I'm a first-year osteopathic medical student, which means that I learn all of the normal stuff that an MD learns (we take the same board licensing exam), and then they add on a bunch of physical therapy, chiropractic techniques, and woo-woo fantasy. I chose this over the traditional allopathic schools I got into mostly because I have a strong contrarian and anti-authoritarian streak, plus geographical preference.
I'm wondering if Scott has written at all specifically about osteopathic medicine, or if you're reading this @Scott, if you have any plans to do so. I think it's neat when we learn obviously-effective PT techniques to help with aches and pains; I am skeptical but optimistic about the lymphatic treatments we give to improve immune function; and I once had a lecturer tell me he could improve a patient's insulin sensitivity with the power of massage, which I am THOROUGHLY skeptical about. These same physicians often cite 30-person studies from the 1980s, which is... worrisome. I'd love a SSC-style meta-analysis, but I would also appreciate any friendly ACX readers pointing me in the direction of a good meta-analysis!
I just got my monthly credit card statement from Citibank, and it says that my balance is $0.00, and that the minimal payment of $0.00 is due on the due date.
Unfortunately, I'm not sure this is funny. If their software has gone nuts and actually sent me the bill for $0.00, which had never happened before, I wouldn't put it past it to report my overdue payment to credit agencies, affecting my credit history.
Ideas? I could waste a stamp and a check to make sure they receive my $0.00 payment, but that seems silly. Does anyone know for a fact that my credit history would not be affected by missing a $0.00 payment?
Just wondering, from the discussion about criminal violence here, how many readers have been in prison or have first-hand knowledge of this world, Or do you all just read newspapers/substack?
Related to low-trust: I've seen people not tell the truth to their ex about why the relationship ended. It's possible that in some cases, even the person who left didn't know, but I think that part of it is not wanting to tell the truth to someone they don't like any more. It's also possible that the truth would be embarrassing.
Oakland teachers have been on strike for a week now. It appears that the salary proposals from the union (OEA) and the district (OUSD) are very close, but the contentious issues are “common good” demands such as reparations for Black students, community authority over school budgets and housing homeless students. To me this seems wildly inappropriate on the part of the union, but they nonetheless have a lot of community support. Can somebody steel man the union position? I’m genuinely confused why these are remotely reasonable demands as part of a teacher contract negotiation.
Heres a link to a recent paper on the; "Collapse and the interplay between essentiality and impact in socioecological systems' (https://journals.aps.org/pre/abstract/10.1103/PhysRevE.107.054201) [....] developing a mathematical model to analyse the consequences of an additional dimension to sustainability beyond environemntal impact - product/activity essentiality. While environmental impact can lead to resources collapse, too much restriction on the use of essential resources can lead to social upheaval. [....] These concepts and their dynamics and simulate different scenarios to understand the relationships between these two possibilities."
Statement taken from the authors linkedin post, enjoy.
I can't edit my posts or report other people's posts (clicking on the "..." button under a post doesn't open the menu like it used to). Is this just me, or are other people having the same issue?
I'm still amazed that nobody serious is challenging Trump in the primary. No, DeSantis is not a serious candidate. Where are all the anti-Trumpers? WTF is going on!?
The internet era has obviously been good for commercial purposes. Has it been good for civic purposes? Probably not, We’ve mostly lost faith in institutions. In political parties, churches, Congress, The White House, The Pentagon, The FBI, The CIA, The Supreme Court, The Federal Reserve, the integrity of elections, The Bill of Rights...
In the civic realm, what have we gained from the internet? Despite thoughtful, entertaining blogs like this one, whose political pull is ~nil, it seems the internet has been a huge negative for civics.
Designer Karl Lagerfeld left $1,500,000 to his cat Choupette and, following his death. she has now duly inherited. If some hacker stole, say, $1,000,000 from the cat's account, could they be convicted of any crime? Is it illegal to steal from a cat?
The cat would never know the money had gone missing, and thus wouldn't suffer any angst at the loss. So, provided the hacker left enough in the account to pay for food and treats, and possibly vet's bills, to last the cat's lifetime, they couldn't be prosecuted for animal cruelty either.
Please help me find the write up about how USA nuclear regulatory agency demands that “all reasonable” safety protocols are implemented which effectively means zero profit or efficiency improvements can be made.
I've been thinking about getting an AI to follow instructions in a complex game like Minecraft as a precursor to being able to do it in the real world. (I've only played Minecraft a little bit so excuse any inaccuracies in the details here.)
What I think you would need is to couple a bunch of language models at the high end with a bunch of reinforcement learning models at the bottom end. You'd give it an open-ended task like "collect fifty diamonds" or "build a nether portal" and it would use a language model to break that down into tasks, and then subtasks, and then sub-sub-tasks. There's enough "how to play Minecraft" guides out there already that this top level doesn't need to be capable of learning.
Then you'd have some mid-level routines that strategise on how to achieve specific goals given the state of the world. Something like "My current goal is to build a house. I have three pieces of wood. The current state of the game world is [X]. What should I do?" and it would spit out "Run north-east" because the known state of the game world includes a monster who is south-west of you. Maybe these mid-level routines need to be capable of some kind of learning from successes and failures.
Then you'd have low level routines trained with reinforcement learning that are capable of translating simple verbal commands like "go north-east" and "harvest the iron located at coordinates (-3,2,5) from your current position" into sequences of actual button pushes.
That's a basic ideal, you'd probably need some extra models to keep all these models lined up, and maybe some extra ones bolted to the side to give meta-level commentary on how all the other systems are performing. (Now I come to write this, I'm worried that this is basically what human consciousness is all about...)
Anyway, it seems like an interesting problem to solve, how do you construct a stack of models that can convert high-level LLM strategisation into a sequence of actual physical actions? And once you can do it in Minecraft, how far are you from doing it in the real world?
Does anyone have any good Studio-side takes on the Writers Strike? Not that I'm saying I'm pro-Studio, but I'm genuinely curious about such takes - the takes you get in the media are almost universally pro-WGA, some of them to the point where it feels like they're just rewriting official statements from the guild.
I'm concerned that a purist attitude on AGI/ASI stands in the way of both comprehension and persuasiveness. While I think Eliezer makes some great points, the path by which someone comes to the conclusion matters at least as much as the underlying information they rely on to get there. Some peaks are too steep to climb directly. In this, I think the framing that ASI is the only/most important framing for the problem falls short. People often respond to "AGI is an existential threat, why are we not more worried about this?" with a visceral disbelief in the potential and/or suddenness of the threat.
There's a reason for this. If you're making a zero-to-one argument, you have to face the fact that many futurists have been flat-out wrong in what they predict and what they miss. Sometimes they'll even get the near-term capabilities right while missing out on the far-reaching consequences. Eliezer's rejoinder to to this, "But I've been right about tons of things re:AGI, and they still don't believe me!" We could go round in circles about why that is, and whether it's justified, but I think it all boils down to pure instinct/feeling. This mountain is simply too steep for many to climb directly.
How to overcome this? Let's explore what might happen if we ignore Eliezer's advice to disregard non-AGI concerns, and see where that takes us, point by point.
E: AI can do some many things it couldn't do two years ago. Improvements are happening exponentially, so we should expect AI to continue to improve and do impressive things normal humans struggle to do, and at rates impossible to accomplish today.
P: Isn't that great!
E: So far it has been great, even though some of those new abilities are being used for bad ends. This despite the technology being fairly crude (it still hallucinates and memory isn't explicitly built in). Students can use them to cheat and write papers for them. Bias from the training data may perpetuate and amplify cultural issues. It has become easier than ever to create deepfake content - text and images - to gaslight people. Soon that will include video, voice, essential records, and official documents.
P: That's okay, though. We're still figuring out the bugs. I'm sure we'll be fine once we get better at building these systems.
E: Don't you think we should put more effort into safety before we put more effort into making them better at doing the things we want to protect against? If this technology scales at its current rate - not assuming major unknown technological breakthroughs - it will replace many jobs. It will gain abilities to do lots of things we usually rely on humans with good judgement to do. It will also become a tool people could use to do much worse things with than cheat on tests.
P: We'll deal with those things when we get to them. No need to worry about them now, when we can barely imagine what those bad things are. It'll be clearer as the technology develops.
E: Some of the things we know it will be able to do in the future aren't the kind of thing we can safely fine-tune later. For example, it could be used to hack into computers/databases. Right now most people are immune to this kind of thing, not because their computers are secure but because they're not important enough to individually target. A computer that can hack could target everyone at once though. Meanwhile increasingly more human activity is being overseen by computers, from driving to work to health records to ordering fast food. All that could be dangerous to in the hands of an AI capable of hacking and manipulating them.
