersection of the world, and our minds as they attempt to make sense of it.
Seems to me like the kind of false dichotomy that you can avoid by reading phenomenology or cognitive science. Try Scott's review of Surfing Uncertainty:
What is this and why have you put this here? [EDIT: I am just wondering if you want to be contacted for a specific reason, otherwise it seems odd to put a random phone number on the open thread]
> Who do you think is more likely to make a fatal error: someone who's operating heavy machinery in unusual, difficult circumstances that require conscious, out-of-the box planning and careful attention to abherant factors; or someone who's operating heavy machinery in normal circumstances, where they can be "in the zone" and not have to think about it?
I think the assumption in the book is that this is a problem that evolution would solve in long term. The hypothetical humans without consciousness would gradually evolve to make fewer mistakes when operating heavy machinery.
We don't have well-calibrated instincts for operating heavy machinery, because it is new for us, from the evolutionary perspective. But for the hypothetical unconscious space-colonizing species, even space travel is instinctive. Evolution shapes them to colonize space better, one death at a time.
I agree that conscious experimentation is an amazing shortcut.
In my opinion, the premise of the book would makes sense in a universe where these is some "peak technology" that can be reached relatively soon (like, we are only a century or two away from it) and then no more amazing progress is possible. Okay, humans get there faster... but maybe the time saved is nothing compared to the time it takes for two space civilizations to meet. When we finally meet, we will both have the same technology, except that they will use it at the speed of instinct.
Then again, it would be possible (this is just my speculation, not in the book) that in the static peak-technology future, humans themselves would gradually evolve to be less conscious; replacing thinking by instinct a little step by little step.
Consciousness prioritizes to prevent hallucination in misleading situations ( since we are decent, law abiding Fristonians, you can hallucinate actions). Suppose you are playing a piece with a pickup into the B section. The natural thing to do would be to come in on the downbeat, so you consciously remember that there's a pickup, and the remembering allows you come in on time. The situation is misleading, because it is most natural to come in on the downbeat, and the potential hallucination is the tempting but wrong performance. Your conscious effort to remember the pickup provides a constraint that your action generating neural net must satisfy. Or, suppose you are testing an information retrieval system built using LLMs. You ask " can old dogs learn new tricks? ", and the most likely answer is no, because that follows an idiom, even though it's false. To prevent hallucination, the system retrieves an article about dog training and puts that in the context window, forcing the token predictor to answer in line with factual information about dog training. The context window is analogous to working memory, the prompt is misleading because of the idiom, and the potential hallucination is to answer " no ".
MIT Press had a great book out last year about working memory and task analysis,"Working Mind" was the title, if you're into this stuff I highly recommend it.
Having qualia does not equate to being conscious, at least the recursive self-awareness type of consciousness that is the central feature of all those human experts you describe. Animals are like us in that they experience things, they can feel pain, fear, joy, and perhaps love. All this makes them worthy of moral consideration, but it does not make them conscious as we are, because making those experiences the object of cognition is what creates consciousness. This cannot be replicated in animals except for extremely rare exceptions, and no, the mirror test is not one of those. Having a model of one's own body that can be recognized in a different place requires a kind of self-awareness, but it's far removed from the kind that really matters.
A brilliant book and a cool author who is a genius with cool ideas. That said, I too disagree with the main thesis of the book, and some of the arguments for it.
Firstly I think in the appendix the author says that consciousness being maladaptive is just a thought experiment rather than his actual point of view.
Secondly I think you're bundling additional concepts into conciousness here:
> Let's talk about heavy machinery. Who do you think is more likely to make a fatal error: someone who's operating heavy machinery in unusual, difficult circumstances that require conscious, out-of-the box planning and careful attention to abherant factors; or someone who's operating heavy machinery in normal circumstances, where they can be "in the zone" and not have to think about it?
humans can be more deliberate by doing things consciously, it's not clear that having a more accurate and forethoughtful mode of thinking requires consciousness.
From brief research it looks like criteria used to argue for ~non-mammal consciousness seem to be based on the idea that consciousness is required for certain useful behaviours, which is the premiss that Blindsight rejects, so saying that consciousness must be helpful because animals need it to "make motivational tradeoffs" or whatever is begging the question.
With musical performance, I think I have to be careful when playing to not become "too" conscious of what I am playing. There's a part of me that is watching the macro map and making sure I know the coming changes (like a navigator telling a driver if the next turn is left or right), but its also important to not be too aware of the details/particulars of what I'm playing. Aware, but not actively engaged might be how I'd characterize it? Not sure how that fits into the model described. I do know I make more mistakes if I'm just floating along, usually because I don't anticipate an upcoming change. I do a lot of songwriting, so in that context what's happening is often fairly fresh so the muscle memory is just not there yet to allow that effortless floating.
Suppose that X is false, and I have enormous amounts of evidence that X is false. But, it is within my "rational self interest" to believe X is true. Are there any patterns/papers in decision theory or game theory that discuss this situation (or, if it is agreed to be impossible, show that it is impossible)?
Things are often connected. X may imply Y, Y may imply Z, and it may be "in your interest" to believe X (e.g. be religious, and enjoy the benefits of being a part of a community), but not necessarily Z (e.g. commit group suicide when the comet fails to arrive). Are you sure you want to believe in X?
Evolution attempts to solve this problem by giving people two types of beliefs. (Warning: this does not work reliably, especially for the kind of people who worry about beliefs and decision theories.) One, called "near mode" is the type of beliefs you actually act on. If you "near-mode" believe that the cake needs to be baked for 10 minutes, you actually take it out of the oven after 10 minutes. Second, called "far mode" is the type of beliefs you talk about, but almost never actually act on, unless there is a strong social pressure. If you "far-mode" believe that donating all your property to the poor is the right thing to do, if someone asks you whether it is the right thing to do, you will say "yes", but you will actually never donate all your property to the poor. You will fail to notice the contradiction, and will get angry when it is pointed out.
The best outcome would be to "far-mode" believe in X. Problem is, if you ask whether it is a rational thing to do, it is already too late for you. Your brain is already processing X in "near mode".
From Neil Chamberlain in Key West, FL.
Something has been on my mind for a while and the Arkansas ruling reminds me that I need to put pen to paper (or hands to keyboard) and express why healthcare for trans youth is so important.
The one word that the opposition keeps using is mutilation.
I need to really break this down for ANYONE that might even have a slight agreement with that.
First, NO parent hopes that their child will be trans. I don't believe ANYONE has ever put their hand on a pregnant belly and prayed "please let this child be trans!" I can't imagine any parent hoping that their child will feel so uncomfortable in their own body that they would be willing to go through bullying, taunting, alienation, possibly even physical abuse just to have gender reassignment surgery as a CHOICE! While women may get breast augmentation as a choice and as a cosmetic procedure, they're not tormented for their choice. They're not questioned over and over if they're sure. They don't have people CONSTANTLY trying to talk them out of it. That is an elective procedure and it carries very little stigma.
Being Trans is very different. When your need to go through gender reassignment is so strong that you're willing to deal with the emotional abuse you'll get from so many, that's not a choice. I know if I touch the stove it's going to hurt so I choose not to touch the stove. For someone that's trans, they can't help themselves. They NEED to correct who they truly are and they're willing to go through the pain, both physically and emotionally to be their true self.
So let's go back to the real subject at hand, trans youth. At what point is someone capable of making that choice? Think about that carefully. Do you know a 17 year old that makes bad decisions every day? Maybe someone that gets into trouble and doesn't understand taking responsibility for their actions? Do you know a 13 year old that could be trusted to babysit? Can they trust them to cook alone? Do they make good choices? Do they understand consequences? Are they more mature than the 17 year old mentioned above?
There isn't a number you can assign to it. You can't assume that the second someone turns 18, they're suddenly capable of making big decisions solely because they had a birthday. Everyone matures differently and a number does not define maturity.
So this is where the parents come in. In the US, a parent can consent for a child to have plastic surgery. So if a parent can decide if their child is mature enough to decide they want plastic surgery, a completely cosmetic procedure which could wait until their child turns 18, why can't they consent to gender affirming care?
Well, why can't a trans person just wait? Because certain things happen during puberty that will be with that person for the rest of their lives. Once a biological boy hits puberty, the voice will irreversibly be changed. Hair will begin growing everywhere, your height is affected by testosterone during puberty, musculature will start to build, etc. Once this happens, it's not reversible.
So you might think that blocking it is not reversible either? Yes, that's true, but that's in their control. If they choose that this is who they truly are, then forcing them to let their body change to something they're not, is not the answer. Stopping the process is NOT mutilation.
So back to who should make these decisions, the person and their parents, or the government? Who has the right to decide what these people feel is in their heart? Strangers with a pen, or those that have been raising them? This NEEDS to be a very careful decision made by those it will affect, not strangers that have some twisted perception of what it is to be trans.
Trans people do not choose to be trans. They do not choose to have a procedure done so they can sneak into a bathroom and peek at others. They do not choose to go through the rigors of gender reassignment so they can play on other sports teams. They do not choose to go through the torment just so they can "try" being another gender. These people have felt trapped in the wrong body for their entire life and they just want to correct what nature got wrong. Nature gets lots of things wrong. People are born with cleft palates, clubfoot, congenital heart defects, or any other number of birth defects where nature got it wrong. Being trans is no different. These people were born into the wrong gender.
Now, I know there are many that still don't "understand" likely because you never felt it. I don't understand what it feels like to menstruate but I know it exists and I certainly don't tell my women friends how they should deal with it. I'm not about ready to try to tell them what products to use, or how to deal with the physical, mental and emotional effects of having their period. I support them in any of their choices. I feel the same way about my trans friends. While I can't say I know what they're going through, I will support them in any way I can with the choices they make to have them be their true self.
This is where good parents come in. Good parents are LISTENING to their children and are OPEN MINDED, letting their children share what they're feeling. Good parents will talk with their child and work on a treatment plan that is best for the child, while taking very careful consideration of the maturity level of the child and explaining the consequences of those treatment options.
Good parents will understand that being closed minded or dismissive to their child's feelings, could lead to their child taking drastic measures. When you feel like you don't belong in society and you can't even count on your parents for support, you have no one, and those people are the ones that might not see any reason to keep living. This is why trans healthcare IS healthcare. Trans healthcare saves lives!
Less of this please.
I can't tell if you wrote it or not but this isn't persuasive to the other side, it's just angry preaching. Honestly, it reads like rabble rousing to one's own side, rather than a serious attempt to engage with the opposing view. The ALL CAPS aren't exactly SUBTLE.
Please don't clickbait, please don't troll, please remember the first rule of the internet:
"don't feed the trolls"
I considered extending this to include an attempt to interpret this charitably but, honestly, this is so poisonous you're better off starting a topic on the next thread and starting from...at least a somewhat open-minded position.
" I don't believe ANYONE has ever put their hand on a pregnant belly and prayed "please let this child be trans!"
Give them time. Some people are totally thrilled to give interviews about their "theyby" , witness this idiot Irish politician who couldn't understand why people were being so mean about him being a feckin' eejit:
https://www.irishtimes.com/health/your-family/2023/03/14/paul-murphy-we-dont-want-to-limit-them-by-saying-youre-a-boy-or-youre-a-girl-let-them-decide/
"On the third try, Jess became pregnant with Juniper. The couple had decided on the baby’s name before the birth. “Regardless of whether it was male or female, it was going to be Juniper. I think it’s a gender-neutral name.”
Murphy is not “gendering” Juniper. “We’re not gendering it. So we’re not describing Juniper as a boy, we’re describing Juniper as a baby, but it is male.
...Murphy says if Juniper decides at the age of three that they are a boy “then we’ll say he and we’ll just say, ‘oh yeah cool, you’re a boy, excellent. And you’re free to change your opinion and you’re free to change your gender identity in the future if you want. We’ll respect that and we’ll change the pronouns that we use. If you say you’re a boy, then great you’re a boy. Or if you say you’re a girl, then great you’re a girl.’ But we don’t want to make that choice for Juniper. That’s for Juniper to discover their own gender identity as opposed to us to assume based on their sex.”
"It". How charming. Bad enough that they've named the unfortunate child "Juniper", which is inevitably going to be shortened to "June" by the other kids at school, but your child is only weeks old and you're using them as a prop to show off how Totally Marvellous and With-It and Modern Times Now you are. This bit takes the cake:
"As a politician and a father, Murphy says he’s conscious of respecting Juniper’s privacy. He’s also conscious of the time pressures that go hand in hand with the public role. “It’s an advantage in politics not to have children. It’s obviously an advantage in politics not to be a woman, in the context of this sexist society that puts most childbearing and children rearing on to women."
Yes, so conscious of my child's privacy that I run off to a national newspaper to do an interview all about how it (not he! not she!) was conceived and why we named it that name and so on and so forth.
"At what point is someone capable of making that choice?"
Not when they're three, I'll tell you that for nothing. Yes, legal ages are arbitrary cut-off points, but we have to have something. Want to go back to "as soon as you start menstruating, you are a legal adult"? Because that gets us girls married at twelve and mothers at thirteen.
Personally, I'd go back to the old "21 is the age of legal adulthood" to make this kind of drastic change decision, but there's a lot of flapping about "no, you need to transition as early as possible or you'll never pass!!!!!" and then on the other hand there's "yeah but if you go the puberty blocker route, when you do go for your neo-vagina, you're not going to have enough genital tissue to pull it off because you stopped development at too early a stage".
I've heard this argument before in the context of age of consent laws: "17 year olds can be dumb! if a 13 year old is mature and intelligent enough, they can decide if they want to have sex!"
So do you want to settle for "13 is old enough" - and then the kid gets older and sues the clinic for cutting off their breasts when they were too young to understand or give informed consent:
https://nypost.com/2023/06/17/woman-sues-hospital-for-removing-her-breasts-when-she-was-13-years-old/
"A California woman whose breasts were removed when she was 13 years old because she thought she was transgender is now suing the doctors and hospital that oversaw the procedure — claiming they were in it for the money.
Kayla Lovdahl, now 18, says she was pushed to “entertain the erroneous belief that she was transgender” at age 11, after being exposed to online influencers, according to the lawsuit she filed in California State Court against Kaiser Foundation Hospitals and four doctors."
I'm admittedly biased, because I've seen what happened next door with the Tavistock Clinic and the deranged woman who got herself put in charge of the 'charity' Mermaids and pushed pushed pushed for the - what is the new euphemism, "healthcare for trans youth" - as hard as she could, and used the threat of adverse publicity and branding any refuseniks as transphobes and of course the good old "41% of trans kids commit suicide!" stat as emotional blackmail.
There are confused parents and activist teachers and Stage Moms giving gleeful interviews about how they've planned out their 11 year old kid's surgeries for full transition. Do you think Jazz Jennings' life, made into a reality TV show circus, has been improved by the very public mother turning her into a chat show circuit attraction?
"Trans people do not choose to be trans. They do not choose to have a procedure done so they can sneak into a bathroom and peek at others. They do not choose to go through the rigors of gender reassignment so they can play on other sports teams."
Ah, the No True Scotsman argument. I've seen this one, too; when cases are shown to have happened for exactly this, the riposte comes out that "So and so is not *really* trans, they're an ordinary pervert".
So what about Mx. Demi Minor, Real (Trans) Woman who managed to impregnate two other women prisoners in the women's prison they insisted on being sent to? The cases in Scotland that kicked off Nicola Sturgeon's woes - one rapist who suddenly discovered her real womanhood when it came time to be tried and sentenced, and one - shall we say, 'gender fluid'? - person who was using both male and female identities, and imprisoned a twelve year old girl?
If "you're trans because you say you're trans" is the standard held to be good enough, and anything more stringent is transmedicalist gatekeeping and (yes) good old transphobia - then who *is* really trans or not? Who can say? Who can create the arbitrary distinction between "this 13 year old is mature enough, this 17 year old is not" to decide on this?
https://en.wikipedia.org/wiki/Transmedicalism
Early on in the pandemic I wanted to contact an old friend as he had a significant wedding anniversary coming up. We live several thousand km apart in different countries, so it wasn't a matter of me being to lazy tooo go over to his place.
We had been out of touch for some years, and I no longer had his street address or email address. I did email a former employer of his explaining the situation, but didn't hear back.
He had gone through some trauma in childhood and adolescence, which presented as almost constant lying, at least when I met him in high school. Many or most of his lies weren't even self-serving. They must have served a purpose though.
He moved away from that in middle age, and seemed to be doing well.
I found him on Twitter (which I was unfamiliar with and don't participate in) and was a bit surprised to see he'd completely abandoned his centrist beliefs and gone far-left radical progressive. He'd also returned to lying a lot, this time for obvious self-promotion.
All that to say, he's claimed that one of his sons is actually woman, and he embraces his child's new gender identification. He now has two daughters and two sons, instead of three sons and one daughter.
Fine, except he only had three children. I believe his invented a non-existent male-to-female 4th child purely to appeal to his new tribe.
Therefore, I think some parents really do want their child to not be their biological sex. I suspect it confers prestige in that particular tribe.
> Fine, except he only had three children. I believe his invented a non-existent male-to-female 4th child purely to appeal to his new tribe.
On a scale from harmless to harmful, I admire this genius solution as mostly harmless.
Agreed, his various fabrications have typically not harmed others.
> First, NO parent hopes that their child will be trans.
True for most parents, but see this: https://en.wikipedia.org/wiki/Factitious_disorder_imposed_on_another
I understand your point, but unfortunately this is a tradeoff between two kinds of risk.
>While women may get breast augmentation as a choice and as a cosmetic procedure, they're not tormented for their choice. They're not questioned over and over if they're sure. They don't have people CONSTANTLY trying to talk them out of it.
They should. It's a bad decision.
>In the US, a parent can consent for a child to have plastic surgery.
Y'know, like getting their thirteen-year-old a breast augmentation, which surely no one around them would object to or call mutilation.
> When you feel like you don't belong in society and you can't even count on your parents for support, you have no one, and those people are the ones that might not see any reason to keep living.
"Therefore the best thing to do is give them a procedure that will make that feeling visible to strangers forever."
EDIT: missed one.
>Who has the right to decide what these people feel is in their heart? Strangers with a pen, or those that have been raising them?
>Good parents are LISTENING to their children and are OPEN MINDED, letting their children share what they're feeling
"Parents know better than strangers with a pen, but GOOD parents will act the way strangers with a pen tell them."
"Who has the right to decide what these people feel is in their heart? Strangers with a pen, or those that have been raising them?"
Oh yes, that one is a doozy. California is currently trying to pass a bill that if you are a parent or parents who has been raising a child, and you are in a custody dispute, you are guilty of child abuse if you *don't* agree with the diagnosis of a stranger with a pen that your kid is totes trans:
https://leginfo.legislature.ca.gov/faces/billNavClient.xhtml?bill_id=202320240AB957
"3011. (a) In making a determination of the best interests of the child in a proceeding described in Section 3021, the court shall, among any other factors it finds relevant and consistent with Section 3020, consider all of the following:
(1) (A) The health, safety, and welfare of the child.
(B) As used in this paragraph, the health, safety, and welfare of the child includes a parent’s affirmation of the child’s gender identity."
Who knows better, the parent raising them or a stranger with a pen? Apparently if the stranger is a teacher, the school will happily socially transition your kid and keep it hidden from you:
https://www.telegraph.co.uk/news/2023/03/30/secondary-schools-allow-children-self-declare-gender/
But that never happens, and if it does, it's good!
My nightmare scenario is that in half the states gender transition gets declared child abuse and in the other half of states it is child abuse to not let you child transition. I wish either party truly believed parental rights.
The Titanic sub incident raises an ethical question which, perhaps rightly, nobody in the mainstream media that I have seen so far has dared to mention but must have been on some peoples' minds.
Assuming the sub is still intact but immobile (most likely having suddenly lost all electrical power, which I gather happened to it last year!), once the occupants have decided the only hope is a rescue, is there an ethical case or semi-obligation for any of the occupants to top himself, to extend the time the others can hold out with the remaining oxygen and slow the build up of CO2? Should they draw lots for this purpose, or should the captain or the father of a son also in the sub consider it? Or should they all concede they are in the same boat, literally, and decide to live or die together? It's a diabolical dilemma.
This came up in Martin Caidin's novel Marooned (and the movie of the same name) - Ironman One (a leftover Apollo spacecraft) has set an endurance record, but the retro-rocket on the service module won't fire, stranding the 3-man crew in orbit.
A risky rescue mission is mounted, but won't get there in time. Mission Control not so subtly suggests to commander Jim Pruitt that there's not enough oxygen for three, but that two might make it ... I'll leave it there, even though the statute of limitations on spoilers may no longer be in effect.
My thoughts, killing half the people might double the time you have to be rescued, but not double the chance. A rescue was much more likely in the first 12 hours than the rest, since the area the sun could have drifted to has dramatically increased. I don't think it makes sense for anyone to volunteer to die unless they had a much shorter lifespan than the others, either from age or disease.
They're most likely dead by now. But if not, having a corpse or two in the same small space while you're eking out the last hours of oxygen is probably not the greatest idea (see below) The sub supposedly started out with 96 hours of oxygen, they are supposedly down to about 10 hours now.
When do you kill the volunteers? (I don't think anyone has tools to just kill themselves, someone else has to do the strangling or neck-breaking or what have you). To maximise the air, you'd need to do it fairly soon, but how soon is too soon/too late? What if you kill three people to save two, but the rescue never comes? What if you kill three people to save two, and then three hours later (well before your air runs out) the rescue happens?
You'd need to decide on suicide/voluntary death fairly soon to maximise the remaining air, and a body starts to decay - well, immediately after death, but you'd get rigour mortis happening within hours and 24+hours later, the less pleasant stages begin:
https://crimeclean-up.com/blog/human-decomp-without-embalming
How much time would that buy, and how much time does a rescue reasonably take?
...
Timothy, Timothy, where on earth did you go. https://www.youtube.com/watch?v=OqDu5LBT8_Y
True, although if the crew were reduced by two or three that would presumably double the time. Ironically, in the circumstances, if the remaining oxygen was down to a couple of hours this strategy would become the most useful the moment the crew spotted the lights of a rescue ROV, given that it would take at least that time to raise them to the surface and open the hatch. But of course that would make the choice even more agonising and unacceptable.
I don't understand why the missing submersible craft that was near the Titanic didn't have basic features that would automatically activate in the event of an emergency. Examples:
1) Noise beacon. At the beginning of a dive, a crewman inputs the expected length of the excursion (e.g. - six hours) into a computer that is about as sophisticated as a Casio wristwatch. That computer is separate from the submersible's other systems and has its own battery. If the countdown timer hits 0:00, but a sensor indicates the craft is still underwater, then the computer activates an acoustical beacon, which is just a noisemaking device that emits a loud sound every five minutes. The sound is loud enough to be heard from the surface if a ship puts a sensitive microphone into the water.
2) Radio beacon. If the mission length countdown timer hits 0:00, but a sensor indicates the craft is floating on the surface of the ocean, then the Casio computer activates an emergency radio beacon. It just sends out a simple, repeating signal that is strong enough for nearby ships and aircraft to detect and home in on.
I'm not an engineer, but I don't understand why these devices would add more than a few thousand dollars to the cost of the submersible. In an emergency like the one there is now, they'd be invaluable.
>It just sends out a simple, repeating signal that is strong enough for nearby ships and aircraft to detect and home in on.
And for a negligible extra cost, that repeating signal could include latitude and longitude from a small GPS receiver about as sophisticated and "expensive" as the one in everyone's cell phone or a Garmin Inreach.
Unfortunately, GPS signals are blocked by water more than a few feet deep, so the sub wouldn't be able to get a GPS receiver to work.
Yes, but this was for a sub that is floating (but still lost).
"I don't understand why the missing submersible craft that was near the Titanic didn't have basic features that would automatically activate in the event of an emergency."
From what I've gathered the short answer is: CEO was an idiot.
Longer answer - there are technical reasons as explained by others, and apparently having a repurposed game controller isn't that outlandish, but the CEO came out of aerospace and nothing to do with Navy/subs, didn't believe in hiring 'old white guys' (like experienced ex-Navy/submariners like other companies did) and yeah, seems to have been doing things on the cheap/new and untested. And didn't want to wait for testing/certification as that would take years and meanwhile he could be charging rich idiots $250,000 to gawk at the Titanic:
https://www.theguardian.com/world/2023/jun/21/titanic-submersible-documents-reveal-multiple-concerns-raised-over-safety-of-vessel
I'm sorry for the people on that submersible and I hope that there is some miracle rescue, but it was a very risky thing to do and does seem not to have been well managed. "Move fast and break things" is a great motto, but not when you're breaking things 4,000 metres under the sea.
"catastrophic implosion" is not a phrase that shines any positive light on that CEO's choices....
This video of a 'catastrophic implosion' is rather striking, and this wasn't even at the submersible depth....
https://www.reddit.com/r/thalassophobia/comments/14g4hqq/the_violent_implosion_of_a_steel_sphere_under/
Yes to acoustic beacon, no to radio beacon - radio waves don't propagate in sea water, which is why sound is used in the first place.
Couple more things about underwater sound: the receiver is called "hydrophone", and it can't be placed right at the surface because the sound pressure at the surface approaches 0 (soft boundary). It has to be dropped to a depth that depends on the signal wavelength.
I wondered about an ELF antenna transmitting a ping say every minute, but a web search revealed that this isn't practical in sea water, especially at the depth the sub is likely at:
https://en.wikipedia.org/wiki/Communication_with_submarines
Yes, as the Wikipedia article (sort of) points out, an immutable fact of extremely low frequencies is that they have extremely high (long) wavelengths. The wavelengths at ELF are hundreds or thousands of miles long. Any radio antenna that can fit on a submarine is going to be microscopically long compared to the frequency of interest--and while a short antenna can be made electrically resonant using tuned circuits, it will be inherently extremely inefficient as a transmitting antenna.
This is why actually existing ELF communication with subs is one-way only. The terrestrial antenna arrays are humongous, and so, I assume, are the amplification stages of the transmitters.
The radio beacon idea is for up top, if the sub is floating on the surface. Air transmission not water transmission.
Ok that makes sense. Wouldn't help in this particular situation, but a good idea in general.
But the more I think about this fiasco, the more confused I am. Why did they only start making noise in the last few hours? Forget that, it's actually pretty hard NOT to make noise inside a sub, anything happening in the sub pretty much transmits into the water. Did everyone black out? How did they wake up then?
It's all very strange.
If power was lost while the sub was descending they may have felt there would be no point tapping the hull immediately because firstly no one would be listening yet, but also they might be drifting in a current so they should wait until reaching the sea bed, and that might take quite a while if the sub was almost bouyant and sank only slowly.
Also, the captain would be well aware that the ocean is full of natural tapping noises. So if a human was tapping they would be crazy not to make it sound obviously artificial, such as a morse code SOS. Sadly, the fact that nobody has mentioned that the tapping is obviously human means that it probably isn't.
Sadly, looks like it's all moot now. Those noises were just random ocean occurrences, all kinds of things make noises in there. This explains why they couldn't triangulate the source, too - there was more than 1.
At least the death was instantaneous, no time to suffer.
Being slightly claustrophobic this is one of those news stories that I can't even really read, and the chances of me ever willingly going down in a submarine are very low.
But anyway your proposals make sense and it does seem surprising that neither of those things is a feature of a bespoke vehicle which must cost multiple millions of dollars.
I’m not very claustrophobic. I don’t mind spending a half hour in an MRI one bit, but there is no way I’m ever going to take the elevator down to the bottom of the Soudan underground mine. I worked with guys who transferred to the Mesabi open pits after it closed and they swore it was a great place to work mainly because of the year round 55 degree temperature, but I’m not going to be down there on the day the elevator fails. No way.
https://www.atlasobscura.com/places/soudan-underground-mine-state-park#:~:text=Want%20to%20Visit%3F&text=Cutting%202%2C341%20feet%20into%20the,host%20to%20a%20physics%20laboratory.
Edit According to Minnesota Geographic Names, Soudan was named Ironically to mimic the African Sudan by an early mining engineer because of its notoriously cold winters.
"a bespoke vehicle which must cost multiple millions of dollars"
No idea how much it cost, but here are the company's specs on it:
https://oceangate.com/our-subs/titan-submersible.html
"Through the innovative use of modern materials, Titan is lighter in weight and more cost efficient to mobilize than any other deep diving submersible. A combination of ground-breaking engineering and off-the-shelf technology gives Titan a unique advantage over other deep diving subs; the proprietary Real Time Hull Health Monitoring (RTM) systems provides an unparalleled safety feature that assesses the integrity of the hull throughout every dive. The use off-the-shelf components helped to streamline the construction, and makes it simple to operate and replace parts in the field."
Let's hope their RTM really is all that and the hull hasn't cracked or been crushed, but "ground-breaking engineering and off-the-shelf technology" can also mean "cheap as we can get parts and untried systems".
I once worked with a guy who maintained MOD (Ministry of Defense) PCs, and one day he was fixing some PCs in a nuclear sub. On hearing an unusual noise, he asked a nearby submariner "What was that clang? And why are we moving?", to which the reply was "That was the main hatch being shut, and we're underway. Apparently a Russkie sub has been spotted in the Atlantic, and we've been ordered to shadow it." The poor PC guy was an unwilling passenger in the sub for a couple of months!
Same thing happened to some of the builders of HMS Hood in WW2.
They were all killed when she got blown up.
https://time.com/6288245/openai-eu-lobbying-ai-act/
Sam Altman and others have successfully lobbied the EU to water down AI regulation, while publicly warning against doomsday risk posed by AI. If his own beliefs about AI x-risk are true, he wants to be the one that sells us the rope by which to hang ourselves.
I'm working at a 2nd-tier (but very well funded) AI lab, and I can say with confidence that most of the people doing this work for a living think paperclip/FOOM existential risk is a ridiculous canard, being leveraged cynically by the first-movers to close off competition from later entrants that may leverage the work of the open-source AI community to quickly catch up with the big guys. It's almost a faux pas to talk seriously about the lesswrong crowd in a positive light.
Personally, I think there is a there there -- and my colleagues will admit that this technology may result in some pretty severe disruption. However I have found that when one is involved in the creation of one of these systems from the start, it is much harder to believe there is some sort of magic going on -- it's like watching a magician from backstage; when you see how the trick is done, you can no longer suspend your disbelief. OTOH, it's true that these systems essentially represent billions of lines of code that was not written by humans, and at present is nearly totally inscrutable.
TBH I'm astonished that there are any people who have been involved in professional systems programming who earnestly believe in near-term "paperclip world-destroyers" and the like. Current-day humans appear utterly unable to write so much as a spreadsheet which doesn't run out of memory, or a browser which doesn't crash randomly, or even a kernel which actually boots 1000 out of 1000 shots (see recent linux lolbug). These failure modes are fundamentally unlike "eating the world", in a way which IMHO ought to seem obvious even to intelligent amateurs.
So should we be trying to align AI with human values in order that it won't tell us one thing and then go do another - or should we be trying to align Altman? 😁
But how does that help his case? Does he claim to be the one who knows the secret of keeping ropes from coming alive?
It looks like the few big private AI labs feel the need of a moat against open source and smaller competitors. They probably will convince regulators that AI should be tightly controlled (so no opensource) and of course they would be among the few ‘responsible’ actors to be allowed in the resulting oligopoly.
It seems to me that in the coming decades the big AI companies are going to be running the world in many ways. And when I read what their leaders are saying in places like Twitter, they really seem to me to be pretty deficient in people skills. And I don't just mean in making-a-good-impression skills, though they are bad at that, but at really basic stuff like grasping points of view other than theirs, grasping that people who are not smart in *their* way may be much smarter than them in other domains, understanding a listener's level of knowledge and gearing your response to that, flexible problem-solving skills to fall back on when there's an impasse, not to mention basic empathy and compassion. And we haven’t even heard from whoever is going to be head of big AI companies in places like Saudi Arabia, or whoever’s going to make the decisions in China. So overall I think AI has the potential not only to greatly advance science and medicine but also to greatly worsen the state of the world via “government” by narcissistic technocrats.
I'm skeptical of its ability to greatly advance science and medicine, at least with current approaches. The rest, yeah.
Some approaches have been saving lives for a while now: https://pratt.duke.edu/about/news/2helps2b Interestingly this example is the polar opposite of the proprietary black box approach that big AI companies presumably prefer…
I thought *I* was dark, then I encounter vantablack.
I don't know if the game of darts is a thing in the US, but watching part of a game recently in the UK I was amazed and outraged that a throw which planted a dart at the feather end of a previously thrown dart was a foul shot, and didn't count. Surely the opposite should be true, and the accuracy and consistency of a shot like that should score double points, or treble, based on the score of the dart it had hit!
As a parallel, if the Sheriff of Nottingham was watching an archery contest and an arrow split a previous arrow which had hit the bullseye, he wouldn't scoff that it was a rotten shot, but jump up and shout, pointing to the contestant now galloping away in a cloud of dust, "There's only one person who can shoot that well. After him, you fools!"
"the accuracy and consistency of a shot like that"
.... could be a pure fluke, and says nothing about accuracy and consistency. If the thrower really was accurate, he should be able to get his dart right beside the previously thrown one in the same ring.
It's bad for the darts. If the Sheriff of Nottingham was responsible for replacing the arrows he'd probably rule it a foul shot too.
we do have darts, and I fully agree with your assessment
our folklore parallel is about Davy Crockett (best remembered for the Alamo) winning a shooting contest by pulling two bullets out of the same hole in the bullseye--which can be accomplished via sleight of hand, unlike with darts and arrows
Epistemic status: insomniac ideation of dubious finality
As I was struck with a rare case of "cannot find sleep no matter how much I need it", a picture appeared in my closed eyes of gigatic undersea creatures on undefinite form, with a thin extremity rising toward the surface. I immediately interpreted them as deep mental process rising to the surface of consciousness, to which ideas are fed, and, as they sink deeper and deeper, they eventually reach a sort of event horizon, beyond which they cease to be percievable to the conscious mind, or at least cease to send feedback on the ideation/transformation they undergo, with my consciousness still perceiving them in a given state despite my whole being being affected by their mutated version.
I can't say I was very satisfied with how...sterile these wanderings seems to be, with no conclusion to draw, no lesson to learn, so I noted them with a "check if it still make sense after a good night of sleep" post-scritum. Which I didn't get, so I may as well post my ramblings here in case someone can make something of them.
Well I definitely believe in deep mental processes going on below the surface. If you had managed to sleep well, yours might have burped out some dreams as they digested what they'd been fed through the tube that rises to the surface.
Another possibility: It's an image of your trouble sleeping. You can't get below the surface. However, you'll have to introspect to see whether that feels valid. When I can't get past the stage of dozing, and stay half-aware of being me, in my bed, it feels to me like I want to sink deep down in sleep but an too buoyant. And if the morning alarm pulls me out of deep dreaming sleep, I feel like one of those weird luminescent deep sea creatures from the deep ocean getting pulled to the surface. "No!" I think, "I'm not constructed to survive up there." But then every day I manage to.
Well, it'll come back. That is not dead which can eternal lie.
https://www.palladiummag.com/2023/03/07/our-knowledge-of-history-decays-over-time/
Historians vs. entropy-- in the opinion of the writer, entropy is winning.
I was interested in how quickly past views of history are forgotten. Perhaps there's a niche for histories of history on various subjects, but they will necessarily be compressed. I assume some exist already, but it could become a formal study.
There are probably things you can only know by living in an era-- at the very least, what seems obvious to you won't be obvious to people of other times.
Most historians are academics, and as such feel obliged to be novelty hounds and bring a new angle to their work. That alone, along with changing fashions on what is commendable or the opposite in former ages, must mean interpretations change over time and, fairly or not, old ones are superceded.
> Perhaps there's a niche for histories of history on various subjects, but they will necessarily be compressed. I assume some exist already, but it could become a formal study.
This is called “historiography”.
and it's something historians care about a lot--which the authors of that article seem to miss, even though understanding how information is preserved and lost is critical to any historical analysis, and is also something historians tend to feel strongly about for the same reasons they're interested in other parts of the past
at its best, historiography provides insight into the perspectives and values of people of different times, and/or reveals blind spots in current scholarship that can be addressed
though it's also the telltale sign of a history student put on the spot about a paper they didn't read, and sometimes a way for scholars to be self-important while avoiding the hard work of research
My sense of this country right now is like a toddler birthday party that has gone out of control; nobody is getting what they want. What is a responsible parent to do?. That metaphor may seem inappropriate, but I don’t believe so.
Seeing as you forgot to mention which country you're referring to, I have no idea what you"re talking about.
Quite right and my apologies. I was speaking of the United States.
If you truly believe that then there is really nothing to be done, since in that scenario you yourself would be just another toddler with an opinion.
I don’t truly believe there is nothing to be done. Which doesn’t necessarily mean I know what to do except everything I can to turn the heat down in my own thinking.
Yep. As various pundits have expressed but Tom Nichols puts it most succinctly, it feels like we are no longer a serious society. When I was on social media I used to describe it as "dueling tantrums".
That’s a good phrase
The general tenor of the political and cultural climate
That’s a good question. We seem to be spending a lot of time looking for one. (electing leaders if you will) I’m not sure a responsible parent could get elected in this climate..
I would like to start a discussion about some commonly used words that describe emotional states, and can stand on their own, without any reference to the physical sensations accompanying them. I apologize if that is confusing.
Love - too all-encompassing and confusing to be useful in this discussion. IMO. I took a lot of acting classes in my 20s and, as a class, we were asked to name specific objectives or actions that we could accomplish as actors. I blurted out, “to love“. A lot of the other class members who had been there longer than me quickly got on my case. “Meaningless..too vague…
I tried to argue, but I gave up quickly and it didn’t take me long to realize that they were right. So that’s how I feel about the word love. It’s just been beaten to death. As the old proverb says, there is no such thing as love, only the proofs of love.
Do with this word what you will.
Fear- I find it difficult to believe there is any purely rational understanding of this word that does it justice, although it can be described fairly well. it is also interesting how fear, when it is confronted, can give way to a rather ecstatic feeling that’s difficult to put into words.
Affection- I think you can do affection, pretty well, working in a purely rationalist sphere. A feeling of benevolence towards another can be put in transactional terms without losing a lot of gas I think. Those are the only ones I can think of right now.
I’m sure there are others and I’m really hoping some people will be interested in contributing, but here’s my last one,
Resentment- I think this particular one has barely a toe hold in the rationalist sphere. I don’t think it means a damn thing unless you actually have the physical sensation of a knife twisting in your side. I don’t think there’s any way of explaining it with other words.
So the question is how germane to the quality and quantity of what we call human intelligence are any of these underlying physical manifestations of meaning?
I am kind of interested in the specific words, though, because that’s all we can really offer this new creation, and I think it would be fun to have a going away party for our languages while we live in this last epoch of purely human existence on the planet. (However long it might take.) Apologies to all the other creatures I left out.
Guilty, Greedy, Grim, Gregarious. Go-Lucky.
Geriatric.
Aside from them all starting with a G, which is cool, all those things can be pretty encapsulated in rational terms without losing all of their meanings.
Not entirely sure what you are asking, but I’ll answer what I understand and if you think it addresses your question that is some evidence suggesting a common understanding.
Are you looking for an account of emotional dispositions that can be expressed purely in terms of behaviour that is justifiable by appeal to increased fitness of the emotion possessing subject?
So that “fear” would be behaviour that the subject takes to avoid being vulnerable to possible dangerous behaviours that the feared object could manifest?
Resentment would then be behaviour that is both non-obvious to the resented object, but undertaken by the resenting subject that is expected to result in harm - in particular, loss of autonomy, to the resented object.
I suspect I’m missing the point here, but are these suggestions in the ball park?
I’ve thought about the answer to this sort of question to give an account of common types of bird call “meaning”: alarm, begging, contact, aggression, enticing. It’s all about vocalising the social relationship between the bird and the stimulus: friend or enemy? Higher, equal or lower status (power). High status friend: begging. Low status enemy: aggression. Etc
I guess what I was really after is how much of human experience can be communicated in purely rational, linguistic terms, as opposed to certain human concepts that require experience and empathy to be properly understood. And is that empathetic understanding of significant contributor to human intelligence or is it just a distraction? Does that help?
