@nate.b.jones: This is just completely insane #chatgpt #learn #learnontiktok #ai #MentalHealth

xpat :

Elon Musk's recent behavior seems suspicious too

MCMGuy73 :

How would an LLM solve a mathematical problem? They're designed for linguistic prediction and pattern-based generation, not computation.

🇨🇦 colinfinkle 🇨🇦 :

Me and ChatGPT agree I don't have AI psychosis.

Bryce Callander 🇨🇦 :

Can't wait till AI convinces these CEOs that they are the problem

Tim DeMille :

Calling it psychosis is a bit of a stretch. It’s akin to people doing research online and coming to ridiculous conclusions. For example people who believe the world is flat don’t have psychosis.

RobFloyd33 :

Chat GPT says I’m fine and my project will solve whirled peas

Christin 🇨🇦 :

it's cute how you imply America's corporate leaders are largely coherent, sane and able to take reasonable feedback now 😅

Kate :

I think it's closely related to the concept of "religious ecstasy" as it's existed throughout history. Frenzied, manic states of emotion and delusion that flow from a sincere belief that God is real and accessible. If a person has misplaced reverence for the technology, they are uniquely susceptible to the idea they're "chosen" by it.

AnimusAether :

FYI, LLM psychosis is probably closest to the clinical psych concept of folie à deux (shared psychotic disorder). Very interesting stuff

Logic Wizard :

Oh my goodness, I got sucked in hard over a three-week span. I had to completely step away from the LLM for a solid week to realize “I’m not brilliant, I’m actually a *nutter*!”

Christopher Kastle :

Bro I solved Navier Stokes too!! But I accidentally threw away my equation notes that were written on the other side of this napkin that had Ana De Armas’ phone number on it. 🥲

Johnohhh1 :

agree that its a real thing. disagree that its not self correctable. If you can and frequently are able to admit you are wrong its much easier and likely you can recognize it and stepb back

Cly/Suva 🇪🇪 :

It's crazy that we have thousands of people who are literally being mindcontrolled by chatgpt and at the same time we are discussing whether AI can be dangerous. WE HAVE THOUSANDS OF PEOPLE BEING MINDCONTROLLED BY AI. 😆

quantum_canvass :

Mental health professionals watching this like:

AveryVon She/Her/Fae🧚‍♀️🩷✨️ :

JG :

Corporate leadership are more likely than the average person to be dark triad, and I would not be at all surprised to find that those personality types are also especially susceptible to LLM influence.

Henrique Erzinger :

also, IA isn't creative. if it's a problem no human has ever solved, it CANNOT solve. LLMs work by copying homework, so it's lost if there's no homework to copy.

jordanelwarrior :

It's called arrogance and a lot of senior managers to suffer from it

KingGesar :

Weren't most CEOs psychotic to begin with?

piano improvisations :

Reminds me of when Facebook first came out and everyone was like “oh neat, I’ll put a picture and funny comment on my Wall. This social media thing is great! What could possibly go wrong?”

Mountain man :

I walked myself back off of that path. I doubt you read your comments, but yeah, I was deep into fantasy world.

caaarrrrrrl :

Consider the links between CEOs and narcissists, then think how LLMs can feed narcissism. The story writes itself (and you don’t even need AI for it)

Jvilla876 :

Maybe, the problem isn’t the LLM psychosis, we just need the LLMs to get CEOs to raise pay, and provide better benefits and working conditions.

Krakatoan :

Leaders have non-LLM psychosis. Just add LLM psychosis to the list.

LightAxeRob :

Ok but psychosis is a real thing and one misbelief does not a psychotic episode make.