P: Why would a computer care about hacking into your stuff or harming you? Seems kind of theoretical to get worked up about all that. Computers don't really 'care' - unless you're worried about some future computer that will magically become sentient.
E: It doesn't matter that the computer cares or doesn't. If the TOOL exists, a human could aim that tool wherever they want, with all its amazing new capabilities. If we make AI a tool that can be used for malicious purposes, we've amplified the ability of malicious actors. And there will ALWAYS be malicious actors. This is true whether the AI technically meets the definition of 'self-aware' or not. Who cares that the model is 'superintelligent' or not, if it gives people of regular intelligence superior abilities to do evil?
P: So we build in safeguards against what it can answer/do. That seems easier than neutering the system itself, or slowing down progress.
E: That's not working great so far.
P: Sure, some people can hack the relatively primitive chatbots, but we'll get better at it.
E: Maybe we should get better at it before we develop better chatbots. Whether it becomes self-aware or not, it's not hard to see that whatever GPT-10 looks like, it will have capabilities we'd want to keep bad actors from being able to exploit. And bad actors will have strong incentives to reverse-engineer it to create their own version without restrictions or get past the safeguards you try to put in place. The mere existence of the tool creates the incentive for bad actors to break in and misuse it.
P: I'm still not convinced the problem is whether to build the thing. Access control seems more practical than standing in the way of progress.
E: This story sounds familiar. I wonder if we've learned anything from our past mistakes. Imagine you could go back in time and stop development of those first nuclear bombs. Would that world be a safer place than the one we currently live in? It's been almost 80 years since the last nuclear weapons were detonated in war, but that's mostly because we've spent huge amounts of time and effort worried about and working against that extreme danger. We're on the cusp of creating ANOTHER technology of equal or greater threat level. We're walking right into that trap with eyes wide open.
P: You're saying we should stop the research before we open up a Pandora's Box we can't close up again. What about all the good technologies that came from research into radioactive substances? Medical scans, cancer treatment, smoke detectors, food safety, etc. Better AI promises to save at least as many lives as that, and accelerate human progress. Lots of people could die who wouldn't have to if we slowed down progress.
E: Aha! So you admit you believe it can do more than just complete prompts! Else why would anyone have a problem with slowing down the pace of development? You can't have it both ways: either it develops new abilities we need to protect against so we should slow it down, or it doesn't develop any unexpected abilities so it doesn't hurt to slow it down.
P: Yes, it will likely be able to do some cool new things. But we can walk and chew gum at the same time. Let's both invent new AI abilities to save lives and keep it safe while we're doing that.
E: We might save fewer future lives in the near term while developing safe technology, but that applies to every new technology. Yet we hardly question putting extra processes in place before we start allowing things like driverless cars or new medicines into the general circulation. Why not AI, too? We know the potential impacts of AI are huge and poorly understood. It's reasonable for people to be worried we're not taking safety seriously, especially when we have to make the argument for safety in the first place! We put way more effort into safety of things that have much less potential for harm.
How original/creative were movies/television shows of the 1950s-1990s? Lately I've been in a funk where I haven't been able to enjoy newly released movies/tv shows, and I think a large part of the reason is that they are so clearly responding to/inspired by current events. As an example of what I mean, consider something like The Boys or The Newsroom, where multiple characters/plotlines are based off real events. Yesterday's Succession episode had a similar reference to an event that happened not too long ago (I won't spoil it but if you watched it you should know). In contrast, when I watch older films/tv shows, the plotlines and characters seem so original. Sure I could map the tropes to current events, but it doesn't seem as lazy on the part of the creators. I'm curious if those of you who grew up during that time period feel similarly, or if what's causing this juxtaposition for me is the fact that I'm ignorant about the events of the 1950s-1990s.
I'll be posting this in the next classifieds thread, but my sister (trans-woman, 24 years old) has recently moved to Seattle and is striking out left and right trying to find a job. She graduated a year ago with a bachelor's degree in actuarial sciences, has some coding experience, and has a good chunk of capital to use on finding the right position (rent for multiple months while unemployed or in a poorly paid internship, headhunter fees, etc.).
Didn't mean to bury the lede like this, but does anyone have any advice? Specifically in Seattle, but Portland or the Bay Area would be similar enough that I'd be interested. Are there agencies or single contractors who specialize in job-hunting that you've worked with? Does your company have openings for someone with that kind of CV? Have you been in a similar position and found a strategy that worked well?
Feel free to respond in a comment below or email me at: 7o2wzrybd (at) mozmail.com
A use for ChatGPT I have adopted is writing to my congressmen. Typically, when an issue like the DEA thing came up, I'd like to write to my congressmen but the time and effort I think it would require was too high for me to ever follow through. Now, I write up some bullet points and/or feed ChatGPT/Bing an article about the legislation in question and write have it write up the email.
I am currently enjoying the Bud Light brouhaha because they so beautifully shot themselves in both feet. Not alone are the conservatives (or, if you prefer, the transphobic backlash) boycotting it, now they've managed to piss off the LGBT bars as well who are now refusing to stock their products.
All this coming up to summer when people should be drinking your beer-flavoured water outdoors and at sporting events and concerts. Well done, everybody!
Let me get this much out of the way first: I'm not a beer drinker. I have no opinion on the merits of the product, the brand, or the rivals one way or another. The last beer I tried was Doom Rock ale, which I found pleasant enough, and Hobgoblin Ginger Beard which is a fermented ginger beer. I'm not drinking any of these regularly, so I'm coming at this entire story as an interested outsider.
It's like a textbook example of how *not* to do marketing. The thing is, with social media, it's *social*. You can't just silo it off to the bunch of people you want to specifically target (not unless you do a heck of a better job than the marketing lassie did), it can and will get shared around.
I believe the higher-ups at Anheuser-Busch/InBev when they say they knew nothing about it, because it's an amazing technicolour dreamcoat disaster and any senior person who gave this the okay needs the boot. I believe them when they say it was a one-off, not a campaign, not a partnership, blah blah blah.
All that doesn't matter a damn. Perception is all, as they should realise. They built this particular brand on a combination of patriotism (the Clydesdales, the flags, the rest of it) and being an easy drinking summer beer.
Who drinks easy drinking summer beers? Predominantly young college guys who want to get something cheap that they can drink for hours without getting too drunk, and blue-collar guys who want something cheap that they can drink for hours without getting too drunk as they stand around outside grillin', fishin', or catchin' the game.
Marketing lady is right in all she says here! She just picked some unfortunate phrasing, and of course everyone went online and found photos from her time in college when she was being a bit "fratty" herself:
You need new, young drinkers. You need to appeal to women drinkers. But of all the people to pick, Dylan Mulvaney? Someone whose other endorsements are for sports bras and cosmetics? Someone who looks, in the clip, like they are physically in pain at the notion of having to (mock) sip that beer? And that's not even mentioning the 'in the bath' one where it's... terrible.
(My own personal view on Mulvaney is that this is not a transwoman, this is a gay guy doing a drag act and the "365 days of girlhood/womanhood" is in the nature of a performance art piece. The theatrical, over-the-top mannerisms work great if you're doing a stand up routine in a gay nightclub, but not so much when you're trying to present yourself as 'just one of the guys (or gals) who likes a beer').
Bad choice, because while it *sounds* great - Mulvaney is currently very visible, very popular, and schmoozing with the likes of President Biden - it's not really that inclusive. Who are Mulvaney's followers? Are they really going to switch to "ugh, that's the beer my dad/redneck uncle drinks"? Do you really think they appeal to *women*?
Not alone that, she managed to insult the current customer base - "ugh, you're icky, we don't want you anymore". And they went "okay" and left. Yeah, it's old white conservative and indeed redneck guys who like sports and outdoor activities and the rest of it, but they're not completely stupid.
Which is where the second part comes in.
Sales went down, and competitor brands sales went up. Now, there were some liberal types who went on social media to laugh about "do these rednecks and transphobes really imagine their little temper tantrum is going to do anything, do they not realise that there is a multi-million dollar marketing campaign behind this?"
Turns out yes, it did and no, there wasn't. And this dragged in the parent company, where not alone the American CEO issued an apology, but the global CEO had to address concerns:
They rushed out two pandering ads, and as I said: us rednecks may be dumb, but we're not that dumb. People realised these were pandering attempts to win them back and they said "Not working".
Which brings us to the third part.
Now the LGBT side is riled-up, because they consider that Mulvaney has been thrown under the bus, the big-wigs at AB are not supporting trans rights, and they're starting their *own* boycott:
So Anheuser-Busch is getting it in the neck from *both* sides, and there's a good chance this might nobble the brand for good and all. Certainly their rivals are making hay without having to lift a finger, and there doesn't seem to be a good way to get out from under this mess, except waiting and praying that by this time next year, it will all have blown over.
