Live data from Hacker News

In Search of AI Psychosis

astralcodexten.com

221–230 of 258 posts

Re: In Search of AI Psychosis

#221

Earlier quoted context omitted.

God you are so smug about something you know essentially nothing about.

I'm lead dev in a start-up that applies coordination dynamics to spatial-syntax. I probably know quite a bit more than you do about what I'm doing.

I don't care what narrow thing you're working on, the brain is simply not as well understood as you think it is.

Re: In Search of AI Psychosis

#222
post #216
post #198

Earlier quoted context omitted.

You've struck at the essential problem of dualism. If thoughts are nonphysical, how can thoughts influence our physical bodies? If consciousness does not interact with the physical world, but merely arises from it, then how can we possibly discuss it, since anything we describe is causally linked to our description of it? Descartes thought the soul was linked to the body through the pineal gland, inspiring a long tra…

I mean it's funny you mention Descartes, because I find the argument that consciousness is the ONLY thing you can really know exists for sure to be pretty compelling. (Descartes then significantly loses the thread, hah.) I agree with you that consciousness is much more fragmented and nonlinear than we perceive it to be, but "I exist" seems pretty tautological to me (for values of "I" that are completely unspecified.)

I definitely share this intuition - it almost, in some sense, feels like the only thing we can really know. It makes it rather tough for me to accept the sibling comments arguing that "actually, the answer is that consciousness is an illusion." That just seems... transparently experientally false, to me.

Re: In Search of AI Psychosis

#223
post #198

Earlier quoted context omitted.

I think my question is more a question of how whether than whether , if that makes sense. There is something about "thought" affecting "matter" that feels spooky if there is a bidirectional relationship. If thought / consciousness / mind is purely downstream of physics, no spookiness. If somehow experienced states of mind can reach back and cause physical effects... that feels harder to explain. It feels like a sort…

You've struck at the essential problem of dualism. If thoughts are nonphysical, how can thoughts influence our physical bodies? If consciousness does not interact with the physical world, but merely arises from it, then how can we possibly discuss it, since anything we describe is causally linked to our description of it? Descartes thought the soul was linked to the body through the pineal gland, inspiring a long tra…

Thank you for the links!

Re: In Search of AI Psychosis

#224
This article is actually quite well written. "Lenin is a mushroom" ties in nicely, and I think it's the biggest factor - that LLMs have somehow ended up with this societal air of authority. I disagree that it stems from the funding and Sam Altman, though. Personally I am not sure what the root of the authority is, but I think it could be the impressive abilities of the LLMs themselves. Package up everything in a box - music, picture and video generation, impressive language skills, et cetera - and people would be fascinated by, and respect the box, even with zero context. Respect and authority for an extraordinary talent is the default case.

Re: In Search of AI Psychosis

#225

Earlier quoted context omitted.

Did he?

Well you said, "The author has a hypothesis and it's looking for evidence" - that's usually the definition of the scientific method - you form some hypoothesis and then try and find evidence to test it. So as far as I can tell, you're describing the scientific method as "bad thinking" and criticizing the author for using it. Which is certainly quite a take.

That's not the definition of the scientific method. That's more like a quote from Interstellar than how science works.

When you encounter an unexplained phenomena (such as an increase in cases of psychosis), you should observe first, then hypothesize on what is the cause.

Re: In Search of AI Psychosis

#226

The description of the risk factors very much jibes with what I have seen in a friend recently. He is quite isolated, and spends most of his evenings writing using AI (he works in a blue-collar trade and wouldn't be typing stuff out by hand usually). He's convinced that he has discovered a grand theory of human connection / relationships / energy / physics, and keeps interrupting in conversation to explain how someth…

Don't get me wrong, but from a relationships perspective, your comment sounds more like a frustrated spoiled ex-girlfriend than a friend. I mean, friends don't go badmouthing their friends on the internet. It's a dick move.

Regarding "theories of everything" and stuff like that. Well, lots of people have those. If I were to call everyone that believes in god or horoscope lone losers, then the asshole would be me, wouldn't it? I know it's different, but also, it's not.

Friendship doesn't require that you buy into the other's vision in order to want them around. That's ideology. Perhaps you misunderstood what friendship means? It's ok, the world is in a weird place right now.

Re: In Search of AI Psychosis

#227
> A large number of Soviet citizens (one estimate puts the number at 11.3 million audience members) took the deadpan "interview" at face value, in spite of the absurd claims presented.

Has no one considered that up to 11.3 million Soviet citizens might have a sense of humor and have been pretending to believe it for fun?

Re: In Search of AI Psychosis

#228

Earlier quoted context omitted.

I'm lead dev in a start-up that applies coordination dynamics to spatial-syntax. I probably know quite a bit more than you do about what I'm doing.

I don't care what narrow thing you're working on, the brain is simply not as well understood as you think it is.

We have a substantial grasp of the allocortex to make significant probes into working memory that relate to semantic memory, episodic thought, emotion and certain senses. The very idea we understand that Sharp Wave Ripples encode this and that various scales of waves integrate to create memories and action is the beginning of bypassing existing computational frameworks.

Re: In Search of AI Psychosis

#229

Earlier quoted context omitted.

It's science challenging CS. If there are no such things as symbols and metaphors in reality, and these are arbitrary, how possibly does it stay referential? It can't. It falls apart, language can't produce itself. It was denoted by Aristotle, the cognitive mapping people Okeefe, Moser, Kandel deciphered this in 1973 and now it feeds from mythological thought into AI "psychosis". Solve the problem at the source units…

Your posts read like someone experiencing LLM psychosis.

Looking at their comment history, it's practically the only thing they ever write, over and over again.

Re: In Search of AI Psychosis

#230

Earlier quoted context omitted.

It's also absurdly wrong, and a quote that only a self-identified rationalist could smugly tout. Of course everyone has world models. Otherwise people would wander into traffic like headless chickens, if they'd even be capable of that. What he likely means is that not everyone explicitly things of possibilities in terms of probabilities that are a function of Bayesian updating. That does not imply the absence of worl…

Yes, everyone has a world model even a toddler has a casual model (“cry → mum comes”).

This is the opposite of a world model and is very much how machine learning works. It's purely correlative without some underlying theory (the world model). The article's (absurd) example of a person not being a mushroom gives a simple example of a "violation" that having a world model about the kinds of organisms that exist would catch. It's not at all pattern matching, nor is it science in any sense about understanding things from first principles, it's about having an understanding of how you imagine things work that you can sanity check against.
Post reply on HN