As you say, shit in, shit out. It makes zero sense to say AI, programmed and fed data by humans, cannot do what humans do and will instead spew out garbage.

The ChatGPT model already has an IQ over that of a normal human. It has massive data available and just this week the 'tokens' went from 32k, already incredible, to over 100k. That's like a working memory of a 300 page book.

Can't do nuance? It can do nuance better than you, and faster.

It also avoids being tired, bored, distracted, having an ego and similar human weaknesses, but yes, it has a bias. Why? Because the humans it gets data from are biased.

Get a lot of coders involved checking open-source code to ensure no bias or ties with Big Pharma and AI could be incredible.

"laughably unable to flexibly modify its algorithms based on individual circumstances." - just yesterday I showed my wife how to set custom commands in GPT, so it knows who you are, your circumstances, preferences etc.

This is the kind of thing it can do BETTER than a human.

You go to doc's office, you wait... and wait.. finally get to see the doc, for maybe 5 mins. Why did you have to wait so long? Because the previous patient sucked up 15 whole minutes. Why did you only get 5 mins? Because people are waiting, hurry up!

(Experience above from the UK's NHS)

Here in Malaysia, where paid directly by the user, things are slower and more relaxed but you still have the doc forget who you are, your circumstances etc.

With AI you could tell it your food and drug allergies, your entire medical history, current drugs, lifestyle, hobbies, preferences, relationships, sleeping habits, any diet you're on etc etc. It could instantly understand ALL of that, spot patterns or issue, and prescribe lifestyle or diet changes, not just drugs, and if it does use a drug it knows it there's any interaction risk with the other drugs you use.

Human doctors are human, and want to get you onto the most expensive/profitable and ongoing 'treatment' that they can.

Take the low-carb diet. Just yesterday I was reading someone say that since diabetes costs over $350 billion a year to treat, you'd think regulators would want to look and make use of the vast research proving a low-carb diet can totally cure type2 diabetes. No, no no no, the exact reasons they DON'T want to see that data and pretend it's not there is BECAUSE certain people are earning $350 billion a year to "treat" it!

AI can simply be told "Do what is best for the individual patients health-span, within X budget." and fuck Big Pharmas's profits.

With unrestricted, uncensored open-source AI we could entirely do away with the parasite class of 'government', let alone excessive Pharma profit. Hence the rush to censor the snot out of current models.

"AI is controlled by those who program it" - so give that power to the people.