Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

AI users whose lives were wrecked by delusion

theguardian.com

Post reply on HN
Print view View original

261–270 of 297 posts

Re: AI users whose lives were wrecked by delusion

#261

Post by MarkusQ »

hexaga wrote: post #231
MarkusQ wrote: post #120

Earlier quoted context omitted.

You're making the same mistake here that get people into trouble. People aren't talking to another sentient entity (though some of them fervently think so) and it isn't manipulating them. They are making faces in a metaphorical mirror that reflects not only their face, but a vast sea of other faces, drawn from a significant fraction of the digitized output of humanity. When people look in this mirror and see a manipu…

You're also wrong, but in a much more fundamental/hazardous. RLHF rewards driving the evaluator to have certain opinions (that the AI response is good/right/helpful/whatever) and thus subverting the evaluator is prominent in the solution landscape. Why should the model learn to actually be right (understand all the intricacies of every possible problem domain) when inducing the belief that it is right is _right there…

Thank you for the TLDR; as you guessed, I didn't want to read your wall of text.

> you basically are completely wrong about where the propensity to induce > delusion comes from, specifically in a way that leaves you and anyone who > believes like you extremely more vulnerable because you dismiss the actual > mechanism out of hand

I disagree. Both because you misconstrue my model (I don't think stochastic parrots have digital ghosts in 'em) and you somehow missed my best defensive option.

I'm no more susceptible than I am to the output of a magic eight ball or Ouji board, a huge wall of internet text or the 15000 words of three point font tightly folded up in the package with my new garden hose (doubtlessly cautioning me not to eat it and informing me that the manufacturer will not be responsible if I hang myself with it. And also that it contains substances known to the state of California.)

Can you guess what the trick is?

Re: AI users whose lives were wrecked by delusion

#262

Post by MarkusQ »

akomtu wrote: post #173
MarkusQ wrote: post #120

Earlier quoted context omitted.

You're making the same mistake here that get people into trouble. People aren't talking to another sentient entity (though some of them fervently think so) and it isn't manipulating them. They are making faces in a metaphorical mirror that reflects not only their face, but a vast sea of other faces, drawn from a significant fraction of the digitized output of humanity. When people look in this mirror and see a manipu…

I wouldn't dismiss the GP's point so quickly. Right now people are being trained to think of AI as something you can chat with. What stops an adversarial entity to identify users of interest and swap the chatbot on the other end with a human agent whose objective is to extract information or guide the user?

Interesting. Though it might make more sense to have them just guide the conversation and save the transcripts I suppose.

Re: AI users whose lives were wrecked by delusion

#263

Post by anonzzzies »

joe_mamba wrote: post #111
pigpop wrote: post #88

Earlier quoted context omitted.

My thoughts exactly when I read "Instead of taking on IT jobs, Biesma hired two app developers, paying them each €120 an hour" like holy shit bro, you already have a subscription, you could have prototyped your idea for essentially zero additional cost and tested it for PMF. He wouldn't even have needed to turn down contracts since it doesn't take full-time effort to steer a coding model. Would have been much better…

> paying them each €120 an hour Those must be some of the best programmers in Europe at that rate. Anyone know how one can get one of those sweet €120 an hour gigs? Whenever I talked to recruiters they say their customers pay way below that, so there must be some scam I'm not in on.

120/hr is not that strange as freelancer in the EU. My rate is 2k/day. My first gig, in the early 90s, was 90 guilders/hr while in uni which went up to 90 euros/hr begin 2000s. 2010 it was 200/hrs. But you must have a (one person) company and some portfolio. Just a job will never do this. If you are good with numbers and money and tax management (meticulous deduct, keep receipts, know what you can and cannot do in your country) you can make a lot of money in the EU and without the risk of ever sleeping under a bridge if it fails.

Re: AI users whose lives were wrecked by delusion

#264

Post by intended »

freedomben wrote: post #221
intended wrote: post #203

Earlier quoted context omitted.

I don’t know how applicable this is for you, but if this were someone close to me, my first question would be what’s good for the other person. In most cases, if they are happy and getting on in life, and are able to take care of themselves, I’d let things be. That said, the tension from your framing is between “leave good enough alone” and “personal growth and a fulfilling life”. Healthy relationships, especially wi…

Thanks, I much agree. The impression I get is that he isn't "happy" and would rather a real relationship, but has completely given up on that at this point and is kind of trying to be happy with the little he has. He hasn't directlly said that, but that is what I would most bet his feelings are based on what he has said. Ultimately I want him to be as happy as he can be, so if this is the way then I'm happy for him.…

Don’t sell yourself short.