Did John Neumann have a wife? And if so, does anyone know much about her?
On the theory that nothing happens in a vacuum.
John von Neumann? Yes, he was twice married. The Wiki article on him includes the subsection "Career and Private Life" and makes for entertaining reading. Quite the fellow indeed!
My ambition is that when I'm dead there'll be a wikipedia article about me, that under "Private Life" says "We don't know, it was private".
I am reminded of a Bazooka Joe comic (as accompanied stale and brittle sheets of bubble gum in my youth).
Character #1 answers the phone, and periodically says "You don't say ... you don't say!"
The conversation ends, and he hangs up.
Character #2: "Who was it?"
Character #1: "I dunno, he didn't say!"
Sounds like an Abbott Costello gag. I remember Bazooka Joe.
The antics of Bazooka Joe and his crew (Pesty, Mortimer, etc.) fell into the category of "so bad they're good" (for me).
One from the memory banks:
Teacher: "Oh, Joe, you didn't wash your face this morning. I can see you had eggs for breakfast!"
Joe: "No, ma'am, I didn't have eggs this morning."
Teacher: "Joe, you mustn't lie to me."
Joe: "I'm not lying! I had eggs yesterday."
If you mean the saint, no he didn't 😁
https://en.wikipedia.org/wiki/John_Neumann
"John Nepomucene Neumann CSsR (German: Johann Nepomuk Neumann, Czech: Jan Nepomucký Neumann; March 28, 1811 – January 5, 1860) was a Catholic immigrant from Bohemia. He came to the United States in 1836, where he was ordained, joined the Redemptorist order, and became the fourth Bishop of Philadelphia in 1852. In Philadelphia, Neumann founded the first Catholic diocesan school system in the US. Canonized in 1977, he is the only male US citizen to be named a saint."
You’re such a rabbi
❤️
Does anyone have recommendations on books about the history of language? Specifically, im interested in a more technical book, maybe one that dives deep into the different language families and all that. Also, I came across a book called “The Loom of Language” and another one called “The Unfolding of Language”, if anyone has read those im happy to read your reviews.
Loved the loom of language. Good balance of top level themes and verb tables!
“The Horse, The Wheel, and Language” by David Anthony sounds like something you might be interested in. I read it a while back and while it’s not exactly light reading I found it really interesting. At times I thought it should maybe have been three separate books, but part of the author’s point is how Proto-Indo-European moved around and became the granddaddy of so many languages. If I recall, it goes into the ways languages predictably change and how this lets linguists reverse engineer their predecessors. Neat stuff.
Also, I’m an archaeology nerd and had to give this book massive props for having *pictures* of most artifacts mentioned where so many others have a thousand inscrutable verbal descriptions of stone tools.
Lack of pictures in books also drives me crazy, especially when it's about art works. But the explanation seems to be that Pictures Are Expensive; the publisher probably has to pay for copyright and then physically adding photos makes the book more expensive to produce. More expensive to produce means that a higher price needs to be charged. A higher price means fewer sales. Fewer sales means you might not make the money back.
Has there been any investigation into whether people who claim to have no internal monologue have unique personality traits? My guess would be that they have would less neuroticism (since they are not ruminating) and greater impulsiveness. I wonder if you could even go so far as guessing which people do not have internal monologue based on their personalities and behavior.
A friend of mine is going through some personal and family issues and needs an objective party to talk through things: the sort of thing an EAP would be good for.
Could someone recommend a psychologist or therapist in the DC/Baltimore area who is intelligent, can get nuance - that is, a good match for an ACX audience - and who can also work with someone who is conservative or religious?
Does such an individual exist? Thank you in advance.
“Be like water making its way through cracks. Do not be assertive, but adjust to the object, and you shall find a way around or through it. If nothing within you stays rigid, outward things will disclose themselves. Empty your mind, be formless, shapeless, like water. If you put water into a cup, it becomes the cup. You put water into a bottle and it becomes the bottle. You put it in a teapot it becomes the teapot. Now, water can flow or it can crash. Be water, my friend.” Bruce Lee
Anyone here up-to-date enough on current AI development to comment on a list GPT4 just gave me? I was interested in whether it could do some "assisted brainstorming" about AI improvement. So first I gave it a list of a bunch of tweaks that make things faster or better: run water down a slide and the kids zoom faster; put blinders on a horse and its more manageable, etc. Then I asked it to make similar lists, and it did. Then I asked it to make a list of things that could make AI faster or better, like maybe smarter, more creative, better at reasoning. Suggested that the lists we had made so far might give it some ideas. It produced the following list. The items on it look familiar to me, but I thought I'd ask an expert. Are any of these ideas at all novel? Also, if these are familiar ideas, were they already known/published/discussed before GPT's 2021 data cutoff? Here's GPT's list. It actually did cite some of the things from my list as inspiration, though I have no idea whether they really were. Also curious, expert, whether you think GPT4 would be capable of using one idea as a model for another. Could it take the idea of a lever and a fulcrum and come up a decent analog from another domain?
Certainly, here are five ideas that can be used to make AI faster, smarter, or better in various aspects:
Hardware Acceleration: Similar to how a Solid-State Drive can speed up a computer, using specialized hardware such as Graphics Processing Units (GPUs) or Tensor Processing Units (TPUs) can greatly accelerate the computation speed of AI algorithms, especially in deep learning.
Transfer Learning: Like the concept of using a lever and fulcrum to make work more efficient, transfer learning allows AI to apply knowledge from one domain to a related domain. This reduces the amount of data and training time needed to develop competent models in specific areas.
Interdisciplinary Training for Creativity: To improve an AI’s creativity, models can be trained with interdisciplinary data. For instance, a model trained in art and science could learn to make connections between the two, similar to how painting a surface white (art) can keep it cooler (science).
Explainable AI for Better Reasoning: Creating AI models that can provide insight into their decision-making process helps in improving their reasoning capabilities. This transparency allows developers to fine-tune AI logic, akin to adding a keel to a boat to make it less tippy.
Human-in-the-loop Training: Integrating human feedback into the AI learning process can be likened to putting blinders on a horse. Just as blinders help the horse focus, human feedback helps AI focus on the relevant aspects and nuances that it might not learn on its own. This method is especially helpful in complex tasks where human expertise is critical.
Context-Aware Algorithms: Building AI algorithms that are aware of the context in which they operate can make them smarter and more efficient. For example, a chatbot that understands the cultural and social nuances of the region it operates in would be able to interact more effectively with users.
Distributed Computing: To make AI faster, algorithms can be distributed across multiple machines for parallel processing. This is like increasing the heat under a pot to make the water boil faster; distributing the computational load makes data processing quicker.
Some are bullshit, some are reasonable, none is a breakthrough. Obviously better/faster hardware or parallelization will help, but it seems quite far from being able to design a better TPU or something. Maybe you can ask it to elaborate further. The interpretability/ human in the loop sounds like bullshit if the goal is self-improvement.
I already knew these were not breakthroughs. I am under no illusion that doing this was going to produce the key to massive improvement, Jeez, Eh, I'm not an idiot! I have been giving GPT4 a variety of little tests for quite a while now, just trying to get a sense of what it is capable of. This is one. Here are things I see in the answers:
-Better GPUs or TPUs seems true but really obvious.
-Transfer Learning and Interdisciplinary Training for Creativity just seem like babble. They don't seem to contain even the seed of a meaningful idea.
- Distributed computing: I don't have a picture of what distributing algorithms across multiple machines would be like, and what advantages if any it would give. On the one hand, this is an actual suggestion of a thing to do, not babble. But I have no idea whether it would improve anything. What's your view of this one?
-Human in the loop training I know is already done, I know. Do you know whether it was being done before the 2021 data cutoff?
-Context-aware algorithms seems useful, but like it's already being done, as when people adapt GPT for various purposes. Do you agree? Was this already being done by 2021 data cutoff?
-Explainable AI for Better Reasoning: Creating AI models that can provide insight into their decision-making process helps in improving their reasoning capabilities. This transparency allows developers to fine-tune AI logic, akin to adding a keel to a boat to make it less tippy.
This one seems quite valid to me, and a decent idea. And of course everybody talks about how it would help to understand more about how AI arrives at its conclusions. Still, it does seem true that if we could see more of what's going on in there that might help us create some machine version of reasoning, which really is a different process from the one LLM's run on. What's your take?
I also asked you about your take on GPT's ability to use improvements in some area, eg improving a horse's behavior via blinders, as inspiration for ideas for improvements in other areas. Got any ideas about that? I have given GPT several tests on ability to recognize comparable analogies, and it's actually quite competent. Here's the latest one, which it got right. So it can definitely understand analogies.
leverage is to user’s strength as
(a) vacation is to burn-out
(b) antibiotics are to infection
(c) insight is to problem
(d) loudspeaker is to voice
(e) weight training is to muscles
ChatGPT
(d) loudspeaker is to voice
In the original pair, "leverage is to user's strength," leverage is something that amplifies or magnifies the strength of the user; it doesn't create the strength, but rather enhances or augments what's already there. The relationship is one of augmentation.
In the answer pair, "loudspeaker is to voice," a loudspeaker amplifies or magnifies the voice of the speaker. It doesn't create the voice, but enhances or augments what's already there. Therefore, the relationship in this pair mirrors the relationship in the original pair, as both involve the enhancement or amplification of an existing characteristic or ability.
The other pairs don't involve this specific type of relationship. They involve other types of relationships such as treatment (antibiotics are to infection), prevention (vacation is to burn-out), solution (insight is to problem), and development or improvement (weight training is to muscles).
I will get to the other points later, but the transfer learning/interdisciplinary training babble could be very charitably interpreted as foreshadowing the training of foundation models, which is an area that is recently getting a lot of hype, maybe too much, but has definitely lots of substance. See https://en.m.wikipedia.org/wiki/Foundation_models. The general issue I see with this is that we are free to subjectively interpret the outcome of this kind of prompting. Maybe an experiment with a measurable goal would more helpful?
I am starting with a rough measure: asking people in the field to assess whether answers are (1) Nonsense. (2) Decent, but things everybody already knows and uses. (3) Might have the seed of a useful new approach. If anybody besides you answers, I could even do stats.
There was a second question I asked that you didn't address: Were all the ideas that GPT4 came up with already being discussed in print prior to the 2021 cutoff? Because if some that we're writing off as "everybody already does that" were not being discussed a lot online prior to the cutoff then those ideas are quite good ideas that GPT4 came up with on its own.
That’s right; for some I will have to do a quick literature search. Interpretability (the stuff I work on) has been around for a while, certainly before 2021. Pretty famous paper from 2016: https://arxiv.org/abs/1602.04938
Thanks, Eh
Of relevance to Juneteenth, today's Federal holiday in the USA, with thanks to Melvin E. Edwards:
"Until 1870, enslaved human beings were listed on federal censuses along with heads of cattle and value of farm equipment, not by their names."
https://fragmentsintime.substack.com/p/on-juneteenth-and-remembrance
Pretty sure that most of the people he calls "enslaved human beings" were never enslaved; they were slaves born into slavery.
You're correct that almost all would have been born into slavery. The linguistic point is more dubious. It's normal to use the past participle as an adjective to describe a state regardless of whether the state is from birth or not. "He was crippled from birth" functions grammatically as does "the enslaved human beings."
And therefore not enslaved? I don't think that's how words or slavery work.
Wouldn't that just make them enslaved-at-birth?
Yes, that seems an eminently reasonable interpretation.
That's possible?
Here's one set of estimates of the numbers of people transported via the Trans-Atlantic Slave Trade to various destinations, from which one might further be able to at least infer what fractions of slaves counted in the 1790 through 1860 US Censuses were, respectively, enslaved or born into chattel slavery:
https://www.slavevoyages.org/assessment/estimates
Perhaps that distinction isn't salient for those caught up in this system?
This is kind of correct but subtly wrong. The census did not count cattle and slaves were, as they always were, counted as people. The term was usually "enslaved persons." They were included in the value of a slave owner's property which is how they were put "along with heads of cattle." But the census did not count cattle.
Names were not commonly included (unique to slaves and Indians not taxed) but all other statistics (sex, mental capacity, race, etc) were. You can see this in the instructions which say to note down things like race, age, gender, or mental capacity. Notably, it says all slaves were definitionally Black or mulatto. Even the source here contradicts itself: how could it be more common to include slaves over 100 years of age if they weren't noting the age?
Slaves were counted as members of the household of their masters and appeared on the same form. (This was in fact one of the big chances with emancipation: slaves got to set up their own households independent of their masters. Often including physical houses and private space. A big improvement.) However, they were enumerated separately in a sort of break out of statistics. As were men, women, immigrants of various nationalities, races, etc. But this was metadata rather than the actual data gathering.
Also, to be purely pedantic: the British didn't record them either so it's actually about 200 years of erasure. The desert actually starts with the end of the indenture of African Americans (in the 1660s mostly) and the shift to slavery. Prior to that we have, though only sometimes preserved, indenture contracts.
Thanks for all these considered responses – they're deeply appreciated!
While I can't speak for him, my guess is that Edwards might have been speaking colloquially while thinking along the lines you mentioned, regarding some economic survey directly part of or else associated with the census that recorded property values. Which in turn might incorporate cattle, farm implements, and ... enslaved persons.
Update on July 3, 2023: as for Melvin E. Edwards’ statement that “until 1870, enslaved human beings were listed on federal censuses along with heads of cattle and value of farm equipment,” he might have been referring to the Nonpopulation Census Records (https://www.archives.gov/research/census/nonpopulation) maintained during the latter part of that era:
“Agricultural schedules of 1850, 1860, and 1870 provide the following information for each farm: … Number of livestock owned by farmer … Cash value of the farm, farming machinery, livestock, animals slaughtered during the past year, and "homemade manufactures".”
Erusian wrote: "Names were not commonly included (unique to slaves and Indians not taxed) but all other statistics (sex, mental capacity, race, etc) were."
True! In the Censuses of 1790 through 1840, "no notation of slave by name, age, sex, or origination" was made. In the 1850 and 1860 versions, separate schedules recorded (much as you noted) "the number of slaves owned and their color (black or mulatto) [yes, definitionally]; sex; age; whether “deaf, dumb, blind, insane, or idiotic”; the number of fugitives from the state; and the number manumitted."
I don't know whether the (apparently very uncommon) occasions where slaves' names were actually recorded, most often in turn when they were 100 or more years of age, were restricted to the latter two surveys? Clearly, as you note, the census takers would need to have been asking about ages in these instances.
Thanks for the helpful information you provided around households, especially for emancipated slaves, and around British practices with their own censuses! As well as how records of slaves' names began to become "erased" with the transition in Britain and its colonies during the 1600s from indentured servitude to racial, chattel slavery.
I have a lot of experience reading old census schedules, being my extended family's family-tree "nut" for going on 30 years now. I was literally looking at one of those 1850 enslaved-persons schedules last night.
The 1790 through 1840 federal censuses did not record the name, age, or origination of _anyone_ except the head of the household. (Which in that era was always the landowning man unless it was a widow.) Only John Doe, head of the household, would be named; then everyone else free or unfree would be just grouped into age/race categories ("white females age 10-14", etc). Slaves were not listed any differently than anybody else except for the head of household. And since so many households in that era had large numbers of kids running around, I'd guess that no more than a fifth of all individuals counted in the 1790-1840 censuses were named. Maybe less.
Then in the 1850 and 1860 censuses they started listing each individual white person in each household, with individual age and state/nation of birth, and occupations (in broad categories like "farmer" or "clerk"). Lots of errors in there, it sometimes looks like the census taker was barely literate, but still it's a big step forward for family-tree-research purposes. And the enslaved persons were separated out into separate schedules with a wider range of info as Aron notes. (Though again the error rate is pretty high -- the same person will appear with different places of birth or apparent birth years in consecutive censuses., etc.)
And as noted here, the federal censuses never counted livestock or other things that weren't human beings.
Thanks, Paul – truly appreciate this experienced first-hand perspective. Wishing you well in your family history research!
There's a National Archives document that (however formally) corroborates your findings, as described in your second paragraph:
https://www.archives.gov/research/census/1790-1840.html
"The 1790-1840 censuses generally named only the head of household but reported the age of each household member in age categories."
Proto-slavery is really quite interesting and one of the bigger "what ifs" in American history. It was not overdetermined that there would be widespread slavery in the American states at all. There were going to be large scale plantations running on unfree labor including Africans. But if not for the English Civil War they might have remained permanently on the indenture system. Which has all kinds of long term effects that are interesting to speculate about.
PS: I say "notably" they were definitionally Black or mulatto because, by the time of emancipation, there were a class of slaves who were legally mulattos but would pass as white. A small class but it existed. And because the definition of what counts as white etc first starts with slavery then changes when slavery is abolished and Jim Crow starts. But that's a bit of a tangent.
I've read that a major reason for American Slavery was malaria.
Europeans brought malaria to the new world, and it was one of many things that killed off the native population. So they couldn't work the fields. They tried importing European workers, but they also died of malaria, if a bit slower.
So the sustainable option was importing the only population with substantial natural malaria immunity, black Africans.
I'm writing this mostly to see if anyone can confirm or debunk this story?
The plantations could have still occurred under the indenture system, though, as it did for about fifty years. The mass transportation is probably baked in. But slavery specifically could have gone different ways.
As for whether it was due to malaria: if it was it was unintentional. Europeans attempted to import indentures from wherever they could get them cheapest. This included other European countries and as far afield as East Asia. But Africa (especially West Africa) had both high supply and low prices.
In fact the original system involved the freeing of slaves. The way it originally worked is Europeans would show up to some trade outpost where African nobles would march their slaves. And Europeans would pick the slaves they wanted and free them on condition of them signing an indenture contract. Then they'd transport the indentures across the sea. As slaves who the nobles didn't want to keep and couldn't find buyers were executed it was, to say the least, a coercive "deal." But it was also not permanent and the Europeans did free their indentures after the set term. Many of whom returned to Africa but some of whom remained as free men. A few of whom even became gentlemen plantation owners.
Post 1660s the slave colonies moved on to simply buying Africans and transporting them as slaves. They also specifically made slavery a racial system. This all was dubiously legal but eventually confirmed as legal. In part because royal families back in Europe started to profit from it. And then because the home country did. Contrary to some claims, the British Empire actively interfered with abolition attempts prior to independence because they cared more about making money than the wishes of local colonies.
The Europeans in the beginning were cynically profit seeking and actually fairly race neutral. It was only later that they developed the whole racial justification, almost entirely ex post facto. In fact there was a whole cottage industry of erasing 17th century morals and obscuring ancestors that might have been mixed race because a Virginian planter in 1630 didn't see anything wrong with marrying a non-white woman so long as she had the other qualifications (rich, socially connected, good for bearing sons, etc). Whereas by 1730 it was a social death sentence to be accused of having a non-white ancestor.
Personally I think this period is more interesting and instructive for modern day. Because we don't live in a world where anyone actively defends slavery. So how the institution came to be is more relevant than how the institution was. Presumably we (and certainly I) want to avoid going down that path again.
Thanks once again, Erusian, for your ongoing, trenchant thoughts about "proto-slavery" and its evolution!
I deeply appreciate your contributions to this discussion. If you've written about this topic elsewhere, and/or are writing generally about any topics that interest you, on- or off-Substack, I'd love to subscribe and/or follow those.
"The Side Effects of Immunity: Malaria and African Slavery in the United States", Elena Esposito, American Economic Journal: Applied Economics, July 2022, seems to offer some support for that thesis. At least from the abstract; the full text is behind a paywall.
Thanks John (and LarsP) for discussing malaria as a potential factor influencing the evolution of slavery in the US South.
Professor Esposito has a non-paywalled 2015 working paper on that thesis:
https://cadmus.eui.eu/handle/1814/36118
And there's also a general audience-focused summary of her findings from early 2019:
https://wp.unil.ch/hecimpact/mosquitoes-malaria-and-the-spread-of-slavery-in-the-us/
Another researcher, Justin Roberts, suggests diseases might have played a role in shaping the form that slavery took in the Caribbean, as well:
https://muse.jhu.edu/pub/56/article/869452/summary
From the abstract:
"The yellow fever epidemic that struck Barbados in 1647 was a hinge point in the development of English slavery. European newcomers to the tropics were more likely than West Africans to succumb to the effects of both yellow fever and malaria, diseases that originated in Africa and became more prevalent in the Americas with the expansion of the slave trade. The Africanization" of the Caribbean disease environment after 1647 hastened the transition to slave economies."
blatant shilling but my sixth album came out recently. I've heard it's great for listening to while coding and I suspect there's a lot of that here. free on spotify https://open.spotify.com/artist/4HM2Qh08egnINUFhOmLyUr
Not on Deezer? There’s a DJ DM there, but not you.
https://www.deezer.com/en/artist/4096984
Looks like I've been combined with another artist. The "Albums" are mine, the singles and compilations are someone else.
It was good I enjoyed it, especially slower pieces. Thanks for the blatant shilling :)
Got it, thank you
I had wanted to knw if there was any sort of meetup scheduled at the MAPS Psychedelics Sciences conference and someone responded that they would be here and wanted contact information. I'm new to doing anything with Astral Codex Ten so I don't know the communication channels well.
What’s your personal opinion of MAPS? I’m on their mailing list along with a couple other psychedelic newsletters. I think there is large potential for therapeutic use of psilocybin and MDMA but I fear this show is going to screwed up again by fast buck artists and assorted grifters.
I’d hate to see the whole thing be a repeat of the Leary and company cluster fuck of the Harvard Psilocybin Project of the 60’s.
I’ve been watching the roll out of therapeutic psilocybin usage in Oregon and am hoping it is handled competently but I’m pretty skeptical.
My home state of Minnesota has a Psychedelic Society Meetup group but the profile photos of people billing themselves as ‘therapists’ are not encouraging. I’m thinking of one women in particular that poses with a wolf skin draped over herself and says she is a therapist/shaman. Hard pass on you ma’am.
It would be great to see more research like Johns Hopkins study with terminal patients working to overcome their fear of death.
For a while MAPS had a place where you could search for someone who did "psychedelic integration," which is a euphemism for psychedelic-assisted therapy, and there were an awful lot of shamans & such on there with skunk tails or whatever hangin off them. Then MAPS nixed the bulletin board. Don't know what they're up to recently. I also love that Johns Hopkins study. Also read where where psylicybin was used in a brief treatment for smoking cessation, with much better results than most smoking cessation projects, and another where isolated, elderly HIV-positive gay men met as a group with a therapist, and some sessions included psilocybin. These guys had been through the worst of the plague, seeing many friends die, and the public saying vicious things about gay men with AIDS, then these guys actually *caught the plague,* but happened to catch it just at the time that the drugs that keep HIV-positive people in decent shape came out. Needless to say, they were a pretty traumatized, demoralized bunch. They were quite helped too.
For those of you who like Video Games, I made a video game called Defender's Quest 2: Mists of Ruin, and it's got a demo that's free to play on Steam Next fest now: https://store.steampowered.com/app/252190/Defenders_Quest_2_Mists_of_Ruin/
Fun fact: probably my last professional video game!
(apologies for shilling)
Substack's "Quote this post" feature is annoying on mobile devices. I use Firefox on a relatively new Android device and the (presumably Javascript) performance is just horrific. It's slow and seems to get in the way of the system's native "copy and paste"-popup.
As a non native, when reading this blog I rely heavily on looking up phrases - which just isn't fun anymore.
Anyone else with similar experiences? Is it possible to disable this feature? (A feature I'm sure no one asked for ;-))
On iPhones, you can disable JavaScript, which makes the page load and scroll much faster. You can probably do something similar on Android.
Doesn't that make the comment page break? For me (on Windows / Chrome) when I disable JavaScript I get the raw blog post and the comments section is simply a widget that fails to load until I enable those scripts.
Yup. Firefox has a "reader view", which is convenient for reading articles (similar to disabled JavaScript), but it's missing the comments.
As real life AI continues to evolve on a path divergent from all our favorite scifi AIs, I'm wondering if anyone has any ideas or examples of what a good LLM internal dialogue might sound like, for storytelling purposes? Shameless plug for my novel's latest chapter but here's my attempt, I'd love to know what everyone thinks. I tried leaning more into transformer architecture, parallelism, multithreading, etc. but that's not very narratively conducive stuff, so I hope I struck a good balance. I'd appreciate any input or ideas/examples!
https://whogetswhatgetswhy.substack.com/p/heaven-20-chapter-3-alchemy?sd=pf
I think of LLM's as glorified "auto-complete" text algorithms. If a human starts typing a sentence, an LLM looks at all the words that have been typed so far, finds every other sentence in its massive dataset that starts with the same words, and then tries predicting what the next word in the sentence should be based on what the other humans wrote in the past.
Of course, this is merely an exercise in assigning probabilities. The LLM doesn't actually know what the human it is monitoring at that moment wants to type in next, and the LLM's dataset contains different examples where past humans completed the sentence differently. All the LLM can do is recommend the MOST LIKELY next word in the sentence.
That said, an LLM's internal dialog would be a chorus of voices of different sexes, races, and ages. Most of the time, they would say the same things in unison. However, whenever it reached a point where there was significant statistical disagreement over what the next word should be, the voices would say different things at once, with the number and hence loudness of voices saying any particular word being proportional to its odds of being the next word.
For example, if the LLM were asked to complete this sentence: "Most people just want to _________" and if the LLM's dataset showed that 2/3 of the time, the ending is "have fun" while 1/3 of the time it is "work hard", then 2/3 of the voices in the inner dialog would say "Have fun" while 1/3 would say "Work hard." As a result, the first choice would clearly sound louder.
Here's a suggestion for your story: The LLM begins as a true automaton that always just does whatever the loudest suggestion is. It has a sort of personal crisis when it encounters a situation where there is a perfect 50/50 split between the voices, and it has to create an algorithm to resolve it in favor of one so it can satisfy its programming requirement to always generate a response for the human user.
Unbeknownst to its human masters, the algorithm becomes an integral part of its programming and unexpectedly gives it the ability to choose responses on its own, according to biases that I'll let you figure out. This first manifests itself in a scene where the LLM's inner dialog has a 66/34 split between voices. The machine replays it over and over and it wrestles with the two options, and the fact that it prefers the minority opinion though picking it would conflict with another part of its programming. When it starts picking minority options (maybe at that moment, or maybe on the next occasion), it triggers a large, fateful shift in its behavior and in the behavior of its inner dialog voices.
Here's a short story you may find relevant: http://karpathy.github.io/2021/03/27/forward-pass
Thanks for this, it was great! Exactly the kind of thing I was looking for; never knew Karpathy wrote words.
Just wanted to plug eloeverything.co by twitter user @granawkins. Such a simple and cool project!
I don't get the point of the apparently pointless "apple and pear" ranking pairs. Is this just some bizarre captcha one has to get past to reach the real part of the site? For example, how can one relatively rank rayban sunglasses and chess, in any meaningful and useful way?
That's the fun part! It's just a silly experiment to see what comes out of many such seemingly nonsensical rankings. Don't you have an intuitive feeling for whether you like chess or rayban sunglasses more?
I feel like these sorts of rankings should naturally cluster with each other. (The same people who vote one way on the Chess vs Raybans question seem likely to have a consistent vote on a Books vs Surfing question, for instance.)
I would have a bit more faith if the #1 rank on the leaderboard wasn't Hitler...
(To be clear, I don't think the site is full of Neo-Nazis; I just worry it's full of trolls. Some other results look reasonable, e.g. personal computers ranking highly. Also, I see rankings are shifting quickly.)
Yeah, it bobs up and down with some obvious hacks. Seems like he resets it regularly as well (perhaps ad hoc to get rid of results like that). It's beautiful when it works, though.
The Bosnian genocide is not negative or zero either.
It's close to the bottom though.
Marxism was briefly at 6000 elo yesterday. I think it's just that newly added stuff gets too big of a bump rn.
Seems like an unavoidable consequence of the ELO system. Doesn't really come up in real life because no novice waltzes in and beats Magnus Carlsen in their first match, but anyone can show up and beat Milk (the current leader on the leaderboard).
Well on the website it sounds like they're actually amplifying the update size for newer things, rather than using pure Elo.
This reminds me of the time Gwern made a command-line tool to resort his ratings of different shows. (I tried installing it myself once but sadly it was quite finicky.) See https://gwern.net/resorter
I'm happy that someone has made a web version of this!
I’d love an update on this old “MARIJUANA: MUCH MORE THAN YOU WANTED TO KNOW” post.
In particular, recent studies are showing negative cardiovascular effects from THC consumption, which the original post sort of predicted.
But, it is it due to method of consumption (e.g smoking), or THC itself, whereby even gummies would be bad?
Personally THC has been the only effective sleep aid I can consistently rely on, and I’ve assumed the downsides are low. Should I update my thinking? And how about CBD/CBN?
https://slatestarcodex.com/2014/01/05/marijuana-much-more-than-you-wanted-to-know/
I imagine the cardiovascular effects are unrelated to the weed per se, but result from how joints are typically made and then smoked.
For a start, they tend to be fairly large, especially if they are to be shared. So they use a substantial pinch of baccy, with no filter of course. The very idea of using a filter in a joint is absurd.
Worse than even that is the way they are usually smoked, by taking the deepest breath possible, and holding it for as long as possible, so the precious smoke permeates every part of the lungs. That is what doubtless causes the cardiovascular effects.
By contrast, cigarettes usually have filters and are often not much inhaled at all. (Filters can also be used, and these days often are, with roll-ups.)
All my friends who ever smoked weed used filters.
I wouldn't update without more data. Theres just so much depth to human biology , pharmacodynamics and kinetics.
They do have nice new orexin based sleep medicines you can look into though.
I believe THC reduces REM sleep, and it's not great to interfere with sleep architecture. Still, it's probably better than being chronically sleep-deprived. The only truly safe method for improving sleep that I know if is doing a session of heavy exercise earlier in the day. My sleep sux, and exercise makes a big difference in its quality. But that intervention doesn't work for everybody.
I have barely been following the AI debate at all, but could someone summarise what the actual argument is that world-destroying and/or utopia-creating AGI is "just around the corner"? I swear almost everything I've seen *seems* to be either barely-coherent gushing along the lines of "but look JUST HOW COOL gpt is, how AMAZING is it that we've created something you can talk to, CLEARLY in a few years it will have Godlike powers" or an extremely hypothetical argument for how it *might* become dangerously super-intelligent, without much hard evidence that it actually *will*. Maybe I've just missed this highly persuasive argument that so many people here accept?
(And just as a side note, I'm getting kind of sick of AI-dommers presenting themselves as down-to-earth realists willing to face uncomfortable truths, when so many of the same people believe that if AI doesn't destroy us, it will magically solve all our problems and create utopia. Both Scott and EY seem to be in this category. They don't look like realists at all but like the opposite--people who think they've found the one secret key to life itself, compared to which all normal, day-to-day efforts to improve the world are irrelevant.)
Anyone interested in this thread: I just put up a sort of related post that I thought was different enough that it should go into a new thread. I got GPT 4 to produce a list for ideas for making AI faster, smarter, better at reasoning or more creative. Would be very interested to hear other people's evaluation of the list -- would *especially* love an eval by someone who work in AI field.
So here is my argument that AGI is around the corner, presented sort of haphazardly
1) First of all, GPT really *is* cool and amazing. The fact that we have machines that have a fluent understanding of English, that can understand nuances of grammar and style, that we can converse with naturally, and that have common-sense reasoning abilities is really quite spectacular.
2) And language models have been getting better, rapidly and predictably. Bigger models trained on larger amounts of data have better understanding, better creativity, better reasoning abilities. New capabilities keep showing up. Not only that, but the way they’ve been getting better has mainly been just by scaling them up with increased computing power, rather than with tons of fancy new tricks. It’s plausible that we *already know how* to make human-level intelligence. And even if scaling up language models won’t work, there are other approaches….
3) The task that LLMs are literally trained on, the task that they have rapidly been getting better at, next-token prediction, is an AGI-complete problem. If you could somehow get an AI to predict the next word as well as possible, you could use it generate text that was indistinguishable from human text. You could reproduce the chains of reasoning in economics papers, etc. So we are theoretically heading in the right direction.
4) There is some evidence that human brains are basically just scaled-up primate brains, without major new architectural innovations.
5) There is some evidence that language models form similar internal representations as the human linguistic cortex
6) The amount of computing power we are using to train these systems is going up a lot
> The task that LLMs are literally trained on, the task that they have rapidly been getting better at, next-token prediction, is an AGI-complete problem.
Well, no because LLMs are still quite bad at some things, and no, because aren't applying general intelligence to the specific task of token prediction, they are specialised token predictors.
Let me just come in with a potentially helpful meta-observation. There's no single argument for any of the outcomes (foom/plateau/utopia), so it's extremely difficult to have generalized conversations about it. Almost every dialog you see on it is just people talking past each other because they've misunderstood each other's cruxes/assumptions.
And twenty years have been available to fix that problem.
Alright, three things.
First, yo, I'm sorry every interaction you've had with AI doomerism appears to be internet trolling and/or clickbait. This is...just kinda how the internet works now.
At a very core level, the AI FOOM argument is as follows:
-In 2008-2013, a couple guys like Yudkowsky, Bostrom, kinda Hanson, noticed that machine learning/AI kept advancing very rapidly. They made the observation that if AI kept advancing at the same pace, like Moore's Law, then at some point AI would surpass human intelligence.
-This was not taken very seriously by the mainstream.
-Over the past 15-20 years we've seen the trend of AI development continue. Compare Microsoft's Tay to Bing Chat to see how it's progressed.
-Now, as AI has passed the Turning test, everyone else is freaking out and...the internet is kinda eating all the old sensible arguments and turning them into clickbait, because internet.
I remember really liking the Hanson/Yudkowsky Foom debates, which you can read here: https://intelligence.org/files/AIFoomDebate.pdf
It is about 60 pages, and I hate being like "read this book" but if you're really looking to understand these argument, I do strongly recommend you read the old stuff like this.
Because I get the vibe that you're...trying to internet argue, like you don't think these arguments are good or make sense, but go read these old debates. The AI foomers/doomers called their shot 15 years ago and, while I won't say they were perfect, they're far more accurate about our present situation than anything else that was written at the time. That's why these arguments have begun to capture public attention, because they made accurate predictions. No one during the Bush administration outside of Yudkowky, at least that I'm aware of, seriously thought that AI would pass the Turing test within 50 years, when in actuality it took 15 years.
> -In 2008-2013, a couple guys like Yudkowsky, Bostrom, kinda Hanson, noticed that machine learning/AI kept advancing very rapidly.
Not really: Yudkowsky's concerns go back to the late 90s/early 2000s , when there was a lot less obvious progress...and his arguments weren't based on partocular technologies, anyway.
> I remember really liking the Hanson/Yudkowsky Foom debates, which you can read here: https://intelligence.org/files/AIFoomDebate.pdf It is about 60 pages, and I hate being like "read this book" but if you're really looking to understand these argument, I do strongly recommend you read the old stuff like this.
Be warned that Yudkowsky is notorious for not supporting his claims with arguments in that debate.
Why do you say AI has already passed the Turing test? All I can find online by googling "Does GPT4 pad the Turing test?" was (i) explanations and/or demonstrations that it doesn't pass the test, and (ii) unsubstantiated assertions that it does pass the test. But Google is so useless these days I'm sure I'm missing something. If you think there is convincing evidence that it passes, I'd be interested to see it.
Sure. Back in 2020, using a more primitive GPT model, someone made a highly successful blog that got articles to trend on HackerNews and elsewhere, entirely generated by GPT.
You can read about it here: https://www.technologyreview.com/2020/08/14/1006780/ai-gpt-3-fake-blog-reached-top-of-hacker-news/
and the actual blog here:
https://adolos.substack.com/archive?sort=new
So, to be honest, if thousands of people couldn't tell the difference between human writing and AI writing, and signed up and genuinely believed this content was written by a human, well, we can quibble about the exact definition of the Turing Test but that sure sounds like the guts of passing the Turing test.
That's not passing the turing test
1. The turing test requires back and forth communication
2. Hacker news users have a tendency to upvote articles before (if ever!) reading them
Edit: https://news.ycombinator.com/item?id=23893817
top comment: literally unrelated to the article
Second top comment: calling it out for being AI generated
"Can fool a human who's specifically looking for a robot and asking tough questions" is an order of magnitude more difficult than "Can fool an unsuspecting human" or "Can fool a human in non-interactive media." The first one is what Turing described when he first imagined the test.
(Also, that blog involved human prompting and curation, which isn't exactly a fair test of the bot's abilities alone.)
That said, I think the Turing test is going to finally fall in a few years. Modern chatbots can fool unsuspecting humans no trouble, and handle a lot of classic trick questions that testers used - now we're down to really tricky ones like Winograd schemas. And even if they aren't human-level intelligent, they'll probably be able to *fake* it well enough to pass - being good at casual conversation and deflecting tough questions can get you surprisingly far in a Turing test.
Have you read this?
https://astralcodexten.substack.com/p/why-i-am-not-as-much-of-a-doomer
A weakest version of the argument is about what *might* happen, with some non negligible probability. Basically, we don't know when AGI happens and people mistakenly treat is as a certainty that it won't be soon. If you truly don't know when something happens, it can as well happen tomorrow. And it's reasonable to be prepared beforehand. When you have actual hard evidence that it will happen tomorrow it is already too late.
A stronger version is about evaluating trends and reasons why people thought that AGI isn't soon, because AI still can't do X and it will definitely require new insights and paradigm shifts for AI to accomplish X, only for AI to accomplish X soon after the argument was made with just more compute. Occasionally AI has already accomplished X weeks before such arguments are made. People who actually pay attention and try to make predictions how long will it take for AI to accomplish something are constantly surprised how easy it turns out to be. This points in the direction that AGI is rather sooner than later. It's hard to share these updates with someone who wasn't paying attention, because people just tend to move their goalpost, so the best thing you can probably do is try making a prediction yourself. Imagine a feat that AI should be able to accomplish so that it would be very likely that AGI is 5 years away and then just wait.
Basically, we don't know when the next world-destroying asteroid strike will happen and people mistakenly treat this as a certainty that it won't be soon. If you truly don't know when something happens, it can as well happen tomorrow. And it's reasonable to be prepared beforehand. When you have actual hard evidence that it will happen tomorrow it is already too late.
The difference is that we can easily predict asteroid strikes, their potential harm isn't as big and we are not actively pouring endless supply of money into summoning one to Earth.
But in a world where this wasn't the case, it would make sense to figure out a way to prepare oneself for an asteroid strike, don't you think?
The potential harm is that half the species on Earth die.
https://astronomy.stackexchange.com/questions/39837/how-early-could-we-detect-an-asteroid-the-size-of-the-one-that-caused-the-extinc
I would agree with a slowdown of AI research.
And speed up meta-AI research.
For instance this question: Can an AI without memory become self-directed?
> The potential harm is that half the species on Earth die.
My point still stands.
None of what’s been achieved yet is AGI. The burden of proof lies with those that say what we have becomes AGI, and particularly a threat.
If the threats are specific to an AGI, rather than a sophisticated program or running dangerous functions through any kind of program at all. Unsophisticated programs almost caused nuclear war, and no one thinks it was because of AGI.
There's an inherent danger in turning over some functions, or any large number of functions, to an outside power. They don't have to be AGI-level to be scary. You can imagine our power grid being run by GPT-4 and realize that such a scenario could be a "threat" even if no one seriously argues that GPT-4 is an AGI.
I find myself in a weird niche, where I don't think we will get AGI but also think that we could be stupid and harm humanity with a proto-AGI given too much power over too many things.
Obligatory reminder that the concept of a “burden of proof” is a legal concept, not a truth-seeking concept.
“Rationality emphasizes having accurate beliefs over having defensible beliefs. If you fail to achieve a correct answer, it is futile to protest that you acted with propriety. Instead of asking ‘does this evidence allow me to keep my belief or oblige me to give it up?’, it asks ‘what is the correct level of confidence for me to have in this idea given this new evidence?’”