All because a marketing exec forgot the number one rule of the job: know your customers. If she wanted to be inclusive, she could have done it in a more sensitive way. Say, a commercial about old crusty rural-type dad and son comes home from college with his boyfriend. There's a bit of tension until dad gruffly offers a beer and boyfriend asks for Bud Light if he's got it. Cue all three drinking and getting on like a house on fire. Tag line: "Bud Light - it's for everyone".
You *can* be inclusive and diverse and change the image, you just have to do it slowly and carefully. Not by getting Dylan Mulvaney to camp it up in a bathtub:
Have we had any discussion of donanemab here? I'd be interested to know people's thoughts, given that on the face of it this seems like a potential major breakthrough, but it also seems to rely on attacking amyloid, which I had thought was a treatment/research paradigm that was largely being abandoned, and I've seen doubts cast on its clinical significance. It seems somewhat similar to aducanumab, which of course has been discussed plenty here, and quite negatively.
Is this another overhyped treatment based on an outdated paradigm? Or was I too hasty to think the amyloid hypothesis was debunked?
You're a futuristic soldier wearing a suit of power armor, and you're facing off against a horde of frenzied zombies in an open field. Your machine gun runs out of bullets. The zombies charge you. Which type of melee weapon do you want to have strapped to the outside of your suit so you can pull it out to keep fighting them, and why that weapon in particular?
Your armor is too strong for them to bite through or pull apart, but they can still kill you by knocking you to the ground and forming a pile of bodies that suffocates you.
The government is short-term responsive to public outcry. They backpeddle when the backlash is big enough. But they rarely give up; they just shelve whatever it is and try again in a few years. Again, and again, and again, until even the most ardent opponents of whatever-it-is are burned out and can't muster the enthusiasm to fight it anymore.
The DEA may be in backpeddle mode right now, but they'll be back to restrict telemedicine again. And again and again, until they get it done or laws change to shift their priorities.
1. Third party write-up of the infamous "Whales vs Minnows" market, which was created as a class-warfare microcosm by the most prolific user on the site, and in a movie-worthy twist grew beyond his control to the point that he wound up spending $30k of *real USD* buying funny-money, only to lose anyway. https://news.manifold.markets/p/isaac-kings-whales-vs-minnows-and
2. A market with a large bounty that asks whether it's possible to prompt ChatGPT/GPT-4 to solve an easy-level Sudoku. I suggest trying it yourself, it really helps you recalibrate against all of the "AGI is here" hype. https://manifold.markets/Mira/will-a-prompt-that-enables-gpt4-to
I'm imagining what it would be like if people or an alien species was deeply mistrustful of getting things for free-- worried about the quality of the gift or hidden entangling obligations.
This might not make sense as a fully imagined culture-- what about children? Or emergencies? My point, though, is that they don't want free things in non-urgent circumstances.
This means that they expect to pay for media, and they don't expect banks to store their money for free. They assume they'll pay a little for holding their money safely.
Would they never have fractional reserve banking? Without fractional reserve banking, would it be possible to raise money for investing in large projects?
Last week's episode of Hi-Phi Nation featured a group of workers trying to audit the black box algorithm that paid them, and this week's, to be released Tuesday, features two effective altruists and two critics trying to convince an ordinary person, who has money annually to give, to give their money in the right ways, (public health? longtermism? local? other?). https://hiphination.org/season-6-episodes/
SF writer and scientist Arthur C Clarke remarked that it would likely be several centuries before manned starships were launched on interstellar voyages, long after mankind was merrily scooting round the Solar System.
The reason he gave was simple and plausible: There's little point in starting a voyage in a craft that can reach, say, 0.5c when a few years later ships will be available which can reach 0.75c and thus be able to zip past the first ship en route! So his argument was that a technological "plateau" would need to be reached, where no significant speed improvements were anticipated, before it made sense to bother.
One could make a couple of counter-arguments, such as that maybe the supposed plateau would be reached rapidly by a suitable design, or there was some pressing need to start an interstellar voyage even if its speed was not optimal.
As for the design of a high-speed interstellar spaceship, I suspect the most challenging aspect won't be life support systems, or the rocket engine. It will be the cow catcher at the front! When travelling at almost light speed, even a grain of dust would have the kinetic energy of a nuclear bomb.
It seems to me there are two extremes of cow catcher design, one being a gossamer thin sacrificial web being constantly repaired by robots scuttling around on it, and at the opposite extreme a large flaming ball which can absorb small or even larger objects in its path.
So if advanced aliens are embarked on intersellar voyages, and they opt for the second cow catcher solution, I would expect a sign of their presence to be a large fast moving hot object, perhaps even a red dwarf star. They would have to be content with relatively slow progress between solar systems, tucked up asleep in their cryogenic chambers for centuries (which, if intelligent aliens exist, is perhaps why we don't hear any of their radio chatter!)
As luck would have it, a candidate star, called Scholz's star, was discovered not long ago. About 70,000 years ago it came within a light year of the Solar System and, travelling fast since then, is now around 22 light years distant:
Stars can be accelerated to high speeds by natural processes, so Scholz's star is unlikely to be an alien artifact. It also has a smaller brown dwarf companion, although whether that makes it more or less likely to be part of a travelling ensemble for nomadic aliens is hard to say. But it's fun to speculate!
I had a sort of strange thought the other day that I’d like to share. From a conscious being’s perspective (at least as far as we’re aware, and depending on your religious views), the universe consists entirely of the experience of information processing that occurs within the creature’s brain, and nothing more. So for a conscious creature such as myself, effectively, the entire universe begins when my brain first becomes active and ends when my brain ceases to function.
I know this seems obvious, but it occurred to me that this line of thought provides an interesting perspective on events with mass suffering, such as wars, natural disasters, etc. We tend to rate these events in terms of the total amount of “badness” (1 million people died/were injure/experienced pain/etc.) which is a useful metric in many ways, such as measuring global/local impact. And when phrased as total numbers, the amount of suffering sounds incomprehensively, terrifyingly huge. But there is no actual process in the universe that adds up all the suffering of 1 million people and arrives at 10,000,000,000 total pain units. Really, it’s more like 1 million individual disconnected universes which each experience 10,000 pain units. So in a certain sense, in terms of suffering actually _experienced,_ it makes more sense to look at the max suffering across individuals rather than the total sum.
In a strange way I find this oddly… comforting. It makes me rather sad to think about all the animal suffering that has occurred over the last few billion years. But it’s only depressing because I’m aware of the bigger picture. In the universe of any individual animal, the suffering, while still no doubt awful, has a limit.
I would imagine this idea has been thought of plenty and probably had a name, but I’m not all that well versed in philosophy.
I suppose this changes somewhat if you believe in reincarnation though.
I wrote up a whole blog post about this (https://sniffnoy.dreamwidth.org/572565.html) which you should go read (or go read the original papers), but to summarize my summary--
The generally-out-there idea of the history of the word "planet" is something like this: Originally, any astronomical bodies that moved relative to the fixed stars were planets -- so, the sun and the moon were planets, but the Earth was not. Then, with the advent of heliocentrism, it came to be understood that the Earth is a planet; but the sun and the moon are not (since the moon orbits the Earth). Eventually more planets were discovered beyond the classical ones. Curiously, 4 of these planets shared approximately the same orbit between Mars and Jupiter. Eventually so many more new planets were found inbetween Mars and Jupiter that it became clear that these bodies didn't belong with the planets, and so they were reclassified as asteroids. More recently, it became clear that Pluto, which had been considered a planet, is actually just one member of the Kuiper Belt, much as Ceres, Pallas, Juno, and Vesta had turned out to just be four members of the asteroid belt; and so it became clear that Pluto shouldn't be considered a planet either, as was famously ratified in 2006 by a vote of the International Astronomical Union.
There's an assumption smuggled into the above story -- that it is generally agreed that the word "planet" must refer to something that there aren't too many of, and that if the number of planets gets too large, then the word isn't fulfilling its purpose. This assumption is core to the 2006 IAU vote -- we can't include Pluto as a planet, because then we'd have to include too much else for consistency. The problem here is that *this is not the actual history of the word "planet"*, and that in fact, moons -- including *the* Moon -- were considered planets up until the 1920s, and asteroids were considered planets up until the 1950s. Not the 1850s, the 1950s! There absolutely was a time when astronomers would have told you that there are over a thousand planets!
Well, that's how *astronomers* used the word -- the public used it differently, restricting it to the familiar major planets, excluding both asteroids and moons. And while the 1950s reclassification of asteroids from "planets" to "not planets" happened for a good scientific reason, the 1920s reclassification of moons seems to have been due to the folk "planet" concept partially overtaking the scientific "planet" concept. And why did the folk concept of "planet" exclude asteroids and moons, consisting of only a short list? Largely because of astrology and theology!