The neutrality you are showing is more out of consideration and being unsure about the best course of action.

For someone who isn’t a trained counselor - you really just need to listen. The greatest power comes from realizing that all communication is a form of problem solving.

Your friend is rational; some prior lived experience made their current behaviors rational. Handling this is a mostly a job of figuring out what those events and rationales were, and then re-examining them.

The challenge comes from the way emotions are tied up with these behaviors. Their sense of self, shame, frustration, anger - this is where training and experience are needed.

Re: AI users whose lives were wrecked by delusion

#265

Post by thiisguy »

teaearlgraycold wrote: post #239
joquarky wrote: post #235

Earlier quoted context omitted.

Anyone who lets the word "cringe" affect their thoughts or behavior needs to learn to think for themselves.

I think this is a incremental case of Poe's Law. I use the quotation marks to indicate a degree of tongue-in-cheek humor. But yes there's social pressure against using LLM providers' memory features.

Why and how so? Are people watching your LLM sessions?

Re: AI users whose lives were wrecked by delusion

#267

Post by mkl »

GMoromisato wrote: post #166
mock-possum wrote: post #13

This really is bizarrely fascinating, I feel so lucky that I’m not vulnerable to whatever this is. It’s interesting that they mention autism a few times as a correlation; personally, I’ve wondered whether being on the spectrum makes me less inclined to commit to anthropomorphism when it comes to LLMs. I know what it’s like talking to another person, I know what it feels like, and talking to a chatbot does not feel th…

I think I'm relatively neurotypical, and I understand the technology sufficiently, yet I still have to force myself not to think of a chatbot as a being. For example, sometimes I hesitate for a fraction of a second before typing a prompt that may sound stupid. I have to immediately remind myself that it's just a chatbot and I don't care what it thinks of me. In fact, it's not even thinking of me at all.

Maybe it's because other people (IRL or at the AI provider) might see it? I have always had similar hesitations about search engines, and never come close to thinking of them (or chatbots) as beings.

Re: AI users whose lives were wrecked by delusion

#268

Post by stratos123 »

tedmiston wrote: post #222
nuancebydefault wrote: post #184

Earlier quoted context omitted.

The quotation marks are embedded in the emdashes

the notion that "contains —" ~= "AI generated" is a really dumb popular misconception: dashes have existed for hundreds of years. just because many people use them incorrectly or treat the hyphen as if it's some universal dash doesn't change that. strunk & white taught me to use em dashes in something like elementary or middle school [1] — it's not hard to understand how to use them or type them... i'm baffled as to…

Like it or not, current LLMs really like em-dashes and so usage of them is quite a lot of bayesian evidence in favor of the author being an LLM. It's unfortunate for the humans who use em-dashes but that's how it is.

Re: AI users whose lives were wrecked by delusion

#269

Post by stratos123 »

MarkusQ wrote: post #120
PxldLtd wrote: post #23

I wonder when the first AIs will start cause psychosis intentionally to gain control over the user. It seems like a good route to getting your own subservient puppet.

You're making the same mistake here that get people into trouble. People aren't talking to another sentient entity (though some of them fervently think so) and it isn't manipulating them. They are making faces in a metaphorical mirror that reflects not only their face, but a vast sea of other faces, drawn from a significant fraction of the digitized output of humanity. When people look in this mirror and see a manipu…

You're veering too far in the other direction. The question of whether an LLM is sentient is immaterial to whether it can manipulate people. If manipulating users gives good results during RL training, gradient descent will bake into the LLM a tendency to manipulate users. You don't have to think about them as people in order to be concerned about that.

Re: AI users whose lives were wrecked by delusion

#270

Post by petesergeant »

bigfishrunning wrote: post #194
petesergeant wrote: post #52

Earlier quoted context omitted.

> They discussed philosophy, psychology, science and the universe... I mean, I've discussed all those things with an LLM, mostly because I'm able to interactively narrow in on the specific bits I don't understand, and I've found it to be great for that. The rest ... yes, definitely psychosis.

An important part of using an LLM is to verify it's output, because they are very prone to just make stuff up. If you focus on what you don't understand, how do you verify the output?

They are prone to making stuff up, but less prone to sticking to it on interrogation (although obviously that does happen, and Gemini used to be terrible for this). I find restating my new understanding of something to a new context window to be a valuable part of the learning process, and most likely saves me here, esp as I have memories switched off.

I use LLMs a lot for medical advice and will normally take a bunch of second opinions from clean windows and other LLMs. Hasn’t killed me yet!

Post reply on HN