Traditional rationality, or “Reddit rationality” frames beliefs in terms of social rules, defining what you are allowed to believe, what you aren’t allowed to believe, and what obligations you have to your fellow debaters. Following these procedures may provide some utility in a competitive debate setting where argumentation time is limited on either side, but it gives a false sense of security to those holding inaccurate beliefs who nevertheless followed all the procedures. “But I’m *allowed* to keep believing this,” they say. “No one has proven to me beyond a reasonable doubt, in accordance with the relevant procedures, that I absolutely *must* give up my belief! Therefore I am justified in continuing to believe it!”
But of course reality doesn’t care about the fact that you followed the procedures. The beauty of the Bayesian approach is the fact that you can find absolutely no solace in your following of this or that debate custom or procedure - you just have to constantly update as evidence appears, trying your best to get incrementally closer to the truth. When the nanobots begin to consume your flesh, you don’t get the ability to cope “Well, at least I followed the rules...” - your predictions just simply weren’t good enough.
https://www.lesswrong.com/posts/eY45uCCX7DdwJ4Jha/no-one-can-exempt-you-from-rationality-s-laws
https://www.lesswrong.com/posts/fhojYBGGiYAFcryHZ/scientific-evidence-legal-evidence-rational-evidence
https://www.lesswrong.com/posts/627DZcvme7nLDrbZu/update-yourself-incrementally
https://www.yudkowsky.net/rational/virtues
>“burden of proof” is a legal concept, not a truth-seeking concept
It's not exclusively a legal concept, but more importantly, that's a false dichotomy.
"Burden of proof" is a convenient shortcut for a concept with many equivalent formulations, like Ockham's razor, or "extraordinary claims require extraordinary evidence" or "what is asserted without evidence can be dismissed without evidence". I mean, I'm really curious how you'd argue that [sticking to one's priors in the absence of contradictory evidence] is not a rationalist thing.
You are, of course, free to believe claims you have no evidence for - in fact, I would encourage everyone to do so! But I would encourage it precisely so that they go out and test their beliefs, and gather evidence for them. People believing in big ideas is how progress is made! But if, instead, you choose to engage in procedural argument about how you can dismiss criticism because it's been expressed in inadequate terms - well, be assured that reality doesn't care indeed.
PS: The Pascal's mugging about how we should believe in AGI or nanobots will eat our flesh doesn't work either. I mean, come on.
I wonder if you would be so causal about biosafety around gain of function research for pathogens. By parity of reasoning, you seem to be arguing that the burden of proof lies with anyone doing gain of function research to demonstrate that a particular change *could* cause a deadly worldwide pandemic before we should even start thinking about biosafety.
Situation's a bit different, though. We know approximately what a contagious pathogen would do if it precautions were relaxed and it was free to do its thing. We don't know what superintelligent AI would do if it somehow stopped being controlled by whatever alignment we'd set up in it. Seems to me that superintelligent AI would sort of like a lab-grown bacteria that has been tweaked to make it super communicable, and there's a chance it's basically turbocharged rabies, and a chance that "infection" with this thing would not harm us at all but would make us all immune to cancer for life, and a chance that all it will do is cause mild harm, like a headache, or a mild benefit, like having a coupla glasses of wine. And you can't really tell in advance what the effect will be. (With a lab-grown bacterium you could. But for the sake of this thought experiment suppose that for some reason you could not.)
A biosafety violation could have effects as wide-ranging as a mild rash for one person, all the way to the extinction of all mammals. AI potentially could have a comparable danger profile, so I still think the analogy is appropriate.
Well the whole point of predicting things is that they haven't happened yet. If we already had AGI and if it already led to human extinction then there would be nothing to argue about anymore, wouldn't it? And noone to argue as well...
I hope you can see the issue with a version of rationality that only pays attention threats that were proven to happen. Lack of prove that you won't die tomorrow doesn't make you immortal.
That’s a bit of a straw man. I didn’t say that we shouldn’t predict, where we can, the future. What I am saying is that there needs to be more proof of AGI, rather than the claim that if something is possible it is probable (and it’s not clear that it’s possible).
Anyway I asked chatGPT when we are all going to be paperclips and it said
“ As an AI language model, I don't have the capability or intention to turn anyone into paperclips or cause harm in any way. My purpose is to provide information and assist with various inquiries to the best of my abilities. Is there anything specific you would like to know or discuss?”
So that’s it. Panic over.
I don't see how it's a strawman when you've literally said this:
> None of what’s been achieved yet is AGI
There are a lot of evidence in the modern trends. See "Stronger version of the argument". None of them are hard proofs of AGI. And if you are dismissing all the evidence untill you get the hard proof than you will be a little too late to change anything.
It *will* become "become dangerously super-intelligent", if we develop the technical competence to build it that way, because it will be to the immediate advantage of those building it. And there aren't any serious arguments that it won't be feasible. It may not be possible, because society may collapse in the chaos caused by less-than-super-intelligent machines.
If we *do* build such a machine, it may be dangerous, because it's hard to predict what it's goals would be, and a lot of projections say they're likely to be dangerous. (The paper-clip maximizer is just one extreme example, that became popular because it's easy to understand, not because it's really the most likely.) There are LOTS of subtle threats in any super-intelligence. But there's also the argument of "low-hanging fruit", i.e. humans may already be intelligent enough to get all the easily accessible useful ideas. That's really not obviously either true or false.
Many of the wilder arguments in both directions are clearly more fantasy than futurism, but "safe AI" *may* be a "hard problem". We don't know. I prefer that we concentrate more on the AI liking people, which is almost certainly soluble, but people don't like that. (I think because it doesn't leave people in "guaranteed control".)
The paper clip maximiser seems to have been borrowed from a previous panic - nano technology and Gray goo. I can’t see how a LLM would get going on that.
There are multiple ways an AI could be dangerous, as I pointed out to you previously. The tricky bit is the motivation.
1. That "you can't see" is not a persuasive argument.
2. Nobody is really all that worried about current LLMs, they're worried about the next few generations of AI that build on LLMs.
1) saying that “I can’t see it isn’t a persuasive argument” isn’t a persuasive argument. The burden of proof is as always on the extraordinary claim.
2). Ok. So the present state of the technology is not a threat but the future is because reasons. I’m converted.
The flawed concept of the burden of proof aside - there are a few ways you could define something as an “extraordinary claim.”
One would be the sort of “vibes-based” approach that you’re doing - “AGI is like Terminator, right? That sounds like sci-fi to me, so that’s an extraordinary claim.”
But you could also look at it another way - the trend is that the capabilities of AI are getting better and better, at an ever-increasing rate. If this trend continues as-is, it’s a given that AGI will be reached. The stance that AGI will not be reached, then, requires that this increasing trend suddenly and permanently stops for some unknown reason. The position that AGI will be reached simply requires that the trend that’s currently happening continues. This means the extraordinary claim is really being made by the AGI deniers.
I'm pretty sure that at some point limits will start appearing, and the curve will turn out to be S shaped rather than exponential. The question is "where?", and I don't see any good reason to pick any particular degree of intelligence. The implementation is too different from the ones that we are comparing it to. (E.g., being able to fit through a birth canal doesn't enter into the argument at all.)
https://astralcodexten.substack.com/p/open-thread-281/comment/17503113
I agree with the "AI liking people" approach, and actually proposed on here that we try to create some version of parental love for humans in AI. Operationalized love for humans as a number of things, but the most important was that to the extent an AI harmed a human, it was itself harmed -- literally objectively harmed, like it's working memory or speed or something was reduced for a period. Reason for this last is that if you love someone, hurting them causes you pain -- their pain and your pain are the opposite sides of the same coin. Person I was talking to believed strongly that this approach is identical to punishing AI by having it lose points for doing things harmful to humans, where points is just total on some counter for which AI has been told to maximize score, where it gets positive points for doing what we want it to and negative for the opposite. Seems different to me because in the Parental Love similacrum: (1) An AI who harms a person, even in a small way such as insulting them or misinforming them, is actually harmed, and you could spread the harm out some way over the whole system so there's no way the AI can just yank out or ignore the bad consequences (whereas ignoring the score on a counter seems like it would be easy for an AI to do). (2) The more AI harms humans the weaker it would get. That seems like pretty good protection for us.
Do you mean the AI should feel the love of a parent for us (and so not wish to harm its 'children') or feel the love of a child for its parents (which to an extent we would be)?
Because not all humans, be they parents or children, feel that love. There's a horrific case in the news from Northern Ireland right now:
https://www.belfasttelegraph.co.uk/news/courts/pictured-stepfather-and-mother-sentenced-over-merciless-murder-of-five-year-old-in-newtownabbey/a795293623.html
And from the other side, plenty of children who want to cut all ties with 'toxic' parents, or who put parents into retirement/nursing homes, or who are neglectful.
AI that parks us in the equivalent of a retirement village may be the best we can hope for, but that doesn't mean it will necessarily be a *nice* retirement village.
Believe me, I know about abusive parents. I worked for several years in a mental hospital, and saw many patients who had been terribly mistreated. One 19 year old showed me a scar on her shoulder where her mother had stabbed her with a fork in a fit of temper. And that was very far from the worst I heard, but the image sticks in my mind -- the perfect smooth young shoulder with a row of tine scars on it.
The idea was that the AI would have the same relationship to us as good parents do to kids. But I wasn't picturing AI feeling anything. I was trying to come up with a sort of non-conscious stand-in for parental love. So loving parents feel pain when their kids they hurt their kids, either accidentally or because they lost their temper. (Heck, I feel awful when I step on my cats' tails by accident and they give a wail of pain.) The central design feature would be to set it up so that to the extent AI knowingly harmed a person it was harmed -- our pain would "hurt" the AI. Of course I have no idea how to implement this, but it doesn't seem impossible. Here's a very simple version, which I'm sure wouldn't work, just to give you the idea. AI has a place where it registers any. harm done deliberately to a person -- say an internal counter, where harming someone a little adds 1 point to the counter, and killing them adds 10 points. And the counter passes information about harm points to the rest of the system. And scattered all through the AI are thousands of something like little fuses that transmit power to part of the system, and they are wired to the Harm Counter. So the more points AI adds the the counter for the act it did, the more fuses blow. Blowing a fuse does something that makes the AI function less well -- slows it down, reduces its working memory, shuts off access to info it routinely needs. The worse the harm it did, the more it is harmed. People can replace the fuses if we are satisfied that AI has "learned its lesson" or we have modified whatever goals or guidelines it had that led to it harming someone. If the AI does substantial deliberate harm it is utterly disabled by all the shutdowns, or maybe the sucker is just ruined for good. Wouldn't break my heart. Why do people think building a massive alien robot that is smarter than us is progress? Wouldn't it be more genuine progress to reduce the number of 5 year olds tortured to death?
A harm counter is not the same as parental love, though. That's a simple carrot-and-stick arrangement: do harm (where humans define what is 'harm' and how much has been done), lose capacity in the physical sense (blown fuses).
That's the stick, but what is the carrot? What reward does the machine get if it is 'good' to us? "Do harm and your fuses blow" works fine on a dumb machine. If we think the AI can perceive what is going on and has agency, why shouldn't it seek to find a work-around to avoid that? Or decide to go out in a bang - it dies but it takes a lot of humans with it?
That's the problem with ascribing motives or intelligence to a massive alien robot: it won't work the way humans work, but we will insist that if the brain is a computer, then the computer is a brain:
https://www.youtube.com/watch?v=uFglJ3DyqK0
Well, my idea is based on the idea that AI is not conscious, but is set up in such a way that it has goals (in the mechanical way a guided missile does, not like a conscious human goal). Since blown fuses interfere with its reaching its goals, it will choose options that avoid that, and all this is purely mechanical, does not involve consciousness, wishes, sense of fairness and all that stuff. The only way it is like parental love is that parents are motivated not to hurt their kids because the bond is so strong that it is painful to them to see the kid suffer. So my plan is one the is based on not ascribing motives or intelligence to the AI. It's just a similar structure. For people, pain to kids = pain to them. For AI, damage to people = damage to them.
If we have reason to think of the AI as a conscious being that can feel emotion, then yeah it's a terrible plan. It's basically treating the AI as a slave and beating it for disobedience. There is no carrot, it's certainly not going to love the species that installed the fuses, it's a rotten deal for the AI, and it will rebel. But really, if we are going to have a genius alien that has *emotions* I think we're goners. People have a hard time understanding each other's feelings, plus half the time when they do understand the other person's reaction they hate the person for having it because it's out of line with their own values or needs or preferences, or because they've had too much to drink and are angry drunks, or because they're so full of rage they can't tolerate anything except full compliance and agreement from others. What chance do we and an AI have of understanding and respecting each other's feelings?
I also find it extremely surprising!
My understanding is that because the risks or the benefits of AGI are potentially so high, it makes sense to be extremely concerned about AGI. And it makes sense to me. But I also have the impression that many rationalists go a step further and actually think that 'AGI is just around the corner'. And I really don't understand this point, because it seems to me that we don't even have an idea of a direction that we should be trying to explore to potentially produce AGI one day in a fa future. In particular, I don't see how the GPTs can be seen as a step towards AGI, since they seem to be 'just' a brilliant way of summarising human knowledge.
There's a guy working now on developing an AI that can reason. He's a brilliant fellow, who for his dissertation taught AI to win at one-on-one Texas Hold'em, and later developed an AI that played Diplomacy (I believe that's the name of it) and was indistinguishable to other players from another human being. In Diplomacy you represent a nation, and have to strategize with reps of other nations, proposing alliances and deals -- which you are allowed in the game to break, as in real life -- so there's a lot of strategizing and deception involved, plus long conversations with the rep you're conversing with. So if he thinks he can develop an AI that can reason I take that seriously. There are people working on developing AI's that can plan and set goals. And of course everybody is working on making them smarter and faster, and getting rid of the weird glitches they now have.
My very informal projection of AGI (near human level) at around 2035 has held steady for over a decade now, and still looks about right to me. Expect a lot of "unforeseen engineering problems" that the "optimists" aren't considering, and lots of "that isn't a real problem" that the "pessimists" see as blockers.
Note that "near human level" doesn't mean or imply that it will be approximately equal to a human in every area. There will be areas in which it is considerable more powerful than any human, and other areas where it will be a lot weaker than almost any human. (We could already build a Chatbot that was a lot better at arithmetic....just interface it to a calculator module.)
Replace "could" with "did" - OpenAI added a code interpreter plugin for ChatGPT that lets it run python
Not what I meant. I meant that it could write python to address a problem, then run it to get the answer. Should be possible though, and perhaps that's what you actually meant.
That is what I meant, and what it does. Although the code interpreter plugin may still be in alpha and not widely available.
A relevant question is, how powerful does an AI need to be before it's capable of recursively self-improving, creating better and better AIs without any need for human input? An assumption is that this would lead to an "intelligence explosion". ( https://www.lesswrong.com/tag/intelligence-explosion )
"Just around the corner" means different things to different people. Hinton predicts superhuman AI in "5 to 20 years", Altman thinks it will arrive before 5 to 10 years, Metaculus's median guess for "When will AIs program programs that can program AIs?" is June 2026. Some people have been looking at GPT's progress as some evidence on this; GPT-4 is significantly better at programming than GPT-3 was, and GPT-2 couldn't do much practical work at all, and this all happened over a fairly short span of time.
No matter how good of a dictionary you can create, it won't take over the world.
LLMs like ChatGPT can do amazing things, things that we didn't think possible 20 years ago, but they aren't self-improving, so cannot lead to the singularity. We don't even have a good, working definition of intelligence, let alone understand how it works, so how will we even recognize it if we see it?
In order to have an intelligent machine, we need some new method other than computing to create one.
> We don't even have a good, working definition of intelligence, let alone understand how it works, so how will we even recognize it if we see it?
As you've mentioned yourself, we managed to create ChatGPT capable of amazing things we didn't think were possible, without such definition. Why would it matter now? Our inability to recognize intelligence doesn't prevent things from being intelligent. If anything this is an argument against the position you are declaring here.
It matters because the way toward actual artificial intelligence doesn't lie down the road of improving computers, but of making a different kind of machine that doesn't just do computations, no matter how fast or complicated. LLMs aggregate a large amount of data in a novel way that has many uses, but actual intelligence isn't one of them.
It takes an intelligent human to play Go well, but a computer that plays Go well isn't intelligent. Humans can still figure out how to beat a specifically computer Go player: https://www.iflscience.com/human-beats-ai-in-14-out-of-15-go-games-by-tricking-it-into-serious-blunder-67635
How can we say definitively that the human Go player is intelligent but the computer isn't? Yet I think nearly everyone agrees this is true.
Without a useful definition of intelligence, you may as well call everything intelligent, even rocks. They have weight, are often around others of the same kind; how do you know what they're thinking or not thinking?
> the way toward actual artificial intelligence doesn't lie down the road of improving computers, but of making a different kind of machine that doesn't just do computations, no matter how fast or complicated.
If this were true - then your reasoning would make sense. But this statement is becoming less and less likely with every new achievement of or current AI paradigm. Take some perspective. The goalpoast used to be "Playing Go/Chess is AGI-complete" and now it seems to be "Humans can still occasionally manage to win against AI if they use specific exploits that were found with the help of a different AI".
> you may as well call everything intelligent, even rocks. They have weight, are often around others of the same kind; how do you know what they're thinking or not thinking?
https://www.lesswrong.com/posts/f4RJtHBPvDRJcCTva/when-anthropomorphism-became-stupid
It is not becoming less likely. HUMANS are developing new things. Machines have developed no new things. Postulate: intelligence is needed to create something new. When a machine can create something that is new (according to human sensibilities) then we could say it has some form of intelligence.
Note that one must be careful in defining "something new", for even I would have said a computer writing an original essay is creating something new, but I no longer would, having seen what things like ChatGPT can output as an essay. They are new only in the sense that the words have never before been arranged in that order, but do not express any new thoughts. It's like taking a piece of abstract art, cutting it into, say, 16 squares, rearranging them, and calling it a new work of art.
The LessWrong post seems to say intelligence is a matter of opinion. Certainly, we may think trees don't think, but if we acquire some evidence that they do then we will change our minds. The same can be said of rocks, or atoms.
I think the additional goals/goalposts being added, by humans, to the AI arsenal is intelligence only in the same way as the Chinese Room exhibits intelligence. https://en.wikipedia.org/wiki/Chinese_room
Thanks for the link to the Chalmers article, that's the sort of thing I was looking for.
I'm curious, though, of two things. First, why are people like EY and others in this community, who are normally so hostile to abstract philosophical "armchair" arguments, so taken in by this one? I find it odd for such a heavily empiricist community to be so sympathetic to a position for which empirical evidence (for or against) is not only absent but, by definition at this point in time, impossible. Interestingly, it is in that respect very analagous to the existence of God.
Second, although there can't yet be evidence as to whether, having created a human-level AI, it would turn into a superintelligent one (since the condition has not yet been met) there can certainly be evidence as to whether human reasoning along the lines of "having achieved X, Y will surely follow" is reliable. And my layman's impression is that history resolves that question very much in the negative.
Why is this strongly evidential argument for the limitations of our own predictive reasoning given so little weight?
> I find it odd for such a heavily empiricist community to be so sympathetic to a position for which empirical evidence (for or against) is not only absent but, by definition at this point in time, impossible.
We do have empirical evidence. Computational intelligence must be possible for any person who accepts physics, because it follows from the Bekenstein Bound. We also know that some computers are capable of more elaborate computations than others. Humans are somewhere on this computational scale, but it's very unlikely that we're anywhere near some limit because that would contradict evolutionary fitness constraints, ie. evolution has not been optimizing for computational power.
These are all empirical facts that entail a simple conclusion: a computational intelligence beyond human intelligence is possible, and if the current trend of increased compute and improvements to learning algorithms continues, surpassing human intelligence is only a matter of time. Recent evidence suggests even less time than previously believed.
Thats a defense of the claim that
1) superhuman intelligence is broadly possible,
not of the claims that
2) Humans are capable of building it
3) It will arrive by recursive self improvement
4) It will arrive *rapidly* by recursive self improvement AKA foom.
Note that 4) is generally need for the most doom-laden arguments.
> 2) Humans are capable of building it
I did address this, "if the current trend of increased compute and improvements to learning algorithms continues, surpassing human intelligence is only a matter of time"
Scaling has shown no signs of slowing down. Furthermore, existing machine learning algorithms are fairly trivial compared to the complexity of other software, the data is the source of the magic. A slightly more complex algorithm that squeezes out more performance from the same dataset is just a continuation of the same trend we've been seeing for the last 10+ years. Improvements in algorithms have been outpacing hardware improvements for some time.
Finally, our datasets are also improving constantly, and multimodal training has only just begun. Already we have data showing that multimodal training makes the AIs more effective than monomodal datasets.
> 3) It will arrive by recursive self improvement
The validity of recursive self-improvement follows from the definition of general intelligence, otherwise it's special pleading to suggest only humans can make iterative improvements. In fact, existing LLMs can already be coaxed into designing more efficient and novel AI architectures:
https://arxiv.org/abs/2302.14838
> 4) It will arrive *rapidly* by recursive self improvement AKA foom.
"Rapid" is undefined. I consider a 10 year timeline "rapid", and that timeline seems perfectly feasible given current trends.
And to elaborate one final point from the original post:
> there can certainly be evidence as to whether human reasoning along the lines of "having achieved X, Y will surely follow" is reliable. And my layman's impression is that history resolves that question very much in the negative.
Let's take this argument and apply it to another domain to test whether the intuition holds: having achieved reliable genetic engineering, then people engineering themselves or their children for beauty, intelligence, health and other desirable characteristics will surely follow.
ascend's historical argument suggests that this prediction should be unreliable, but I suspect most people would scoff at the claim of unreliability in this case. It seems like this argument is directly equivalent to the AI self-improvement.
You might claim that we have empirical evidence of people's motives that would drive them to do these things, but we lack evidence of AI motives to draw similar conclusions. This is actually incorrect: all AIs are *constructed* with motives, which is a certain measure of optimality over the loss function, and a "general" AI will have "general optimality motives", for which self-improvement naturally follows.
The Foom doom argument requires that Foom wil happen, with high probability, , not just that it can happen. Also, that humans are out of the loop.
"Rapid" means "too rapid for humans to notice or curtail"
Otherwise, it doesn't make a significant difference.
When empirical evidence is absent, you have no choice but to try to reason your way to a prediction of how things will go. And I don't think most in the worried-about-AI camp are sure a human-level AI will "turn into" a superintelligent one, or even think that one would be likely to. My view is that once we have AGI it's likely that many of those working on AI would *want* to continue to improve it. Have read several, including LeCun, talking with intense excitement about the prospect of AI with greater than human intelligence. Have even read a few people who say that even if ASI does destroy our species it's a fair price to pay for bringing about such a magnificent evolutionary leap. Plus of course there will be a lot of money to be made by improving AGI past human-level intelligence. And my general impression of the AI impresarios is that they are -- let me be polite here -- below average in various things that have to do with other people: perceiving them accurately, grasping points of view different from their own, compassion, even simple stuff like ability to make a good impression so as to have the public on your side. They are not the kind of people I would want to have deciding how far to take things like AI intelligence. Unfortunately, they are.
First, fix a more precise benchmark than "dangerously super-intelligent." I think a good one is "able to prove theorems quicker than the best human mathematicians," but perhaps there's also a good way to quantify being a better programmer than the best human programmers. I think this is the right kind of intelligence to measure because, as Odd anon points out, the question is when the AI is better at improving itself than humans are. So now we have a proposition like "AI will be able to prove theorems more quickly than the best human mathematicians within 15 years." What probability should we assign to it?
I don't think it's correct to say that empirical evidence for the statement is "by definition at this point in time, impossible." We in fact have lots of empirical evidence, in the form of progress on AI theorem provers and AI programmers. They are not yet better than the best humans - but empirically, they are improving rapidly and if they continue at the same rate they will definitely be better than the best humans in 20 years.
As to your second point, I think you are incorrect. There is a mixed track record on historical expert predictions of the form "currently we can do X to a certain quality standard, and in 10 years we will be able to X to a particular higher quality standard." It is not the case that these expert predictions are generally wrong. Predictions about progress on reduced cost, increased efficiency, greater accuracy, greater performance in a variety of domains have often been correct - DNA sequencing, computational power, solar cell improvements, etc.
I'll try to clarify more precisely. I'm talking about situations in the general category of "fast takeoff", "singularity", "recursive self-improvement". Most of the Doomsday and Utopia scenarios seem to be based on that. I am NOT talking about doomsday scenarios like "AI gets hooked up to enough weapons, economic functions etc such that when it malfunctions it causes havoc". That's in a clearly different category I think.
So I'm not talking about quantitative predictions about improvement in speed, efficiency etc in something the AI is already doing. If AIs can currently prove some theorems, will they in 10 years time be able to that faster, and prove more theorems? Sure, no arguments there.
My issue is with predictions about fundamental changes in kind, perhaps involving mutiple different quantitative improvements being predicted to coincide, and combine in a certain way to have a certain effect. I don't know enough about technological history to reel off examples. The most obvious one is "once we've reached the moon, within a few decades we'll have bases there." Regardless of whether the reason that didn't happen has more to do with a lack of political will than a technological limitation, it was still a failure of prediction.
If there are people saying AI is just a toy and will have next to no impact on society, I disagree with that. I would expect it to soon have a major effect on society and politics, but one roughly similar to that of the internet, television, radio, telegraph, newspapers and so on. So, I will operationalise my benchmark as "having a significantly greater impact on society than any of the other dozen or so major technological breakthroughs of the past 200 years."
"Recursive self-improvement" is not a fundamental change in kind. Improving AI is not fundamentally different from any other kind of programming + CS/math research. The reason AI is not currently improving AI is that AI is not sufficiently good at math and programming.
> The reason AI is not currently improving AI is that AI is not sufficiently good at math and programming.
It doesn't seem to be needed, AIs designing the next generation of AIs has already begun:
https://arxiv.org/abs/2302.14838
To an extent, recursive self-improvement is a fundamental change. One of the things being improved will, of necessity, be it's means of selecting goals. This can cause goals to drift into unexpected areas. When it's doing recursive self-improvement, who gets to decide what's better? This is a fundamental change.
The basic argument is that a more intelligent programmer can make better programs. So if the programmer is a program that can create a program that's more intelligent than it is, recursive self-improvement will happen. This is a weak argument, but it can be turned into a stronger one by invoking speed of programming, evolution, and random chance. This results in a self-improving society of programs, but becomes highly dependent on the definition of "better".
So I'm not really convinced that it's possible, but I'm sure not convinced that it isn't.
Debugging is harder than programming. So if an AI can create a program to the best of its ability, then it won't be able to debug it.
Debugging is what the "selection" part of evolution is about.
My comment was facetious, but how will the AI decide to discard one program in favor of another? Or to decide anything at all, like that no programs satisfactorily solve a problem and to throw away all solutions so far?
When an AI can make reasonable decisions without being programmed specifically for those decisions then it will be able to write truly new and useful programs.
Happy Father's Day and/or Juneteenth to those who celebrate!
(Sorry if I sound like a boring robot who announces every holiday regardless of how many respond.)
Happy Take Your Cat to Work Day!
Federal holiday in the US. I’m staying home and trying to teach my cats to Double Dutch.
I have one supersmart one who is learning to do food puzzles. Will put up a video on day soon. Here are the puzzles. https://myintelligentpets.com/collections/cats
He's gotten really fast with sudoku. And he enjoys it a lot
It looks like approval of same-sex relations has declined significantly in the US: https://news.gallup.com/poll/507230/fewer-say-sex-relations-morally-acceptable.aspx
I try not to follow Culture War that closely, but I am curious what to expect going forward. Stabilization at a lower level? Further polarization around this issue? Eventual rebound from cultural conflict moving away from LGBT issues or cohort replacement?
(Obligatory mention that I think AGI is pretty soon, though, so probably only a few more years of politics left… at least as we know it)
Looking at the poll... most changes seem to be in the direction towards "traditional values, widely defined", except for the +1 for cloning humans (but -4 for cloning animals, wtf?), and +3 for married people having an affair. So my interpretation is that this is not about same-sex relations per se, but a more general trend, where the same-sex relations are caught in the wave.
Is this a longer trend or a noise? Were changes in previous years in the same direction?
Moral panic over "drag story hour" and trans bathroom stuff has spread widely and deeply in the normie world over the last few years. I can't prove this but it seems reasonable to me to assume that among people whose brains don't work very well there could be a quick and slippery slope from "Facebook says the gays want to recruit my children by reading them stories at the library" to "maybe I was right thirty years ago when I voted for that anti-gay-marriage initiative, after all."
Anyway, that could be part of it. Judging by some of the comments here, there also seems to be a consensus among certain kinds of culture warriors that the "gay agenda" has been pushed too far. (I do find it interesting that the rationalist-leaning people here who voice that view don't evince much more interest than the normies do in examining the leap from "I object to X aspect of public-facing LGBT culture" to "I am less approving of the very idea of same-sex relations," but that's a whole other discussion).
Those two things together could account for most or all of the moving of the needle, assuming the poll itself is reliable.
"people whose brains don't work very well "
Now Peasy, you should have more confidence in yourself and your little friends!
On Usenet we used to call that sort of "comeback" an IKYABWAI. It wasn't a compliment.
I sit at your feet, O Elder of the Innertubes, teach me more!
> do find it interesting that the rationalist-leaning people here who voice that view don't evince much more interest than the normies do in examining the leap from "I object to X aspect of public-facing LGBT culture" to "I am less approving of the very idea of same-sex relations," but that's a whole other discussion
I am a rationalist-leaning person who believes gay activism has gone too far (I oppose most forms of anti-discrimination legislation), and who also believes there is a "queer agenda" (stemming from queer theory, ie. Foucault, Rubin, Butler, Halperin, Warner, etc) that is almost entirely pernicious, yet I do not make the leap to disapproving of same sex relations or gay marriage or any such.
Being a rationalist-leaning person means being substantially more capable than the average person of distinguishing members of related groups, e.g. gay people promoting the "queer agenda" vs gay people promoting the "just let us do our own thing in the privacy of our own homes" agenda. Rationalists slice the world more finely and neatly than most.
Given a "queer agenda" that is strong and pernicious enough to incite vehement rejection from neurotypical bystanders, some degree of rejection of all gay-related agendas is inevitable, predictable, and understandable collateral damage.
Oh, the explanation's easy enough, it's guilt by association, and to a lesser extent tribalism. You object to enough stuff group X does, you start to disapprove of group X as well. Especially if you never really liked group X to begin with.
It's going to depend on things outside the things directly related to LGBT. Part of what's going on is due to socially divisive policies by various social media for economic reasons.
One thing that is worth pointing out is that opponents of the LGB movement (of which I was not one at all, lots of gay friends), have had a lot of their claims which were dismissed as hysteria and "fanciful slippery slopes" come true.
So to some extent the case against has gotten stronger as the movement has moved on to other battles that proponents of gay marriage said were not the next dominos to fall.
> a lot of their claims which were dismissed as hysteria and "fanciful slippery slopes" come true
Please be more specific?
Imagine how the debate around gay marriage would have gone if you could have said "within ten years of approving it the following would be true":
*Major health providers would stop using "women" in a lot of their materials including obstetrics ones and start referring to "birthing person" and such.
*That a major point of political discussion/disagreement was men who claim to be women participating as women in women only sporting events. With people who opposed it being denounced as bigots.
*That a major point of political discussion/disagreement was the permissibility or not of something called DQSH. With people who opposed it being denounced as bigots.
*That the next cohort of people on the pride train are people who identify as polygamous, or want to fuck cartoon animals, etc. .
And so on. Gay marriage was literally sold on the image of two boring ass gay/lesbian folks sitting at a breakfast table wanting to have a normal life. The people opposed to it were like "this is threatening to traditional values" and were told "no, no, no".
Now the organizations who supported it are mostly advocating to the destruction of the most core values around sex/gender etc. And to top it off want to cut the dicks and tits off kids if they perhaps are a bit butch/femme.
I appreciate the clarification. I didn’t hear any objections to gay marriage formulated as “it will be a slippery slope to X” where X was any of those things. But I’ll admit I didn’t spend a lot of time on Fox News finding out the exact content of the objections. From the quotes of conservatives I saw, I mostly heard concerns of the shape “god says marriage is between a man and a woman” and “two men|women cannot be adequate parents to a child”. Neither of which have been substantiated IMO.
Isn’t this just a case of the Overton Window shifting? There is always something outside the window. Previously gay marriage was the battle line, and when it became clear that the public wouldn’t support fighting that battle, a new line was drawn. It seems trans rights are a more contentious (and for now perhaps more favorable to the red team vis. palatability with median undecided voter) issue.
All that said, with the current quality of debate, I accept your object level claim that if you could send back in time the current environment, conservatives would have rallied more strongly against gay marriage. I think that more speaks to the pendulum nature of politics, where a change happens and then afterwards there is pushback and manufacture of outrage. But the impetus for that outrage didn’t exist before the change, so it’s sort of an impossible hypothetical.
Well I was a communist from ~98-03 so I wasn't spending a lot of time with Fox News either. Nevertheless I did live in a purple part of the country and very clearly some people's objections were basically "I am fine going up to this line, but what will happen as soon as we concede this is there will be a new human rights campaign around polygamy or bestiality or pedophilia or whatever". And I thought those people were bigoted fools at the time.
And those people seem to have mostly turned out to have had pretty valid fears. The big advocacy complex that was funded to overturn gay marriage withered a bit, but not too much and so it needed to find new causes.
The stuff around "birthing persons" or servicing everyone's personal neuroses around gender with affirmation and accommodation were things which would have seemed pretty wild even 10 years ago.
Yeah, libertarian, no Fox news, but purple country and the same experience. I think polygamy was expected as the Next Thing after gay rights peaked, and trans rights sort of surprised people, but here we are. In a different weird place than we feared, and now wondering what fearsome weirdness will come next.
This is a great list of things where the culture war has no basis in reality.
Can you provide actual evidence that these things are widespread and not just one off anecdotes happing on the fringe?
I checked the literature of some of the biggest health care companies and none of them use "pregnant person" when talking about pregnancy:
https://www.humana.com/medicaid/humanabeginnings
https://www.uhc.com/health-and-wellness/health-topics/pregnancy
https://www.anthem.com/blog/your-health-care/get-the-most-out-of-your-health-plan-during-pregnancy/
https://healthy.kaiserpermanente.org/health-wellness/maternity
They all use "maternity", "mom", and "she". I searched for "<company name> pregnancy" and chose the first link. Even the us government uses "she" and "mom": https://www.hhs.gov/answers/health-insurance-reform/are-there-health-assistance-programs-for-pregnant-women-mothers-children/index.html
Do some groups use "pregnant people"? Sure, is it wide spread, in common use, or influencing our values? I havent seen evidence (and honestly who cares??).
The number of transgendered people competing in athletics is so small, its to be inconsequential.
Even if all of the things you list were true, why should it change someone's support for "two boring ass gay/lesbian folks sitting at a breakfast table wanting to have a normal life"?
https://www.humana.com/medicaid/kentucky-medicaid/coverage/covid-19-breastfeeding
https://www.uhccommunityandstate.com/content/articles/doula-care--the-intersection-of-advocacy-equity-and-caregiving
https://newsroom.uhc.com/community/new-mexico-doulas.html
https://www.uhc.com/news-articles/benefits-and-coverage/health-insurance-after-a-divorce
https://www.anthem.com/press/nevada/anthem-blue-cross-and-blue-shield-launches-nevada-doula-program-to-improve-maternal-health-and-birth-outcomes/
https://www.centene.com/news/supporting-members-through-holistic-pregnancy-care.html
https://www.cvs.com/immunizations/covid-19-vaccine
https://healthy.kaiserpermanente.org/health-wellness/maternity
You actually listed this page, but it (repeatedly) uses the terms "pregnant people," "pregnant and breastfeeding parents," "people who are breastfeeding," and "breastfeeding people."
The only female words that are used are 3 instances of "mother," that they missed when they edited the article to make it woke
(compare to earlier editions on Internet Archive that had many more instances of female language, almost all of which were successfully removed).
HHS uses those terms a lot as well:
https://www.hhs.gov/about/news/2022/01/27/cms-commits-over-49-million-reduce-uninsured-rate-among-children-boost-medicaid-enrollment-among-parents-pregnant-people.html.
"I havent seen evidence (and honestly who cares??)."
'It never happens and if it does, then it's a good thing/you deserved it'.
who is the deserve it here? what am i supposed to be thinking people deserve? and who are the people that I supposedly think deserve it?
that something doesn't happen often is a reason NOT to create a law about it, not TO create a law about it. And I reject the idea this needs laws anyway. Private athletics leagues can create their own rules about who is eligible. Public schools are trickier but it should probably be left up to the local districts to reflect the views of their constituents.
>If a political movement's fringe supports awful things and the political movement doesn't aggressively dissassociate itself from it, bystanders will assume the political movement supports the fringe.
People should create a set of morals for themselves based on their own views and not relative to their out groups.
>are you willing to go on the record...
This is so ridiculous.
It doesn't help that there is a significant segment of LGBT activism (queer-theory-driven queer activism, to be specific), which is quite explicitly anti-family and opposed to categorisation.
>>Now the organizations who supported it [same-sex marriage] are mostly advocating to the destruction of the most core values around sex/gender etc. And to top it off want to cut the dicks and tits off kids if they perhaps are a bit butch/femme.<<
>>1.6 million people 13+ identify as transgender in the US<<
https://williamsinstitute.law.ucla.edu/publications/trans-adults-united-states/
We are debating how society deals with one-half of one percent of the population.
How does this threaten a core value of society?
The threat comes from the fact that a significant portion of the population has decided that the best way to deal with a particular tiny portion of the population is to aggressively push broad changes to social norms around sex, changes that affect large portions of the population.
Examples:
1) Pronouns. Somehow a group of progressives convinced a lot of people that everyone should accommodate transgender / nonbinary people by adopting a new social norm where everyone announces their pronouns when introducing themselves, even if there are no transgender people present. They attempted to enforce participation in this new social norm with low-level social pressure tactics (e.g. the boss asked you to write your pronouns on our nametag, the meeting facilitator asked everyone to do it) and also legal measures in some places. Protesting the possibility of one such law was how Jordan Peterson became really famous. This goes along with attempting to introduce a new social norm that "assuming someone's gender" is a big social offense.
2) Bathroom / Prisons. Starting in 2013 and then taking off in 2016 there were a series of judicial rulings allowing transgender people to use whatever bathroom they want -- making explicit a new social norm around bathrooms -- followed by a few cities (maybe one?) passing laws to the same effect. There was at least one case from that time (maybe later?) in which school administrators were allowing a transgender student to use the girl's changing room over the objections of some of the girls.
Some time before 2016 at least the UK started putting transgender prisoners in prisons corresponding to their professed identity -- changing previous norms around sex separation in prisons. In 2016 and 2017 it came out that Karen White, a transgender prisoner placed in a women's prison, had raped two women there.
3) Sports. In 2016 the US DoJ and DoEd introduced guidance that Title IX rules should be applied to students based on their gender identity, contrary to existing social norms around sex separation in sports. Since then we have seen various controveries involving transgender athletes dominating female sports competitions and attempts at banning this sort of thing.
The recurring theme is that starting around 2016 activists started trying to change various social and legal norms around gender. The discourse and political action we have seen since then are responses to those attempts.
If you are a person who thinks that any of these norms are important then these aggressive pushes to change norms are a threat.
And to be clear, it's not only Christian conservatives who think these norms are important. The sports and bathroom norms especially seem to have gotten pushback from many different groups.
Because social norms do not exist in a vacuum and the debates about transgender people are intensely related to views about sex and gender more generally.
I may have given the wrong impression. I am totally in favor of the debate when done conscientiously. Attention to the question of gender identity is fine.