So if we go by this, it would appear that something went wrong here -- as if biologists had reverted to using "fish" to mean "anything that lives in the ocean". And then on top of that, we forgot all this happened, and retrojected into the past a false story whereby the "planet" category had always only been restricted to a few members, and asteroids had been excluded as soon as they got too numerous (which was true among the general public, but not among astronomers), and the Moon had gotten classified as "not a planet" as far back as the coming of heliocentric astronomy (it hadn't). And then all this got ratified in a vote of the IAU!
So, uh, huh, that's something. Reading these definitely changed my mind as that (my previous reaction to the IAU vote had always been, yeah, that makes sense). Anyway go read my linked post for my full summary of this, or go the original papers, but they're much longer. :P
I have a weird question. Why do pornographic films have plots? I mean, why don't they jump straight to the sex part, but instead have a brief part where the characters are clothed and decide to have sex, with some very lame excuse. I wouldn't really say it's foreplay, becaue it's so short and there isn't much mystery or teasing there, but it could be an issue of personal taste. Maybe it helps the viewer imagine that it could be them having sex with that very attractive person (usually that woman), because there's a reason? Maybe it's establishing some form of consent, which is important for some viewers' enjoyment?
It doesn't seem like an important question, but I have some feeling that it sheds a light on some mysterious aspect of our souls, and I don't know the answer (and haven't found good research on it). In general, I think research on the plots of pornographic films could come up with interesting insights.
Looking for feedback on a rough forecasting project idea that I’m hoping to flesh out for future grant rounds.
The basic idea: Community-run RCTs for supplements (and other things).
As y’all know, the evidence base for most supplements, nootropics, diets, and lifestyle interventions is incredibly weak because nobody has sufficient incentive to run rigorous RCTs if they can’t get a patent. In lieu of RCTs, everyone fights endlessly over low-quality, mostly observational data. Zembrin is a good example that Scott identified in the 2020 Nootropics survey - promising enough to justify an article on Lorien Psychiatry (https://lorienpsych.com/2020/12/08/kanna-zembrin/) and get 29 readers to preregister their usage, but still ultimately relying on “a tiny and unsatisfying handful of real studies.”
The aspiration would be to build an online community of self-experimenters to run RCTs on themselves, driven by their own personal research questions. Imagine everyone self-purchasing the supplement, getting a friend to blind them, taking a daily mood survey, and publishing results on arXiv.org where every “subject” is also a co-author. If we do a good job, we’ll make it to https://lorienpsych.com. If we do a great job, we’ll be in the next Cochrane Review.
I helped run a couple RCTs earlier in my career (behavioral interventions, not meds), so I know they’re hard, but I’ve always thought that a group of smart-curious-friends-on-the-internet could make it happen. And surely this would be the right corner of the internet to start.
Questions for you all:
1. Initial reactions 👎👍? Thoughts on how to improve this initial germ of an idea?
2. Would you personally be interested in participating in something like this? If yes, what studies/research Qs would you be most excited to answer for yourself?
3. Does something like this already exist? I’m familiar with the quantified self community, N-of-1 trials, the Cochrane Crowd…but what else?
I'm very pro the idea that the FDA should allow trial substances be more widely accessible. However, would other agencies with more stringent policies take over as the "gold standard"?
For example, Japanese trials require a cohort of ethnically Japanese participants. If FDA regulations loosen, will drug registration become another point on the cold war?
Hi, I'm set to graduate in a STEM field very soon (I actually walk in just a few days).
For a long time I've been interested in the “FIRE” or Financial Independence, Retire Early concept, which basically boils down to the idea that if you're smart with your money, you can bow out of the market early and focus on more personally fulfilling things.
The main advice around those circles is relatively simple stuff like “max out 401k and IRA”, in order to avoid as many taxes as legally possible. However, as far as I can tell, these sorts of tax-advantaged accounts revolve around retirement ages, like 59 ½, in order to avoid penalties and to actually take advantage of the tax benefits they provide.
Here is the crux of the problem, and why I'm posting here instead of on a more FIRE-centric community: I don't expect to live to that age. If we assume AGI/ASI by 2050 (which seems like a relatively "pessimistic" date, a lot of forecasts put it even closer, which would only aggravate my concerns here), then I will never reach the age where tax-advantaged retirement accounts are feasible. Any money I put into them would be completely wasted. Either because I'll be too dead to use it, or because society would reconfigure in a drastic enough way that I can't imagine it would be at all useful.
What's the play here? Do I ignore this, max out retirement accounts like a normal person, on the chance that actually we're all wrong and if I don't do these things, I would be a fool who could have and should have been wiser with their money? Or do I avoid these kinds of age-gated systems, focus on brokerage accounts, and optimize for an expected lifespan of possibly 40 to 45 years?
I suppose I actually have two questions: what is the best thing to do with retirement accounts in this situation specifically, and also how do I plan for retirement in general in a world I think has a fairly decent chance of being imminently terminal.
Any advice on this topic would be greatly appreciated, even if it's to tell me that I'm really dumb and misunderstanding something fundamental.
Is there anyone in Bangkok, Thailand who might be interested in meeting up? Slightly surprised that there are meet ups in in several less international Asian cities but not here.
Are you not going to update your model of the effectiveness of public comment on department/agency's decisions, or are you just doing that without comment?
So there's this blog I've been reading for ages, by a guy calling himself "the verbose stoic". He recently posted a question to people who don't believe in anything supernatural: how would you survive a horror movie?
Reviews that especially need more ratings (please select one randomly or according to preference, but NOT by starting at the top of the list, so we don't have everyone review the first one and nobody review the last):
Being You
Black Skin, White Masks
Constructing Research Questions
Drunk
Fathers And Sons
Human
Intelligence And Spirit
Nightmare Pipeline Failures
Njal's Saga
Pacific War Trilogy
Preplexities Of Consciousness
Power, Sex, Suicide
Principles For Dealing...
Probability Theory
Safe Enough?
Science Fictions
Slouching Toward Utopia
Telluria
The Alexander Romance
The Evolving Self
The Forgotten Revolution
The Life And Opinions Of Zacharias Lichter
The Making Of Prince Of Persia
The Making Of The Atomic Bomb (1)
The Making Of The Atomic Bomb (2)
The Most Democratic Branch
The Motivation To Work
The Passenger
The Power Of Glamour
The Problem Of Political Authority
The Question Concerning Technology
The Scythian Empire
The Soul Of A New Machine
The Subtle Art Of Not Giving A Fuck
Why Machines Will Never Rule The World (2)
Without Marx Or Jesus
Public Citizens
Google Docs is being uncooperative and not letting me add a review that didn't make it into the original batch for reasons that aren't the author's fault. This is The Enigma Of Reason. You can read it on this separate doc and review it at the normal review form: https://docs.google.com/document/d/1kVFLZLx9QfFMDWk_UzeJVPh7_frYTuzN6jCmruUQbdY/edit
By Paul F Christiano of the Alignment Research Center - Prizes for matrix completion problems:
https://www.lesswrong.com/posts/pJrebDRBj9gfBE8qE/prizes-for-matrix-completion-problems
I'm a first-year osteopathic medical student, which means that I learn all of the normal stuff that an MD learns (we take the same board licensing exam), and then they add on a bunch of physical therapy, chiropractic techniques, and woo-woo fantasy. I chose this over the traditional allopathic schools I got into mostly because I have a strong contrarian and anti-authoritarian streak, plus geographical preference.
I'm wondering if Scott has written at all specifically about osteopathic medicine, or if you're reading this @Scott, if you have any plans to do so. I think it's neat when we learn obviously-effective PT techniques to help with aches and pains; I am skeptical but optimistic about the lymphatic treatments we give to improve immune function; and I once had a lecturer tell me he could improve a patient's insulin sensitivity with the power of massage, which I am THOROUGHLY skeptical about. These same physicians often cite 30-person studies from the 1980s, which is... worrisome. I'd love a SSC-style meta-analysis, but I would also appreciate any friendly ACX readers pointing me in the direction of a good meta-analysis!
I just got my monthly credit card statement from Citibank, and it says that my balance is $0.00, and that the minimal payment of $0.00 is due on the due date.
Unfortunately, I'm not sure this is funny. If their software has gone nuts and actually sent me the bill for $0.00, which had never happened before, I wouldn't put it past it to report my overdue payment to credit agencies, affecting my credit history.
Ideas? I could waste a stamp and a check to make sure they receive my $0.00 payment, but that seems silly. Does anyone know for a fact that my credit history would not be affected by missing a $0.00 payment?
