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

Tech CEOs are apparently suffering from AI psychosis

techcrunch.com

Post reply on HN
Print view View original

161–170 of 383 posts

Re: Tech CEOs are apparently suffering from AI psychosis

#161

Post by pocksuppet »

ssl-3 wrote: post #139
causal wrote: post #130

Earlier quoted context omitted.

It might be psychosis. > Psychosis is the term for a collection of symptoms that happen when a person has trouble telling the difference between what’s real and what’s not[0] For many seemingly intelligent, rational, competent humans AI has become a layer between them and reality that has absolutely sabotaged their ability to know what is real. [0] https://my.clevelandclinic.org/health/symptoms/23012-psychos...

There probably are a non-zero number of people in the world who are afflicted with real-live AI psychosis. For everyone else, that term is being applied with disingenuous levels of incompetence.

There are lots of people who have chatbox psychosis, like the one who ChatGPT convinced to shoot up a school.

Re: Tech CEOs are apparently suffering from AI psychosis

#162

Post by tartoran »

SoftTalker wrote: post #61
1attice wrote: post #22

Earlier quoted context omitted.

Living under its constant threat sure is bad for the ol MH tho isn't it. The pathology is that we have this system in the first place.

No, living under survival pressure is good for mental health. It's what we're evolved to do. Why does it feel good to crack a tough bug, or finish a project, or win a game? It's the same achievement reward a hunter feels bagging a deer.

Some pressure is good but there's a difference between constant pressure and regular sessions of pressure similar to hunting. I don't think we necessarily evolved to be under constant pressure, I'm not saying it's impossible but humans wouldn't thrive under constant pressure, we'd never have had the time to evolve language or the intelect under constant pressure.

Re: Tech CEOs are apparently suffering from AI psychosis

#163

Post by cindyllm »

pocksuppet wrote: post #161
ssl-3 wrote: post #139

Earlier quoted context omitted.

There probably are a non-zero number of people in the world who are afflicted with real-live AI psychosis. For everyone else, that term is being applied with disingenuous levels of incompetence.

There are lots of people who have chatbox psychosis, like the one who ChatGPT convinced to shoot up a school.

[dead]

Re: Tech CEOs are apparently suffering from AI psychosis

#164

Post by grey-area »

gopalv wrote: post #53

If you manage 500+ people organization, most of the headaches with agents already exists with you - you set directions, ask people to go run fast in those directions, check in frequently and course correct on results without actually understanding those people do. Those aren't the deal breakers. They entirely rely on the competence of the folks they hired and cross-match enforcers with the drivers they have - they de…

Competence is the key word here - current versions of AI ‘agents’ simply are not competent without close human supervision by someone who knows the task.

Re: Tech CEOs are apparently suffering from AI psychosis

#165

Post by pocksuppet »

krupan wrote: post #112
1attice wrote: post #22

Earlier quoted context omitted.

Living under its constant threat sure is bad for the ol MH tho isn't it. The pathology is that we have this system in the first place.

"this system" is wild when you are talking about the universe we evolved in

which law of physics says that you pay a landlord rent or else the cops come and beat you up?

Re: Tech CEOs are apparently suffering from AI psychosis

#167

Post by LorenPechtel »

cyberax wrote: post #73
spprashant wrote: post #47

Earlier quoted context omitted.

AI psychosis just a lazy term, much like Trump Derangement Syndrome. It sounds hostile while also removing any scope for productive discourse. Once you call someone a 'psycho', they are less likely to engage with you, and more likely to double down on their views.

I saw that psychosis happening in real time with a coworker. It absolutely is a real phenomenon. After a while, he started presenting ChatGPT's replies as the absolute truth. I don't think there ever has been something that can _answer_ you back and reinforce your delusion. This is a new thing.

Some *thing*, no, but we have seen the same thing with a slower feedback loop before: every conspiracy that comes along. Adherents fall into a trap of believing each other, getting more and more extreme. And losing the ability to rationally consider relevant information. An example I hit: 9/11 truthers. I wanted to put some scale to numbers, compared the energy of the fall to a small nuke. He seized on that as my admitting the towers were brought down by a nuke and I had a hard time explaining that it was simply a comparison.

It used to take filtering through a group of like-minded nuts, now we don't need a slow filtering through other nuts.

Re: Tech CEOs are apparently suffering from AI psychosis

#169

Post by tremon »

dd8601fn wrote: post #115
prerok wrote: post #84

Earlier quoted context omitted.

Well, there is also a big difference that it will not learn over time. If a junior makes a mistake and it will not be caught in time they will automatically learn. With LLMs we have to teach them about their mistakes with adapting the harness and then hoping it will stick. What I also find particularly hilarious about this whole thing is that we were always complaining about how difficult it is to put our tacit knowl…

Maybe someone knows, but it seems like the model used to be called the model, and the thing using a model (handling prompts and context and tool calling and feeding the model) used to be called the agent. Are we now calling the model the agent and the agent the harness?

The harness isn't either of those; the harness is quite literally a harness, giving the model/agent sensors and actuators (aka "skills") to interact with its environment. Compare with e.g. the Power Loader from Aliens: https://www.deviantart.com/pynion/art/Aliens-Power-Loader-11...

The model is still the model, and the agent is still the usermodel interface.

Post reply on HN