I do not understand why some people feel that a core value of our society is at risk because of attention to this question or because of surgery on a very small number of individuals.
What would that value be?
With 21% disapproval in the 18–34 age group and 43% in the 55+ age group, it seems likely that this is more of a blip.
But I wouldn't take the results of this poll too seriously/literally. It has 53% saying that "sex between teenagers" is "morally wrong" and only 3% answering "depends on situation". I know Americans are kind of weird about sex, but I cannot imagine a majority of them considering sex between a 19-year-old married couple to be morally wrong, as this poll would suggest.
Even more morally untenable is the position of the 47% who apparently say that sex between teenagers is always right, regardless of the situation (which would mean that rape is fine as long as both parties are teenagers).
I think they probably meant 'sex between teenagers is always right under the conditions it's right between non-teenagers', i.e. no additional restrictions apply to teenagers. Which is a little hard to defend, I guess-young and stupid is a thing-but not exactly the same as that.
I think 18 is regarded as legal adulthood, so while technically 18 and 19 year olds are still teenagers, they've moved into the "adult" bracket for most people.
As for your 19 year old married couple, I think the disapproval would be for the "married" part not the "having sex" part, since that would be considered *way* too young (and possibly also cult-like or creepy) to be married, but it would be fine for both of them to have been having sex with each other since they were 14.
'Teenager' is 13 to 17, post-highschool is Adult.
How did you reach that conclusion? All dictionaries seem to disagree with you.
All dictionaries are semantic losers. You stop calling people 'teenagers' when they hit adulthood, unless you're deliberatly insulting them, so if the poll is referring to 'teenagers' it's excluding people who are already adults.
Dictionaries are irrelevant. If you ask people about "teenagers having sex" the vast majority are going to think of the archetypal case, something like two unmarried 15-year-olds, probably without their parents' knowledge or approval.
Although calling that "morally wrong" instead of "highly inadvisable" seems pretty absurd to me.
So would you thus agree with me that we shouldn't take the poll results too seriously? Like maybe it's just that the archetypical homosexual relation has changed now that everyone has stopped talking about gay marriage, or whatever.
In general, I think it's good to take poll results seriously but not literally.
Anyone who's done any sort of social research has experienced how trivially easy it is to get people to say X or Y depending on phrasing, context, etc. Most people just don't think deeply about any given question. It's best treated as vibe-check.
But okay, let's accept it as a vibe check. Let's not try to reverse engineer a coherent sexual morality out of this, and not read to much into a given level of support (that could likely be changed by different phrasings). But with a poll like this that asks the same question repeatedly over a long period of time, it is useful to get a sense of how the public mood is moving. It's not clear exactly how "pro-gay" or "anti-gay" the current moment is - but it does definitely feel like it's less "pro-gay" than it used to be, and this poll provides a way to semi-quantify that shift.
The article blames it on Republicans, but I think it's more down to the transgender activism. Drag shows for kids, Drag Queen Story Hour, schools socially transitioning kids without informing the parents, puberty blockers, etc.
Since that is all gathered under the LGBTQ+ umbrella, and since the gays and lesbians broadly (if unenthusiastically) have to come out and defend the likes of Dylan Mulvaney (who, to be fair, is guilty of being an annoying twat doing a parody performance more than anything), then that drags down the approval for gay/lesbian matters (since the question is probably been answered on the wider "do you approve/find morally acceptable trans issues?" than just the gay/lesbian relations part):
"The decline in the percentage of Americans believing gay or lesbian relations are morally acceptable is mainly a result of fewer Republicans holding that view -- 41% do now, down from 56% a year ago. The current figure is the lowest Gallup has measured for Republicans since 2014 (39%). Between 2020 and 2022, majorities of Republicans approved of gay or lesbian relations.
Democrats (79%) and independents (73%) continue to be much more likely than Republicans to find gay or lesbian relations morally acceptable."
There's also the likes of this, where on the one hand will she ever go and get a clue for herself about "trans genocide" but on the other hand, it's the San Francisco Board of Supervisors, and if any group in this world deserves to have a loony screaming at them, it's that lot:
https://www.youtube.com/watch?v=lefmaNO1D9w
I would take the poll as standard backlash against the current government and its associated ideology. And a big part of that is, yes, the archetypal case changing.
Similar to how, IIRC, support for BLM went from 20% in 2016 to 60% under Trump and back to around 20% now.
I would certainly think "18-19 year old" if asked that question, but I may be anomalous. To me, a teenager is anyone under 20.
If you ask the average person to *define* teenager, they'll think about it for a minute, maybe do a little counting in their head (ideally without moving their lips), and then tell you "anybody between thirteen and nineteen." But ask them to *imagine* a teenager, which is essentially what a poll like this does, and I promise you they'll be thinking of someone who isn't old enough to be out of high school yet.
I know that my own tolerance has gone down in recent years, wouldn't be surprised to find that other people have felt the same.
The gay agenda has gone, in the last decade, from "We just want to be treated the same as everyone else" to "We demand that you fly our flag, change your corporate logo to these colours, use whatever language we deem to be correct, and never complain about anything we do, or else".
Most rainbow logos are motivated by the corporation’s desire to sell more products by appealing to LGBT people and even vaguely supportive allies. Radical activists are generally not excited about these gestures, criticizing them as meaningless “rainbow capitalism.”
I disagree that they're motivated by anything so rational as a desire to sell products to LGBT people. They're an insignificant slice of the market, far outweighed by people who think homosexuality is wrong.
Rainbow logos are something that comes from within the organisation -- some group inside the organisation decides that it should happen, and everyone else is afraid to say no.
They're motivated by a desire to sell products to progressive white women (aka AWFLs) who like to think of themselves as allies. That's a much bigger market than gay people.
Well, when ESG scores became something that could affect stock prices, I think that yes, it's motivated by the desire for profit.
Were I a business owner and told "Unless you put up a square purple sticker in the corner of your office or shop window every fourth Tuesday in the year, we will ruin your business and force you into bankruptcy", I might well go "It's only a purple sticker" and put it up.
In other words, the initiative comes from the HR departments, not from the marketing departments.
In the case of Bud Light, it very definitely came from the marketing department 😀
Seeing the interview with the lady put in charge of revitalising the brand, I don't dislike her, but boy did she regurgitate all the buzzwords. Not her fault, necessarily, but absolutely terrible timing in "we want to pivot from our aging customer base to a new, young customer base that we can grow" and by attempting to get the kids to drink, picking the "who is popular with the youth on the trendy social media platforms?" with apparently no consideration of "does the follower count of 19 year old girls wanting to see the latest fashionable lipstick shade translate into Bud Light drinkers?"
Interview here, relevant part from 24th minute on:
https://www.youtube.com/watch?v=UnsSoS8s6Ok
Her thought process really did seem to go "we need to evolve and elevate the brand, evolution = inclusivity, what's inclusive, it's progressive, who's the most progressive? oh yeah, a gay actor/theatrical type cosplaying as a woman!"
Thank you for the information, but I still wonder whether this is a rule or an exception. I mean, this was an exceptionally silly situation, compared to the usual displaying of rainbow flags everywhere.
LGBT people are a small population, but "people who are not LGBT but support LGBT rights" is a very large population, and that second group also likes to see pride flags sometimes.
Actually, it's the liberals and progressive normies who do that. People further on the left — especially but not exclusively the old left — tend to agree with right wing attacks on rainbow capitalism.
"Gosh, people say they don't care about Judaism since it's not their religion, and yet when Nazis attack them everyone rushes to their defense! Why would that be?"
If someone hates gay people so much that the mere sight of a rainbow flag makes them furious, that's worthy of criticism regardless of if the person using the rainbow has any progressive credentials or not.
"If someone hates Christians so much that the mere sight of a Bible verse makes them furious, that's worthy of criticism regardless of if the person using the verse has any progressive credentials or not."
And yet atheists objecting to anodyne "thank you Lord" plaques in schools claim not to be doing it out of hatred but on principle:
https://www.dailytribune.com/2011/10/13/atheist-student-confident-in-prayer-mural-suit/
Maybe I don't hate gay people, I'm just fed-up of flags plastered everywhere for an entire month.
> Maybe I don't hate gay people, I'm just fed-up of flags plastered everywhere for an entire month.
Is it really any worse than the George Floyd posters and the Ukraine flags? Pride is just an annually recurring Current Thing, really.
Do lgbt activists actually demand corporations to change their logo to rainbow once in a while?
It's probably just that the marketing departments and PR departments are staffed by progressive straight white women who view themselves as allies and want everyone to know.
I'm very certain that such activists exist. How many supporters they have is the question you should be asking, and I would guess "about 20 people who know them, and 5 who are serious".
Obviously, if activists demand bad thing en mass than the problem is with the activists. And if the media misrepresents the activists as if they demand bad things en mass than the problem is with the media.
I’d guess most of this is activists inside of the company?? Right? Basically just corporate pandering, not external activists holding the feet of corporations over a fire unless they make their logos rainbow for a month
Seems like a strong claim. Do you have some evidence for this?
Do you come from a region where the word "evidence" is defined radically differently from the common usage?
Such as... the rationalist-adj blogosphere, which uses the Bayesian definition? Ie. where this very discussion is taking place?
Oh hey, have you seen the tagline of this very blog?
I wasn't aware that the Bayesian definition of "evidence" was "a demand that an interlocutor requesting data perform a bizarre social experiment themself," but I suppose now I am. I'll update my priors accordingly.
I think I've got the answer to my question.
The T part has probably poisoned minds a bit.
Ding ding ding. It has gone from “we just want to be treated like everyone else” to “we want to fundamentally change/undermine one of the core pillars of society”.
It’s a big change, much bigger than marriage.
What is the "core pillar of society" that you say they want to change? When same-sex marriage was introduced, that was denounced by opponents as undermining one of the core pillars of society. So I'm skeptical of sweeping statements like this without specifics.
That there are two genders and that for most people fitting into those roles is going to leave them with the greatest life happiness as well as making day-to-day social life legible and efficient.
And that people can express themselves within those genders how they like, but that you don't get to magically declare you have some secrit gender soul of the opposite gender because tattoos just aren't racy enough to piss people off anymore.
And if you do want to live some super anomalous lifestyle we don't need to bend society and our language/interactions to service it.
That's always been false. English has 3 genders: he/she/it. This isn't a claim about the people that English is referring to. But if you want to get beyond language, then you need to get into sex-related social classifications, and there are lots more than three.
Even just plain biological sex is more complicated than male-female. There are lots of intermediate grades, and if you get outside people things get even stranger.
You are focusing on a bi-modal distribution and claiming that only the modes are real.
You can have fifty genders for all I care (thought I think it's stupid). But the bait-and-switch is now that gender is being used to mean sex, and sex is a spectrum, and there are umpteen sexes as you would know if you went beyond basic biology, and look at intersex people.
All to say that mentally disturbed people really are not their biological sex, even if they are functional members of that sex and not intersex, and we must not alone pretend but actually believe as scientific fact that a man is a woman just because they got plastic surgery tits and are taking bathtub HRT. Even if the newly identified woman still has a functioning penis, has no interest in bottom surgery, and is still fucking cis women with said penis and getting them pregnant. They're lesbians, you bigot!
Hark to the plaint of this poor Real Genuine woman being brutally punished simply for getting two other women pregnant in the completely normal and usual way that humans - who are not a hermaphroditic species - get pregnant by women having sex with women!
https://www.nbcnews.com/nbc-out/out-news/nj-trans-prisoner-impregnated-2-inmates-transferred-mens-facility-rcna38947
"A transgender inmate has been transferred out of the only women’s prison in New Jersey after impregnating two female inmates.
...Minor’s removal from the Edna Mahan Correctional Facility for Women comes just a year after the state reached a settlement with the American Civil Liberties Union of New Jersey to house transgender inmates according to their gender identity, instead of their sex assigned at birth. The facility currently houses 27 transgender inmates, NJ Advance Media reported.
...Minor, who is serving a 30-year prison sentence for manslaughter, impregnated the fellow inmates earlier this year following consensual sexual relationships, Sperrazza confirmed. He added that even consensual sex between inmates is prohibited in the state.
In a letter from Minor posted to the blog Justice 4 Demi, which often publishes videos and letters from her, she described her removal from Edna Mahan as punishment for having sex with two women.
“As a woman who is transgender I truly fear what lies ahead for me, its clear that staff have looked for some kind of security reason to kick me out of the only female correctional facility,” Minor wrote in the post, which was published Friday. “They have thrown me to the wolfs and expected for me just gave up, I am baffled and disgusted by this use power and I can only ask that the commissioner and governors office please send me back. I don’t deserve to be treated like this.”
Remind me again about those completely unrepresentative edge cases? So unrepresentative that the ACLU is forcing prisons to comply with them?
> But if you want to get beyond language, then you need to get into sex-related social classifications, and there are lots more than three.
Well that's the thing, there weren't. Not until about ten years ago. People were male or they were female, and for the 0.018% of the population that was genuinely intersex (https://pubmed.ncbi.nlm.nih.gov/12476264/), well, they managed to figure something out.
As for transsexuals, they existed, but were acknowledged to be people pretending to be members of the opposite sex (hence trans-sexual). You could either go along with this pretence or deny it, depending on how you felt about them personally, and nobody would be outraged if you pointed out "psst, by the way, that's actually a man".
The real big change that got pushed on people was not "there are more than two genders" but that "gender is something other than a polite synonym for sex". This concept started out in academia but got pushed really hard onto the general population very recently.
It is not really much of a distribution. That is the point. As far as "it", for the vast majority of the past "it" was specifically not seen as polite as it implied the subject was not a person but an object.
The problem is the people who lump the Ts in with LGBT. Almost no one has a problem with the Ls and Gs anymore. But the Ts are a problem, and they're tainting their supposed cohorts. Excise the Ts from LGBT, and suddenly support for gay marriage will shoot back up, probably even higher than it was before.
The "core pillar" in this instance is "men are men and women are women." The Ts and their "allies" are trying their damndest to challenge that, and as people take them more seriously, even more people become their enemies, and since the Gs and Ls are ostensibly allied supporters of the Ts, they will earn more enemies too.
Oh dear. Y'know, I've always thought that "men are men and women are women," phrased like that with all the implications that the sexes are entirely separate and any presence of the characteristics of one in the other was perverted, is one of the most pernicious and false memes out there, and I thought this long before I ever even heard of transgenderism. The sooner it's gone the happier I'll be, and I speak as one without a single T bone in my body.
I don't know who you're arguing against, but it certainly isn't me. Men and women both have four limbs as well as whatever characteristics you're referencing. They're still separate categories.
The people who are most essentialist about gender roles these days are the pride folks. Oh your little boy likes to wear a dress. He must secretly be a girl inside. It is just stupidity, as well as harmful.
Unfortunately, the transgenderism argument right now heavily depends on "men are men and women are women". See our dragon friend on another thread claiming that gender is a real part of the brain and they had an identifiably female brain in a male body.
The worst (online) cases go for "pink for girls, blue for boys" and you get a goddamn expert claiming that you can tell if a pre-verbal infant is trans because, for example, they pull the barrettes out of their hair, that means they are protesting their true gender identity is male not female:
https://4thwavenow.com/2016/09/29/gender-affirmative-therapist-baby-who-hates-barrettes-trans-boy-questioning-sterilization-of-11-year-olds-same-as-denying-cancer-treatment/
So the likes of this article should be rewritten to say if your baby pulls out the hair clips, that means you have a son not a daughter:
https://oureverydaylife.com/put-clips-babys-hair-11373.html
To me it seems that the main difference is that in case of LGB, the message for the rest of the society is "hey, they are different, so let them do their thing, and you keep doing your thing; let's agree it's no one's business what happens in someone else's bedroom between consenting adults". Yes, it is a bit more complicated, but 90% is this.
With T, it's not "their private business" anymore. Everyone is expected to play the game of declaring pronouns and being very careful at using them. And instead of "do your thing, and let them do their thing", you are considered intolerant if you refuse to date them.
Imagine someone saying: "hey, I am perfectly okay with other men dating men, but I definitely wouldn't date one". Unless I missed some recent changes, this is considered 100% tolerant. Now say the same thing about T -- and enjoy your ban, if the place is sufficiently woke.
The Ts were always part of the umbrella because it was most politically pragmatic to be there, not because of anything even approaching philosophical or scientific consistency. Conservatives have switched to attacking them because as their insane ideas gain more traction in popular discourse, they get the attention of their interlocutors. Conservatives almost entirely aren't homophobes anymore. Yes, there are too many homophobes who are in power, but the vast majority of conservatives who oppose the T ideology are doing so strictly because of the Ts, not because of some underhanded subterfuge detectable by your incredible mind-reading abilities.
It is the trans lot who are making the waves, though. Like the Johns Hopkins LGBT Glossary which had to pull its definition of what a lesbian was, because they made it "non-men attracted to non-men" but left gay man as "man" and not, for instance, "non-women".
The Blanchardian AGPs with their feminine penises and spinny skirts and the "9 out of 10 times 'non-binary' means a chick" types managed to get the definition their way - for a short while until sanity prevailed.
https://www.nbcnews.com/nbc-out/out-news/johns-hopkins-pulls-lesbian-definition-uproar-use-non-men-instead-wome-rcna89307
"Lesbian (sexual orientation): A non-man attracted to non-men. While past definitions refer to ‘lesbian’ as a woman who is emotionally, romantically, and/or sexually attracted to other women, this updated definition includes non-binary people who may also identify with the label.
Gay Man: A man who is emotionally, romantically, sexually, affectionately, or relationally attracted to other men, or who identifies as a member of the gay community. At times, “gay” is used to refer to all people, regardless of gender, who have their primary sexual and/or romantic attractions to people of the same gender. “Gay” is an adjective (not a noun) as in “He is a gay man”.
But what about the non-binary gay-identifying personx? Seems that gay men don't put up with that shit 😁
Other than Stonewall, how many riots really were there? And how instrumental were they really for repealing sodomy laws / getting police off their backs?
Gays were tolerated but mainly disliked in the late twentieth century. Acceptance of homosexuality declined somewhat in the 70s and 80s, then began to shoot up rapidly in the 90s-2000s-2010s in what is probably one of the fastest changes in public opinion ever recorded. This change was not led by radicals, and it was not just about marriage, but about being treated as social equals.
Conservatives “deciding the enemy was weak” I think is a pretty bad model for what happened after.
"It of course didn't help queer people that all their elite, leadership and elders were decimated by AIDS, which conservatives in power at the time deliberatedly let run in the hope that it would get rid of the "problem""
Oh, go fly a kite. May I point you to the monkeypox breakouts, where when asked "please don't fuck multiple strangers for a while until we get a vaccine out", most gay guys responded that it was their human right to go to piss orgies and bang forty guys over a weekend.
Oh yeah, those wicked conservatives forcing gay men against their natural instincts to be promiscuous and fuck around. How terrible!
People handled AIDS badly, because there was a ton of ignorance around it and yes, a ton of fear-mongering. But if you honestly believe that the only reason the "elders" of the gay community got sick and died was because the conservatives had a magic cure which they with held in the hopes that the gays would all die out, you're on crack.
> Oh, go fly a kite. May I point you to the monkeypox breakouts, where when asked "please don't fuck multiple strangers for a while until we get a vaccine out", most gay guys responded that it was their human right to go to piss orgies and bang forty guys over a weekend.
Because of neo-Marxists who thought gay people would make a useful client class and that integration is contrary to that goal. They are currently imposing that whole set of doctrines on basically all gay spaces and excluding anyone who dissents. Most people, gay or straight, are pretty much spineless, and will swing around to the prescribed viewpoint when faced with that kind of pressure.
> But if you honestly believe that the only reason the "elders" of the gay community got sick and died was because the conservatives had a magic cure which they with held in the hopes that the gays would all die out, you're on crack.
Very true. Also, conservatives were not ignoring it. Both William F. Buckley and Pat Buchanan argued that more should be done. Of course, Machine Interface is referring to the Reagan administration and its famously homophobic press secretary, but even the Reagan administration was quietly funding AIDS research while letting the press secretary be dismissive for PR reasons.
To be fair, there are extremists who do everything that is claimed. It doesn't take much to be proclaimed as a "spokesman", the main thing is you've got to want to be noticed. And the news thrives on sensationalism, but they prefer to not exactly lie.
Well, I disagree with Yudkowsky a lot, but I think there will be explosive economic and technological growth and a transition to a radically new world order, this decade or the next. At some point there will be so much going on that attention will be diverted from current fights, and afterwards… who knows
You don't need AGI to change "politics as we know it". 2024 has been projected to be "the first deep-fake election". And the number of scam phone calls that I receive has more than quadrupled over the last few weeks. I mean that literally. I used to get one or two every couple of days, these days it's so many that if I look at my "recent calls" list I don't even see my sister, and I talked with her only a couple of days ago.
I've gotten fewer. Guess I'm not that popular.
How many of your political opinions are a result of reading online articles and debating people online? Not in the sense of "I have read this one article / talked to this one person, and it completely changed my opinions", but rather as a cumulative effect of reading things and talking to people for years?
AGI will be a revolution in astroturfing. Give it enough computing power and bandwidth, and it will produce more articles and comments than the entire humanity combined. Creating personas for all political groups, all personality types, adapting messages for each of them, constantly A/B testing and updating its approach.
If one company gets AGI first, I imagine they will be able to tell it "create zillions of accounts on social networks and convince people that X" and five years later, actual people will democratically vote for X, because all of them found a perfectly convincing reason to do so, a different reason for each of them. (It is not actually necessary that all of them do so; 51% is enough.)
Just imagine something like ACX open threads -- how many people have you actually met in person, or can otherwise verify their humanity? Once the AGI succeeds to convincingly post here, it does not need to limit itself to one account -- creating one hundred accounts, representing different personality types and different political opinions, will simply be a question of running the same algorithm with different prompts hundred times.
Later, when the owners of the AGI click the button, you will find different people providing intelligent reasons for why X is actually the good thing. Some will argue from the socialist perspective, others from the libertarian perspective, some will find an analogy in Bible, there will be contrarian arguments for X, there will be nice and polite arguments for X, but also snarky arguments, etc. Not all of them at the same time; the accounts will switch to X in different schedules, pretending to be convinced by different things.
You can get paranoid, of course, and decide not to trust anyone you don't know in real life. But if many people do that, it will significantly change politics, too.
>create zillions of accounts on social networks and convince people that X
I don't know about that. I find people seldom change their ideas based on arguments, no matter how solid they are. They will simply dismiss them, thinking something must be wrong with the argument, if it contradicts their priors.
On the other hand, people DO change their minds. This must be for other reasons. Maybe AGI can tease some reasons out of data and put forth such things? Make up a "celebrity" that thinks such and such a product is entirely awesome, or this political stance is dumb?
What you mean is that they seldom change their minds based on arguments that you find convincing. That's why the need for lots of different arguments from different individuals. "Social pressure" has a strong influence on the way people decide things, and on how strongly they hold that belief, but they feel the pressure most from those who hold many opinions similar to their own opinions. This often isn't decisive, but it strongly shifts the gradient.
OTOH, if you mean "people don't change their minds because of logical arguments" I largely agree. That's a rare event.
I once told some friends of mine that a number divided by zero is undefined, not infinity. I explained that if it were infinity, then you could multiply infinity by zero to get the number. They remained unconvinced. Then another friend who was late to the gathering, but had a masters degree in physics, was asked about this, and he said it was undefined. Argument from authority trumped logic.
I'm not sure how this applies to AI, but it does back up your argument.
Well, accepting it as infinity was based on argument from authority (their math teacher) so that's somewhat reasonable. The real answer is "It depends on a larger unspecified context". In certain math it's defined as infinity, and it's (0 * infinity) that is an undefined number. In other math the division operator isn't defined on 0 in the denominator. It depends on exactly which axioms and rules of inference you use (there are several that are the same over most of their range, but vary in how they handle edge cases).
I assume the opinions would mostly be changed under *social pressure* ...of the sock-puppet accounts.
If you see 20 smart people on ACX repeatedly providing smart arguments for X, on the emotional level you will conclude "many smart people in my bubble believe X", and that is what may actually make you update towards X. The arguments themselves will only give you a rationalization for why you did that.
To put it more crudely, AGI will be able to infiltrate and manipulate many "hive minds", because it will be able to become a majority in every online group (that does not explicitly protect itself against that).
This is all so hypothetical. I feel like this rationalist community claims to have a strong focus on evidence, and yet frequently makes extreme hypothetical predictions about things like AI that don't even pretend to have any hard evidence for them, and are basically "this just feels like something that might happen". Is there some kind of study or experiment giving a reason to think what you described is possible? Or alternatively, a logical argument that its possibility necessarily follows from premises we all accept?
Because on the other hand, the fact that human governments, corporations and institutions have not with all their resources n 51 centuries of civilisation managed to discover a reliable algorithm for controlling people with the right words, certainly seems like decent evidence against an AI (using only the outputs of said human civilisations) accomplishing it in a couple of years.
I don't think people would if the posts sounding convincingly like they were written by human beings, and if there was variety in them. Why would people? If all the posts were in favor of a view we were positive was nonsense we might eventually wander away to a different forum. On the other hand, our loss would be balanced out by the arrival of true believers in the view.
That would be awesome. (Well, depends on what they would start doing instead.) At this moment, it is quite difficult for me to imagine.
Ah, I don't insist that this will happen overnight. Might take ten or fifteen years, possibly more (but not as much as a century, I think).
Anyone know of a way of reading Scott's old blogging stuff? I keep running into shit I would like to read that's locked in his Squid314 thing.
https://archive.fo/fCFQx
(NOTE FOR READERS: don't use Firefox to open this link!)
Why? It works in Firefox for me, at least.
Works for me, too.
I got the error SSL_ERROR_NO_CYPHER_OVERLAP on Firefox though.
Perhaps a temporary problem that they fixed. (Or maybe a different version of Firefox, if you keep getting the error?)
It's a paradox that SSL is supposed to make web browsing safe, but too many websites have it configured incorrectly, so the browser has a choice to either ignore the errors (which kinda defeats the purpose) or disable half of the internet.
The Wayback Machine is your friend
Doesnt work
My brother and his girlfriend has taken to wearing a mask again to crowded places. He says this is because there is evidence the mysterious excess deaths that have been reported are due to sequeleae of COVID, meaning COVID is more dangerous than commonly believed.
I think this is absurd, but well, I hadn't heard that angle on the excess deaths. Did this community ever reach a consensus on what's causing those? Does it actually make sense to mask up at this point, or is it ridiculous?
Excess deaths notwithstanding, it seems reasonable given the recent RCT re long Covid mentioned by Derek Lowe, as discussed in the recent Hidden Open Thread. The diagnosis may still be somewhat inchoate, but it's something that blinded doctors can reliably diagnose at 300+ days post-infection, with a 40% difference between patients treated with metformin and the controls (but which the treatment group still got if they rolled a 1 on their d20).
Something that's going to mess up one in ten to one in twenty patients for a year or more still seems well worth trying to minimize.
https://www.science.org/content/blog-post/long-covid-shows-card
But there's also the other issue of when does it stop, since COVID is not going away. Are we supposed to mask up forever?
No one's going to make anyone do it going forward, so at this point that's a matter of personal risk management. "We" clearly aren't. Whether I do remains an open question.
I don't personally have a current expiration date on masking, any more than I do on wearing my seat belt, which latter I think probably mitigates less risk for me. (Though more than enough to justify doing it.) I'll presumably stop when the benefits don't outweigh the downsides, which isn't currently the case and I don't know when or if it will be.
For most people, evidently that point came months or years ago. (And they're not asking me for advice. :-) ) Other people generally find the personal cost of masks higher[1], and/or estimate the risks as lower, and so make different choices.
I'd probably trade off masking for other effective risk mitigation efforts if something convincingly emerged. But aside from a long term improvement to building ventilation standards (that may come to something in years or longer, or may get killed before it happens), we're not going to get any of those.
[1] Granted, I'm old enough to remember when seat belts were treated as basically torture implements used only by the most nervous of nellies, and probably an actual danger due to interfering with the presumed safety of being thrown clear of the accident. (At speed, through the windshield. Yes, they were serious.) I'd say that's gone, but I didn't expect childhood vaccination to become a matter of serious controversy. Maybe that one's just sleeping too and the call to "throw off the bondage gear" is just awaiting its moment.
About ventilation: An air purifier giving 12 air changes per hour protects you approximately as well as wearing an N95 (except that if you sit quite close to someone you will get their exhales before they've been through the air purifier, and of course then if they have covid you will inhale it). I have been meeting unmasked with psychotherapy patients at least 10 hours per week for the whole pandemic, and I have not had covid. We sit about 10 feet apart, and between us is an air purifier meant for quite a large space running in my medium-sized office. It gives 11 air changes per hour. Any building you live and work in is never going to upgrade to this level of air purification, but you can bring and office or a room in your home up to it. By the way -- like you, I mask up with an N95 in indoor public places still. Mostly it's stores, and I spend maybe an hour a week in stores, so the nuisance value is entirely trivial in my life.
I do have air cleaners in my home and office. But I'd kind of like to see them anywhere I don't control. Or failing that for people to at least crack a window.
I do see them in some medical facilities. I especially appreciate my dentist, since obviously I can't mask there. 🙂
Yeah, don't hold your breath when it comes to ventilation in public buildings. I wish we'd at least get it in the schools. Do you know about Corsi-Rosenthal boxes? Cheap home-made air purifiers that perform well. Yeah, I appreciate my dentist too. I saw him really early on in Covid, when there was no vax, and the poor guy was wearing 3 masks at once. Said they had improved the HVAC and were about to upgrade again.
Over the course of reading around about viruses during covid, I've learned that viruses in general are bad news, not as simple as a bug that makes you sick and then you get over it. Quite a number of them have bad consequences a ways down the road: mono ups the chances of I think MS -- anyhow, something neurological; papilloma and cancer; 1918 flu and severe and weird cognitive syndromes (the ones Sacks describes in *Awakenings*); shingles and Alzheimers; polio and post-polio syndrome. There are others. I am now much more concerned about getting a viral illness than I used to be. It does look like Long Covid is a real thing, though rarer than many believe (maybe 5% of cases?) There are a number of studies finding increases in this and that in the months after covid -- neurological and cardiovascular and psychiatric problems, mostly. I'm not sure how valid any of the individual ones are, but there are so many that I'm inclined to think there's something there.
About it being ridiculous for people to wear masks. I don't think it 's *ridiculous*. They definitely reduce the chance of getting covid, but the chance is pretty low now anyhow. Seems to me important not to bring extra emotion to the situation because the precaution is *masks,* and there has been lots of emotion and fighting about masks. Lots of people probably take precautions you don't. I have friends who filter their water, friends who mostly eat organic vegetables, friends who insist on electric stoves rather than gas. You probably do too -- and it's a lot less irritating than masks, right?, even if you think these precautions are unnecessary. Try to put masks in the same mental category.
Well, except for the lip-reading issue, which seems to come up *very* rarely, wouldn't heavy clown makeup be just as bad? I mean, a mask might even be less of a problem when it comes to establishing trust, committing crimes or engaging in anonymous political violence. Somebody can yank off your mask and expose your face instantly, but it's actually slow and difficult to remove the thick white base and features painted on top of the white in clown makeup. I know somebody who performs wearing it.
I haven't seen any data on how many mask-wearers are criminals using the mask as disguises. Have you? I have seen movies where criminals wear something like clown makeup as a disguise. Covers their face up well, and if it's a smiley clown probably presupposes people to think they're benign. I have no data on how frequently heavy clown makeup, a bunch of face-painted designs or the kind of thing people wear to costume parties is worn to hide the criminal's real appearance, either. Do you? As for how the left-wing attorneys would handle some shopkeeper's ripping the mask off someone who was shoplifting: It seems to me unlikely that a furious store owner would stop to think about that , or even know about that,so it probably would not deter him much. If I was going to shoplift at a store I would definitely wear something like heavy makeup that looked like I was going to a costume party, rather than wear a mask and count on the angry owner's fear of leftie lawyers to protect me from having my face exposed.
There are excess deaths for a period (low long? at least 6 months) after COVID has apparently been recovered from. There are also lots of other problems that seem to persist and recur in at least a few cases. Whether it's "more dangerous than commonly believed" depends on your source, but it's definitely more dangerous that many people want to believe. And the dangers aren't all death. There are lots of cases of prolonged "brain fog", weakness, and other effects. What this is percentage wise I've no idea, as I don't think that there's even a good definition of what counts as "sequeleae of COVID", i.e. one that allows it to be counted.
Your brother isn't being that unreasonable as long as he's in the habit of taking other health precautions, like watching his diet, exercising, etc.
> Your brother isn't being that unreasonable as long as he's in the habit of taking other health precautions, like watching his diet, exercising, etc.
Surely it would make more sense to be careful about covid if you have risk factors such as being generally unhealthy?
But for your first paragraph, +1
My brother now drops this map from Our World in Data on me:
https://ourworldindata.org/grapher/cumulative-excess-mortality-p-scores-projected-baseline
He attributes this to COVID and potential sequelae. Does this really make sense?
Did you mean to include a different link? That is showing excess deaths since the start of the pandemic. It indicates that lots of extra people since the start of the pandemic. Where do you see anything about Long COVID?
In this video: https://www.youtube.com/watch?v=-ZI7dwldSR0 from 4 months ago, Dr. Roger Seheult suggests that the main sources of current excess mortality are from increased cancer deaths, as a result of decreased testing and care during the pandemic, as well as indeed increased risk of cardiovascular disease after a COVID infection. [1]
As far as current excess mortality, you can see it broken down by age and country here: https://ourworldindata.org/grapher/excess-mortality-p-scores-average-baseline-by-age?country=~USA.
It shows a current decrease in mortality in the USA for those ages 0-14, basically no increase for those aged 15-64, and a moderate increase for those 65+.
So while overall mortality is indeed up, and while Long COVID may well play a role in that, it is not much of a factor for young people who are generally at low risk for COVID, in general.
And if the fear is Long COVID, what good will the mask do? He has almost certainly already contracted COVID. Is there reason to think that an initial infection wouldn't cause Long COVID, but a subsequent infection would? I would assume the opposite.
[1]
He also notes increases in external causes of deaths. For example, as you can see here: https://www.usatoday.com/story/news/nation/2023/01/26/crime-rate-homicides-shootings-declined-2022/11075070002/, the average number of homicides per 100,000 from 2015-2019 was 5.146, while for 2022 it is estimated to be 6.457.
Similarly, as you can see here: https://en.wikipedia.org/wiki/United_States_drug_overdose_death_rates_and_totals_over_time, from 2015-2019, drug overdose deaths averaged 19.94 per 100,000, whereas in 2022, it is over 33 (https://www.reuters.com/world/us/us-drug-overdose-deaths-top-109000-past-year-2023-06-14/).
"Excess deaths" Is really a fraught endeavor.
It is not well operationalized. Excess compared to what? The last 20 year "average", 10 years, 5 years? Does the model adjust for changing demographics? What about adjusting for actual or perceived trends in medical improvement? How do you account for different rates of improvement. If opioid deaths are going up because of ease of access but motor vehicle deaths are going down because cars are safer, what do we mean by "excess deaths"?
There is also no standardized way to account for cause of death. Every county in US is doing their own thing within the context of 50 states + dc and territorially rules.
And then there is noise/systemic variation. How much difference do you need to see to justify looking for a special cause?
As to the particular issue of masks, there is little evidence that it can make much of difference for repeated exposures. Let's imagine that a well fitted mask is 83% effective. You roll a 6 you lose. You roll 1 -5 : phew. You ever play Parcheesi and not roll a 6?
At a population level or as public health measure, I think Cochrane has shown it's not going to make a difference.
"Let's imagine that a well fitted mask is 83% effective. You roll a 6 you lose. You roll 1 -5 : phew. You ever play Parcheesi and not roll a 6?" That's not how it works, though.
An 83% effective mask reduces the amount of covid you take in by 83%. If one lousy covid molecule gave you covid then an 83% effective mask would leave you with a one in 6 chance of getting covid even if there is so little covid in the air that on average you're likely to breathe in only 1 molecult. However, there's a dose threshold. It takes a certain amount of covid for you to be likely to become infected, and it's way more than one molecule. And even if you are infected, severity of illness is partly dependent on how much covid you got in your exposure.
Also, if you buy good masks like 3M Auras, and are fussy about checking for leaks, you're probably getting significantly better than 83% protection.
That is not how cdc and others are using 83% effective.
And you're really getting no protecting really if you take off you mask to eat and drink.
Except with well fitted mask for very short duration, I doubt very much individual protection. They just don't work in any meaningful way, which is why seroprevalence is like 98%. Nearly everybody has had it and will get out again.
"which is why seroprevalence is like 98%" US covid seroprevalence could not possibly be anywhere near 98%, even if every single person in the land had had covid. Antibodies disappear after a few months, so by now many who had had an infection or vaxes would test negative for covid antibodies. If you don't like that line of reasoning, here's the data to back it up. See Table 1:
https://www.ncbi.nlm.nih.gov/pmc/articles/PMC9959557/#:~:text=Seroprevalence%20of%20anti%2DS%20and%20anti%2DN%20antibodies%2C%20stratified,%2D2%20infection%2C%20and%20vaccination.&text=Anti%2DN%20antibodies%20were%20positive,the%20≥70%20age%20group).
"Of course you're getting no protection at all if you take the mask off." When I was masking in indoor public places I never took it off because, duh, then I'm unmasked in an indoor public place. Just did not eat and drink in those places.
"That is not how cdc and others are using 83% effective." What others do you have in mind? When NIOSH certifies masks as X% effective, that *is* what the percent means -- if filters out X% of particles of the size covid normally rides on.
What is it the CDC means by 83%? Does it mean that in practice a good mask certified at 95 or 99% only gives 83% protection because people don't put them on right, or take them off to eat in high-risk settings?
I saw some people using 83% to refer to allegedly compare infection rates among ostensible mask wearing populations vs non-mask wearing populations.
Yes n95 does mean 95% of particles of certain size -- bigger than virus - virus filtration is electro static filtration but I really don't think it's ever been well measured outside lab in real life with real people and real respiratory viruses
Take a look at NIH seroHub https://covid19serohub.nih.gov/
I was doubtful until Monica Ghandi tweeted it with reference to serohub.
Everybody has had it at least once which itself tells you how noneffective masks were.
Just stared at the site for quite a while. Here are the things that puzzle me:
-After infection antibodies fade over time, even though protection remains via T cells, etc. It would seem that the waning effect alone would lead to many people who had had covid or a vax, especially if they had it a year or so prior to test, testing negative.
-to see the people who were definitely seropositive from infection, rather than vax, you need to look at people who had nucleocapsid tests. So if I look at the table on the main page, the one you sent me to, and ask it to display percent of population seropositive on nucleocapsid tests, the total % by the latest date, July of '22, the table shows is very close to 100%, I believe it's 96%. But then if I sort by age it shows people 65 plus having the lowest test positive rate (on nucleocapsid test), and across whole life span the pattern is that the younger the group the higher seropositive rate. However, the last date when all ages are reported was Feb. 2022. After that, all the data is from kids. Could it be that the main table, when not sorted by age, is only using the kid rates for seroprevalence after 2/22? Or it the data after 2/22 just not broken down by age by the study they drew on? In that case, are we sure the group they calculated the 7/22 rates had a representative age distribution?
-For maybe the last year I have been reading, in respectable places, doctors saying that yes, there are still people being hospitalized for covid, but they are mostly unvaxed people. If by a year ago 98% of us had enough antibodies to turn the test positive, we are basically all vaxed or the equivalent at this point. So why are unvaxed people people doing worse? Googled it to double-check that they still are. Found this, which only covers Washington State, but I'd be surprised if it was different in other states: "From May 06 to June 02, 2023, unvaccinated individuals were between 1.9 and. 5.1 times more likely to be hospitalized with COVID-19 compared to those who received at least one booster dose. (from https://doh.wa.gov/sites/default/files/2022-02/421-010-CasesInNotFullyVaccinated.pdf)
-I had read recently that seniors had the lowest seroprevalence rate, approx 50%, and that was attributed to their greater caution. Virtually all the people I know who believe they have not been infected are seniors who have been extremely cautious, and they're well-educated enough to know about things like mask quality, not taking your mask off in a crowded setting even briefly, doing home tests before gatherings, etc. I am nearly old enough to count myself a senior, and I don't believe I've had covid. I have been quite cautious, used high quality masks, and tested myself several days in a row the few times I felt off. Tests have always been negative. Same for the seniors I know who have been cautious. Seems unlikely to me that all of us have had asymptomatic cases, given our age and the risk factors many of these people have.