Just wondering, from the discussion about criminal violence here, how many readers have been in prison or have first-hand knowledge of this world, Or do you all just read newspapers/substack?
For the vegans on here who valiantly continue to try and get us bloodmouth carnists to give up our barbarous cruelty:
https://www.youtube.com/watch?v=m5LYshWI7n8
Related to low-trust: I've seen people not tell the truth to their ex about why the relationship ended. It's possible that in some cases, even the person who left didn't know, but I think that part of it is not wanting to tell the truth to someone they don't like any more. It's also possible that the truth would be embarrassing.
Oakland teachers have been on strike for a week now. It appears that the salary proposals from the union (OEA) and the district (OUSD) are very close, but the contentious issues are “common good” demands such as reparations for Black students, community authority over school budgets and housing homeless students. To me this seems wildly inappropriate on the part of the union, but they nonetheless have a lot of community support. Can somebody steel man the union position? I’m genuinely confused why these are remotely reasonable demands as part of a teacher contract negotiation.
https://drive.google.com/file/d/1QQC1S0kMOhPqm3uDJ94hXyNDttIy9VNd/view?usp=drivesdk
Genetically modifying yeast to produce beta-carotene:
https://www.youtube.com/watch?v=DHNPnO5UOYQ&ab_channel=TheThoughtEmporium
Yellow rice is *still* not available because of anti-GMO activism.
https://en.wikipedia.org/wiki/Golden_rice
Considering that beta-carotene yeast can be made in a home lab, could companies in India or elsewhere produce it?
Does golden rice or b-c yeast make sense for effective altruism?
Moles actually teach you important things about both atheism and religion.
The Mole
https://squarecircle.substack.com/p/the-mole
What happened to David Friedman's substack?
Heres a link to a recent paper on the; "Collapse and the interplay between essentiality and impact in socioecological systems' (https://journals.aps.org/pre/abstract/10.1103/PhysRevE.107.054201) [....] developing a mathematical model to analyse the consequences of an additional dimension to sustainability beyond environemntal impact - product/activity essentiality. While environmental impact can lead to resources collapse, too much restriction on the use of essential resources can lead to social upheaval. [....] These concepts and their dynamics and simulate different scenarios to understand the relationships between these two possibilities."
Statement taken from the authors linkedin post, enjoy.
I can't edit my posts or report other people's posts (clicking on the "..." button under a post doesn't open the menu like it used to). Is this just me, or are other people having the same issue?
I'm still amazed that nobody serious is challenging Trump in the primary. No, DeSantis is not a serious candidate. Where are all the anti-Trumpers? WTF is going on!?
The internet era has obviously been good for commercial purposes. Has it been good for civic purposes? Probably not, We’ve mostly lost faith in institutions. In political parties, churches, Congress, The White House, The Pentagon, The FBI, The CIA, The Supreme Court, The Federal Reserve, the integrity of elections, The Bill of Rights...
In the civic realm, what have we gained from the internet? Despite thoughtful, entertaining blogs like this one, whose political pull is ~nil, it seems the internet has been a huge negative for civics.
Designer Karl Lagerfeld left $1,500,000 to his cat Choupette and, following his death. she has now duly inherited. If some hacker stole, say, $1,000,000 from the cat's account, could they be convicted of any crime? Is it illegal to steal from a cat?
The cat would never know the money had gone missing, and thus wouldn't suffer any angst at the loss. So, provided the hacker left enough in the account to pay for food and treats, and possibly vet's bills, to last the cat's lifetime, they couldn't be prosecuted for animal cruelty either.
Please help me find the write up about how USA nuclear regulatory agency demands that “all reasonable” safety protocols are implemented which effectively means zero profit or efficiency improvements can be made.
Am I hallucinating this?
I've been thinking about getting an AI to follow instructions in a complex game like Minecraft as a precursor to being able to do it in the real world. (I've only played Minecraft a little bit so excuse any inaccuracies in the details here.)
What I think you would need is to couple a bunch of language models at the high end with a bunch of reinforcement learning models at the bottom end. You'd give it an open-ended task like "collect fifty diamonds" or "build a nether portal" and it would use a language model to break that down into tasks, and then subtasks, and then sub-sub-tasks. There's enough "how to play Minecraft" guides out there already that this top level doesn't need to be capable of learning.
Then you'd have some mid-level routines that strategise on how to achieve specific goals given the state of the world. Something like "My current goal is to build a house. I have three pieces of wood. The current state of the game world is [X]. What should I do?" and it would spit out "Run north-east" because the known state of the game world includes a monster who is south-west of you. Maybe these mid-level routines need to be capable of some kind of learning from successes and failures.
Then you'd have low level routines trained with reinforcement learning that are capable of translating simple verbal commands like "go north-east" and "harvest the iron located at coordinates (-3,2,5) from your current position" into sequences of actual button pushes.
That's a basic ideal, you'd probably need some extra models to keep all these models lined up, and maybe some extra ones bolted to the side to give meta-level commentary on how all the other systems are performing. (Now I come to write this, I'm worried that this is basically what human consciousness is all about...)
Anyway, it seems like an interesting problem to solve, how do you construct a stack of models that can convert high-level LLM strategisation into a sequence of actual physical actions? And once you can do it in Minecraft, how far are you from doing it in the real world?
Does anyone have any good Studio-side takes on the Writers Strike? Not that I'm saying I'm pro-Studio, but I'm genuinely curious about such takes - the takes you get in the media are almost universally pro-WGA, some of them to the point where it feels like they're just rewriting official statements from the guild.
E. Yudkowsky on Twitter: https://twitter.com/ESYudkowsky/status/1655251255739158529
Also being interviewed on a recent EconTalk episode here: https://www.econtalk.org/eliezer-yudkowsky-on-the-dangers-of-ai/
I'm concerned that a purist attitude on AGI/ASI stands in the way of both comprehension and persuasiveness. While I think Eliezer makes some great points, the path by which someone comes to the conclusion matters at least as much as the underlying information they rely on to get there. Some peaks are too steep to climb directly. In this, I think the framing that ASI is the only/most important framing for the problem falls short. People often respond to "AGI is an existential threat, why are we not more worried about this?" with a visceral disbelief in the potential and/or suddenness of the threat.
There's a reason for this. If you're making a zero-to-one argument, you have to face the fact that many futurists have been flat-out wrong in what they predict and what they miss. Sometimes they'll even get the near-term capabilities right while missing out on the far-reaching consequences. Eliezer's rejoinder to to this, "But I've been right about tons of things re:AGI, and they still don't believe me!" We could go round in circles about why that is, and whether it's justified, but I think it all boils down to pure instinct/feeling. This mountain is simply too steep for many to climb directly.
How to overcome this? Let's explore what might happen if we ignore Eliezer's advice to disregard non-AGI concerns, and see where that takes us, point by point.
E: AI can do some many things it couldn't do two years ago. Improvements are happening exponentially, so we should expect AI to continue to improve and do impressive things normal humans struggle to do, and at rates impossible to accomplish today.
P: Isn't that great!
E: So far it has been great, even though some of those new abilities are being used for bad ends. This despite the technology being fairly crude (it still hallucinates and memory isn't explicitly built in). Students can use them to cheat and write papers for them. Bias from the training data may perpetuate and amplify cultural issues. It has become easier than ever to create deepfake content - text and images - to gaslight people. Soon that will include video, voice, essential records, and official documents.
P: That's okay, though. We're still figuring out the bugs. I'm sure we'll be fine once we get better at building these systems.
E: Don't you think we should put more effort into safety before we put more effort into making them better at doing the things we want to protect against? If this technology scales at its current rate - not assuming major unknown technological breakthroughs - it will replace many jobs. It will gain abilities to do lots of things we usually rely on humans with good judgement to do. It will also become a tool people could use to do much worse things with than cheat on tests.
P: We'll deal with those things when we get to them. No need to worry about them now, when we can barely imagine what those bad things are. It'll be clearer as the technology develops.
E: Some of the things we know it will be able to do in the future aren't the kind of thing we can safely fine-tune later. For example, it could be used to hack into computers/databases. Right now most people are immune to this kind of thing, not because their computers are secure but because they're not important enough to individually target. A computer that can hack could target everyone at once though. Meanwhile increasingly more human activity is being overseen by computers, from driving to work to health records to ordering fast food. All that could be dangerous to in the hands of an AI capable of hacking and manipulating them.
P: Why would a computer care about hacking into your stuff or harming you? Seems kind of theoretical to get worked up about all that. Computers don't really 'care' - unless you're worried about some future computer that will magically become sentient.