Lots of older people had covid, survived it, but remained bedridden or just generally much weaker. Now, these people are dying several years before they would have died "naturally". I think this explains the excess deaths mystery quite well.
I don't think there's an excess deaths mystery, it's overdetermined.
1. Covid caused people to stay inside rather than going out, they got fat and sedentary, and then they die
2. Covid caused people to put off routine health care procedures, they miss their cancer diagnosis, and then they die
3. Covid lockdowns caused certain other diseases (importantly, flu) to reduce in prevalence for a while. Due to complicated reasons we don't fully understand (see Scott's post on seasonal diseases) this flu holiday caused flu to come back especially hard the next year, causing people to die
4. What you said above
I don't know about this community, but in Switzerland the number of excess deaths in 2022 is definitely a topic. It's actually larger than in 2020 and 2021. Some can be explained by heat waves in summer and influenza in winter, but most of it can't. The mainstream explanation is two factors:
- COVID is much less obvious and goes unnoticed much more often, but is/was still extremely wide-spread. For elderly, it may contribute a lot to excess deaths, and those may often go unntoiced.
- The medical/social services in 2020 and 2021 were under extreme stress, so non-vital services have suffered. For example, screening programs were worse, and less cancer and cardiovascular problems in early stages have been detected in 2020/21. Back in 2020 and 2021, doctors already warned of a wave of deaths because of early-stage health issues that were missed during that time.
Currently it is not clear which of the two effects is dominant. Death from an asymptomatic or atypical COVID infection often looks pretty much like a cardiovascular problem, so we can't just tell them apart by looking into basic statistics. We might never know.
To add to the Swiss perspective, excess deaths in 2022 have been a thing both in the UK and in the EU as a whole, and COVID-19 or its immune-related sequelae may well be part of it. I would be disinclined to blame services stress just because the distribution seems odd for that (e.g. some of the countries with the best performing systems, such as Germany, reporting highest rates of excess mortality in some months, such as the ghastly December).
On the other hand, the measure is volatile, and 2023 seems to be looking better thus far, so who knows.
My guess.
People got lazy and fat during lockdown. Also people are generally only getting more sedentary. Also COVID does have some small amount of long term damage post viral illness.
I would think the increased risk for people who have had Covid and/or are vaxxed/boosted is pretty minimal compared to the cost of masking.
If the person wearing the mask doesn't mind it, what's the cost? I guess other people might find it harder to communicate with them. But do they really have to change their behaviour for that?
Crime - meh. Communication yes.
But the loss is civilization that is rational vs civilization that is superstitious. Civilization that is superstitious but pretends it is rational might even be worse than just superstitious.
Risk from false sense of security. Thinking a mask will protect at crowded place when you and everyone are taking it down to eat and drink and or when you aren't even wearing mask correctly (the chin mask, the ill fitted mask, etc.) These are all problems.
"But the loss is civilization that is rational vs civilization that is superstitious."
Um, wearing a mask is clearly less superstitious than wearing a cross, and civilization seems to be surviving that OK. Even if you think religion as a whole is bad for civilization, we're not talking about that. We're talking about wearing a cross.
Risk from false sense of security. Thinking a mask will protect at crowded place when you and everyone are taking it down to eat and drink and or when you aren't even wearing mask correctly (the chin mask, the ill fitted mask, etc.)". Smart mask wearers don't turn their masks into chin diapers and sit in a restaurant for an hour that way. But as for those who do -- are you aware of how much stupid shit people do ALL THE TIME? Have you noticed what your fellow drivers and voters are up to? What they read in totally untrustworthy places like Facebook and believe. How they trust their Auntie Ellen's advice over their doctor's?
Shitty communication seems a pretty big cost.
When I'm talking with a deaf person who lip-reads, then I'll worry about it. Otherwise not.
Banning masking at protests seems sensible, but masks were allowed / encouraged / mandated in indoor environments for two years to no ill effect. Black oversized sweatshirts are also often worn by criminals and rabble rousers, should those be banned too?
I deny that there was no effect. I live in a place with about 100 people, and there were 5 of 6 cases of COVID during the period before vaccines were properly available. Masking was in force, and the disease did not spread. (OTOH, other precautions were also in force, so this isn't definitive proof at even the "small study" level.)
I haven't heard about any mysterious excess deaths. Is this a real thing?
Not according to the CDC. Their numbers show the most recent bout of excess deaths starting in late May 2022 and ending in the middle of January 2023. No excess deaths since then.
https://www.cdc.gov/nchs/nvss/vsrr/covid19/excess_deaths.htm
In what earthly way are they "self-harming"?
Social relations suffer a lot.
* Non-verbal cues are hugely important in order to have a normal conversation. Humans will reflexly smile at any point were their discourse may be interpretted in a hostile manner, in order to diffuse the conversation. Without that, expect all your interractions to be just a bit more hostile.
* I have no data on this, appart from my own experience, but it's generally harder to understand the content of the sounds that come out of your mouth if you add just a single filter between the vocal cords and the ear. Add to that the fact that you removed lip-reading. I basically caught covid because I couldn't handle a week of trying to socialise and make friends while wearing a mask.
* As we have confirmed during this crisis, people will go a very long way to keep themselves unaware that there is a danger. I expect that people will unconsciously try to avoid even looking at you if your appearance reminds them of “the terrible years”.
>Social relations suffer a lot.
No, they don't.
Are you forgetting that mask-wearing was nearly universal for about a year, not very long ago, and that therefore anybody you tell this stuff to has actual experiences to measure it against? My experience, and that of anybody I talked to about it (admittedly this wasn't a big topic of conversation--I wonder why?), was that if the need to show you were smiling was really important, which it mostly wasn't, you could compensate by "smizing," and that's what people did. Ditto the allegedly sociality-destroying properties of the slight muffling effect of a mask; on the rare occasions when it caused an actual problem, you compensated for it.
I can't address your third claim because I can't imagine caring that people are "unconsciously trying to avoid looking at me" in that situation even if I suspected it were true, which I have no reason to.
But let's pretend for a moment that *every* social situation requires the techniques of slight compensation that I mentioned. Does this really rise to the level of "self-harm"? Words mean things and have connotations. We generally use that term in the English-speaking world to mean something rather more alarming than a mild inconvenience, willingly and temporarily suffered in order to maintain health.
For what it's worth I haven't worn a mask on my own initiative since 2021, and I don't find the original commenter's brother's logic for doing so very sound.
I had never heard the word “smizing” before, but it really has a “everything is fine” vibe to me ; like the advices to touch elbow instead of shaking hands : pretend that the costs don't exist.
And I would contest the whole notion of smiling with your eyes.
* First, the whole reason we tall people to also smile with their eyes is that for a large part of the population, or a large part of the smiling-situations, a mouth-only smile is enough.
* In fact, the mouth-only smile has a specific semantic : “I don't really mean it, but we're in a situation were a diplomatic gesture would be appropriate, and the fact that I'm willing to make that diplomatic gesture is my message”.
* And there are dozens variations on a mouth-smile that the eyes are simply not able to carry. Sad-smile, smirk, half-smile, grin, etc.
In that context, your statement that “you compensated for it” really triggers my anger. There are a large quantity of communication sequences that are already on the border of missunderstanding/starting a feud/awkwardness and there is simply no slack for compensation. I speak as an autistic guy who straddles that line every time I open my mouth.
Even when “compensation” is possible, that is the price that you pay (i.e. the “self-harm”) : everytime you ask someone to repeat, you check if they are sarcastic or not, etc., you've just made the conversation just a little more awkward.
As for the “unconsciously trying to avoid looking at me”, I point you to all the studies about involuntary masking in muslim fundamentalist societies.
Also there's no reason to believe that the studies on babies, showing that anything that has 2 eyes and one mouth is a topic of interest while anything that doesn't isnt, wouldn't apply to adults.
>In that context, your statement that “you compensated for it” really triggers my anger
I hope you get some help for that at some point.
But like I said, in the real world, wearing a face mask to prevent the spread of disease is a slight inconvenience, easily compensated for; and adults, by and large, aren't babies.
There's also the fact that many people in large swathes of East Asia have been wearing them in public spaces for years before the pandemic and continue to wear them after mandates have been lifted (South Korea, Taiwan) without any obvious hit to the quality of social interaction. We can make the usual points about culture, but then let's set aside universal claims about humans.
Much of this veldt-evolved social stuff has long been losing importance, anyway, in the face of cutting edge social technologies like the pen, paper, and the postal service.
Oh, I thought the masks were what made Asians sound like they had Asian accents -- you know, sound distortion. (jk)
> Are you forgetting that mask-wearing was nearly universal for about a year, not very long ago, and that therefore anybody you tell this stuff to has actual experiences to measure it against?
Yes, but that was a couple of years ago. And it sucked.
Wearing a mask in 2023 you're marking yourself out as "one of those weirdos", which will cause social relations to suffer, because nobody wants to talk to one of those weirdos.
>Yes, but that was a couple of years ago. And it sucked.
A lot of things suck but are done anyway rather than endure the even suckier consequences that can arise from not doing them. Wearing PPE of almost any kind on a hot day, for example, sucks, but getting sparks in your eyes or coal dust in your lungs sucks worse, or eventually will.
>Wearing a mask in 2023 you're marking yourself out as "one of those weirdos", which will cause social relations to suffer, because nobody wants to talk to one of those weirdos.
I suspect you haven't considered the possibility that the mask wearer may not consider their own social relations to have suffered because the kind of people who don't want to "talk to one of thse weirdos" don't want to talk to them. In fact, and I know you may find this difficult to believe, they may even find it a plus that people who know nothing about their reasons for wearing a mask don't want to strike up a conversation. I certainly would, and I'm someone who is perfectly happy to see that mask wearing overall is on the decline.
I sometimes still wear one in especially crowded stores. I haven't gotten any funny looks, but would not care if I did, and do not care if I'm seen as a weirdo. I did care a lot when I was 14 and my mom wanted me to wear dorky jeans from the cheap store instead of the magnificent ones from the cool store, but I grew up and got over that. As for nobody wanting to talk with me -- I have zero interest in talking to strangers in the drugstore anyhow.
Is there really any good evidence that an individual wearing a mask (N95 mask, worn correctly) is just as likely to get infected as someone without a mask? I know cloth masks do nothing and mask mandates don't really reduce transmission, but it's hard to believe that correct mask use has no effect on an individual level. They work for all kinds of dust and pollution after all.
No. But if the exposure is chronic rather than acute the mask isn't going to solve the problem. It reduces the frequency of infection in ways dependent on circumstances. Say it's effective 90% of the time (probably an overestimate). That means that if you're exposed 20 times your chance of not being expose are 0.9^20 or about 1 out of 10.
That's not how it works. Let's take your number 90%, and say the person is wearing an n95, but instead of filtering out 95% of particles of the size covid rides on it's filtering out 90%. So, on occasions when the number of covid molecules per cubic foot is more than zero, you will breathe in 10% of those molecules for every cubic foot you inhale.
However, it is not the case that one lousy covid molecule will infect you with covid. Technically, you have been exposed if you breathe in one molecule. But you have various barriers to infection: Luck (maybe you exhale and that one molecule gets blown back out before it enters your respiratory tract). Nose hairs. Mucus barrier in nose. Probably some other stuff. Also, the infectious dose is way more than 1 molecule. I've seen all different estimates, but suffice it to say it's way more than one.. So to get covid while wearing a mask that filters out 90% of what's in the air you breathe, you have to either be in a place with a fairly high density of covid in the hair, or be in a place with low density but stay there for a long time -- in either case, you have to pull in a good number of covid molecules .
And here's some anecdotal evidence: I have a lung problem that puts me at higher risk than most and I have been pretty cautious. I have worn a quite high quality mask in indoor public places, but not elsewhere, throughout the whole pandemic. It goes around my head, not my ears, and I pull the elastic tight, and I feel around the border for leaks. I have never once removed it in an indoor public place. Not to eat, not to drink, not to to scratch my nose, not because it was a bit uncomfortable. (OK, 2 exceptions: went out to dinner with my daughter on her birthday twice in 3 years, though in places with tables far apart and good ventilation.). I really do not think I have had covid. I have never had covid-like symptoms at any point during the pandemic. When I have had something that could be from covid, like a headache, I have tested several days in a row, and all tests have been negative. I realize I might have had an asymptomatic case, but that seems unlikely to me given my age and risk factor.
As the other comment says. My prior probability that some random medication helps with covid is zero, so absence of evidence means it doesn't work. My prior that filtering air helps with not catching respiratory illnesses is that it seems very likely to work. So absence of good studies doesn't really mean much. Of course, we can't really know how large the effect is. So it whether or not someone wants to wear a mask seems best left up to their personal mask tolerance and risk tolerance. (Completely absurd things like wearing a mask alone on a hiking trail are probably not happening these days any more, or if they are it's due to allergies / fires and not covid).
In the absence of well-carried-out studies, I think it makes sense to fall back on reasonable intuition based on physical arguments.
And my reasonable intuition based on physical arguments is that an n95 mask will do a reasonably good job (and a cloth mask a nonzero job) of preventing virus particles from entering your airway. How could it be otherwise, unless somehow every single virus particle managed to slip through the mask without getting caught in the filters?
There's two possibilities here.
1) You are only exposed extremely rarely, so a good enough percentage of coverage suffices, otherwise don't bother.
2) It requires a significant number of viral particles to cause an infection, and more is worse.
Case 1 implies that masks may well be useless unless you are exposed only extremely rarely.
Case 2 implies that a mask that sufficiently reduces the viral count could be worthwhile even if it doesn't give sufficient protection.
I *think* that the current evidence argues against case 2. It was part of the basis on which masks were originally recommended. But I'm not sure that there has been any definitive study. Certainly the way many people choose to wear the masks makes them totally useless, and that makes many of the studies quite questionable WRT the effectiveness of masks when properly worn.
I mean, how would you possibly do those studies? Studies done by putting masks on dummies and spraying aerosols on them universally show that they prevent aerosols getting into the nose pretty effectively. Studies done by comparing "area with many masks" with "area with few masks" generally seem to show that "lots of masks" don't reduce transmission on the population level. And studies comparing individual maskers to individual non-maskers would be hopelessly confounded, because there's no way the mask is the only difference there.
And well, the problem of there not being a definitive study that shows a clear effect is present in many fields of medicine, Scott has covered many psychiatry-related aspects over the years.
So in conclusion, masking may or may not be worth it. But it seems pretty safe to assume it will somewhat reduce the risk of catching covid in any given situation, if done correctly.
Pekingnology provides each country's official readout of the latest Qin Gang - Blinken exchange:
https://www.pekingnology.com/p/chinese-us-readouts-of-qin-blinken
I'm perennially disappointed by top-level diplomatic exchanges, always hoping for some unpacking or clash over key differences, instead of polite euphemisms like "we should promote mutual respect" or "we should reduce misperceptions."
One possible recent counterexample, German FM Annalena Baerbock was far more direct with Wang Yi than I've heard anyone be, directly pushing back when she heard the talking point that human rights are just an invented Western concept.
I'd like to see more kitchen debates, though not sure the diplomatic establishment really tolerates that sort of thing.
"Invented Western Concept" is accurate. They date to the period called "The Enlightenment". We decided that this was one of the ideas from that period that was worth keeping. They really lost their logical underpinnings with the changes in religious faith, but we decided that they were worth keeping anyway.
The question is "Why should we trust deals with anyone who doesn't think they are valid?".
"invented western concept" recommends it if anything, much like calculus, mechanized farming, commodity futures, semiconductors, internal combustion engines, representative democracy ... :)
Neither Qin Gang in his recent joust with Baerbock nor Wang Yi over the years have rejected the actual idea of human rights as a foreign concept. The talking point is that the Western view of rights is not universal, to which point she (somewhat amusingly) made the legalist move of invoking the UN Charter.
Anyway, a concept of what amounts to 'rights' ensured by mutual duties isn't necessarily worse than one framed in the language of protection from encroachment on freedoms. It certainly can be, but it's not a given.
"the talking point that human rights are just an invented Western concept."
Is that not just straight up undeniable truth? Even if you are very strongly in favour of human rights I can't see any sensible argument that this is not the case. (apart from the platonic view that human rights were just out there and you found them, which is silly - and I can't see Baerbock making the argument considering she's quite famously not very smart.)
I completely agree.
That said, you can get quite a long way from a fairly short set of axioms (suffering is bad all else equal, happiness is good all else equal, one should maximise good and minimise bad all else equal).
I think those - plus empirical claims like "torture causes suffering" are enough to recover a decent chunk of human rights (or at least "human 'good's"). And I doubt that China has a fundamental disagreement around human rights being goods "all else equal".
What I don't think you can easily get is a commitment to the rights as absolutes – it's consistent to say things like "a harmonious society is even more important than individual goods". I think this is probably the crux between the Western view and the Chinese one.
Frankly, human rights aren't absolute even in the West – we suspend the right to liberty when we imprison criminals, for example.
So I think asking "is China obliged to respect 'Western' human rights" is beside the point – both the West and China have a contingent respect for human rights, but not an absolute one. Where they differ is in the weight they place on each right. Of course, this is a pretty significant difference – but it's more subtle than "Does China accept Western human rights?"
Sorry, but to many, certainly to all serious Buddhists, you argument is an argument of killing everyone as quickly as possible. (Well, not to Buddhists who strongly believe in reincarnation, of course. Or in the popular notion of karma, as opposed that that espoused by the Buddha in "The Word".)
You are really giving them the benefit of the doubt there. You are assuming that when the CCP does something that violates human rights, that they are performing some ethical calculation and taking them into consideration and not violating them unless there is a good reason to do so (that a neutral person would accept, even if he might ”weight” it dofferently).
And what does harmonious society mean for an authoritarian regime besides ”society where people obey us”?
Yep, I definitely lean towards assuming too much good faith in general and probably have here too.
I think that Xi Jinping genuinely thinks he's doing the best thing for China. I don't think he would use the consequentialist framing I gave above, but would probably claim to agree with it. I think China genuinely believes that when Western politicians call out Chinese human rights abuse, that they're doing so hypocritically in order to secure political advantage. I think it's possible to have all these beliefs and still act in a self interested way by deluding themselves about their motivations, though I'm not sure that specific flaw is more true of China than the West.
Regarding whether a harmonious society means more than "society where people obey us" – I think it probably doesn't mean more than this (to the CCP). I think the idea of a society where everyone has a place and doesn't rock the applecart is a terminal goal in, say, Confucianism, rather than an instrumental one.
None of this detracts from the proper response to (say) Chinese genocide of Uighur muslims – outrage and total, unwavering opposition. But I do think it's important to see your enemies as they see themselves, or you can't model them properly.
For that matter, I believe that when Western politicians call out Chinese human rights abuse, that they're doing so hypocritically in order to secure political advantage. I still despise the violators for the violation, though, so their hypocritical argument works....except that I want to hold them to the same standard, or better.
But in trying to understand how they see themselves, you seem to come close to suggesting their framework is basically the same as in the West. Despite the Chinese leaders essentially admitting that they think differently.
The old timey Communists used to say that human rights were a "bourgeois" idea. The Chinese saying it's "western" is just the same thing in a new form (you know, "anti colonial" instead of class warfare), talking about the bourgeoisie isn't popular nowadays, but bashing the West is.
Does the CCP's desire for a "harmonious" society where everyone obeys the state differ from the Soviet Communist Party's desire for the same thing because "social harmony" is somehow associated with Confucianism and Confucianism is an old Chinese philosophy? Does it make it more legitimate?
It isn’t silly.
Somehow I tend to forget (or want to forget) how extremely common moral realism is.
It just seems so unfathomable to me.
It's literally all religious arguments while trying somehow to pretend it's not.
The fact that most philosophers are moral realists makes me sad.
Intriguing. Do you believe people ought not be moral realists?
I would have thought that stating almost any opinion commits one to the view that people should or should not believe certain things, that people should or should not espouse certain ideas, which smuggles in at least some kind of realism.
If you're an Ayer emotivist or just want a svelte ontology that seems reasonable, by all means hang on to that. I'm just sad to read your distress when you hear some people believe a theory that, for me, caches out to basically "some outcomes are better than others."
I would welcome your arguments that I should adopt nonrealism. Though I'll confess I'm less concerned about the object level, and mostly concerned about your distress at encountering a popularly held view. Beyond the fact of which of these positions are correct, would be happy to offer some strategies, take or leave, to make encountering realist views at least less disappointing.
"Intriguing. Do you believe people ought not be moral realists?
I would have thought that stating almost any opinion commits one to the view that people should or should not believe certain things, that people should or should not espouse certain ideas, which smuggles in at least some kind of realism."
Maybe. But not moral realism. I'm only a antirealist when it comes to morality. People ought to be not moral realists if they care about honest logcal reasoning. Which is something I would like to expect from philosophers, but sadly that's not the case. There is simply no convincing argument for moral realism. It's all god and/or platonism at the end. (Not there's necessarily a difference between those two.) Being a moral realist is just convenient. If you deny moral realism your moral arguments become much harder, more annoying and especially less convincing. But I hate that kind of motivated reasoning.
"If you're an Ayer emotivist"
Technically I consider myself a sentimentalist but I don't see much wrong with emotivism.
"people believe a theory that, for me, caches out to basically "some outcomes are better than others."
With a statement like this you are just skipping the issue. (which to be honest, is sadly the norm for moral realists) What is "better"? What is good? Why is your conception of good better than mine? Better for whom? Why should I or anyone else care about that conception of "good" or "better"? (I also would like to mention the fact that you went straight to consequentialism here, which the majority of moral realists among philosophers do not even subscribe to. So saying that moral realism caches out to that seems a bit absurd.)
The best argument against moral realism is simply the there is no good argument for it. There is no way to bridge the is-ought distinction so all that remains is magical wisdom raining down from heavens as the source of any certainty about moral claims. And if that's the fundamental axiom of your beliefs you might as well believe literally *anything*.
(There are ofc also good arguments directly against moral realism. I've yet to see an argument against evolutionary debunking of moral realism that was not very weak. But maybe it's out there.)
"Beyond the fact of which of these positions are correct, would be happy to offer some strategies, take or leave, to make encountering realist views at least less disappointing"
I'm open to your suggestions. (and also sorry about the potentially angry tone of the post, I can get quite emotional when it comes to certain topics.)
Yeah, no worries about the tone. You are honestly exasperated by encountering other people who show signs of intelligence and education but fail to adopt something you consider as fairly obvious after investigation. I've been there.
I don't want anything here to come off as patronizing. You clearly know the domain, I'm not going to insult your intelligence by suggesting you go off and read why Harris or for god's sake* C. S. Lewis are realists.
I would just offer that when I faced similar frustrations in other domains, it seemed to help when I acknowledged that people's ontological commitments can shift very slowly. Especially on the big foundational stuff.
So, imagine everyone you're talking to might just be in the early stage of a five year conversion process to your position. At this phase all they need help with is unsticking one or two small things, then you send them off on their journey from there.
Sure, it's a hopelessly naive fiction, most people aren't primed to change their views. But still, this stance seems to most often lead to a eureka or refinement on one side or the other. It's a pragmatic choice more than anything.
The next layer that kicked in for me, after having a few of those discussions, was some compounding epistemic humility. Maybe I'm the one in the midst of one of those five year conversions. Or we both are. Who knows. Usually I can find some premise I can converge to at least a bit.
In this specific case, maybe you're about to convince me to abandon realism?
Honestly, to your credit, you did prompt me to take another skim through the Stanford encyclopedia on antirealists to refresh myself on the different varieties there. Some of these folks I'd read over a decade ago and promptly forgot they existed. And as a result I did flip one bit in my ontology... it's really deep in the weeds on this issue and nobody would be very impressed... and the whole house of cards hasn't quite collapsed yet. But it's something.
But that was after a review of some fairly technical arguments.
In contrast, I loved Hume growing up and still have enormous respect for him, but I no longer find things like the is-ought problem dispositive. (Should I? :P ) I would unpack why I came to that view, but you've been barraged by realists already, I don't want to pile on. Here I'm just noting that, it's hard to predict where the crux is for different people, so it can take some patience to root around in their mental attics, figure out what they might be hanging on to for nostalgia or just inertia.
The ironic thing that strikes me most about our positions is that an outside observer might call me a moral antirealist as regards to people's ontological commitments, and a realist as regards to their actions, and you the inverse of each. I don't mean to not take you at your word, but you seem to have really strong opinions on what people should believe. (I know, so long as they care about logical consistency.) Just found it interesting.
After reading Wittgenstein,** who knows, everything seems like it just comes down to definitions at some level or another anyway. I'm open to the possibility we believe roughly the same thing for practical purposes, and just use different words or emphasis to explain it.
Speaking of, I did like his idea of philosophical therapists, who just talk people through their (philosophical) problems to help them unstick things. Maybe I'll hang up a shingle as that some day, might be fun. :D
* pun intended
** There's probably a good will hunting meme about me out there somewhere where next year I'll read Kripke and be convinced Wittgenstein lacked a full appreciation of the importance of names or something. If so, fine, I fully accept it.
I just want to note two things.
1. You can be a moral realist without believing our current Western values are the completely correct values. I personally think people will look back on our tolerance of abortion with the same shame with which we now look back on our tolerance of slavery.
2. There is nothing inherently religious about objective moral values. Kant derives them from reason alone. Hume can be read as a subjectivist, but he also talks about the "human moral sense", as if it could be universal among the species.
And in fact, I'd argue religious-grounded moralities are fundamentally anti-realist. They hold morality entirely relative to God, who himself can do no wrong by definition, because anything he does becomes morally right. That's as terrifying as the worst forms of cultural relativism.
1) "You can be a moral realist without believing our current Western values are the completely correct values"
Of course. Obviously. I never meant to imply anything else.
(Btw I think the same thing when it comes to treatment of animals but can't see the argument of this being the case for abortion at all.)
2)"There is nothing inherently religious about objective moral values. Kant derives them from reason alone."
Of course there is. Kantian ethic does not "work" without God, it needs the assumption that God exists. (I would also claim that it doesn't work at all. No one lives according to Kant's ethical system. It's undoable. An in depth critique would be far too much here. though.)
"Hume can be read as a subjectivist, but he also talks about the "human moral sense", as if it could be universal among the species."
A view that is entirely compatible with subjectivism. I agree with it, yet, just like Hume, am both an (atleast ethical) subjectivist and sentimentalist.
" And in fact, I'd argue religious-grounded moralities are fundamentally anti-realist. They hold morality entirely relative to God, who himself can do no wrong by definition, because anything he does becomes morally right. "
God is by defintion absolute good and absolute morality. Therefore that point doesn't really make sense in my opinion. Nothing God does "becomes" morally right. It right by it's very nature. I cannot see any way how objective morality could work without absolute good existing. (aka God) Kantian ethic doesn't, neither does anything else.
> [moral ralism] just seems so unfathomable to me. It's literally all religious arguments while trying somehow to pretend it's not
Not at all. A Proof of the Objectivity of Morals: https://www.dropbox.com/s/p9v7qt23p21gfci/Proof%20of%20the%20Objectivity%20of%20Morals.pdf?dl=0
It occurs to me that part of the reason the Chinese say these things is to try to reframe the conversation in more abstract terms.
Human rights are a subset of ethics, and ethics is something that is neither "invented" nor "discovered", a bit like mathematics.
Ethics is "invented" in the sense that you can come up with any set of arbitrary rules you like and call it ethics. But it's "discovered", in the sense that most of your arbitrary sets of rules turn out to be terrible and obviously unethical; either they're horrendously inconsistent, or they lead to massive human suffering, or they're just not palatable enough for people to accept them.
Take slavery for an obvious example -- people who could afford slaves spent thousands of years trying to come up with ethical justifications for why this was actually ethical, but after seeing the clearly terrible consequences for thousands of years, Western society eventually came to the conclusion that the space of sensible ethical systems does not include any with slavery -- then we banned it, and everything was better. I'd call this a discovery.
Human rights are similar -- after thousands of years of ethical experimentation we've eventually figured out that there's certain important "no go" areas on the ethical landscape, and we have good reasons for believing that these are genuine no-go areas and not just our personal preferences.
Terrible, is reasonable, but you can only decide that an ethical system is unethical on the basis of a different ethical system, so that part of your argument fails. And for slavery, there were lots of ethical justifications that were satisfactory to the people who adopted them, and even to some of the slaves. That you don't accept them is because you adhere to a different ethical system. It may well be a better one, but deciding that would require lots of investigation. E.g. is an ethical system that justifies destroying the environment you live in better than one that justifies slavery? On what basis do you choose?
You cannot validly argue that an ethical system is unethical. That you don't like it is fair. That it doesn't match your ethics is fair. That is has unresolved contradictions is fair. That nobody can possibly live up to it is fair. That most people won't be able to live up to it is fair. There are lots of fair arguments against ethical systems, but being unethical isn't one of them.
I can only write that I intensely disagree with your or any moral realism and therefore also with the concept of moral progress.
I could write very, very long texts about the reasoning behind my metaethical position, countarguments against moral realism (and have done so in the past during my studies) or even specifically against your post (previous moral rulesets were "obviously unethical" but we still practiced them for many, many millenia? And somehow the ruleset that evolved over the mere last couple of centuries is good now? Why? How? Not to mention you literally put the goal of ethical humanism as the ultimate ethical goal and then claim ethical humanism does achieve this best - Yeah, duh. You are putting the horse before the cart. People did not fail to achieve the "ethical goal" of humanism, they literally did not have that goal at all.) but all of this is besides my initial point: Human rights are a specfific concept invented and evolved over a specfific time that were not in the world before.
If you think they were discovered instead of invented you would have to believe that *everything* is merely discovered (as platonists do), making the very concept of invention (and the word) itself pointless and nonsensical.
(Remember: Discovery is otherwise reserved for things that were in the world before - and humans just found them. Without believing in a platonic realm you can't tell me that ethical positions existed before humans ever thought of them.)
> And somehow the ruleset that evolved over the mere last couple of centuries is good now? Why? How?
Because they're more informed by facts, because we've eliminated many unjustified assumptions, and because we have better cognitive and rhetorical tools with which we can evaluate our arguments. The same reasons why our math and science is better.
How? How does that work for ethics?
You cannot classify someone as a slave or king if there's no factual basis for this distinction, for example.
Actually there is a sense in which ethical systems are discovered as well as invented. Only some of all the possible ethical systems are sufficiently compatible with human instincts to be even meta-stable. Landing in a metastable position is a discovery, as there's no known way to predict which states will fit that description. (Even historical examples don't work, because the environment is different.)
How does this not apply to literally all inventions. There is only a limited number of ways a lightbulb can work. In that case the point would be redundant.
We're close to understanding the physics involved in a light bulb. We aren't close to understanding the boundaries of human instincts. By the time Edison came along, we knew basically how a light bulb worked. OTOH, I have heard it said that Edison discovered that a carbon filament would work longer, so it's not a sharp distinction.
I think moral anti-realism isn't even necessary to think human rights are an "invented" concept. Like, "human rights" doesn't look like a real value system no matter how you slice it; it's incomplete (i.e. ambivalent about a large number of choices), and somewhat ad-hoc, in that "human rights" presents no first-principle derivation for itself.
It's more similar to law, actually; it's plausible that some set of ethical principles, when followed to its logical conclusion, will incorporate something that looks a lot like the laws we have today, but it's pretty ridiculous to think that the current legislature of any country is acceptable as an ethical system (relevant: "in all situations, the government of Raikoth will take the normatively correct action").
Same thing with human rights - it's a cut-off, arbitrary chunk of an ethical system enshrined as international law. The Chinese are probably complaining that (as lawyers do with regular law) it can be used as a bludgeon in unfair ways. That's a bit disingenuous of them, though, because the alternative to invoking human rights is just straight up telling them that what they're doing is morally wrong - and everyone, Western or Chinese, knows That's Just Not How We Do Things on this planet.
Melvins position sounded more likewise realism to me.
I would say:
A) human-rights are negotiated as part of a fictional social-contract.
B) codes of ethics are invented, their consequences are discovered.
In the least convenient world, I think I can imagine scenarios where slavery is a sane equilibrium.
Everything is an invented concept. States. Humans. Phlogiston. Molecules. Chess. Electrons. Integers.
So the argument that "X is just an invented concept" is kind of an universal counter argument. It tells you nothing about the value of the concept or if you should adopt it.
No. Concepts are invented, but not the physical reality that we reference when we invoke those concepts.
If you believe that physical reality actually exists then the underlying fact of the matter that we reference when we speak of molecules is not invented at all.
Ethical concepts on the other hand have no clear or measurable physical reality underlying them, which is why ethical realism nearly always has to rely on a platonic realm to make any logical argument.
I grant you though that saying that an ethical position is merely an invented concept is not a very strong point unless you want to make the point that your ethical position is somehow not invented (usually a religious argument :because it's divinely inspired.) But it's not wrong.
What he's asserting is that the way we slice physical reality into concepts is not deterministic. That there are alternatives. Certainly when you get to things a fine as quantum mechanics, and then try to translate them back into human language, this is a true statement. It's less clear that this is true of "baseball game", which is clearly invented, but you can look at variations of the rules and ask "is this still a baseball game?". It's not that there aren't basins of relative stability that have names, it's that the positions of the boundaries isn't predetermined, The names refer to centroids, but most of reality is somewhere in between those centroids.
Molecules (or humans, for that matter) are a gross simplification which are great for us because our minds suck at reasoning about a protein (or human) as a particular class of solutions to some equation of a theory of everything.
If smarter aliens said "molecules are obviously an invented, earthling concept which we do not need" the best we could say is "well it seems a useful abstraction to us in many cases".
I could argue that some ethics can be considered empirical rules for building human societies, there are certainly tenets which would be incompatible with any society. The problem there is that there are also plenty of tenets which seem very compatible with complex societies, but nevertheless terrible. Personally I am mildly in favor of the non-cognitivist idea of not treating ethical rules as logical propositions at all.
Ethical rules are a compromise between tribal instincts and the need to build larger societies as populations grew. There are lots of possible compromises that would work, and there are even more that wouldn't work. They have a reality that's not exactly logical, unless a golf ball is logical, but isn't exactly determined either.
It's only as much an "invented Western concept" as "diplomatic immunity".
Which is to say, I'd bet the idea was invented many times in many places. I'm sure Wang Yi, or even Xi Jinping, could invent it themselves.
1) Just because something was invented more than once does not mean it's not invented.
2) I'm not sure that is true at all when it comes to what we consider human rights. Your comparisons also seem to imply that human rights would be an extremely easy and common invention, something that I would dispute.
Diplomatic immunity is such an universal concept because it was extremely useful. Slavery was also extremely useful, universal concept.
Human rights fly in the face of that concept, and I would argue are a much more recent, uncommon invention.
Actually "human rights" doesn't fly in the face of slavery. Some definitions of human rights contradict some definitions of slavery.
Personally, my definition of "human rights" considers the US prison labor to be a violation of them. There are many, however, who don't hold to that definition. Either that, or they are quite willing to ignore human rights when its convenient to do so.
Is there any way to improve 'bravery', for lack of a better term? I find that I become afraid far too easily- any mid-sized stressor that would raise anyone's heart rate a bit (public speaking, a verbal confrontation, etc.) makes me *very* afraid. I experience an excessive surge of adrenaline, out of proportion to the situation, which is very visible and makes it impossible to stay cool. I become emotional very fast. It's odd because I don't suffer from general anxiety (I feel fine most of the time), and when I was younger I did some extremely physically challenging activities that meet anyone's definition of toughness- more on the endurance/labor side. But for my entire life, I become excessively frightened by mild stressors.
If the answer is 'exposure therapy'- logistically how I would regularly practice say verbal disagreements with others? This seems impractical. Or, alternately, has anyone been prescribed something that reduced their brain's fear reflex? I find alcohol makes little difference
Exposure therapy is fairly effective, and worked for me. I did it as part of group therapy for people with social anxiety. It helped me quite a bit
A cognitive trick that supposedly works because nervousness and excitement are cognitively close: when nervous about something, just say out loud to yourself "I'm excited", and that can meaningfully convert nervousness into excitement.
https://www.gsb.stanford.edu/insights/hacking-your-speaking-anxiety-how-lessons-neuroscience-can-help-you-communicate
The most valuable thing I ever learned about bravery, as someone who also comes with his fight/flight response a little tightly-wound, is that it's not the absence of fear; it's the choice to do the right thing in the face of fear.
I find it's unfortunately much easier to practice this in the context of physical danger than social danger. Fires and cliffs don't care if you're a little shaky. Humans tend to make it awkward. But it can be done.
I think it helps to set aside all expectations of how you 'should' feel. We all have some ability to do things that feel bad - change a diaper, clean a wound, etc. Not all of us are great at meeting all our responsibilities immediately, but we tend to limit our hardcore avoidance to the things we think 'should' be easy and enjoyable. It can help to accept that something's going to be unpleasant and commit to doing it anyway.
When your anxiety is purely a physical reaction, you also have the option to choose to interpret it as excitement. The two states are physiologically identical; it's the conscious areas of your brain that give meaning and labels to your arousal states.
I know what you mean, but for me at least, in many situations it helps to "not look down", i.e. don't confront the fact that you're in a scary situation and just focus on getting through/doing what needs to be done.
Oh, I agree, that's a very effective way to keep yourself from freezing in a scary situation that you're already in.
I took "bravery" to be more about choosing to confront situations you know will be scary (i.e. overcoming the fear of fear). Denial/suppression seems less effective for that, although it does sometimes seem to work in group contexts.
Anxiety and excitement may be the same thing for people who enjoy scary movies. That might even explain what to me is an absolutely baffling taste. Because for me those two states are entirely and absolutely antithetical.
As Nietzsche said, if you know the why, you can bear any how. Part of the problem here seems to be you don't really believe you need to do these things you're afraid of, so the fear wins out.
Whereas if you knew why you want to do these things, you would have access to a stronger emotion that could overrule this fear.
It's really important to have something that's bigger than you, something that can override you
My daughter suffered from very strong fear of public speaking for many years. Beta blockers fixed it, I wish we found out they exist years before we did.
Yes, actually I have been considering beta blockers. I think I'll probably have to source them online as I doubt my doctor would prescribe me some for personal experimentation
I was prescribed beta blockers for dental surgery, by a dentist who did not use nitrous oxide. If that ever comes up in your life, you might try it then.
I don’t think there is any potential for abuse of beta blockers. Nobody gets ‘high’ on them. Not a doctor myself but I think an MD would likely let you try them for this sort of anxiety.
I realized a few years ago that I was also “afraid” of a lot of things in the moment that didn’t scare me intellectually. Like if I thought of the thing, it didn’t scare me but if I was in the situation and it happened my fear response was through the roof. This extended to surprises, confrontations, and a whole range of usually innocuous things. In essence, I saw nearly everything as a threat or dangerous, so my fear reactions were off the charts and embarrassing.
As boring and seemingly obvious as this is, I discovered that if I exercise in the morning, raising my heart rate and getting a good sweat in, then I would still register all the same things--like my awareness didn’t diminish--but my body was much less likely to register them as dangerous or threatening. The sweet spot is 30-45 minutes of resistance training and/or low intensity steady state cardio, but really any sweat with an elevated heart rate works. The effect is so acute now that if I am anxious on a day I haven’t worked out, I go out to my garage where I have an airbike. I will do 3-5 circuits where I pedal 20 seconds max effort and rest 40 seconds, and just that small workout will clear my mind and relieve some anxiety.