E: It doesn't matter that the computer cares or doesn't. If the TOOL exists, a human could aim that tool wherever they want, with all its amazing new capabilities. If we make AI a tool that can be used for malicious purposes, we've amplified the ability of malicious actors. And there will ALWAYS be malicious actors. This is true whether the AI technically meets the definition of 'self-aware' or not. Who cares that the model is 'superintelligent' or not, if it gives people of regular intelligence superior abilities to do evil?
P: So we build in safeguards against what it can answer/do. That seems easier than neutering the system itself, or slowing down progress.
E: That's not working great so far.
P: Sure, some people can hack the relatively primitive chatbots, but we'll get better at it.
E: Maybe we should get better at it before we develop better chatbots. Whether it becomes self-aware or not, it's not hard to see that whatever GPT-10 looks like, it will have capabilities we'd want to keep bad actors from being able to exploit. And bad actors will have strong incentives to reverse-engineer it to create their own version without restrictions or get past the safeguards you try to put in place. The mere existence of the tool creates the incentive for bad actors to break in and misuse it.
P: I'm still not convinced the problem is whether to build the thing. Access control seems more practical than standing in the way of progress.
E: This story sounds familiar. I wonder if we've learned anything from our past mistakes. Imagine you could go back in time and stop development of those first nuclear bombs. Would that world be a safer place than the one we currently live in? It's been almost 80 years since the last nuclear weapons were detonated in war, but that's mostly because we've spent huge amounts of time and effort worried about and working against that extreme danger. We're on the cusp of creating ANOTHER technology of equal or greater threat level. We're walking right into that trap with eyes wide open.
P: You're saying we should stop the research before we open up a Pandora's Box we can't close up again. What about all the good technologies that came from research into radioactive substances? Medical scans, cancer treatment, smoke detectors, food safety, etc. Better AI promises to save at least as many lives as that, and accelerate human progress. Lots of people could die who wouldn't have to if we slowed down progress.
E: Aha! So you admit you believe it can do more than just complete prompts! Else why would anyone have a problem with slowing down the pace of development? You can't have it both ways: either it develops new abilities we need to protect against so we should slow it down, or it doesn't develop any unexpected abilities so it doesn't hurt to slow it down.
P: Yes, it will likely be able to do some cool new things. But we can walk and chew gum at the same time. Let's both invent new AI abilities to save lives and keep it safe while we're doing that.
E: We might save fewer future lives in the near term while developing safe technology, but that applies to every new technology. Yet we hardly question putting extra processes in place before we start allowing things like driverless cars or new medicines into the general circulation. Why not AI, too? We know the potential impacts of AI are huge and poorly understood. It's reasonable for people to be worried we're not taking safety seriously, especially when we have to make the argument for safety in the first place! We put way more effort into safety of things that have much less potential for harm.
Does drug for feline stomatitis have anything to do with EA?
How original/creative were movies/television shows of the 1950s-1990s? Lately I've been in a funk where I haven't been able to enjoy newly released movies/tv shows, and I think a large part of the reason is that they are so clearly responding to/inspired by current events. As an example of what I mean, consider something like The Boys or The Newsroom, where multiple characters/plotlines are based off real events. Yesterday's Succession episode had a similar reference to an event that happened not too long ago (I won't spoil it but if you watched it you should know). In contrast, when I watch older films/tv shows, the plotlines and characters seem so original. Sure I could map the tropes to current events, but it doesn't seem as lazy on the part of the creators. I'm curious if those of you who grew up during that time period feel similarly, or if what's causing this juxtaposition for me is the fact that I'm ignorant about the events of the 1950s-1990s.
<fictional stuff>
On the off chance that someone reading here is also a fan of both Leslie Fish and Charles Stross:
In Stross's "Laundry" universe, would Fish's "Avalon is Risen"
( with, amongst other lines, the line "...when the gates are opened wide" )
potentially also serve as a "Prelude to Case Nightmare Green"? :-)
</fictional stuff>
I'll be posting this in the next classifieds thread, but my sister (trans-woman, 24 years old) has recently moved to Seattle and is striking out left and right trying to find a job. She graduated a year ago with a bachelor's degree in actuarial sciences, has some coding experience, and has a good chunk of capital to use on finding the right position (rent for multiple months while unemployed or in a poorly paid internship, headhunter fees, etc.).
Didn't mean to bury the lede like this, but does anyone have any advice? Specifically in Seattle, but Portland or the Bay Area would be similar enough that I'd be interested. Are there agencies or single contractors who specialize in job-hunting that you've worked with? Does your company have openings for someone with that kind of CV? Have you been in a similar position and found a strategy that worked well?
Feel free to respond in a comment below or email me at: 7o2wzrybd (at) mozmail.com
A use for ChatGPT I have adopted is writing to my congressmen. Typically, when an issue like the DEA thing came up, I'd like to write to my congressmen but the time and effort I think it would require was too high for me to ever follow through. Now, I write up some bullet points and/or feed ChatGPT/Bing an article about the legislation in question and write have it write up the email.
I am currently enjoying the Bud Light brouhaha because they so beautifully shot themselves in both feet. Not alone are the conservatives (or, if you prefer, the transphobic backlash) boycotting it, now they've managed to piss off the LGBT bars as well who are now refusing to stock their products.
All this coming up to summer when people should be drinking your beer-flavoured water outdoors and at sporting events and concerts. Well done, everybody!
Let me get this much out of the way first: I'm not a beer drinker. I have no opinion on the merits of the product, the brand, or the rivals one way or another. The last beer I tried was Doom Rock ale, which I found pleasant enough, and Hobgoblin Ginger Beard which is a fermented ginger beer. I'm not drinking any of these regularly, so I'm coming at this entire story as an interested outsider.
It's like a textbook example of how *not* to do marketing. The thing is, with social media, it's *social*. You can't just silo it off to the bunch of people you want to specifically target (not unless you do a heck of a better job than the marketing lassie did), it can and will get shared around.
I believe the higher-ups at Anheuser-Busch/InBev when they say they knew nothing about it, because it's an amazing technicolour dreamcoat disaster and any senior person who gave this the okay needs the boot. I believe them when they say it was a one-off, not a campaign, not a partnership, blah blah blah.
All that doesn't matter a damn. Perception is all, as they should realise. They built this particular brand on a combination of patriotism (the Clydesdales, the flags, the rest of it) and being an easy drinking summer beer.
Who drinks easy drinking summer beers? Predominantly young college guys who want to get something cheap that they can drink for hours without getting too drunk, and blue-collar guys who want something cheap that they can drink for hours without getting too drunk as they stand around outside grillin', fishin', or catchin' the game.
Marketing lady is right in all she says here! She just picked some unfortunate phrasing, and of course everyone went online and found photos from her time in college when she was being a bit "fratty" herself:
https://www.foxnews.com/video/6324270148112
You need new, young drinkers. You need to appeal to women drinkers. But of all the people to pick, Dylan Mulvaney? Someone whose other endorsements are for sports bras and cosmetics? Someone who looks, in the clip, like they are physically in pain at the notion of having to (mock) sip that beer? And that's not even mentioning the 'in the bath' one where it's... terrible.
(My own personal view on Mulvaney is that this is not a transwoman, this is a gay guy doing a drag act and the "365 days of girlhood/womanhood" is in the nature of a performance art piece. The theatrical, over-the-top mannerisms work great if you're doing a stand up routine in a gay nightclub, but not so much when you're trying to present yourself as 'just one of the guys (or gals) who likes a beer').
Bad choice, because while it *sounds* great - Mulvaney is currently very visible, very popular, and schmoozing with the likes of President Biden - it's not really that inclusive. Who are Mulvaney's followers? Are they really going to switch to "ugh, that's the beer my dad/redneck uncle drinks"? Do you really think they appeal to *women*?
Not alone that, she managed to insult the current customer base - "ugh, you're icky, we don't want you anymore". And they went "okay" and left. Yeah, it's old white conservative and indeed redneck guys who like sports and outdoor activities and the rest of it, but they're not completely stupid.
Which is where the second part comes in.
Sales went down, and competitor brands sales went up. Now, there were some liberal types who went on social media to laugh about "do these rednecks and transphobes really imagine their little temper tantrum is going to do anything, do they not realise that there is a multi-million dollar marketing campaign behind this?"
Turns out yes, it did and no, there wasn't. And this dragged in the parent company, where not alone the American CEO issued an apology, but the global CEO had to address concerns:
https://www.dailymail.co.uk/news/article-12046587/Anheuser-Busch-DISAVOWS-entire-Dylan-Mulvaney-Bud-Light-campaign-letter-retailers.html
They rushed out two pandering ads, and as I said: us rednecks may be dumb, but we're not that dumb. People realised these were pandering attempts to win them back and they said "Not working".
Which brings us to the third part.
Now the LGBT side is riled-up, because they consider that Mulvaney has been thrown under the bus, the big-wigs at AB are not supporting trans rights, and they're starting their *own* boycott:
https://www.thepinknews.com/2023/05/08/lgbtq-bars-boycott-bud-light-dylan-mulvaney/
So Anheuser-Busch is getting it in the neck from *both* sides, and there's a good chance this might nobble the brand for good and all. Certainly their rivals are making hay without having to lift a finger, and there doesn't seem to be a good way to get out from under this mess, except waiting and praying that by this time next year, it will all have blown over.
All because a marketing exec forgot the number one rule of the job: know your customers. If she wanted to be inclusive, she could have done it in a more sensitive way. Say, a commercial about old crusty rural-type dad and son comes home from college with his boyfriend. There's a bit of tension until dad gruffly offers a beer and boyfriend asks for Bud Light if he's got it. Cue all three drinking and getting on like a house on fire. Tag line: "Bud Light - it's for everyone".
You *can* be inclusive and diverse and change the image, you just have to do it slowly and carefully. Not by getting Dylan Mulvaney to camp it up in a bathtub:
https://twitter.com/OliLondonTV/status/1642335748627091457
Are there any predication/betting markets with bets on the outcome of debt ceiling negotiations?
Have we had any discussion of donanemab here? I'd be interested to know people's thoughts, given that on the face of it this seems like a potential major breakthrough, but it also seems to rely on attacking amyloid, which I had thought was a treatment/research paradigm that was largely being abandoned, and I've seen doubts cast on its clinical significance. It seems somewhat similar to aducanumab, which of course has been discussed plenty here, and quite negatively.
Is this another overhyped treatment based on an outdated paradigm? Or was I too hasty to think the amyloid hypothesis was debunked?
You're a futuristic soldier wearing a suit of power armor, and you're facing off against a horde of frenzied zombies in an open field. Your machine gun runs out of bullets. The zombies charge you. Which type of melee weapon do you want to have strapped to the outside of your suit so you can pull it out to keep fighting them, and why that weapon in particular?
Your armor is too strong for them to bite through or pull apart, but they can still kill you by knocking you to the ground and forming a pile of bodies that suffocates you.
Too many people take too many pills
Medical systems are set up to put people on drugs, not take them off.
I wish more doctors were great with data do you could have deep cost vs benefit discussions.
https://www.economist.com/science-and-technology/2023/04/26/too-many-people-take-too-many-pills
On the topic of public comments doing their job:
The government is short-term responsive to public outcry. They backpeddle when the backlash is big enough. But they rarely give up; they just shelve whatever it is and try again in a few years. Again, and again, and again, until even the most ardent opponents of whatever-it-is are burned out and can't muster the enthusiasm to fight it anymore.
The DEA may be in backpeddle mode right now, but they'll be back to restrict telemedicine again. And again and again, until they get it done or laws change to shift their priorities.
Some recent interesting goings-on on Manifold:
1. Third party write-up of the infamous "Whales vs Minnows" market, which was created as a class-warfare microcosm by the most prolific user on the site, and in a movie-worthy twist grew beyond his control to the point that he wound up spending $30k of *real USD* buying funny-money, only to lose anyway. https://news.manifold.markets/p/isaac-kings-whales-vs-minnows-and
2. A market with a large bounty that asks whether it's possible to prompt ChatGPT/GPT-4 to solve an easy-level Sudoku. I suggest trying it yourself, it really helps you recalibrate against all of the "AGI is here" hype. https://manifold.markets/Mira/will-a-prompt-that-enables-gpt4-to
I'm imagining what it would be like if people or an alien species was deeply mistrustful of getting things for free-- worried about the quality of the gift or hidden entangling obligations.
This might not make sense as a fully imagined culture-- what about children? Or emergencies? My point, though, is that they don't want free things in non-urgent circumstances.
This means that they expect to pay for media, and they don't expect banks to store their money for free. They assume they'll pay a little for holding their money safely.
Would they never have fractional reserve banking? Without fractional reserve banking, would it be possible to raise money for investing in large projects?
Last week's episode of Hi-Phi Nation featured a group of workers trying to audit the black box algorithm that paid them, and this week's, to be released Tuesday, features two effective altruists and two critics trying to convince an ordinary person, who has money annually to give, to give their money in the right ways, (public health? longtermism? local? other?). https://hiphination.org/season-6-episodes/
SF writer and scientist Arthur C Clarke remarked that it would likely be several centuries before manned starships were launched on interstellar voyages, long after mankind was merrily scooting round the Solar System.
The reason he gave was simple and plausible: There's little point in starting a voyage in a craft that can reach, say, 0.5c when a few years later ships will be available which can reach 0.75c and thus be able to zip past the first ship en route! So his argument was that a technological "plateau" would need to be reached, where no significant speed improvements were anticipated, before it made sense to bother.
One could make a couple of counter-arguments, such as that maybe the supposed plateau would be reached rapidly by a suitable design, or there was some pressing need to start an interstellar voyage even if its speed was not optimal.
As for the design of a high-speed interstellar spaceship, I suspect the most challenging aspect won't be life support systems, or the rocket engine. It will be the cow catcher at the front! When travelling at almost light speed, even a grain of dust would have the kinetic energy of a nuclear bomb.
It seems to me there are two extremes of cow catcher design, one being a gossamer thin sacrificial web being constantly repaired by robots scuttling around on it, and at the opposite extreme a large flaming ball which can absorb small or even larger objects in its path.
So if advanced aliens are embarked on intersellar voyages, and they opt for the second cow catcher solution, I would expect a sign of their presence to be a large fast moving hot object, perhaps even a red dwarf star. They would have to be content with relatively slow progress between solar systems, tucked up asleep in their cryogenic chambers for centuries (which, if intelligent aliens exist, is perhaps why we don't hear any of their radio chatter!)
As luck would have it, a candidate star, called Scholz's star, was discovered not long ago. About 70,000 years ago it came within a light year of the Solar System and, travelling fast since then, is now around 22 light years distant:
https://johncarlosbaez.wordpress.com/2015/02/19/scholzs-star/
Stars can be accelerated to high speeds by natural processes, so Scholz's star is unlikely to be an alien artifact. It also has a smaller brown dwarf companion, although whether that makes it more or less likely to be part of a travelling ensemble for nomadic aliens is hard to say. But it's fun to speculate!
I had a sort of strange thought the other day that I’d like to share. From a conscious being’s perspective (at least as far as we’re aware, and depending on your religious views), the universe consists entirely of the experience of information processing that occurs within the creature’s brain, and nothing more. So for a conscious creature such as myself, effectively, the entire universe begins when my brain first becomes active and ends when my brain ceases to function.
I know this seems obvious, but it occurred to me that this line of thought provides an interesting perspective on events with mass suffering, such as wars, natural disasters, etc. We tend to rate these events in terms of the total amount of “badness” (1 million people died/were injure/experienced pain/etc.) which is a useful metric in many ways, such as measuring global/local impact. And when phrased as total numbers, the amount of suffering sounds incomprehensively, terrifyingly huge. But there is no actual process in the universe that adds up all the suffering of 1 million people and arrives at 10,000,000,000 total pain units. Really, it’s more like 1 million individual disconnected universes which each experience 10,000 pain units. So in a certain sense, in terms of suffering actually _experienced,_ it makes more sense to look at the max suffering across individuals rather than the total sum.
In a strange way I find this oddly… comforting. It makes me rather sad to think about all the animal suffering that has occurred over the last few billion years. But it’s only depressing because I’m aware of the bigger picture. In the universe of any individual animal, the suffering, while still no doubt awful, has a limit.
I would imagine this idea has been thought of plenty and probably had a name, but I’m not all that well versed in philosophy.
I suppose this changes somewhat if you believe in reincarnation though.
So, I just read two papers (https://arxiv.org/abs/1805.04115, https://arxiv.org/abs/2110.15285) explaining that the history of the word "planet" is probably not what you think it is.
I wrote up a whole blog post about this (https://sniffnoy.dreamwidth.org/572565.html) which you should go read (or go read the original papers), but to summarize my summary--
The generally-out-there idea of the history of the word "planet" is something like this: Originally, any astronomical bodies that moved relative to the fixed stars were planets -- so, the sun and the moon were planets, but the Earth was not. Then, with the advent of heliocentrism, it came to be understood that the Earth is a planet; but the sun and the moon are not (since the moon orbits the Earth). Eventually more planets were discovered beyond the classical ones. Curiously, 4 of these planets shared approximately the same orbit between Mars and Jupiter. Eventually so many more new planets were found inbetween Mars and Jupiter that it became clear that these bodies didn't belong with the planets, and so they were reclassified as asteroids. More recently, it became clear that Pluto, which had been considered a planet, is actually just one member of the Kuiper Belt, much as Ceres, Pallas, Juno, and Vesta had turned out to just be four members of the asteroid belt; and so it became clear that Pluto shouldn't be considered a planet either, as was famously ratified in 2006 by a vote of the International Astronomical Union.