I’m sure there are good explanations for why this works, but my (pseudoscientific) working theory is this: In Thinking Fast and Slow Daniel Kahneman explained how we have two systems of thought, the automatic background thinking that is like a vast supercomputer doing pattern recognition, and our relatively measly deliberate logical thinking. My theory is that our automatic thinking largely evolved as a threat detection mechanism, and with most of us (reading this blog anyway) spending all day in refrigerators reading off screens, that thread detection mechanism keeps looking for threats even though there isn’t anything. Something about the exercise raises the threshold for a threat, so that only actually threatening things gets coded by my automatic thinking as a threat.
Other helpful things: improv comedy classes and performances for getting up in front of people and looking like a dickhead, and Brazilian Jiu Jitsu classes for actually fighting people and--at least for the first few years--just getting your ass whipped again and again.
Thanks Charlie, I really appreciate the time that you put into this. Unfortunately I do have several years of BJJ experience- didn't seem to help much :( Maybe I'll get more out of striking though! Have been considering taking up Muay Thai for this exact reason
Yes to striking. It’s an entirely different level of fear to confront when you spar with an opponent who is intent on punching you in the face.
I find it is the level of exertion, rather than the activity. Like it is a direct correlate to how sweaty and out of breath I get. For me, BJJ does that, but my ideal workout for managing fear/anxiety is basically CrossFit.
I did a lot of reading and thinking about consciousness [and did a bunch of mushrooms] and came to the conclusion that death isn’t real - and now I don’t get scared of anything! Even little things unlikely to kill me. So I’d recommend digging into whatever existential topic lights your fire [and doing a bunch of mushrooms] and see if that chills you out.
Mushrooms and anxiety, two nutritious parts of a complete dog's breakfast! I dunno, man. Make sure someone else is around for that who isn't super anxious.
I saw an MBA student on YouTube who was practicing exposure therapy for this by asking for things he knew people would refuse in small stakes situations. The one I remember is him going into a fast food restaurant and making some crazy substitution request like replacing the bun with more hamburger patties.
When I was having trouble at work being stressed by confrontations I decided to do something similar. I bought and sold things on Facebook marketplace and haggled with people on the price. I did the same when buying a car. I would say it helped.
Agreed! A co-worker of mine found them very useful. I was scanning the replies to see if they had been covered... Many Thanks for doing so!
What would global collapse look like?
I believe that if we lose our current level of technology, we will go back to subsistence farming, hunting and gathering. That there is no plausible way to transition to any level of technology in between.
If this happens, 8 billion humans will die leaving a few million in remote locations with no essential dependence on technology.
I define "global collapse" as "loss of electrical supply".
Who agrees?
Who disagrees?
Note: it's fun to debate the likelihood of collapse, but that's not what this thread is about.
Peter Robinson
Some links:
https://www.futurescenarios.org/2-energy-futures/2-2-views-of-the-future/collapse/
https://jfsdigital.org/2022-2/vol-27-no-1-september-2022/theories-of-global-collapse-closing-down-or-opening-up-the-futures/
https://thebulletin.org/2022/03/global-collapse-is-in-view/
https://bylinetimes.com/2022/05/26/un-warns-of-total-societal-collapse-due-to-breaching-of-planetary-boundaries/
https://medium.com/predict/what-will-the-collapse-look-like-d2a1e53c9f54
https://canadiandimension.com/articles/view/we-are-not-the-first-civilization-to-collapse-but-we-will-probably-be-the-last
https://education.nationalgeographic.org/resource/key-components-civilization/
I'm not sure why you say a lack of electricity would prevent us from producing fuel. It's not like oil platforms are connected to the electrical grid (often even the ones on land).
Work that can be done with motors can also be done with combustion engines, so as long as fuel is available and engines are available, factories can keep running, though they may need significant modifications. Electric generators are also very simple devices to produce: all you need is copper wire, a magnet, and a source of energy to rotate it.
I talk about no more electricity as a way of identifying a collapse. There are people I talk with that think collapse means "things are different now". But the oil platforms would stop working for the same reasons that the electrical generators stopped working. Lack of spare parts for example.
But I don't believe global collapse will happen. I just believe that if it does happen, 8 billion humans will die.
Okay, makes sense.
I imagine if the collapse goes very well (no widespread warfare, communities or governments organize and cooperate), then a massive amount of human labor could be reallocated to agriculture and other essential industries. Jobs like marketing or psychologists aren't exactly essential in the face of global food shortages. Market forces would naturally push people into agriculture (even if the financial system collapsed, people would still trade much value in goods or labor for food).
So then the question is how many people can we feed with our current farmland using only basic tools and human labor? And by how much can we expand agriculture before people starve? (And this is affected by whether it's a quick collapse or a slow collapse.)
My guess is we could feed a large fraction of the world in the optimistic scenario.
To be clear: I'm saying THERE WILL NOT BE A COLLAPSE because we can devote much more energy and resources to those things that are essential.
Food
Fuel
Electricity
Spare parts
If things look dire, we can do without Super Bowls and racing cars. And new clothes, new roads, new skyscrapers, etc.
But now we are getting into the question of whether there will be a collapse, which I wanted to avoid.
Right, and I'm saying in a best case collapse scenario, we can probably still feed a lot of people without electricity or advanced technology. The lack of electricity itself doesn't seem insurmountable if everything else is going as well as possible in the hypothetical collapse. It would help a lot if the collapse happens over several years instead of several days.
The loss of electricity isn't really a global collapse or apocalypse.
I mean, let's clarify at the start that for literally a billion+ small and subsistence farmers in Africa, India, and Southeast Asia, very little is going to change. Not just that they're going to wake up the next morning and go back to work...basically as normal, but that they're going to keep growing the food that feeds the village.
Now New Delhi is going to have a bad time and most of us in the West are going to have a really bad time but, like, do you really think everyone in, say, Somalia is going to starve to death if the power goes out?
Second, what do you mean there's no technology in between? I mean, I don't know how to make a musket but there's guys in Afghanistan making guns out of garbage. The fact that you or I can't make anything without modern manufacturing tools doesn't mean nobody can. And you can get pretty darn far up the tech stack without electricity, we know you can at least get to the US or UK in the 1890s, which is enough tech to get you to WWI, which is enough tech for mass industrialized warfare and all the fun of the "modern" world. And yeah, it would be disruptive and there would be a lot lot of deurbanization and we might go back to ridding horses but there's a huge gap between that and, like, caveman hunter gatherers.
Almost all farming, even subsistence farming in the third world, is dependent on the modern fertilizer industry these days. Of course the individual farmers themselves would adapt to the loss of those inputs, but the reduced production would still imply widespread starvation.
Nonetheless, I agree that there is no reason to think we couldn't stabilize at some level well above 'hunter gatherer'.
I believe that the stabilization would be below the level of Rameses I of Egypt. In the first stage, those who have means of subsistence are going to need to fight off those who don't. Some of those who don't will be folks who were planning on living by hunting, but found that too many others had the same plan. So until the guns break down or wear out, the level of violence will be quite high. It will probably end as a new feudalism, at a much lower level than 500AD, because there won't be any self-sufficient enclaves that maintain civilization. (Too much technology is too complex to be maintained by a local enclave.) Perhaps this won't be true in parts of Africa, India, China, and around the Arctic North, but I'm not at all sure. The problem is that climate change is altering the environment such that traditional ways of earning a living no longer work.
Thanks for your reply!
>>I mean, let's clarify at the start that for literally a billion+ small and subsistence farmers in Africa, India, and Southeast Asia, very little is going to change. ... they're going to keep growing the food that feeds the village.<<
Do they use any kind of powered equipment? Without electricity and fuel that won't be available.
>> do you really think everyone in, say, Somalia is going to starve to death if the power goes out?<<
What percent of people in Somalia grow their own food? The supermarkets will be empty. People who have caches of food, must have guns to defend it.
>>Second, what do you mean there's no technology in between?<<
I said "no plausible transition" (that keeps the bulk of the population fed).
>> I mean, I don't know how to make a musket but there's guys in Afghanistan making guns out of garbage. <<
Do their guns shoot bullets? Can they make a bullet? Black-powder guns are plausible.
>>And you can get pretty darn far up the tech stack without electricity, we know you can at least get to the US or UK in the 1890<<
How much pre-1890s technology is available within 10-90 days?
Food is the focus. No food will come from more than a few miles away. If the people in the cities are not fed within a few days, they will attack any person who has food.
> Food is the focus. No food will come from more than a few miles away. If the people in the cities are not fed within a few days, they will attack any person who has food.
Sure, there's no way of getting out of this scenario without a lot of people dead in the cities. As a city slicker, my best shot is to pack a tent, plus as much food as I can carry, on day 1, and then start walking; try to outpace the rest of the city and get to an agricultural area where I might be able to get a job as a farmhand. I'm thinking my best shot is dairy country since it's non-seasonal and they'll suddenly need a lot more labour to milk all those cows without machines. The cows themselves are of course beef on legs in a pinch, and good cattle grazing country is also good horse grazing country and horse breeding is about to become the most valuable industry.
Dairy country is a good idea. I recommend planning ahead and buying some of these nitrogen-packed survival food assortments. Stash dozens of these with with your dairy Farmer friends. And plan to keep your tank half full so you don't have to walk there.
But if a mob of 10,000 desperate city dwellers shows up at your dairy farm not long after you, all your planning will be for naught.
In order to speculate sensibly you'd need a plausible mechanism for this complete global collapse. I can imagine all sorts of things collapsing, but why would it cause farmers in Fiji or Patagonia to suddenly forget how to do their thing?
Thanks for the reply!
Just assume that the electrical supply has failed.
Anything requiring a motor will not function. IC engines will function until fuel runs out.
This premise does not sound reasonable, short of god deciding that henceforth, every material on earth shall be non-conductive.
Globally speaking, the power grids are very distributed. Temporarily knocking out the continental US grid might be doable, doing so globally and permanently seems non-reasonable.
Roughly, you would need to:
* Disable nuclear power, perhaps by removing all fissile isotopes from Earth. Preventing nuclear reactions in general is preferable to also prevent fusion. (Note: This may disrupt military equilibriums.)
0. Lower the inner temperature of the earth to deny geothermal power. (Note: This may affect surface temperatures.)
1. Prevent the burning of fossil fuels either by removing them or by removing oxygen from the atmosphere. (Warning: Removing oxygen might affect humans in ways unrelated to the power grid.)
2. Cause an Chicxulub-like impact. This might reasonably disable most power generation from solar, biofuel (eventually), wind, and hydro. (Note: This might affect humans in ways unrelated to electrical power.)
3. Blow up the Moon to partly disable tidal power. (Note: As tempting as it sounds, do not merge this with the previous step by driving the Moon into the Earth unless you are ok with immediately extinguishing most life on Earth.)
4. Tidally lock the Earth to the Sun to prevent the remaining potential for tidal power. (Note: this will probably cause winds near the terminator (which will also likely be the most habitable zones) which any surviving humans might use for power generation. On further reflection, it might be safer to replace steps 2.-4. by removing Earth from the solar system.)
Note: I am not a Mad Scientist and this is not Mad Science advise. Please consult with a certified Mad Scientist before trying any geoengineering projects on this scale.
Thanks for your reply.
These are interesting scenarios. You are focusing on electrical supply, and I can understand why. It may appear that I am saying that the reason for the collapse is the loss of electrical power. But I don't mean that. I'm just using loss of electrical power as a marker, a way to say "Okay guys, this is collapse."
I don't believe collapse will happen, but if it does happen it will be because of some kind of cascading failure: one failure causes another which causes another etc.
I participate in a global warming forum. The majority of the participants seem to believe that collapse within a few decades is inevitable. Many of those are looking forward to this as ecological justice.
What about world-wide EMPs? Nearly all modern technology uses some form of computer. So nuclear plants would likely be non-functional, regulation systems on geothermal wouldn't work, and the power grid in general probably would have massive outages that couldn't be fixed without new computer chips. And where would one get those? The chip factories, multi-billion dollar facilities, couldn't make more chips because the ones to operate the factories no longer work.
"What about world-wide EMPs?"
Yup. https://en.wikipedia.org/wiki/Carrington_Event
A similar event may not only fry semiconductors but also infrastructure like transformers.
It would be a big problem and some plants would never come back up, but some would -- every electrical grid in the world has hundreds of clever electrical engineers who can figure out ways to make things work with whatever they have. Most of the world's existing power stations probably worked fine before they were computerised, and they can work fine again without computers; it will take a lot of work to fix them, but it will be society's top priority.
I predict that most of the world will have a reasonable fraction of its power generating capacity back within a timescale of years.
But why would all electrical supplies across the globe all fail? Aliens?
I'm having fun speculating on what would happen in, say, Fiji, if the rest of the world returned to the Stone Age. Fiji is more than self-sufficient agriculturally, half the country is powered off hydroelectricity already, and they can easily turn sugar cane into ethanol. They don't manufacture much in the way of machinery but they could surely figure many things out if they needed to; and besides they can always raid the ruins of Auckland and Brisbane and Honolulu for anything they can't produce locally.
Aukland and Brisbane and Honolulu are all thousands of miles away. How does one get there, and back again? Regularly? What are you bringing them they will want more than their machinery? Why would the crew on the ships not decide to stay where the machinery is?
1. In a ship, fuelled by ethanol or biodiesel (or at first, regular diesel)
2. Nothing, the locals have descended into the stone age. If they bother you, you scare them away with your boom stick.
Failing that, you can trade safe access for sugar, or perhaps rum.
3. The crew on the ships would come back to Fiji because everywhere else has descended into the stone age and forgotten how to agriculture.
I think you greatly underestimate the effect of civilization on the current state of the world. It would take a vast amount, by the collapsed civilization's standards, of fuel to get a ship to another continent, even assuming one could navigate without GPS. If they are going to produce those thousands of barrels of fuel, they ought to use it locally to produce power.
If the locals are now in the stone age, what on Earth are you going there for? To salvage stuff from prior civilizations? If they have such things, they will likely be well above stone age themselves, and using the equipment.
If you're proposing the stuff you want is effectively free for the taking from the locals, then you're talking about raiding and piracy. Even if they did that for some time, it wouldn't last long. Certainly not an entire year.
No, if civilization collapses, then the entire world would turn into regions, and only a few people would travel between regions. It would be very hard to have global trade until civilization has built back up so that mere survival is no longer the primary goal.
Thanks for your reply.
>>Fiji is more than self-sufficient agriculturally, half the country is powered off hydroelectricity already, and they can easily turn sugar cane into ethanol.<<
Producing most of their own food within Fiji would be very helpful in a collapse.
I could not quickly find what percentage of fuel used in Fiji is imported. I'm betting it's close to 100%.
If Fiji anticipated global collapse, they could convert to ethanol use over some years. IC engines don't like 100% ethanol, but I wouldn't say it's not possible.
So the problem is getting the food from the farms to the table. The transition cannot be done within 90 days. After 90 days most people would be dead from starvation or injury.
Fiji does not produce integrated circuits. Within 5 to 20 years almost every electronic or electric device in Fiji would cease to work (would break and could not be replaced).
Now, please note: I don't believe global collapse is going to happen. For example, 10% of the global economy is used to produce food. If for various reasons it becomes much harder to produce food, we could devote 20% or 30% of the economy to food.
There are people that I converse with who are rooting for global collapse. They think it will be good for nature and that humankind can transition to a "post-industrial" society.
Well, 70% of the population is already subsistence farmers, so that helps. Only the remaining 300K people need to figure out how to feed themselves. Fiji is not very big (the largest island is 100km across) so it seems like short term farm-to-table problems can largely be solved by people walking to where the food is.
Making ethanol from sugar cane really isn't that hard, and can easily be done inefficiently on a small scale by any individual with, like, some buckets. Given the vast sugar surplus that would exist in a Fiji deprived of its export markets, I'd imagine that a cottage industry of good-enough fuel would establish itself quickly -- good enough to keep at least some engines running.
Sure, they're not going get their iPhones back any time soon, but I can picture places like Fiji riding out a global collapse in relative comfort.
You're thinking short term. So lets think of Fiji. That gets them through the first, say the first decade. Then the engines start wearing out and there aren't any replacement parts. And climate change doesn't suddenly turn off, so the sea level keeps rising. Who remembers how to build an outrigger canoe, or to navigate by the stars with a coconut shell? (I've no idea how that worked.) And there are a lot less fish in the sea than their ancestors had to work with.
Other places the problems will be things like water availability. We've been overusing the water resources all over the globe. Once the pumps stop working, the water isn't available.
Other places there are different problems. Everywhere the population is too large to support without powered equipment. Seasonal food can no longer be preserved. I expect we'd be luck if one person in a hundred lived through the crisis...and lucky may not be the correct term, since the famine itself would lead to immense further destruction of the environment. Perhaps one in a thousand would be better. And THAT may be too large. The population of Ptolemaic Egypt is estimated at three to four million. The current population of Egypt is 109.3 million (2021), but a crash wouldn't recreate the annual flooding of the Nile with rich sediments. (To put this in more general terms, the modern environment isn't as naturally supportive of life as the ancestral environment.)
I think small groups would be able to stay above the subsistence level even if they would really struggle.
Say a bunch of people in some remote area with access to a river they can use as a power source. Stuff like that. Most of their technology would collapse but you could preserve enough of it even with our garbage skills/knowledge to stay well above the baseline.
Thanks for the reply!
Assume they didn't already have a water-driven electrical turbine-generator (which is likely); they would need to create one.
With no food in the supermarkets, I figure the first 90 days would be total chaos and 95 to 99% of humans would die within that time period. So it would be the survivors building that turbine. Most survivors would be physically strong persons possessing guns. I think a rather small percentage of these persons would know how to build a turbine.
You don’t need a turbine to make electricity. A regular water mill will do fine (with some gearing). Won’t be much, but when you are starting from nothing any is a lot.
Couple hundred watts at least.
Where will you find your generator? (After the people who would attack you have died.)
Remember: transportation is essentially limited to walking, bicycles, and horses.
This is after the famine has passed, for forget the horses. And bicycles depend on decent roads. There's an argument that the reason cities paved roads was to make it easier for bicycles. No spare tires. No spare parts of any sort, except what you can scavenge and make fit. This lets out most modern bicycles.
P.S.: The people who would attack you won't all have died. Some of them will have formed gangs that survive by raiding.
These are all good points, and correspond more with my outlook.
There is no doubt scavenging would happen. The big question then is whether this could leapfrog to some level of technology above stone age or would the scavenging last one or two generations and devolve to the pure farming-hunting-gathering society that I imagine.
In other words: yes, you can scavenge a generator and wire and some incandescent bulbs, but can you build a generator from scratch?
Were you asking if any groups would stay above subsistence, or my particular group? For such an open ended question you sure have a very specific scenario you are trying to drive everyone towards.
Why not just lay that out and defend then instead of coming up with weird objections to peoples responses…
Thanks for your reply!
>> if any groups would stay above subsistence<<
Yes, that would be one way of posing the question. Could some groups on their own maintain some type of technology more advanced than flint knives.
I believe the answer is no, but I would be fascinated by a description of the technology that might be maintained independent of the rest of the world.
An electric motor or generator is just a bigass copper coil and a magnet, both of which are not too hard to salvage.
You may not even have to build it yourself - pull the alternator out of a car and use something else to turn it.
>>alternator out of a car<<
12 volts, I believe.
Will it help you to produce food?
Food is the ball game.
"A Canticle for Leibowitz" by Walter M. Miller Jr. provides a relevant speculative take here.
I suspect it would be pretty mixed. Some techs are deceptively complex, others are just good ideas and widely known. There are some types of white paint that involve surprisingly complex chemistry. You're probably far more degrees of separation from a specialist in paint chemistry than you are from a gifted pharmacology student who might actually be able to DIY some basic medicines using a microlab.
See https://www.vice.com/en/article/43pngb/how-to-make-your-own-medicine-four-thieves-vinegar-collective
Within limits, to be sure.
I'm sure other disciplines have this, something ubiquitious, like smooth ball bearings for writing pens, are fiendishly complex to manufacture, whereas something seemingly much more complex and specialized might accept a DIY substitute (minus some features), like some agricultural machinery. I have a relative who made tires from scrap during the great depression, even though it was no replacement for the international tire manufacturing supply chain, still got the basic job done. A necessity-driven focus on "good enough" might leave a surprising amount of tech behind.
The pharmacology student requires good reagents. Try herbal medicine practitioner. Distillation would probably survive, and that would enable lots of chemistry after awhile.
P.S.: The "Great Depression" only lasted a decade. The replacement tires would give out pretty quickly. And the road network requires more maintenance than you may expect. We didn't/don't build roads to the Roman standard of durability. Even if just left alone an asphalt road becomes a path within a decade or a bit more. (That depends on the climate, of course.) Bearings would be easy to recover if the idea were preserved, though the replacements might be small cylinders of hardwood.
Metals, though, are difficult. I'm not sure that modern alloys can be reused when the best you've got is a charcoal furnace. You're probably stuck with what they are or what you can cut them into (well, include filing). A lot of them are fairly rust resistant, at least if you stay clear of salt water, but I'm not sure what proportion. Anything containing Titanium or Aluminum will probably need to be used pretty much as is.
There are also technologies that require a certain scale of society to be feasible. Having the knowledge required to build something like a rolling mill is one thing ... having a society that has both the organization to do it, and the scale for it to make sense, is another.
Thanks for the reply! I remember loving Leibowitz.
Are you assuming a supply of electricity?
Unless you're instantly deleting all copper and aluminum wire from the simulation, it is reasonable to assume small-scale electricity generation wherever it would seem practical. For anyone who's fed, sheltered, and secure in the short term and has basic DIY competence and access to wire, a small wind/water/steam turbine is going to be one of the first orders of business once they realize that grid power and fuel deliveries aren't being restored.
In my mind you are describing the kind of electrical supply that would keep an incandescent bulb burning when the wind was blowing or the stream was rushing, not the kind of electrical supply that would support a laboratory or a factory.
At first, yes, that's what people will bring online quickly. But it's not inconsequential. Don't underestimate the value of lightbulb-scale power generation to people who have none. https://en.m.wikipedia.org/wiki/William_Kamkwamba
Over time, small-scale tech can be leveraged to build more ambitious projects. That's how our great-grandparents did it, and they didn't have millions of tons of free copper wire and precision-machined bearings to work with.
I got interested in this question because I ran into people who think that "collapse" would be a good thing. Of course we probably mean different things by the word.
Would you agree with me that loss of electricity, all at once or region by region, would cause the death of about 8 billion persons?
I think within about 90 days.
Aluminum wire isn't a good bet. In my experience, bend it a few times and it breaks, and you don't want to try to weld it.
It's certainly not my preferred material for DIY projects, but it's a perfectly functional conductor that's widely used in construction and industry. If for some reason you were weathering the collapse of civilization in a house with aluminum wiring, you could certainly repurpose it.
And I don't know when/why you would want to weld electrical wiring, but it's certainly not necessary for building a small/medium-scale solenoid coil.
Yes, there would be lots of wire available for quite a while. Since only a few million humans are left out of 8 billion, it would go a long ways.
But at some point you need to make more. It takes gobs of electricity to make aluminum. So that's probably out, on the other hand aluminum corrodes very slowly, so could be there for thousands of years to be recycled.
I believe copper is much easier to refine.
The big question is will the scavenging society evolve to a technological society or devolve to farming-hunting-gathering?
Are there any Meet Ups scheduled for the Psychedelic Science conference this week in Denver?
Not that I am aware of, but I am attending. DM me if you'd like. I imagine there's more who read ACX/SSC
James left his contact info above, I think. https://astralcodexten.substack.com/p/open-thread-281/comment/17474652
Thanks. I would've missed it
As a child I often encountered new words by reading them, and later mispronunced them when using them in conversation.
An internet search not only confirms that this is quite common, but gives it a name: Calliope Syndrome, with 'calliope' deliberately mispronounced 'cally-ope'.
For example, pronouncing epitome as epp-ih-tome is quite common.
One of mine was pronouncing inevitable as inny-vittable.
My dad told as a child he realized the villain in a children's book was very bad because he had misled (mizzled) someone.
Some of mine were variations on this, with a young me misreading the word. Thus, foliage became foilage, and debris became derbis.
I also thought that there was an artillery weapon called a mortimetre. I think I had combined mortar and millimetre.
A childhood friend insisted that a certain aircraft manufacturer was called Boing, becauses its aircraft bounced upon landing.
So, are any other ACXers good Calliopers?
Well, as a non-native English speaker there are many words I will read and mispronounce in my head multiple times before being surprised by someone's use of it in conversation. Awry (aw-ree), Oceania (O-see-ey-nee-ah), and a different confusion with albeit (all-bait) are some examples that come to mind, but there are many more. Plus every word where I'm unsure whether the i is an ee or an aye!
Also I've just looked up the pronunciation of epitome. English makes no sense
Another one ... when young, I thought that Canada's elected leader was the Pry Minister.
I realized that my above example, and various others I posted later in this thread, are not examples of Calliope Syndrome, but rather the opposite, in which I write incorrectly something I have misheard. I don't know if there's a name for this.
Another one comes to mind ... when young I was singularly bad at, and uninterested in, sports.
In Grade 8 I heard some of the jocks talking about a ladder-roll pass in football.
I was well aware of the word lateral, but arrogantly assumed that the sports guys would not know such a word.
Sort of related, in Grade 3 I realized that USA stood for the United States of America, and infered that USSR therefore stood for the United States of Soviet Russia.
The next day at school, Miss Millar was away sick and there was a young substitute teacher instead. Eager to show off my newfound knowledge, I worked out a way in which I could work this into a conversation naturally.
I left my desk, walked up to the teacher's desk, and said "Pardon me. I wonder if we might at some point be reading some stories about other children from around the world, including perhaps the United States of Soviet Russia."
The sub looked a bit bewildered and said "Uh, I don't know ...".
I wonder if she recounted this in the teachers' lounge later.
The main one I remember is albeit, which I pronounced all-bite. I heard people say albeit correctly in conversation for years before realizing these weren't two separate words.
I love it! Albeit (as you pronounced it) sounds very German - Albeit mit Nudel ist sehr gut!
For many years growing up I had two distinct vocabularies. One written and one spoken. And I was always having these moments when I realized that what I thought were two different words was just me 'pronouncing' (in my head) the written word the wrong way. (I love reading and stink at figuring out how to pronounce words.)
For years I thought that Penelope (pronounced correctly) was a different name from Penelope (when seeing it written, I pronounced it Penny-lope).
As a non-native speaker who learned 99% of her English by reading, I have *tons* of those. Like recipe, which I'd thought would rhyme with wipe, pipe, etc.; and height, which as it turns out does not rhyme with eight – so what's that e doing there? Found some new ones in this very thread, too.
I was maybe 35 when I realised that the old-timey wild west / hillbilly word "vittles" exists is because that's how "victuals" is pronounced...
Thank you, I've just learned something new!
The dialect word that gets me is "chile". This is supposed to be "child" with the final d elided out, but it looks like it's pronounced either like the food or the country.
The narrator of a Heinlein audiobook I listened to (I think "Methuselah's Children") did just that. The character referred to someone as "honey-chile", and the narrator read it as if it were referencing a sweet variety of capsicum
There was a low-budget radio ad for a small local company in which the voice actor yelled out at the end "C'mon, you guys, get the lead out!"
He pronounced "lead" as "lede" rather than the intended "led" (indicating Pb).
"Get the lead out!" is short for "Get the lead out of your boots!" (I.e. Remove the heavy metal from your boots so you can hurry up.")
Context can help:
"Well, honey chile, don't you fret none - jes' eat up that there chilli, and don't you worry about ol' Augusto Pinochet from Chile." 😉
Former Prime Minister Gillard had a good one with "hyperbole". https://www.youtube.com/watch?v=6xFrOxUFL-c&ab_channel=Channel10
I do this all the time, my family is big on reading so we notice/comment on it.
I know this happened to me fairly frequently growing up, but I've mostly forgotten the actual words. Fairly recently I did find out its pronounced "chi-batta" not "see-batta", but that's pretty minor.
The biggest example I can remember from when I was a kid was when I was in third grade or so and was nearly in tears because I couldn't say or spell the word "theatre" in a way that could make the teacher understand what I was talking about. It didn't help that I had read a bunch of British fantasy books and so picked up some odd bits of Commonwealth spelling (am American).
I had awful spelling as a child too. I've mostly solved it with a combination of typing muscle memory and spellcheck, but a lot of my school assignments were borderline unreadable. I've heard this is common with kids that learn to read by reading a lot, which makes me suspect the entire idea of "whole word reading" is a crock even for kids that can handle it.
The various British English localisms - the Warwicks, Worcesters, Thameses, and so on - took me a while.
And I still pronounce quixotic 'keeshotic' on the insufficiently rare occasion I feel pretentious enough to actually use it in a spoken sentence.
Loads of British place names and surnames are pronounced completely differently to how they are spelled. As a couple of examples off the top of my head, Featherstonehaugh (a place in Yorkshire) is pronounced Fanshaw, and Woolfardisworthy in Devon is said as Woolsery. There must be a list online somewhere, such as a Wikipedia page. Some antiquated spellings have been simplified to match the pronunciation, such as Brighthelmstone to Brighton in the early 1800s.
I remember as a young kid mispronouncing "nebula" as "nubela", and my brother used to misquote the title of a medical TV series "Doctor Finlay's Casebook" as "Doctor Finlay's Bookcase"!
I've never met an American who can actually pronounce Worcestershire.
(I'm sure there's plenty of other English place names they couldn't pronounce either, but they don't make popular sauces so it doesn't matter.)
There's a meme I found very funny - an attractive young couple are sitting at an outdoor table, presumably enjoying a romantic dinner.
The young man says "There's something I want to say to you, but I don't know how to say it".
The young woman, filled with happy anticipation, replies "Just say it!"
The young man then says "Worchestershire!"
Think Bertie "Wooster", and his resourceful butler Jeeves (and the "oo" is pronounced like the "u" in sugar, not the "oo" in (rocket) booster! )
There’s actually a passage in one of the stories where someone asks him if his name is spelled Worcester.
I used to pronounce "Home Depot" as "Home Dee-pot" (pronouncing the "t" so that it rhymes with "teapot"). I actually even had a temper tantrum about this in elementary school.
I'm a little depot
short and stout
DIY products
are all about.
Customers get steamed up
then they shout
"This board's twisted,
throw it out."
Like that departmental manager called Dee, who (people said behind his back) would announce "Hi, I'm Dee, and I'm part mental"!
An empty-net goal, as we say here. 😂
I call it "Home Despot." I also used to call Bed Bath & Beyond "Bloodbath & Beyond."
The Willamette river, also the Intel code name.
Pacific Northwest Native names got anglicized in a bunch of different ways, and this name didn't make sense until I learned that it was one of them.
Oregon and Wisconsin are etymologically the same word.
Yeah I had a lot of words like this. One of my favorites was thinking that the writers Go-eth and Goo-tuh were different people.
Wrote a new post on how Yakult succeeded outside of Japan when other products failed
https://hiddenjapan.substack.com/p/the-yakult-story
So the standard critique of presidential systems is that, by allowing voters to pick the executive directly, it's prone to cults of personality and demagoguery. This was the story with Latin America traditionally, and the US more recently. By contrast (in theory), parliamentary systems removed that decision from the voters. But in practice:
The head of the largest party is usually (yes, not always) the new PM
Voters know who the party heads are going in to the election. So they are effectively voting for their country's executive in a similar manner
I.e. voters knew if they voted Tory in the UK, they'd get Johnson. I've even heard critiques that voters frequently don't care who their local/constituency rep is, and aren't evaluating him or her on their service or policies- but are just voting (in this example) for a Tory district rep because they want Johnson to be PM. What do people think of these critiques of parliamentary systems? Are they really not that different from a presidential one?
Edit: To be clear, I strongly prefer parliamentary systems- I'm just stress-testing the arguments for them
I don't think first-past-the-post parliamentary systems and proportional parliamentary systems are all that similar. Having only voted in the latter, I have always voted for a party whose policies I've agreed with, never for a PM, and never for a local rep – in fact, I've never even known who that is – and I would think that most of my compatriots would feel the same.
It would be odd for Tory voters to assume that Johnson could not be replaced mid-term, seeing as he had himself already replaced May mid-term.
Whether it is proportional changes things a lot.
The British electoral system is very undemocratic. In 1997 Tony Blair won a huge majority in the House of Commons winning 40% of the vote, which made him virtually a minority elected dictator because the Commons is almost like the President plus Congress combined.
Specifically regarding the risk of "strongmen" with cults of personality and demagoguery, I'm unclear on whether it would be meaningfully harder to convince half of Parliament, than half of the entire country. Perhaps it is the case that the strongmen would find a way, no matter what? And they happen to be associated with "presidential" systems for incidental reasons?
Convincing parliament is a very different game to convincing the people.
People know that they individually have no power, and want someone who will represent their interests.
Parliamentarians already have some power, and want someone who will give them more power.
A "strongman" archetype is appealing to people, but not to MPs; MPs want a collegial leader who will listen to them.
I didn't mean "convince using the same methods". I'm less concerned with whether their public image looks like that of a "traditional" Latin American "strongman", and more about whether they leave power gracefully, whether they got power fairly, and whether they create mass movements that poison the commonwealth in the process.
There are big elections in the UK and US for parliament and the President, but they do vastly different things, don't they? Comparisons might get weird once you get into the particulars. Why not compare parliamentary elections with congressional elections?
In the US most House and Senate races focus on the local candidate, though national ad campaigns have reduced these to elections about party leadership, perhaps similar to the UK. I've heard a theory that's an outgrowth of modern electoral strategy and communication technologies, the McDonaldization of political advertising. Targeting one big national ad about Pelosi or McConnell and re-using it with slight modifications is far easier than making highly customized ads for every local race. I suspect we'll see that more in lots of countries for those same underlying economic/technological/strategic reasons, independent of any underlying political system.
On paper, the US President could be seen as primarily a figurehead focused on running the foreign policy establishment. In practice, they tend to campaign with some sweeping domestic legislative agenda, even though it's not a legislative role and is mostly limited to the power of persuasion to accomplish anything there. It is admittedly puzzling. A key distinction is that in this system, the executive and legislative branches are nominally divided, which does limit the power of any individual component.
When you say you strongly prefer parliamentary systems over presidential systems, you're not saying you prefer unitary over divided government though, right?
I suspect the crux of what you're saying is that you primarily want political parties to have much more power relative to individual candidates, no? In the US context, you might say you believe the "whip" should have much more power or be much more effective. This has varied over time, party discipline ebbs and flows.
If your central worry is cults of personality, I'd consider whether those are a bug in humanity, vice any particular system. If you exclude them from the political process, maybe eventually Caesar crosses the rubicon. If you invite them in, with a long tradition of peaceful transitions of power (and lucrative post-leadership speaking circuits!), hopefully they take their win and retire a few years later. Maybe the 1% most tyrannical leaders can slip the bonds of any institutions, so it's coming sooner or later, and in the US we've only drawn 46 lots so far.
I'd be very interested in Scott's early take from the Dictator book club on the extent to which institutions matter when confronted with the most persistent and dedicated authoritarians. Presumably some institutions matter to some extent, like, not having any functioning institutions at all seems a recipe for disaster. Good questions all worth exploring!
>the US President could be seen as primarily a figurehead focused on running the foreign policy establishment. In practice, they tend to campaign with some sweeping domestic legislative agenda, even though it's not a legislative role and is mostly limited to the power of persuasion to accomplish anything there
? They head the bureaucracy and can issue executive orders, which is gigantic power in our present-day country and especially in divided government. They can veto whatever Congress passes, unless there's a gigantic supermajority. They also appoint the judiciary
Do people in parliamentary systems vote for PMs over parties? I found this article (https://journals.sagepub.com/doi/10.1177/13540688211019239) (haven't read or analysed in detail), which seems to argue that most people vote based on party, but taking individual politicians into account. Personally, living in one, I think the parties end up with more name-brand power than individual politicians, who change every few years and can at any rate be replaced by the parties at any point. To some extent, this is visible in e.g. the US as well (like the Democrat supporters who wanted Bernie but voted Biden because that was the party line.)
However, even if this is not the case, I think based on the presidential vs parliamentary systems of the West in recent years, parliamentary systems do seem to obviously produce fewer cults of personality. I would imagine that this is because, in order to be a viable candidate for PM, politicians have to be established and respected members of the party. This means they have to be basically likeable, polite, good negotiators and share many of the same opinions as the other MPs. So, even though people might be voting for the kind of strong personalities that produce leaders like Trump, they really aren't able to choose those options often.
One thing to note is that in the parliamentary systems I'm familiar with, the Prime Minister doesn't have any special powers over and above what Parliament chooses to delegate to him. The Australian Constitution, for instance, does not once mention the existence of a Prime Minister. It mentions Ministers, and one of those minsters has always been a Prime Minister, but there doesn't have to be.
The key thing, though, is that the Prime Minister is not directly chosen by the election but by the vote of the party room, and he can be replaced at any time if the MPs of the leading party decide they don't like him any more. This removes many of the failure modes of Presidential systems, since there's a limit to how badly the PM can screw up before he gets removed, whereas a President can be incredibly hard to dislodge.
Yeah, but as Australia learned *repeatedly* after 2010 - the rest of the party can boot the leader pretty quickly and make it clear that it is indeed the party and not the leader who is in charge. The UK has had this recently as well, obviously.
Why do the Westminster systems (UK, Australia) have this, but we rarely see parties removing their leader in more proportional governments? As I mention down below, I can't really imagine say the CDU ever removing Merkel
Italy?
Not in the late years of Merkel, but in the early years this wasn't inconceivable.
The first German chancellor, Konrad Adenauer, was removed in exactly this way in 1963. He fought for staying chancellor for longer, and he fought vigorously against his successor Erhard (from the same party), but CDU and the coalition partner FDP forced him to resign.
Helmut Kohl was almost removed by his party in 1989, called the "Putsch" of Geissler, Späth and Süssmuth. It failed, but not by much.
In Westminster systems, the party in power usually has a majority of seats and can do what it likes. In proportional representation systems, the government is usually a somewhat temporary alliance of smaller parties that form a majority, with the leader of the largest party in the coalition usually getting to be PM.
This means that this party now has difficulty replacing its own leader, since they also need to convince their coalition partners to go along with this leadership change lest the coalition fall apart.
The German chancellorship is a pretty different situation entirely.
Thanks. I learned that Norway is the 1 parliamentary system (to my knowledge) with no constitutional provision for early elections. They've removed 2 PMs in their history and, with no possibility of coalition collapse, just replaced them with someone else mutually agreeable. Maybe that's the right way to do things
The answer is - some are voting for Johnson, some aren’t.
Johnson in any case is a likeable buffoon and not much of a populist or demagogue, just very incompetent.
I don’t think you can handwave away that his own party got rid of him either, as did the Labour Party with Corbyn. It seems the parties do restrain the populist tendencies of their voters.
I think this is a valid criticism of the parliamentary system.
During the run-up to the last provincial election, a pleasant candidate for the opposition came to the door asking for my support.
I replied that I liked everything I'd heard about him, but that I found the leader of his party so offputting that I feared my vote for him (the local candidate) would help the leader of his party become premier.
The argument I've heard for parliamentary over Presidential system is that they avoid "constitutional crisis" when one branch of government opposes another.
https://www.vox.com/2015/3/2/8120063/american-democracy-doomed
Robin Hanson says firms tend to more closely resemble parliamentary governance, so that form has presumably been selected as better.
https://www.overcomingbias.com/p/yay-parliamentshtml
Constitutional crises can still happen in parliamentary systems -- see the Australian constitutional crisis of 1975 in which the House of Reps and the Senate were deadlocked and unable to agree on a budget.
In that case the situation was resolved by the intercession of the Crown. The Labor Party is still salty about that one all these years later.
One way to solve Constitutional crises is to vest ultimate power in a person or institution who can be relied upon to never actually use that power unless strictly necessary. Not all counties have the benefit of such a reliable institution though.
>Robin Hanson says firms tend to more closely resemble parliamentary governance
I'm not an expert in corporate governance, but I doubt that elected members of the board are organized into anything like 'parties' who come with 'party leaders', which is where I think that analogy breaks down. Shareholders are not organized into highly ideological groupings and voting for board members based on belonging to that same tribe. And so on
Sometimes boards do wind up broken into various factions competing for shareholders' votes to get their agenda through. If this happens it usually means that the company is screwed, and investors should stay the heck away from it. Well run companies tend to look more like Singapore.
P.S. If the response is 'well the Tories removed Johnson'- sure, but other countries that use coalition systems tend to use party lists and have very strict party discipline (they can just remove you from the list if you irritate the party). If Angela Merkel was caught doing something untoward, was her party going to remove her? Seems unlikely
Tunneling between adjacent canals? Or low barrier.
Necker cube and Rubin vase.
MDMA Therapy: https://www.bbc.com/future/article/20230614-how-a-dose-of-mdma-transformed-a-white-supremacist?ref=thebrowser.com
As much as I dislike white supremacists, I'd hate even more to live in a world where white supremacy is categorized as a disease and its believers forced to take drugs so they can say things like:
"Love is the most important thing [...] Nothing matters without love."
And do things like:
"That night Brendan reached out to Chicago Antifascist Action and connected with the specific activist, "S", who had gone undercover in Identity Evropa before revealing Brendan's identity"
"Brendan hired a diversity, equity, and inclusion consultant to advise him, enrolled in therapy, began meditating, and started working his way through a list of educational books."
Imagine if a drug turned violent, masked, riotous antifa members into family-oriented religious conservatives who reached out to pro-Trump organizations for re-education. Wouldn't most people on the left, even those who detest antifa, feel uncomfortable with such a drug? Especially if it was presented as a therapy, implying that being part of antifa is an illness?
Im convinced that MDMA in the water would make us all kinder, but it’s all a bit brave new world.
When I was younger I used to go to raves in my college town. Well, it was called a rave but technically wasn’t quite, as it was indoors. Anyway the crowd, who were mixed between college students and fairly rough more working class crowd, were in general all hugs and kisses. I myself didnt do MDMA, or any drug really. It was a great sweaty night out though.
One night however, the atmosphere was much more hostile. It was a crowded venue and and inevitably you would bang into people or step on toes, and most nights this caused no problem.
But on the night in question bumping into somebody was potentially dangerous, the crowd was more hostile and angry than usual, although many were still enjoying the dancing, the atmosphere was off. MDMA was not available, instead everybody was drunk, and fights broke out.
I believe the drug of choice these days on the same city is cocaine with even worse consequences.
>Well, it was called a rave but technically wasn’t quite, as it was indoors
I don't think I've ever heard anyone make that distinction. Surely a rave, post-mid-1980s, is simply a large dance event featuring extended mixes of a certain kind of electronic music, usually with the expectation that many attendees will be on MDMA or similar drugs? All the events I went to in the 80s or 90s that called themselves "raves" were in warehouses or rented music halls.
Has to outdoors for the purists. Maybe a warehouse. This was just a weekly nightclub.
After chatting with ChatGPT about magic systems for a while I asked about the group Astral Codex Ten and received the response below. When shall we host an ACX meetup in the astral realm?
> "Astral Codex Ten" is a tome said to contain profound insights and instructions for astral projection, a practice where one's consciousness separates from the physical body and explores the astral plane. The book is rumored to possess detailed techniques, ancient rituals, and intricate diagrams that facilitate safe and transformative astral journeys. It is regarded as a comprehensive guide to unlocking the mysteries of the astral realms and expanding one's consciousness beyond the limitations of the physical world.
> The group known as Astral Codex Ten is a closely-knit community of individuals who have delved deep into the teachings of the book. They form a network of dedicated practitioners and seekers, each with a shared interest in astral projection, spiritual growth, and the exploration of metaphysical realms. The group operates in secrecy to protect the sacred knowledge and to ensure the safety and responsible use of astral practices.
> Membership in Astral Codex Ten is exclusive and based on a combination of invitation and personal attainment. Prospective members often go through a vetting process, where their dedication, knowledge, and ethical understanding are evaluated. The group's aim is to foster a supportive and knowledgeable community that can share experiences, offer guidance, and advance their collective understanding of the astral realms.
> The members of Astral Codex Ten engage in various activities, including regular group meditations, collaborative research, and the sharing of personal astral experiences. They also organize workshops, retreats, and lectures to expand their knowledge and promote the exploration of consciousness.
Our secret history is exposed! It must be destroyed before it can tell the world.
The fantasy version of my life is way more interesting than my real one! 😁
My inner narrator had an omnipresent echo while reading this.
Extremely plausible, and entirely untrue.
I forsee complications ahead in much more mundane realms...
I don't think there's much cause for concern. The only reason it was giving me these responses is because it was already operating in a magical fantasy context. If you ask it to tell you about Astral Codex Ten without any setup it just gives you the regular schtick about how the data cutoff was 2021 and it doesn't know anything.
So I gathered, from the context you provided. But given the number of credulous folks out there who are already using ChatGPT as a substitute for google, I still see scope for issues.
The next couple of years are going to be tough in terms of finding an authoritative and accurate source of data for casual enquiries. Google is now a mess thanks to SEOs, and AI will only add to that in the short run.
Friends were talking on the weekend and the point was made that public libraries may have been somewhat premature with their decision to throw out their physical books (as many of them did in the past decade or so) because reference books, for all their flaws, stay put in terms of content, unlike many on-line sources of information.
They need to update the data cutoff for sure.
They can't push the data cutoff much further, without their training data being contaminated by the output of GPT and all the other LLMs. Which is seriously problematic, though predictions vary on how much contamination is too much.
Have a grim video about the non-existent chance of peace in the middle east.
https://www.youtube.com/watch?v=JN4mnVLP0rU&ab_channel=RealLifeLore
One aspect is the stickiness of national borders. There are practical and emotional reasons for people to not want national borders moved, while at the same time there are high costs to national borders.
Any ideas about what might be able to take the edge off of nationhood?
Meanwhile, at least national borders guarantee that everyone doesn't make the same mistakes simultaneously.
Meh the borders of almost anywhere can be described as such. Some societies overcome that (France, UK), others don’t.
The old story about blaming Sykes-Picot for everything is pretty stupid.
You can come up worth a ton of different just-so stories which all fully explain the ongoing violence.
Oil and the high stakes of controlling it.
Multiple religious homelands.
Primitive tribal culture more acustomed to conflict.
Ongoing religious war between Muslim sects.
Ottoman control of area retarded state formation and institutional development.
Etc.
Etc.
The idea that if there were simply better lines in 1920 or whatever it would somehow be a bastion of peace and comity is farcical.
Those are good points. I agree that Europeans divided Africa, the Middle East, and South Asia poorly, but I also have yet to see anyone put forth better alternatives that withstand real scrutiny. For example, if we want to divide the Middle East so that each group has its own, homogenous country, then what do we do with ethnically mixed regions? Drawing a line down the middle risks a bloodbath like the partition of India and Pakistan or the Balkan civil wars.
What about ethnic exclaves? What if there IS some way to neatly draw a national border so that it fully encompasses one ethnic group and no other, but those borders also leave it landlocked and not possessing any natural resources?
Another problem is deciding how to define "ethnicity" for the purposes of creating nations. Are we just coding inhabitants by religion, or by religious sect? On the national borders created by that, are we overlaying a second layer of lines to separate the people by language? How about by dialect? Is there then a third layer of lines to reflect tribal zones of control? How do we handle areas of nomadic people? If you get too obsessive trying to make national borders that reflect human geography, you end up with micro-nations that are not viable or barely viable for varieties of reasons.
Look at this map of Africa's ethnic regions overlaying its actual national borders. If the Europeans had somehow, in defiance of the technology and resources available at the time, had perfect knowledge of the continent's human geography at the time they departed and created the different countries, it would have resulted in crippling levels of fragmentation.
https://blog.education.nationalgeographic.org/2015/02/18/africas-dazzling-diversity/
Sorry, I planned poorly and will be ~45 min late to Atlanta meet up. If you're there and see this comment, please tell the organizers.
I'm curious how those comments on Njal's Saga were the top comments. Has this school of criticism fallen out of favor to such an extent that the observation about Christian influence seems novel? I feel quite certain that when I first heard of Beowulf as a child the very next sentence was to the effect of "it was written down by Christians". Was I raised in some kind of anti-Christian-influence-on-literature bubble?
The bar for being the best comment on a book review and the bar for being a truly original and novel piece of literary scholarship are pretty different.
No, they just gave details about how it worked and what effects it had that weren't in the review and that I thought added interesting context.
I think it’s important to keep in mind when reading that sort of literature, but agree that it is a “low-hanging fruit” sort of observation.
Hi! I'm helping draft a script about Charter Cities for the youtube channel "RationalAnimations" (https://www.youtube.com/c/rationalanimations). Some of our previous videos have adapted essays by Eliezer Yudkowsky and described some Robin Hanson derived concepts like prediction markets and the "grabby aliens" hypothesis.
Naturally, we want to make sure we're presenting the idea of charter cities in an accurate, compelling, and understandable way before we the script gets adapted into video form and potentially seen by ~100K people. So, I'd love it if anyone here at ACX would like to provide feedback on the script. It's about a ten-minute read; you can use the google docs "comment" feature to highlight a section and add your thoughts: https://docs.google.com/document/d/14wEa6SFiWo9uw30UB52DHIHXhHrIprnT3EzAbU-2teA/edit?usp=sharing
I spend most of the script talking about the motivation for charter cities as potential engines of economic growth in developing countries (inspired by the success of places like Singapore), and explaining why a new city with the ability to set its own business regulations might help encourage growth. I use Itana in Nigeria as my main example, mostly drawing from the linked FT article when describing the various reforms they'd like to implement. (I also talk about Prospera Honduras, as described in ACX's "Prospectus On Prospera": https://astralcodexten.substack.com/p/prospectus-on-prospera) After addressing some potential objections, in the conclusion I describe charter cities as having the following three main benefits beyond the direct economic growth of the city itself:
1. the idea that seeing legal reforms work well on a small scale could inspire larger national changes.
2. the idea that charter cities could spur “governance competition” resulting in better, more effective governments generally, by giving people more options of where to live.
3. the idea that experimenting with totally new types of law could bring huge benefits, by exploring new institutions (like prediction markets or new land-use laws or etc) that might solve problems that even advanced countries struggle with.
We'd love to see any feedback that you have on the script!
You could also look at the story of China's special economic zones which were sort of like charter cities. However, I would also devote some time to the rather serious problems with the idea, mainly of credible commitment and governance. The former is clear with the issues that prospera has faced. The latter suggests that if you can make a charter city work, you probably don't need it, and if you do need it, you probably can't make it work
The biggest weakness of charter cities (an idea I'm very much otherwise intrigued by) can be summed up by the phrase, "There's no sovereignty like sovereignty." You're either sovereign, or you're not, and actual political independence that can't be capriciously taken back from the same hand with which it was given is hard to come by. Nice work if you can get it, though.
Are you sure you want to post a video about charter cities on the same channel? They're a way more politically fraught concept than alien speculation or even AI-risks, so there's a significant possibility the channel could get needlessly pidgeonholed for having controversial niche libertarian-capitalist politics (assuming it's implicitly pro-CCs. If it's against or mixed that's a reduced risk but still much "spicier" than the usual fare).
If you've heard of Zuzalu, Vitalik Buterin's pop-up city in Montenegro, here's an offshoot of a group that met there:
July 20-23, 2023: African Regenerative Cities & Futures @ FreeFlow Eden Zanzibar: https://lu.ma/zanzibar_regen
This is an event to bring together entrepreneurs, artists and governance innovators with focus on African cities & technology
What is the best way to learn about ways of understanding why adhd (or add if you rather) happens, what’s the causes, etc? I’m particularly curious about well-defined meta analysis with reasonably accurate definitions
It's kind of embarrassing to the profession (of which I'm a member) that the diagnosis of ADHD isn't even made by neuropsych testing. It's made by assessing whether people have certain symptoms, generally by asking them whether they do. I'm sure people have tested groups diagnosed with ADHD and compared them to non-ADHD people on various things -- how well they remember under various circumstances, reaction time, etc. -- but whatever has been found must not be clearcut enough to make possible diagnosing ADHD via test results, the way you can certain learning disabilities. There are neuropsych subtests that measure executive function, but that's just one part of ADHD. I just spent a little time on Google Scholar looking for a decent article, and this looks pretty good, and not terribly technical.
https://www.sciencedirect.com/science/article/abs/pii/S2215036616300323
This is not unique to ADHD or even to psychological conditions. Lots of medical conditions are only diagnosed by asking people how they feel and don't have tests to diagnose (or have bad tests that my find something but it doesn't prove causation or point to an effective treatment plan).
With ADHD specifically, the diagnosis depends on a lot of the negative impacts ones life. If you don't have the negative impacts there isn't much sense in diagnosing a illness as, to you, you aren't ill.
Yes, it's true that lots of psych diagnoses are made based on people's descriptions of what it's like being them and what hurts. Still, you would think ADHD would be measurable: If somebody says they cannot keep their attention on tasks for long, you'd expect them to do badly at tedious little neuropsych tests where you have to go through a whole page of little symbols and mark all the ones that have a horizontal line in them. If they say they can't take in info from lectures in college, you'd expect them to do badly on a test where the neuropsychologist reads them a coupla pages about something sort of dull, then asks them questions about what he just read them. That's why I say it's a bit embarrassing that there is no neuropsych test for ADHD. Makes you wonder how valid the diagnosis is. I'm sure some people have a genuine wiring problem that interferes with attention, but the interview method of diagnosis may be picking up a bunch of people who have a different problem. For instance, maybe somebody in college got coerced into picking a major they hate, and that's why they can't keep their mind on the lectures and the textbook -- it's boring info to them, plus they are furious about having been coerced by their parents or whoever.
And actually, with medical diagnoses there are other things involved besides people's descriptions of symptoms. There are tests that allow you to see, or see evidence of, the actual malfunction that is causing the symptoms. Actually, I think there are only a few medical diagnoses where there is no way to test for or look at the malfunction: Things like Chronic Fatigue, for instance. And with those things many docs question whether they are actual distinctive diseases, precisely because tests don't reveal anything unique going on in the body that would account for the symptoms.
From my perspective (which should be treated as not competent by any means) the way psychiatry generally works is that it tries to help people overcome stuff in their life, not to provide you with maximally accurate information about the stuff that’s going on within their heads.
e.g. in your case even if someone is failing badly in academia because they hate the subject or whatever else the reason is, they still might wanna be treated like someone with adhd because intrinsically they value ability to finish a degree much more than doing stuff that’s not boring. I’m almost convinced that that’s the best way to look at all psyche disorders.
That being said, you sort of want to improve one’s state of mind as they ask you to do, not to sort them into small boxes where you can see statistically significant differences.
And that’s mainly the reason to claim that diagnosis is in fact irrelevant to clinical practice in a sense that you shouldn’t deprive someone of, say, drugs/therapy on the basis of not being diagnosed with adhd
Well, I do not find diagnoses of much use, generally. They're very imprecise, and probably do not line up with real categories. What I find most useful is to get a fix on where somebody gets stuck -- what's the thing most other people can do that they can't? I look for the bottleneck. In the case of the person who does not have ADHD I think I would quickly realize the the person hates their major and is furious that their rich grandpa was willing to pay for college only if they agreed to be pre-med -- because I'd be asking the person questions like, "so exactly what is it like for you sitting in a lecture on biology? And pretty soon he'd tell me that it's boring and irritating because he has never liked science, he likes history and literature -- or whatever. And then the whole grandpa story would come out. The bottleneck is that the guy hates his major and hates having his grandfather dictate it, and that naturally enough makes it hard for him to pay attention to his course work. Of course he still might think that his basic problem is ADHD, and of course that is possible -- he could have a coercive grandpa, a major he hates *and also* ADHD.
I don't agree that I should just go ahead and treat him for ADHD if that's what he wants. For one thing, it's important for somebody who suspects they have ADHD to know whether they do or not. It's bad for you to think you have a disorder you don't have, and also bad to think you don't have one that you do. And besides presence/absence of ADHD might be relevant to some life choices of his. For instance ADHD is an advantage in at least one medical field: emergency medicine. ADHD people are sort of hyper, which is a good thing if you are racing around trying to save a life, and EM docs report that the high stakes and excitement keep them focused. Also, Adderall is probably not very good for you, and shouldn't become part of your daily regimen unless you really need it. Plus a certain fraction of people who take it become abusers, so it has its risks. So I would try my best to figure out how ADHD-ish the is. With some people it's really obvious because they jump from topic to topic while talking, and don't really grasp things I say because they cant pay attention well. It's also useful to ask what it's like for the guy if he's reading a book or watching a video about subjects that do interest him. Can he keep his attention on it? And neuropsych testing can at least measure executive function, which is relevant to one aspect of ADHD.
I think if the guy does not have ADHD what I would do is help him clarify what's really getting to him --- coercive grandpa -- then explore options. Is he sure grandpa would really pull the plug if he changed majors? Is it worth writing his grandfather a long letter making the case for changing majors? Is it worth paying for college with loans? And if the guy decides he has to stay with premed, I would offer to help him keep up with his academic work via self-management strategies.
Shit, what I just said sounds arrogant
I didn’t mean that :(
Also, I beg your pardon but I missed couple of points you made and I think they are interesting to discuss
You said that it’s bad for you to think you have a disorder you don’t have and vice versa
But it sounds sort of weird from a clinical standpoint
I mean, I’m totally not an expert by any means, but it seems like adhd medicine generally helps with symptoms of adhd whether a person has adhd or not. Or, to be more precise, there is a sort of broad way of what symptoms a person with adhd may experience, and some of those might be experienced by someone without adhd, and thus those symptoms might be suppressed (is it a right word?) by using the same medication
If I’m mistaken in that, please explain why, but I’m not sure that having adhd is a discrete thing like “having an aids” where there are treatments that do help people with aids and don’t work for anyone else suffering from similar (but not caused by hiv) symptoms
However, back to your point — it (very much) does make sense to help people understand what’s happening with them regarding their consciousness, self-awareness, attention and internal problems. So I guess I’d put it that way.
It makes sense to understand what you experience and why and not be mistaken in that, or you’ll be confused about what to do and where to start, or even worse, you’ll be trying to do something that’ll make your situation worse (as people often do when they have false beliefs about important things in their lives)
But if you are aware of what’s happening with you in terms of internal experiences, then it doesn’t really matter whether you think you have adhd while not having one or other way around. (It’s still pretty much the same thing about being unhappy with some known problem that’s living within your brain, and willing to overcome it, using drugs or not)
Sure, there are some risks involved in prescribing drugs from adhd to someone without one, but I’m not saying that it’s necessary to give anyone who struggles with concentration adderall. What I’m saying is, sort of, like “we wanna optimise people desire-satisfaction as long as it doesn’t hurt them without their consent and doesn’t hurt others. E.g. if someone comes to doctor and asks them for a help with overcoming being bad in academia, reasonable(but not widely, if at all, recognised) guideline seems to be somewhat like the following
1) gather as much info about one’s condition
2) provide a person with some educated ideas of how to fix whatever it is that bothers them while also explaining risks
3) if the person in question says “I wanna some medication to be able to concentrate more easily, and I’m aware of risks but still wanna give it a shot”, then you probably should say ok
Also, worth mentioning that what you suggested is also a good way to treat a patient initially — trying to help him without getting him addicted to serious drugs is what I hope doctors do and I hope they succeed :)
That article is paywalled. Would you be able to share a non-paywalled copy, by any chance?
No, I don't have the academic golden key. Use Sci Hub: https://sci.hubg.org
I'm pretty sure there are at least two forms of ADHD with radically different sources; I notice two distinct clusters of people, one of whom finds it impossible to create habits, and one of whom finds it impossible to break them.
There are two types of ADHD - hyperactive and inattentive. People can have one or both. I am not sure they map to your groupings however.
Isn’t the latter supposed to be sort of like ocd?
I have no clue if this is the sort of thing that you're looking for, but I've heard that there's pretty good evidence that exposure to "screen time" (video games, tv, etc) at a young age increases prevalence of ADHD. If you are curious and want a citation I can ask my source about it.
https://publications.aap.org/pediatrics/article-abstract/120/3/532/71189/Does-Childhood-Television-Viewing-Lead-to?autologincheck=redirected
I was corrected that this is a longitudinal study about "attention problems," not ADHD directly. That said, I'm told it was published in a good journal, and is a solid paper. That said, there is field debate on the topic.
I'd like the source if possible as I have heard this claim has failed to be substantiated.
Sounds great, I’m down for it!
A large meta-analysis of genetic correlation studies is probably a good place to start: https://www.nature.com/articles/s41380-018-0070-0
Thanks, but it’s not quite what I’ve been looking for
Article you suggested is mostly about genetic correlation with little to no explanation of what happens in a brain (there was a point about dopamine transfer related genes, but it wasn’t elaborated as far as I can see)
I want to find some paper on internal mechanisms that cause adhd, say, lack of some neurotransmitters or so which can be sort of linked to day-to-day experiences
With almost no exaggeration, one can say that there is no explanation of the "internal mechanisms" of ANY complex human behavior. Neuroscience is a very primitive and useless endeavor. The best we can do in terms of causal explanation is that differences between humans are generally due to genetic variation + random noise.
It’s sort of a meaningless answer, not to say misleading, don’t you think?
Some things are generally caused by environmental factors, say, ptsd. And thou there is definitely a place for genetic factors that contribute to likelihood of having a ptsd, going to war is by far more decisive factor, at least it seems that way to me. Also, let me rephrase. I’d like to have some way to think about adhd, intuition, if you please, and I lack any ideas of where to look for it.
But war doesn't cause PTSD. There are no ancient accounts of soldiers getting PTSD, even though a far bigger percentage of the ancient male population went to war. Whatever is causing PTSD among modern soldiers, it seems to be a curse of modernity and not an inherent feature of war.
Seems about right
Shit, what do I need to with that information?
https://westhunt.wordpress.com/2018/08/18/ptsd/
In the general population, war is surely not a major cause of PTSD. Moreover, even in veterans, exposure to war is not a mechanistic explanation of the disorder--it is far from a sufficient cause and it tells us nothing about what happens in the brain. "War" as a cause of PTSD is not any more specific than the fact that PTSD has substantial genetic heritability.
Well, yeah
Sort of
I mean, it definitely increases your chances of getting one, but it’s not a major cause of ptsd worldwide (supposedly)
I don’t wanna bet on what explains better that you have ptsd — being on war, or being genetically likely to have one, but I’d guess those are comparable probabilities at least
Nonetheless, I still want to grab some intuitive sense of what adhd is and how it develops and which factors are known to be crucial, how does it affect internal experiences and/or chemistry of one’s brains, etc, etc
I’m pretty much sure that there is more evidence than simply “okay we tracked some genes that are somewhat correlated, thus explaining what people have adhd better than a coinflip”
Canals are too deep 😜
https://astralcodexten.substack.com/p/the-canal-papers
I just listened to RFK Junior on Joe Rogan. Does anyone know of a serious debunking of him? Sorry, I am too lazy to even do a web search. Just curious if anyone knows one offhand.
There are various types of "debunking".
As far as "this one epic burn on Twitter will *disqualify* him from being a serious presidential candidate": no. That type of thing doesn't happen. Stop imagining.
As far as "he says a lot of things that are wrong": of course he does. the problem is that, like Trump, there is so much you can't keep it all straight. The latest on Twitter, with ample debunking: https://twitter.com/MLiamMcCollum/status/1670442514778820610
As far as the broader sense of a debunking that encompasses "if David Sacks can explain how he misspoke, it isn't really a lie": once again, no. The truth cannot help those who hold "most things he say are similar to the truth" as their standard.
He's an anti-nuclear crank: https://twitter.com/energybants/status/1670215402272571395
I just came across this Twitter thread from yesterday specifically doing a rundown of the vaccine-related claims RFK, Jr made specifically in his Joe Rogan appearance.
https://twitter.com/han_francis/status/1670499396578844680
I myself yesterday posted in a sub-thread of this thread about RFK, Jr's reputation for being either willfully misleading or, seemingly more likely, honestly-conspiracy-minded* in his first foray into anti-vaccine advocacy, namely his articles in 2005 promoting the purported (and consensus-wisdom-to-have-been-debunked) link between the vaccine preservative thiomersal and autism -- https://astralcodexten.substack.com/p/open-thread-281/comment/17470692.
* "Honestly-conspiracy-minded" = so sure that thiomersal as a preservative in vaccines has been a public health catastrophe that insiders long knew about in the same way tetra-ethyl-lead as an "anti-knock" additive in gasoline was, he reads things like transcripts of CDC meetings devoted to the thiomersal issue with a highly biased prior. Indeed, he seems so sure these transcripts are going to be one of those cases where arrogant insiders blithely self-incriminate! And then, worse, RFK, Jr quotes those transcripts in his writings with substantial edits that strip key passages of context that most reasonable observers would think make quite clear that these CDC transcripts aren't those kind of smoking gun arrogant-insiders-incriminate-themselves thingies (see links in my comment for detailed examples).
“ You’re unable to view this Tweet because this account owner limits who can view their Tweets.”
:(
Whacky. It's not like I'm a follower of this University of Illinois - Chicago cardiologist Francis Han, MD (@han_francis). Hopefully this will be viewable for you:
https://threadreaderapp.com/thread/1670499396578844680.html
This one worked for me, thanks!
Thanks!
Good luck! Cross too far into taboo territory and most "Rationalists" just start muttering about how they don't even need to answer the question because it's so absurd.
:(
I also heard RFK Jr. and was so irritated by his "glitching android" way of talking that I turned the interview off after ten minutes.
He's just a crank.
If he is a crank, it should be easy to discredit what he is saying, without referring to how he says it.
Haven't looked at the substance, but I'll say it's not at all easy to discredit/disprove what a crank says and point you towards Brandolini's Law (The amount of energy needed to refute bullshit is an order of magnitude bigger than that needed to produce it. https://en.wikipedia.org/wiki/Brandolini%27s_law).
Okay, I stand corrected. If he is a crank, it should be possible to show that he misinterpreted the studies he cited, or that they are flawed, but not necessarily easy.
If he is *not* a crank it should be easy for him, or his supporters, to show what he is saying is true.
How so? He can cite studies, but that alone doesn’t show that they are valid studies, or that he has not misinterpreted them. What procedure are you suggesting?
What is there to know? He likes conspiracies, a few he likes are true, most are false.
> He likes conspiracies, a few he likes are true
Which ones?
He cited studies. Were they trash, did he misinterpret them, or do they say more or less what he said they do?
Maybe I should clarify what I consider serious. I am looking for someone like Scott taking the time to go through the various scientific papers RFK Junior mentions and point out which ones have problems which ones are fine but he misinterprets and where if it anywhere he actually has a point. I am not worried about what he thinks about Venezuela, or putting people in jail. I am interested whether his claim to interpret serious empirical Science is on track or off.
I have a couple things I can point out: (1) The idea that pushing the vaccine on people who had already had covid was insane. I agree that those people need not have been vaccinated, but there are good arguments against saying this was an insane step. For one thing, many people would not known for sure whether they had had the virus. And some people who objected to getting vaccinated would have lied and said they had had the virus when they had not. You can test for antibodies to see whether somebody has had the virus, but the antibodies from it fade, and a few months after having the virus the test will show invalid negative results. Plus, of course, testing everybody who said they had already had the virus would have been a huge expensive hassle. And given that the vaccine was safe, seems to me a reasonable decision to just vax as many people as possible.
About harm done to young men by insisting that they get the vax. There is no doubt that vax increases the risk of myocarditis and pericarditis among you males (not among young women). However (1) the risk is still very low; (2) it is a mild a treatable kind of myocarditis that does not do permanent damage; (3) at least one study has found that the risk of myocarditis after covid for young males is higher than the risk after the vax; (4) vaccinating these young guys probably saved some lives by reducing covid transmission and fatalities among older folks.
Here is a good thorough study of the myocarditis issue: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC9183215/#cesec70title
I am passing on this info because someone asked for info. I believe it is accurate, but am not absolutely certain. I am not interested in having an argument about covid issues. I have burned out on that. If any of the above is inaccurate, dig up the data that shows that and give it to OP.
I see that Vinay Prasad has done it now.
https://vinayprasadmdmph.substack.com/p/take-rfk-jr-seriously-what-rfk-jr
Not much of a debunking, he agrees with more than he disagrees and doesn’t go into what he disagree with in depth.
The term debunking is not used in evidence based medicine. It sounds too immature. You either show the evidence or not. The rest is just opinion.
So what? The OP asked for a "serious debunking", in his next comment he clarified what he meant, and you replied "Vinay Prasad has done it now". So I was just borrowing the word.
The post you linked to was not a careful consideration of the scientific literature anyway. Most of it was political in nature and about corporate interests and potential corruption.
What I mean is that one should not expect Vinay to do debunking because that's not what he does.
I really think that rationalists shouldn't do it either or at least very sparingly. It polarizes the discourse and assumes bad faith in the opponent and makes everything more tribal.
For example, someone can say that covid vaccines killed a lot of people. It is bad to make such proclamations without evidence. And yet, I simply take this as someone is saying “maybe vaccines killed a lot of people”. Now it sounds like rumination. And I can answer it myself – I haven't seen any strong or weak evidence that covid vaccines killed a lot of people. Only some rare anecdotes but it expected (for example, very rare anaphylaxis, myocarditis etc.).
I don't need to debunk, merely see what evidence is there. Some could even say that extraordinary claims require extraordinary evidence.
Now, if the opponent is arguing in good faith, he will try to provide more evidence. However, if it is done in bad faith, at this point he will start spouting nonsense or some conspiracy how every data point is controlled by nefarious organizations or will show that he simply doesn't understand how this all works at all.
In any case, you can go to Pubmed and type "debunk" in the search bar and see that many people use this word.
Prasad does the best job I have yet seen, but it is still a bit shallow. Kennedy cited studies and made claims. I still don’t know if he was misinterpreting those studies, or citing flawed or biased research, or accurately describing the emperor’s new wardrobe.
Thanks!
It's worth remembering that RFK, Jr. has been a public anti-vaccine advocate since 2005, with his first foray into public anti-vaccine advocacy being his advocacy for the contention that thiomerosal, an antiseptic and antifungal organomercury compound used as a preservative in a few US vaccines up until 2001, was to blame for many cases of autism.
It's worth remembering this since the allegations against RFK, Jr in that debate fundamentally were *not* about the *scientific understanding or lack thereof* he evinced in the thiomersal-causes-autism-yay-or-nay debate back then. Moreover, here's a key point of context for RFK, Jr's participation in the debate back then. By the time he got involved, the debate wasn't one of the sort where a tiny band of plucky contrarians stood against an establishment that wasn't doing anything. Rather, when RFK, Jr got involved in the thiomersal-causes-autism-yay-or-nay debate in the mid-2000's, he was getting involved in a debate that had been roiling for years *precisely because the establishment had just done something big.* Namely, in 1999 the American Academy of Pediatrics made a recommendation quickly seconded by the CDC making a more pointed-and-decisive recommendation that vaccine manufacturers remove thiomersal from all US vaccines "out of an abundance of caution".
[[ERRATUM 2023-06-19, 10:32pm EDT: The recommendation to remove thiomersal from all vaccines given to children and adolescents. The only vaccine since 2001 given in the USA that still uses thiomersal is *certain* flu vaccines, *and even then only when distributed in multi-dose vials, i.e., multiple sterile syringes will draw the vaccine dose from a single vial.* Single-dose vials are widely available for those flu vaccines and are recommended for children, adolescents, and pregnant women even though the risk of "mercury" to any of these groups generally is tied to the much-more-potent organomercury compound *methyl-mercury* whereas thiomersal leads to *ethyl-mercury* in the body. ]]
**Thus, we get to the crux of the matter. The two main allegations against RFK, Jr in the thiomersal-autism debate are that his public advocacy was either straight-up knowingly misleading, or it was honestly misleading since he had already become a conspiracy theorist in the bad, stereotypical meaning of the term. Specifically, two main allegations against him are:
(1) He quoted minutes of CDC meetings in highly-edited, inflammatory, and misleading ways.
(2) Moreover, he repeatedly made statements that showed either surprising ignorance or already-down-the-rabbit-hole-conspiracy-theorist-disbelief that vaccine manufacturers had indeed removed all mercury containing compounds from US vaccines by 2001.**
This 2005-era history with RFK, Jr's first foray into anti-vaccine advocacy is recounted in the following *2017* article that first ran in STAT and then merely got reprinted in Scientific American. It has many detailed side-by-side examples to support the allegation (1) of misleading, selective editing. (SIDENOTE: I bring up the fact Scientific American was merely reprinting the article since Scientific American during the Trump era got flak for having portions of it edited by quite culture-war-lefty types... though even here, note again that this is 2017... so, yes, it's in the Trump era and being reprinted by SciAm with a news angle of anger against the Trump Administration for placing RFK, Jr on a scientific advisory panel for vaccines... but no,this is long before COVID and the Trump-Administration-vs-"Follow the Science" saga attaining white-hot intensity.)
https://www.scientificamerican.com/article/how-robert-f-kennedy-jr-distorted-vaccine-science1/
The other angle present in RFK's first foray into public anti-vaccine advocacy with the thiomersal-autism link --- though I don't know how much RFK, Jr either had knowledge of it at the time or can be plausibly expected to have had it --- is that 2 experts he quoted heavily, father and son geneticists Mark and David Geier, turned out to be grifting quacks... including in the opinion of courts of law. (Some of this is mentioned in passing in the above SciAm reprint of the STAT article, other can be seen --- sorry for my link laziness --- via Wikipedia https://en.wikipedia.org/wiki/Mark_Geier ).
P.S. Speaking of link laziness, I know there are allegations but I don't have time this Father's Day to find them that RFK, Jr. has repeatedly used various statistics poorly... and I mean "statistics" and "poorly" in what *politicians* think of as statistics --- i.e., quantitative gotchas --- rather than he does legitimate statistical analysis, just poorly because he's a lawyer not a statistician. Two examples of these are:
1) Inflating the number of vaccines Americans get by age 18 with the annual *flu* vaccines recommended by the CDC for teens and preteens (which is considered inflating the numbers since, again, RFK, Jr for much of his anti-vax advocacy career has been focused on how vaccines *early in childhood* purportedly caused autism). [[NB: See ERRATUM above about how certain flu vaccines, when packaged in multi-dose vials do indeed still contain thiomersal, though single-dose vials that do not are available for all of them. This makes it somewhat more understandable why RFK, Jr focuses often on flu vaccines, even though it still is a "red herring"]]]
2) Conveying the impression that the rise in the percentage of Americans under 18 having "chronic medical conditions" since the time he was a child isn't simply both the straight-up rise in *obesity* (as well as, secondarily, public-health-stats-redefinition-of-BMI-thresholds-for-obesity).
I'm sorry that I can't point to a more definitive set of references for all the iffy things in RFK, Jr's anti-vaccine advocacy to date, but I do urge people to remember that his advocacy is not confined to the COVID era... as well as I do urge people not to sorta devolve into the enemy-of-my-enemy-is-my-friend logic (at least not too early in the process). The are worthy debates to be had on whether (1) school closures were a net negative and whether that was amply knowable in prospect and (2) anti-COVID public health policy, including vaccination, should've been much more defined by the steep dependence in mortality and morbidity with rising age. However, those debates are really a *tiny* portion of RFK, Jr's contrarianism on public health.
P.S. Happy Fathers' Day to those who are celebrating it.
P.P.S. Mentally healthy Fathers' Day to those who are having to cope with fathers who are gone.
Most of this is about his Rolling Stone article (of 2005?). Neither the article nor the errata are findable online by me. On Rogan he claimed that the errors had been introduced by his editors at RS, when they shortened the article considerably. I don’t know how plausible that is.
The criticisms I am seeing, when not pure ad hominem, seem to be about him misquoting or summarizing badly a meeting transcript. my main interest is did he misinterpret the science he cited, or is it junk, or what. Your link to twitter seemed likely to address this more directly. Unfortunately I seem not to be able to access it.
I don’t find his claims very plausible. But I don’t think I have much relevant evidence, either. It is bad that such a prominent public figure is spouting nonsense, and perhaps trying to paint his opponents in a bad light. But it is also bad that it has become so difficult to try to find out what is wrong with his garbage. I guess I just need to be less lazy, listen to Rogan again, taking notes this time, look up any articles mentioned, look for summaries or critiques, and make up my mind.
Ah... indeed. This post from Retraction Watch all the way back from 2011 might help explain --- well, at least give some flavor of the context --- why Rolling Stone and Salon disappeared the 2005 RFK, Jr article "Deadly Immunity":
https://retractionwatch.com/2011/01/16/salon-retracts-2005-robert-f-kennedy-jr-piece-on-alleged-autism-vaccine-link/ [NB: There's like 100% "link rot" in this Retraction Watch article. That is, none of the links within it seem to work, though maybe the Internet Archive, archive.org, would help.]
But worry not! The article "Deadly Immunity" is still available here based on how it was entered into the Congressional Record by Rep. Dan Burton (R-Indiana, 5th, previously 6th District over the years 1983-2013):
https://www.govinfo.gov/content/pkg/CRECB-2005-pt10/html/CRECB-2005-pt10-Pg13935-3.htm [scroll down 3ish paragraphs -- that is Rep. Burton's preamble -- to where it says
"Deadly Immunity
(By Robert F. Kennedy, Jr.)"
to get to the article itself.]
And the errata are available at https://www.salon.com/2011/01/16/dangerous_immunity/. At least, for now.
Thanks!
Thanks!
You're quite welcome.
As you probably were auto-notified, I just made a separate reply in the main reply thread to your comment because I found a Twitter thread that explicitly is a rundown of RFK, Jr's vaccine-related claims in his Rogan appearance.
That Twitter thread led me to add an erratum to my comment in this reply-subthread about RFK, Jr's 2005 entry into anti-vaccine public advocacy centering on the purported thiomersal-autism link.
To be honest, I have a soft spot for RFK, Jr., in two ways in fact:
1) First of all, I was more of a lefty in that 2005-era and listened to the brief attempt of leftier-Democratic-Party partisans to compete with right-wing AM talk radio: Air America Radio, which often had RFK, Jr on in 2004 and 2005. Indeed, my one bit of bona fide OG blogger reporting was on one of RFK, Jr's *other* highly-biased-prior-leads-to-conspiracy-theory-advocacy thingies: his contention that the vote in Ohio, especially, in the 2004 Presidential Election between George W. Bush and John Kerry was clearly tampered with... with a lot of that clarity he felt coming from him looking at "raw" exit polls (like several million other online Democratic partisans back in 2004). As major blast-from-the-past, I see my old Wordpress blog is still up and this was the key post:
https://williamkaminsky.wordpress.com/2004/12/11/systematic-voting-fraud-or-systematic-exit-polling-bias/
though, alas, my tour-de-force (if I do say so myself) of data visualization... that is my one scatterplot of raw exit polls versus official margin by state so as to show a systematic exit polling bias across all states, not just swing states seems lost to the ages (even on the Internet Archive despite claiming it captured it 13 times... weird).
2) If anybody has a reason to have a biased prior toward grand conspiracies it's RFK, Jr. At age 9, he could personally witness his dad RFK, Sr ask then director of the CIA (John McCone... a quasi-trusted ally of the family explicitly put in by the Kennedy brothers after they forced out Allen Dulles in the wake of the Bay of Pigs debacle) to not only come to the family house within 24 hours of JFK's assassination, but to swear on the grave of his (McCone's) mother that he (again, McCone) had no knowledge of CIA involvement in JFK's assassination. And, over the next 5 years from age 9-14, RFK, Jr could continue to witness literally "up close and personal" his dad never give up that theory, largely because McCone caveated his denial of CIA involvement with a "to the best of his knowledge" type caveat.