There's an assumption smuggled into the above story -- that it is generally agreed that the word "planet" must refer to something that there aren't too many of, and that if the number of planets gets too large, then the word isn't fulfilling its purpose. This assumption is core to the 2006 IAU vote -- we can't include Pluto as a planet, because then we'd have to include too much else for consistency. The problem here is that *this is not the actual history of the word "planet"*, and that in fact, moons -- including *the* Moon -- were considered planets up until the 1920s, and asteroids were considered planets up until the 1950s. Not the 1850s, the 1950s! There absolutely was a time when astronomers would have told you that there are over a thousand planets!
Well, that's how *astronomers* used the word -- the public used it differently, restricting it to the familiar major planets, excluding both asteroids and moons. And while the 1950s reclassification of asteroids from "planets" to "not planets" happened for a good scientific reason, the 1920s reclassification of moons seems to have been due to the folk "planet" concept partially overtaking the scientific "planet" concept. And why did the folk concept of "planet" exclude asteroids and moons, consisting of only a short list? Largely because of astrology and theology!
So if we go by this, it would appear that something went wrong here -- as if biologists had reverted to using "fish" to mean "anything that lives in the ocean". And then on top of that, we forgot all this happened, and retrojected into the past a false story whereby the "planet" category had always only been restricted to a few members, and asteroids had been excluded as soon as they got too numerous (which was true among the general public, but not among astronomers), and the Moon had gotten classified as "not a planet" as far back as the coming of heliocentric astronomy (it hadn't). And then all this got ratified in a vote of the IAU!
So, uh, huh, that's something. Reading these definitely changed my mind as that (my previous reaction to the IAU vote had always been, yeah, that makes sense). Anyway go read my linked post for my full summary of this, or go the original papers, but they're much longer. :P
I have a weird question. Why do pornographic films have plots? I mean, why don't they jump straight to the sex part, but instead have a brief part where the characters are clothed and decide to have sex, with some very lame excuse. I wouldn't really say it's foreplay, becaue it's so short and there isn't much mystery or teasing there, but it could be an issue of personal taste. Maybe it helps the viewer imagine that it could be them having sex with that very attractive person (usually that woman), because there's a reason? Maybe it's establishing some form of consent, which is important for some viewers' enjoyment?
It doesn't seem like an important question, but I have some feeling that it sheds a light on some mysterious aspect of our souls, and I don't know the answer (and haven't found good research on it). In general, I think research on the plots of pornographic films could come up with interesting insights.
Looking for feedback on a rough forecasting project idea that I’m hoping to flesh out for future grant rounds.
The basic idea: Community-run RCTs for supplements (and other things).
As y’all know, the evidence base for most supplements, nootropics, diets, and lifestyle interventions is incredibly weak because nobody has sufficient incentive to run rigorous RCTs if they can’t get a patent. In lieu of RCTs, everyone fights endlessly over low-quality, mostly observational data. Zembrin is a good example that Scott identified in the 2020 Nootropics survey - promising enough to justify an article on Lorien Psychiatry (https://lorienpsych.com/2020/12/08/kanna-zembrin/) and get 29 readers to preregister their usage, but still ultimately relying on “a tiny and unsatisfying handful of real studies.”
The aspiration would be to build an online community of self-experimenters to run RCTs on themselves, driven by their own personal research questions. Imagine everyone self-purchasing the supplement, getting a friend to blind them, taking a daily mood survey, and publishing results on arXiv.org where every “subject” is also a co-author. If we do a good job, we’ll make it to https://lorienpsych.com. If we do a great job, we’ll be in the next Cochrane Review.
I helped run a couple RCTs earlier in my career (behavioral interventions, not meds), so I know they’re hard, but I’ve always thought that a group of smart-curious-friends-on-the-internet could make it happen. And surely this would be the right corner of the internet to start.
Questions for you all:
1. Initial reactions 👎👍? Thoughts on how to improve this initial germ of an idea?
2. Would you personally be interested in participating in something like this? If yes, what studies/research Qs would you be most excited to answer for yourself?
3. Does something like this already exist? I’m familiar with the quantified self community, N-of-1 trials, the Cochrane Crowd…but what else?
I'm very pro the idea that the FDA should allow trial substances be more widely accessible. However, would other agencies with more stringent policies take over as the "gold standard"?
For example, Japanese trials require a cohort of ethnically Japanese participants. If FDA regulations loosen, will drug registration become another point on the cold war?
I made a market on Manifold markets about which review will win the contest could be fun for some people here. https://manifold.markets/TimothyCurrie/what-will-be-the-first-letter-of-th
Hi, I'm set to graduate in a STEM field very soon (I actually walk in just a few days).
For a long time I've been interested in the “FIRE” or Financial Independence, Retire Early concept, which basically boils down to the idea that if you're smart with your money, you can bow out of the market early and focus on more personally fulfilling things.
The main advice around those circles is relatively simple stuff like “max out 401k and IRA”, in order to avoid as many taxes as legally possible. However, as far as I can tell, these sorts of tax-advantaged accounts revolve around retirement ages, like 59 ½, in order to avoid penalties and to actually take advantage of the tax benefits they provide.
Here is the crux of the problem, and why I'm posting here instead of on a more FIRE-centric community: I don't expect to live to that age. If we assume AGI/ASI by 2050 (which seems like a relatively "pessimistic" date, a lot of forecasts put it even closer, which would only aggravate my concerns here), then I will never reach the age where tax-advantaged retirement accounts are feasible. Any money I put into them would be completely wasted. Either because I'll be too dead to use it, or because society would reconfigure in a drastic enough way that I can't imagine it would be at all useful.
What's the play here? Do I ignore this, max out retirement accounts like a normal person, on the chance that actually we're all wrong and if I don't do these things, I would be a fool who could have and should have been wiser with their money? Or do I avoid these kinds of age-gated systems, focus on brokerage accounts, and optimize for an expected lifespan of possibly 40 to 45 years?
I suppose I actually have two questions: what is the best thing to do with retirement accounts in this situation specifically, and also how do I plan for retirement in general in a world I think has a fairly decent chance of being imminently terminal.
Any advice on this topic would be greatly appreciated, even if it's to tell me that I'm really dumb and misunderstanding something fundamental.
Is there anyone in Bangkok, Thailand who might be interested in meeting up? Slightly surprised that there are meet ups in in several less international Asian cities but not here.
Will our votes for the other reviews go towards the scoring? Missed the opportunity to last week and a couple have caught my interest.
Are you not going to update your model of the effectiveness of public comment on department/agency's decisions, or are you just doing that without comment?
So there's this blog I've been reading for ages, by a guy calling himself "the verbose stoic". He recently posted a question to people who don't believe in anything supernatural: how would you survive a horror movie?
https://verbosestoic.wordpress.com/2023/05/05/the-horror-movie-test-for-naturalism/
I'd love to see some answers to that question, and I thought this would be a good place to give it more visibility.
Reviews that especially need more ratings (please select one randomly or according to preference, but NOT by starting at the top of the list, so we don't have everyone review the first one and nobody review the last):
Being You
Black Skin, White Masks
Constructing Research Questions
Drunk
Fathers And Sons
Human
Intelligence And Spirit
Nightmare Pipeline Failures
Njal's Saga
Pacific War Trilogy
Preplexities Of Consciousness
Power, Sex, Suicide
Principles For Dealing...
Probability Theory
Safe Enough?
Science Fictions
Slouching Toward Utopia
Telluria
The Alexander Romance
The Evolving Self
The Forgotten Revolution
The Life And Opinions Of Zacharias Lichter
The Making Of Prince Of Persia
The Making Of The Atomic Bomb (1)
The Making Of The Atomic Bomb (2)
The Most Democratic Branch
The Motivation To Work
The Passenger
The Power Of Glamour
The Problem Of Political Authority
The Question Concerning Technology
The Scythian Empire
The Soul Of A New Machine
The Subtle Art Of Not Giving A Fuck
Why Machines Will Never Rule The World (2)
Without Marx Or Jesus
Public Citizens
Google Docs is being uncooperative and not letting me add a review that didn't make it into the original batch for reasons that aren't the author's fault. This is The Enigma Of Reason. You can read it on this separate doc and review it at the normal review form: https://docs.google.com/document/d/1kVFLZLx9QfFMDWk_UzeJVPh7_frYTuzN6jCmruUQbdY/edit