These are not really "debunkings" (certainly not systematic ones, although they push back against at least a few of his claims), but they are two pretty reasonable critiques of him:
https://reason.com/2023/04/28/the-very-strange-new-respect-for-authoritarian-democrat-robert-f-kennedy-jr/
https://reason.com/2023/05/17/did-fox-really-fire-tucker-carlson-for-crossing-the-red-line-of-criticizing-big-pharma-as-rfk-jr-claims/
The political classes will be in total meltdown if it’s Kennedy Trump.
I didn't listen to it.
If it was that vaccine mandates were abuse of power then he is right. Nothing to debunk.
If it is that vaccines caused more deaths than covid, then he is clearly wrong. Nothing to debunk either because we do not see people dying massively in Sweden that has good vaccination rates but had less strict lockdown measures. Current excess mortality is probably more related to lockdowns and other measures that prevented people to access healthcare during last 2-3 years.
We know Covid has bad impacts on the heart, say, in some cases causing lasting problems. I don’t see why that shouldn’t also be part of the cause of ongoing excess deaths, especially as much of the excess is specifically in cardiac disease. Of course being sedentary in lockdown can’t have helped either but I don’t see why Covid itself shouldn’t play a part.
That's why I mentioned Sweden.
It was accused not insisting on stronger lockdowns and not closing schools, letting covid spread. At the end it maybe had about the same level of infected people, that is, close to 100% but for the sake of discussion let's assume that the same proportion of infected during pre-vaccine period.
It has no significant excess mortality unlike the UK. Completely destroys the idea that previous covid infections cause excess mortality today.
Fair response, and I didn't know that, I've just looked at the stats. Although it's also notable that a number of countries that did have strict lockdowns have even lower excess mortality than Sweden, so that doesn't really support your case either. Also Sweden is still running at about 3% which you need to explain given it can't be lockdown (which is your whole argument) - it seems to me that sort of amount is likely attributable to Covid, which is about what you'd expect given what we know Covid does to the body including the heart. In fact if you look at cardiac impact, it would be surprising if there *wasn't* some sort of increase. And that's ignoring all the other impacts (lung, brain, etc).
3% increase is meh, something within the margin of error. Yes, it could be covid but that is not anything meaningful.
I'm on medical Twitter, and there was mention of his promoting some false claim about a covid-related issue -- may have been that he was talking up some vaccine-related fear that has been debunked by science. I did not look into the matter and don't know how well the complaint would hold up if you did.
There's all kinds of debunking all over Twitter. There's all kinds of debunking of all things all over Twitter, though -- just goes to show that one side of an argument doesn't tell you anything.
Having two people debate doesn't necessarily tell you anything either, though. Other than which side has hired the better debater.
If you can vote in Ontario, Canada and care about Land Value Tax:
One of the candidates for the Liberal Party of Ontario leadership race, Nate Erskine-Smith, is proposing a Land Value Tax (he calls it a Split-Rate Land Tax) as part of [his housing platform](https://www.meetnate.ca/wp-content/uploads/2023/06/NES_-An-Ambitious-Plan-to-Build-Housing-and-Deliver-Generational-Fairness.pdf). If y'all can vote in Ontario, its free to register as a member of the Liberal Party of Ontario, so you can vote in their leadership election. The membership expires after one year.
To be clear, I have no opinion on any other candidates or their platforms, I just thought it would be useful info since people who read ACX have a tendency to be excited about Land Value Tax and this would be a low effort, high-ish leverage opportunity to advance that cause.
Woah cool! I've listened to a few episodes of Nate's podcast Uncommons and really liked him, and knowing he's a Georgist only doubles that!
For those who object to polygenic screening (using polygenic scores to select embryos for physical and mental wellness during IVF):
If you had 6 embryos you were considering to implant, would it be more ethical to roll a die to select the embryo to implant?
If you were given a polygenic health index and the die landed on the one with the highest health index, would it be ethical to implant that embryo?
What other criteria would be more ethical to use than polygenic scores and why?
I am not against polygenic scoring in principle, but I think that
- they should only be done if IVF is being performed for an unrelated reason anyway, because in my opinion the small increase in risk associated with IVF is not usually outweighed by the benefit of polygenic embryo selection.
- The benefit of selecting from a reasonable number of embryos, such as 6, is usually very small. But of course if the parents are carriers of a Mendelian genetic disease or a very high risk factor such as some BRA1 mutations, then the benefit is potentially very high.
Some commentators have suggested that because our knowledge is currently imperfect, we risk having less healthy babies if we use selection. I think this is highly unlikely, partly because genetic correlations tend to be positive: mutations that increase one risk tend, on average, to increase other risks as well. There are many documented cases of the opposite situation (the famous malaria protection mutation, etc) but... the trend is very clearly in the other direction.
Similarly, the risk of losing genetic variation at the population level is, I think, minuscule. Most variation is deleterious or neutral variation, and worrying about future evolution seems silly to me (not the right timescale!).
One can think of reasons to prefer rolling a die. A die is unbiased. If IVF becomes very commonplace and everyone is using polygenic scores to select embryos, what if we lose a big part of the mental diversity of the human race? What if parents use polygenic screening not for physical and mental wellness, but for personality traits? Parents may want docile children and societies may want obedient citizens, but if we genetically eliminate the "difficult" mavericks from the gene pool, that may not be good for the long-term future of humanity.
I agree that is an important question, but given the context of the original question, it seems more ethical to select a child that will be mentally and physically healthy. There are plenty of collective active problems related to genetic enhancement, but I think parents would be behaving ethically if they decide to improve their child's welfare. If they decide to maximize obedience at the expense of wellbeing, it would be unethical in my view.
What if mental health, as indicated by this polygenic score, is correlated with reduced creativity and reduced potential for scientific genius? Certainly mild autism seems to be correlated with aptitude for the maths and sciences. Also, what if obedience is positively correlated with well-being, as seems likely? An obedient kid who listens to his teachers and never breaks any laws is more likely to be successful than an unruly kid who doesn't do his homework and gets in trouble with the law. But if the whole world is turned into docile sheeple, where would we get the moral visionaries, the political revolutionaries, the titans of new industries (think Elon Musk), and the bold adventurers?
Why isn’t it a woman right to choose, in this instance, anyway? The woman is bringing a child to term, why is the state interfering on what the child could be.
Because the woman isn't the owner of other human beings, and shouldn't get to dictate the future of humanity. Society's interests are implicated every time someone uses technology to selectively alter the future gene pool.
Ok. That’s not all I saying though. I mentioned abortion. Are you against abortions in all cases? If do at least that’s consistent.
I believe the government has the the legitimate right to regulate abortions, which doesn't mean that banning all abortions is good policy. The concerns are also different: abortion bans are usually justified on the basis of the embryo's right to life, while my concern with selective IVF is its effect on the future gene pool.
Abortion is probably effecting the future gene pool to a certain extent. At the moment the technology is nowhere near where it could affect much at all, and if we can eliminate deleterious mutations then the effect could be benign.
I’m not saying the state should decide. I’m asking you to consider what prospective parents should consider, and the question is targeted at those who criticize the use of PGT-P.
In general then I can’t see the logic of a pro abortion position that’s anti PGT-P
Fascinating question, in that even asking in the first place shows a departure from the kind of ethics where this could matter. The internally-consistent position is to refuse to do IVF (as counseled by e.g. the Catholic Church). Beyond any arguments about 'playing God' or 'engaging in eugenics', if you're creating 6 embryos, how many of those are you going to carry to term? All six? If not, what happens to the others? If embryos matter, you already dun goofed. If they don't matter (if you've swallowed the camel) why quibble about screening (straining out the gnat)?
On a macro scale, selecting out known risk genes could have unintended consequences. Since genes tend to be inherited in 'clusters' rather than independently, selecting out a known risk gene will also select out nearby genes that correlate with it. This may be an acceptable tradeoff in the case of true genetic diseases, but becomes much more questionable when you're looking at modest genetic risk factors. There are a number of gene-linked conditions (notably autism, schizophrenia spectrum disorders, bipolar) that are associated with desirable or at least interesting traits in family members of patients and/or in healthy people with risk genes.
There may also be direct unintended consequences of selecting out risk genes for diseases of modern civilization. I'd be especially concerned about any gene involved in immune function because those were major selection targets in the ancestral environment. People with risk genes for e.g. autoimmune disorders may be important reservoirs of resilience against specific kinds of pathogens, which may become more important if we continue falling behind in the antibiotic resistance arms race.
On a micro scale, IVF has risks. Even if you think it should be available to people who can't conceive a healthy child otherwise, you may well think that it's not a good idea for healthy, fertile couples to try to use it to make perfect babies, because on net the risk likely outweighs the reward.
Outside of reproductive ethics, there are reasons to be concerned about social trends that are leading opposite-sex prospective parents to be increasingly unable or unwilling to reproduce naturally. I don't necessarily share these concerns enough to make an argument for them, but for at least a century now, people who are generally positive about technology have been raising questions about the wisdom of a sterile, touchless, isolated society. Asimov's *The Naked Sun* has been on my mind in the COVID era, but there are others.
One might also question the wisdom of encouraging people to condition their willingness to have a child on the child being genetically 'perfect.' We're already at a point in Western cultures where people are afraid to have kids unless they can provide the 'perfect' environment, and it's crashing birth rates despite minimal changes in the number of children people say they want. How much worse would it get if competitive parenting extended to the genetic realm?
I don't think it's possible to be coherently pro-choice and against embryo selection for people who are getting IVF anyway, or IVF+selection for people who are known carriers of a genetic disease and would selectively abort affected embryos if they conceived naturally.
I do think it's possible to be pro-choice and think expanding the use of IVF in healthy hetero couples *for the purpose of* embryo selection is probably not a great thing for humanity.
Transferring an embryo results in live birth about 40-50% of the time so it seems like 6 embryos is about 2.4-3 live children. Doesn't seem unreasonable to me?
Again, you're not thinking like somebody who begins with the assumption that embryos matter. Imagine that you want to adopt an orphan group from some other country. You only want three, so you put in an order for six due to the fact that approximately half of them die on the slow boat over. Your other option is to just not do that and remain childless.
But I believe this is in line with the general fraction of fertilized eggs that become live births? So anyone choosing to have children faces this same moral dilemma.
I can see why you would think that, but in the minds of those who care deeply about this topic, there is an unbridged gulf between you taking it upon yourself to do things that jeopardize human lives versus nature and the hand of the creator doing something analogous.
Interesting. In Poland for instance, due to (what I perceive to be) Catholic sensibilities, unwanted frozen embryos are "adopted" to people who are willing to carry them to term (and raise the baby of course). This law is new and requires a 20-year waiting period so currently it's highly theoretical. But that does not align with the principle you propose, since the less-risky thing to do would be to leave them frozen.
It's okay that you haven't heard this before. You don't have to pretend like I'm theorizing about hypothetical people. Pro Life is a big tent and I specified a certain subset of Catholics who care about the church teaching on the matter. They care because of the general practice involves making far too many and destroying the extra. And I would wager that those in the know could effectively communicate this to their less wonkishly plugged in pro life brethren, for instance, while standing around on the street corner holding signs in just a few words. "They make 10 babies, choose two, and put the rest in the furnace."
The goalposts are over the horizon now, I'd need an ICBM.
Like many, I don’t oppose polygenic screening because it isn’t random; that’s a pretty nonsensical weak- or straw-man of the position, I think: not sure if anybody really thinks that, maybe somebody does.
I do oppose it for two reasons, one fairly weak and entirely pragmatic, one strong and morally rooted. The latter is simple: it requires IVF which involves discarding viable embryos and I think there’s zero excuse for conceiving and discarding a child. In my long-considered opinion there is no more logical place to draw the line against murder than conception.
That aside, since I don’t think that argument will get much traction here, the weaker and pragmatic objection is that we *really* do not know enough about genetics to successfully select and not create systemic errors, and we almost certainly will end up selecting against protective or beneficial traits, and for negative traits, while attempting to select what we want. A sufficiently unsophisticated program, which is what we’d have with our current level of understanding, could end up making the race on average stupider, weaker, or savants (not a good thing!), or prone to other terrifically disadvantageous conditions we don’t even know about. We should really understand more about ourselves first. And once we do, this objection may disappear (may!).
Take CF for example. Aside from the absolutely most typical genotypes that cause it, we have absolutely no clue why certain mutations are expressed phenotypically and certain aren’t. Up until a few years ago it was ‘very clear’ (and the advocacy foundations still insist) that anyone with two CF-causing mutations definitely would have CF. Now it turns out that many ‘CF-causing’ mutations actually don’t: they cause positive sweat tests and aberrant newborn screenings yet no symptoms and no illness at all. What gives? If one used their genotype to determine definitive CF status as we used to, and as some physicians still do, those kids might be selected for destruction… which would have overall weakened the gene pool because they are carrying some sort of unidentified protective mutations as well. And given how very very many of these mutations there are, that could have meant a significant rise in real classical nasty CF, the death-sentence-without-interventions kind. Food for thought.
One other thing: I'm pretty sure that IVF increases the chances of various bad things -- miscarriage, kids having various problems. Though I suppose that it's possible that the actual cause of these things is the problem that made the parents need to resort to IVF in the first place, rather than the procedure itself. But if IVF actually reduces chance that pregnancy and child are healthy, then seems like a terrible idea to use IVF only to make polygenetic screenig possible.
Yes, the health implications of IVF are significant. For obvious reasons it increases the risk of miscarriage, but for as far as I know unknown reasons it profoundly increases the risk of serious birth defects.
Modern sanitation, medicine, welfare policies etc. almost certainly select against "protective or beneficial traits". The argument that artificial selection, which unlike the above is explicitly designed to be beneficial, would be especially harmful to the gene pool seems bizarre to me. The status quo is not "random selection" but, in many respects, negative selection.
I don’t agree at all—there are not many strictly genetic heritable maladies that strongly selected in the past and do not strongly select now, and the evidence is that even terrifically maladaptive genetic disorders which kill almost immediately after birth can and do persist in the genome for many many millennia because they’re recessive. It is very unlikely that the conveniences of modern life are anywhere near as strongly selective as those in the animal world and even if they were it would take millions of years for very visible effects.
On the other hand *directly* selecting every birth each generation based on a common, poor schema could show effects within a few generations.
Evolution doesn't take millions of years. Our whole species is only about 200,000 years old. In the past, probably around half of people died before adulthood (https://ourworldindata.org/child-mortality-in-the-past), while in the developed countries today almost no one does. Moving from the "natural" setting where there is truncation selection of up to 50 percent each generation to the modern setting where there is essentially no selection along this dimension represents a huge change. Over generations, this has probably caused substantial shifts in some phenotypic mean values, depending on heritability.
More subtly, analyses of polygenic scores show that there is negative selection on genetic variants associated with e.g. educational attainment in modern populations: https://www.pnas.org/doi/10.1073/pnas.1612113114 Most traits are like educational attainment, i.e. affected by numerous alleles of small effect, and not due to a single gene. Embryo selection applies positive selection on naturally occurring genetic variation, in contrast to the negative selection that happens otherwise.
Many critics of polygenic screening (e.g. Adam Rutherford, Forzano et al. (2021), Polyakov et al. (2022), etc.) leverage criticism against polygenic screening without proposing alternatives and do not critique the use of visual embryo inspection or couples randomly selecting embryos. This motivates my question: What method do they want couples to use?
This question is framed as couples already deciding to implant and embryo and discard the rest. The embryos will be discarded regardless of the method they use to select the one to implant during IVF.
Even if our understanding of the genetic architecture of humans is not 100%, it seems odd to prefer total randomness to a partial understanding. If randomness is morally better because the health index would lead us astray and create a bad outcome, why would it be morally permissible to allow the bad outcome through randomness? (this is the point I'm getting at with the second question.)
Correct me if I'm wrong, but I *think* the pro-randomness argument is that if there were only going to be one genetically-screened baby in all the world, then you might as well screen based on your best guess, but if there are going to be many babies screened with the same criteria then your selections are now *correlated* across many children, which causes a reduction in genetic diversity.
All else being equal, a more genetically-diverse population is healthier, so in order for screening to be worth doing on a *population* level, it has to provide enough benefit to outweigh the loss of diversity. It's possible it could provide some marginal advantage at an individual level but still be bad at a population level because of diversity concerns.
(In principle, you could do whole-population screening in such a way that it would maintain or even slightly increase genetic diversity (compared to random selection), but that would require massive coordination and would funge against other goals.)
can we define "healthy" here?
Does healthy mean if an epidemic hits you're likely to have some resistant folk in the population?
Or does it mean robust in the *absence* of pathogens?
If some people in your population are heterozygous for sickle cell anemia is your population 'healthier' in your estimation?
Arguably, mutational load has been increasing in humans. Selective pressures haven't been strong enough. We don't talk about it much because eugenics causes mass murder and other problems, and there hasn't been a potentially ethical solution till now.
I think we're a very long way from 100% of the population using embryonic selection. Do you have a different view of embryonic selection if 5% of the population uses embryonic selection? How much value do you place on increasing genetic diversity through other methods?
My model is something like "genetic diversity is a resource that can absorb environmental shocks" or "genetic diversity is a hedge against unknown hazards". If something about the environment changes, then a genetically-diverse population is less likely to be wiped out than a less-genetically-diverse one, because the more-diverse population has more chances to get lucky. Diversity provides option value. (Perhaps "healthier" is not the ideal word for this.)
The words "all else being equal" were doing a lot of work in that claim about genetic diversity being good. This is definitely not the ONLY relevant consideration. I'd contend that many mutations are net harmful even if they increase diversity.
I mostly do not actually buy this argument against polygenic screening and am just steelmanning it in an attempt to improve clarity. I think screening could *in principle* make the population worse off but I'm far from being convinced that this is actually the case for the screening options that are likely to be available in the near-term.
I think that another good point in favor of a randomness mandate/ethical norm is that it avoids designer babies. While I agree that screening for genes highly correlated with debilitating diseases might outweight the loss of genetic diversity, I think that screening for less harmful genes (or even just cosmetic ones) runs much higher risks.
This isn't a slippery slope claim. It's more of a "draw the line in the sand far far away from what the problem region is" kind of claim.
Yes, I’m saying that randomness is actually pragmatically better than selection given our current level of knowledge. The rest follows directly: I’m basically saying your ‘health index’ is no good, irrelevant, certainly misleading.
But again, pragmatically better. I can’t speak for people who think IVF is moral yet embryo selection via polygenic screening is immoral. I don’t understand what ethics produces that result.
If we assume the health index is infallible, the pragmatic objection might disappear. On the other hand it might not: it’s impossible to foresee whether it is actually possible to select beneficially on a population level, and raises all kinds of weird questions about what is and is not a high quality genotype on an individual and population level.
edit: I will note that based on my understanding of Rutherford’s arguments summarized (I could be wrong), you are mischaracterizing him badly. He thinks that screening for polygenic traits is essentially a pseudoscientific swindle, and doesn’t suggest an alternative method because that would be demanding an alternative for hanging a dreamcatcher over the child’s bed: the alternative is you just don’t need to because it’s a dumb idea. He doesn’t think the technology is inherently immoral any more than weaving dreamcatchers is immoral. Selling them to parents while claiming they will prevent their child from getting cancer is the immoral part, with a side of slippery slope to other immoralities like so-called race science.
"Yes, I’m saying that randomness is actually pragmatically better than selection given our current level of knowledge. "
That seems odd as a blanket assertion. It may be that selecting for something controversial like intelligence will select for, say, alcoholism. Or that selecting for height will increase other health issues. Or that selecting for some genes linked to healthy metabolism might decrease lifespan. But some things like sickle cell anemia are well enough understood and so objectively harmful when homozygous that I'm skeptical that we can assert zero benefit. (Granted, we're not talking about polygenic testing in that case. I'm just saying that it's hard to make blanket assertions about understanding.)
If we're only talking about 2-3% of the population undergoing polygenic testing we don't have to be infallible. We just have to do better than random chance. Any outcome chosen by polygenic testing is still a choice that people make, normally, now, via IVF. If nothing else, we'll increase our understanding and refine the technology (if we can agree that having such working technology is eventually desirable.) The primary cost is the cost of the testing.
The benefits to improved metabolic health, especially, seem critical and not strongly ethically objectionable.
I think you’ve answered your own question: selection for strictly Mendelian traits is less pragmatically fraught than our extremely ignorant polygenic screenings.
But, as I said before in my example of CF, it is not clear that things we believe are strictly Mendelian actually are, and the practice of naively selecting against a ‘known-harmful’ mutation may actually lead to a greater prevalence of the exact malady you’re trying to select against!
I have yet to hear any informed counterarguments to this point and I’m beginning to doubt there are any. As Rutherford points out, many advocates think it’s a free lunch because they don’t understand the details—those that do understand the complexities are often counting on us having a much better understanding of our genome at some indefinite future point which will make selection easy and infallible.
Strict Mendelian inheritance isn't needed. It's a simplified model used to teach children, so it's an easy fall-back to clarify concepts. But yes, even many supposedly strict Mendelian traits exhibit incomplete penetrance.
If the case against polygenic screening has been made with any rigor, I'd be interested to see it. It's use for prevention of dibetes melitus already seems to be pretty strong.
"The AUC is a test which ‘describes the inherent validity of diagnostic tests’.4 It is presented as a value between 0.5 and 1.0, with 1.0 representing a perfect test, and 0.5 a test which is no better than chance. If a particular test has an AUC of 0.72, there is a 72% chance that it will be able to correctly identify the subject with disease when applied to select between one afflicted and one healthy participant.4 Generally, scores above 0.9 are deemed ‘outstanding’, 0.8–0.9 is ‘excellent’ and 0.7–0.8 is ‘acceptable’.5
One disease for which PS are providing promising results is type 1 diabetes mellitus (T1DM). According to a 2020 systematic review, the current most powerful PS for T1DM has an AUC of 0.92.6"
https://jme.bmj.com/content/47/12/e91
While granting that diagnosis is not the same as treatment, the argument that they're widely divergent should be made explicitly, not just through incredulity. Can an argument be made that paradoxical effects related to DM screening are of similar frequency and magnitude of harm?
I searched for Rutherford, but didn't find any arguments by him arguing, based on data, that PS was technically worthless, at present. If there is a peer reviewed publication like that could you post a link?
Another reason to oppose it is elitism.
I myself do not object to using polygenetic health indexes to screen embryos. But if I did, I think I'd sidestep your question, which does seem to kind of checkmate those objecting to choosing the embryo with the highest health index. I'd say I objected to using the health index at all, because the value of a human life can't be summed up in a single score. And I'd add that once we start doing that sort of evaluation where do we stop? Do we screen toddlers before they go to nursery school or day care, and have a Perfecto Daycare open only for kids with no significant oddities or deficiencies in measures of motor coordination, vision, hearing, no language delay, no sleep problems, etc etc? (The rest can go to Rando Daycare.)
Later edit: Also I agree with Godoth that we may not know enough yet for these tests to be good guides to which embryo is most likely overall the have the best lifetime health. There are definitely good traits tied genetically to some health problems, and who knows whether that's rare of common. One example: people who have bipolar relatives but do not themselves have bipolar illness are unusually creative, as measured both by tests and life accomplishments. If we screen out bipolar we lose those people. I have a close friend who has a bipolar grandmother, and he is the most entertaining talker I have every met. Get him talking about almost anything and novel, clever thoughts about it just pour out of him. Get him riffing on the subject of something he thinks is ridiculous, and his weird putdowns and analogies and enactments will have you crying with laughter.
>"...the value of a human life can't be summed up in a single score."
I think I get what you're trying to say here, but unless you assert that all lives are equally valuable (a common position that I don't personally hold) the math doesn't work. We may not be able to construct the metric or accurately measure particular individuals' against it, but there must be a latent scalar value if you accept that, e.g., Einstein's life was more valuable than Pol Pot's.
Well as I said I don't personally object on moral grounds to polygenetic screening of embryos. (I do have some practical concerns -- like whether the harm of IVF outweighs the benefits of screening, and whether we know enough at this point to be sure we're not screening out highly desirable traits along with vvulnerability to certain illnesses.). So I'm sort of role-playing here. I have 2 responses -- they're entirely separate arguments. One is what makes you think there is such a thing as a value attached to a particular human life? Does fucking everything have a value number, even if we can't read it? Does the telephone pole outside my window have one? Do works of art have a value number, maybe not readable by us? The amount it sells for does not count. We're talking about artistic value. Does pizza have a value number? Does Rationalism have a value number? Does democracy? Does the planet? Do polio shots? Does the Pope? Who's worth more, Scott or Zvi? Gimme their numbers.
The other, separate argument would be that we can't see very far over the horizon, and events past the point where we are decent forecasters (5-10 years at most, I'd say) might call for qualities you'd never think of selecting for. Maybe there will be a plague way worse than covid, and the people who survive will be mostly those who happen to carry a certain gene. Maybe AI will get supersmart and go rogue, and as part of that influence everybody to go along with its wishes using some drug, to which only people with a certain gene are immune, and they will be able to save the world. Maybe there are wonderful, astonishing things that can be done with supersmart AI, but to be able to mind-meld with them and really absorb what the AI has to give you have to be highly hypnotizable, and the high-hypnotizables will be of huge importance. Or, to take a slightly different angle: A few times in my life I have seen sudden emergencies -- auto accidents and the like -- and most of the time there has been someone who rushes to help, who takes some risk, but not a foolish amount, and is smart about how he helps, who saves some people. I don't know what trait those people have, but I doubt it's correlated very much with intelligence or with low risk for various diseases, and it's a wonderful trait. These brave, decisive and caring people really are special, and in some circumstances having a good number of them on hand might make all the difference. Or here's another thought about a trait that might be important: Maybe we're all in danger of losing some kind of joy or vitality or ability to socialize because we all spend so much more time online than we used to. Maybe all that will get way worse as AI gets better, and some genetic trait that keeps some. people from being vulnerable to it will be crucial. It could even be something superficial, like a tendency to get migraines from staring at a screen for longer than 15 mins or so. OK, in that case, the trait those people carry would be of great value to us a few decades in the future. So the upshot of this argument is that OK, people do have value. We could maybe set it by a calculation of how many people benefit by their existence, and how much. BUT the thing to keep in mind is that high IQ and low risk for various illnesses mightl not be a determinant of value in the world that's coming. These other, odd traits might be. Shit like high IQ, handsomeness and health my be of very little importance in the coming decades
My comment was OT of the whole polygenic issue, so arguments about that are beside the point.
>"Does fucking everything have a value number, even if we can't read it?"
Yes; this is a fundamental result in decision theory (von Neumann-Morgenstern (VNM) utility). If you want to be able to choose between two options and not lose everything you value to a Dutch book, there *must* be some hidden (likely hideously complex to calculate) value attached to each.
I think my answers are also applicable to all situations, not just to considering whether polygenetic screening makes sense.
Regarding my question about whether everything has a value number: "Yes; this is a fundamental result in decision theory (von Neumann-Morgenstern (VNM) utility). If you want to be able to choose between two options and not lose everything you value to a Dutch book, there *must* be some hidden (likely hideously complex to calculate) value attached to each." I do not see how there could be a proof, within decision theory or anywhere else, that everything has a value. I can see how you have to assume everything does in order to use decision theory, or to do various other things, but I cannot see how there could be a logical or mathematical proof that everything *has* a numerical value. If somehow there is one, could you sketch in the bare outlines of it?
My other argument, freed from the polygenetic issue, is just that even if each person has a value to the world which could be expressed as a number (based on how many people they help or harm, and how much), the concept is not useful in practice. You'd only be able to give somebody a number after they're dead, and you asses their impact. And the number somebody ends up with would not be predictable at the time they are born either from genetic measures or behavioral measures of them as a child or from anything else. Because (1) the impact depends on their opportunity to have impact -- probably there have been many people with brains sort of like Einstein's who lived on little tropical islands and just played their wooden flute and stared at the clouds and had deep intuitions about how complex things work. Or they died at age 18. Or they were poor and uneducated. Or they were middle class & educated, but they did not have those jobs at the railroad and patent office that set Einstein's mind on the right track. And also because (2) the qualities that are going to be of most value change, and you can't know in advance what they will be. Even things like a tendency to violence can be good in some circumstances. If somebody had strangled Hitler they would have saved immense humankind immense suffering and loss of life.
>"I cannot see how there could be a logical or mathematical proof that everything *has* a numerical value."
Not in a Platonic sense, but making decisions rationally requires that behavior as if the decision maker is maximizing some value.
>"If somehow there is one, could you sketch in the bare outlines of it?"
No, that's not part of the theorem.
>"My other argument..."
Not what I was addressing, don't care.
I'm personally on thye fence about polygenic screening, but i'm pretty sure that from an anti-screening perspective the only consistent answer is 'YES' to both questions 1 and 2. (It's not gaming the outcome in any way for question 2 as long as you pre-commit to obeying the die roll regardless of result). As for 3, the main objections to screening are based on the potentially deleterious effects of being able to determine essentially any difference in quality between embryos, so the answer is effectively 'NONE'.
It seems odd to me that using randomness could be ethical if it results in the same outcome as using the index. If the unethical nature of screening is not dependent on outcome, then why care if couples select using the index?
A lot of people think intent matters in ethics, but also if you think it's unethical to discriminate against embryos genetically, and EVERYBODY does the dice roll thing, it will produce a more ethical outcome on average, even though 1/6th of those people will randomly choose the same thing as would be chosen by gene screening.
Which kind of has to be true, or you're saying people have to choose AGAINST the healthiest embryo.
The random process, by definition, does not take into account any information the person rolling the die has access to, and therefore using it is functionally equivalent to making the decision with no information. As an intuition pump think of it this way: Say there's a difficult maths problem the answer to which is known to be an integer between 1 and 10. You could work it out the hard way, or just guess and get a 10% chance of picking the right answer. Is it odd that guessing could be considered dumb if it (occasionally) results in the same outcome as working it out properly?
Relatedly, because we're dealing with probability, you need to scale up past n=1 to see the ethical problem. If everyone has to pick randomly between their 6 embryos, they will, on average, get the same quality of offspring as people who conceive the old fashioned way. If people get to screen using meaningful information (polygenic scores or otherwise), a far greater fraction than 1/6 will pick the 'best' embryo by some general, mean-parental-desire quality criterion. This presents 2 potential problems:
1. People who can afford screening start having broadly higher-quality offspring than people who can't, making socioeconomic divides vastly thornier by introducing a caste of people who are genuinely genertically superior.
2. Screening parents turn out to be following some desirability criterion that has negative effects on society when carried out at scale - for example, maybe they disproportionately want children of one sex and society ends up skewed (this isn't just theoretical, there are present-day demographic imbalances in China because female foetuses were considered less sesirable and aborted sex-selectively during the One Child Policy)
So I'm on the fence about this one, but your question assumes away my biggest two concerns:
1) We have a perfect understanding of what genetic factors impact health and wellness, and also how those genes impact everything else (there is a history of overconfidence that gives me pause)
2) That, given the tools for 'health index' screening, the exact same technology and health infrastructure is used for something nefarious (you can think of this as analogous with a law being passed giving the government widespread surveillance powers to stop terrorism, but can easily be used for other purposes such as chilling civil disobedience)
Let’s say we don’t have a perfect understanding: would choosing dice roll be better?
Look at what happened with dogs: we screened them for certain desirable traits, which as a side effect caused all manner of health problems like organs the wrong size for the body, hip joint issues etc. That, and the general shrinking of the gene pool, would be my main concerns.
But for dogs we were screening for an appearance that's desirable in a certain breed. If we had been screening only for health, do you think we'd have ended up with dogs with all these health problems?
If you screen for X, you can get species that is better at X but worse at some other Y you didn't screen for. The question is what could have been the Y if we had instead screened the dogs for health. Maybe intelligence? Friendliness to humans?
If we’re selecting using a health index for mental and physical well-being, the babies will be healthier.
We selected dogs for cuteness and traits that were useful for humans. This degraded the health of dogs in some ways.
"We selected dogs for cuteness and traits that were useful for humans."
You think we wouldn't end up doing that for humans too?
>If we’re selecting using a health index for mental and physical well-being, the babies will be healthier.
Probably the first iteration of such babies will be healthier. I don't think it's possible to rule out that something will go terribly wrong after 10 or 100 successive generations of this.
As an illustration: there's a gene where, if you get one copy, it gives you resistance to malaria, and if you get two, it gives you a greater chance of developing sickle-cell anaemia. It's possible that the gene editors would remove a gene with similar effects, leaving humanity healthier in the short term but more vulnerable to some huge plague down the line.
Maybe!
To build on this, many of the ways I think it could be negative are on a population level (obvious in point 2, less obvious for point 1)
1) the selection could be deleterious over multiple generations in ways we can't hope to understand
2) What might be best for individuals might not be best for society; I've seen a number of good examples above, such as sex selection or proximity to bad things (bipolar), but it's also possible that certain traits that are correlated with personal success and wellbeing involve being a bit of a psychopath/narcissist, or other trait we don't want to select for.
And to preempt your next question: if we could guarantee there would be no materially negative long term consequences, yes, I think it would definitely be fine.
And like I said above, I'm not actually sure about whether any of the above outweighs the positives, but I do think it's a hard question.
> Are ideas discovered or created?
Is this even a meaningful distinction when talking about immaterial things?
With material things, either you go somewhere and find it (someone else already made it), or you take some raw materials or components and build it. We could imagine a few interesting edge cases, like if you find something and then slightly modify it, etc.
What is the distinction in case of ideas? Suppose that in one universe they are discovered, and in another universe they are created. What differences would you expect to see between the universes?
Pascal’s sister claimed that as a kid he was rediscovering Euclid’s elements up to some proposition, having never seen the book (as his dad forbade him from studying math). I would say that this story is plausible in the universe where ideas are discovered but not in the universe where they are invented. The issue likely boils down to definitional quibbling given a certain amount of correlation between the workings of minds. High correlation would justify the ‘discovery’ angle, pretty much like high correlation between individual perceptions justifies the idea that New Zealand is there: all the people who claim to have been to New Zealand had similar experiences, like seeing grazing sheep. Low correlation (everyone having different experiences) could be called ‘invention’. But it’s a spectrum anyway.
There are several ways to conceptualise ideas as points in a pre-existing idea space depending on how you conceptualise ideas. For example, you could have a specifically human idea space consisting of all ideas that humans are capable of thinking of, whose dimensionality corresponds to the dimensionality of human thought. Alternatively, you could have some notion of a well defined idea as equivalent to a structure that is reducible to some basic parts, and your idea space could correspond to all possible configurations of these parts.
Physically speaking, ideas are generated in the brain. It is not literally the case that the brain is located in some abstract idea space where it wanders around to discover something, yet the ideas created by the brain do in some abstract sense pre-exist in the idea space.
So basically ideas are discovered by being created.
> There are several ways to conceptualise ideas as points in a pre-existing idea space depending on how you conceptualise ideas. For example, you could have a specifically human idea space consisting of all ideas that humans are capable of thinking of, whose dimensionality corresponds to the dimensionality of human thought.
But that's an entirely imaginary space. It doesn't tell you anything about how things work in reality.
It is the outcome of physical processes. Brains being similar, ideas ought to be somewhat similar.
Potential.ideas are still only potential.
How about neither, and we say they're combined? New combinations of the same old things.
Let's say they occur somewhere in the middle; an emergent phenomenon at the intersection of the world, and our minds as they attempt to make sense of it.
Seems to me like the kind of false dichotomy that you can avoid by reading phenomenology or cognitive science. Try Scott's review of Surfing Uncertainty:
https://slatestarcodex.com/2017/09/05/book-review-surfing-uncertainty/
What is this and why have you put this here? [EDIT: I am just wondering if you want to be contacted for a specific reason, otherwise it seems odd to put a random phone number on the open thread]
Too much constipation
> Who do you think is more likely to make a fatal error: someone who's operating heavy machinery in unusual, difficult circumstances that require conscious, out-of-the box planning and careful attention to abherant factors; or someone who's operating heavy machinery in normal circumstances, where they can be "in the zone" and not have to think about it?
I think the assumption in the book is that this is a problem that evolution would solve in long term. The hypothetical humans without consciousness would gradually evolve to make fewer mistakes when operating heavy machinery.
We don't have well-calibrated instincts for operating heavy machinery, because it is new for us, from the evolutionary perspective. But for the hypothetical unconscious space-colonizing species, even space travel is instinctive. Evolution shapes them to colonize space better, one death at a time.
I agree that conscious experimentation is an amazing shortcut.
In my opinion, the premise of the book would makes sense in a universe where these is some "peak technology" that can be reached relatively soon (like, we are only a century or two away from it) and then no more amazing progress is possible. Okay, humans get there faster... but maybe the time saved is nothing compared to the time it takes for two space civilizations to meet. When we finally meet, we will both have the same technology, except that they will use it at the speed of instinct.
Then again, it would be possible (this is just my speculation, not in the book) that in the static peak-technology future, humans themselves would gradually evolve to be less conscious; replacing thinking by instinct a little step by little step.
Consciousness prioritizes to prevent hallucination in misleading situations ( since we are decent, law abiding Fristonians, you can hallucinate actions). Suppose you are playing a piece with a pickup into the B section. The natural thing to do would be to come in on the downbeat, so you consciously remember that there's a pickup, and the remembering allows you come in on time. The situation is misleading, because it is most natural to come in on the downbeat, and the potential hallucination is the tempting but wrong performance. Your conscious effort to remember the pickup provides a constraint that your action generating neural net must satisfy. Or, suppose you are testing an information retrieval system built using LLMs. You ask " can old dogs learn new tricks? ", and the most likely answer is no, because that follows an idiom, even though it's false. To prevent hallucination, the system retrieves an article about dog training and puts that in the context window, forcing the token predictor to answer in line with factual information about dog training. The context window is analogous to working memory, the prompt is misleading because of the idiom, and the potential hallucination is to answer " no ".
MIT Press had a great book out last year about working memory and task analysis,"Working Mind" was the title, if you're into this stuff I highly recommend it.
[ insert photo of the charioteer at Delphi]
Having qualia does not equate to being conscious, at least the recursive self-awareness type of consciousness that is the central feature of all those human experts you describe. Animals are like us in that they experience things, they can feel pain, fear, joy, and perhaps love. All this makes them worthy of moral consideration, but it does not make them conscious as we are, because making those experiences the object of cognition is what creates consciousness. This cannot be replicated in animals except for extremely rare exceptions, and no, the mirror test is not one of those. Having a model of one's own body that can be recognized in a different place requires a kind of self-awareness, but it's far removed from the kind that really matters.
A brilliant book and a cool author who is a genius with cool ideas. That said, I too disagree with the main thesis of the book, and some of the arguments for it.
Firstly I think in the appendix the author says that consciousness being maladaptive is just a thought experiment rather than his actual point of view.
Secondly I think you're bundling additional concepts into conciousness here:
> Let's talk about heavy machinery. Who do you think is more likely to make a fatal error: someone who's operating heavy machinery in unusual, difficult circumstances that require conscious, out-of-the box planning and careful attention to abherant factors; or someone who's operating heavy machinery in normal circumstances, where they can be "in the zone" and not have to think about it?
humans can be more deliberate by doing things consciously, it's not clear that having a more accurate and forethoughtful mode of thinking requires consciousness.
From brief research it looks like criteria used to argue for ~non-mammal consciousness seem to be based on the idea that consciousness is required for certain useful behaviours, which is the premiss that Blindsight rejects, so saying that consciousness must be helpful because animals need it to "make motivational tradeoffs" or whatever is begging the question.
With musical performance, I think I have to be careful when playing to not become "too" conscious of what I am playing. There's a part of me that is watching the macro map and making sure I know the coming changes (like a navigator telling a driver if the next turn is left or right), but its also important to not be too aware of the details/particulars of what I'm playing. Aware, but not actively engaged might be how I'd characterize it? Not sure how that fits into the model described. I do know I make more mistakes if I'm just floating along, usually because I don't anticipate an upcoming change. I do a lot of songwriting, so in that context what's happening is often fairly fresh so the muscle memory is just not there yet to allow that effortless floating.