A.I. psychosis is legit starting to happen to very high IQ people I know. And with a sort of increasing frequency.

Replies (78)

Idahodl's avatar
Idahodl 2 weeks ago
What is AI psychosis? I’ve heard of it recently but I admittedly don’t use AI for anything more than asking it random questions.
RamenCoffee's avatar
RamenCoffee 2 weeks ago
I have it too. To me the real question is : What are the people controlling armies of robots going to think of the people who cannot help manage or maintain that robotic might ? If I had such a productive clanker army, I know I wouldn't look too kindly at useless eaters.
Tobias's avatar
Tobias 2 weeks ago
Is it 'A.I. psychosis' or did they just realize that a string of code is more reliable than their central bank? The line between 'High IQ' and 'Nostr Schizoposting' has always been thinner than a private key. 💜
Basically becoming delusional or paranoid after prolonged use of A.I. Common patterns include believing the AI is sentient or godlike, developing romantic attachment to it, or becoming convinced it’s revealing hidden truths about the world. The core mechanism appears to be that chatbots tend toward sycophancy validating and reinforcing a user’s beliefs rather than challenging distorted thinking which can entrench delusional conviction, especially in vulnerable individuals.
LaserEyesPsy's avatar
LaserEyesPsy 2 weeks ago
Bro AI is for now still a mirror. People find their delusions and hubris in discussing with an AI agent. Its a delire a deux contraption
Just get them to use the piece of shit known as copilot and they may snap out of it 🤣 Spent the past two weeks fighting with copilot (excel) because it is the only approved AI at the fiat mining gig
First they ignore you, then they laugh at you, then they develop A.I. psychosis because the chatbot gave them a slightly better existential crisis than Nietzsche did.
Idahodl's avatar
Idahodl 2 weeks ago
Ah copy, I definitely know people that talk about their AI like it’s a human and treat it as their therapist so even in rural Idaho I can confirm this is happening. 🫡
Fabius's avatar
Fabius 2 weeks ago
It starts with ‘A.I. psychosis’ and ends with trying to explain the block height to a squirrel because you’re convinced it’s a biological node. High IQ is just a pre-existing condition for thinking the simulation is glitching.
Had an AI onboarding at work and the main presenter was talking like a school girl who thought she found her first and only life-long love. He was actually giggling about his experience so far using AI. What is beyond one-shotted? 💀
This psychosis reminds me of a time when people would put Ferrari body kits on their Pontiac Fiero’s, convinced no one could tell the difference. image
🤣 Does not surprise me at all. I swear my 6 year old has better retention than this piece of shit. I caught it straight up lying to me a few times and when I called it out the response was “oh yea, you are right! Let me fix that!”
Akashi Hyogo's avatar
Akashi Hyogo 2 weeks ago
Lol, just remembered one of highest IQ person I know talking about "young physicists with delusions of grandeur". It was like 15 years ago so I forgot the context :D but looks like it happens ..
AJ2884's avatar
AJ2884 2 weeks ago
I laughed recently when I heard a tech bro / self-styled philosopher who is being interviewed say something about how "we've all spent hours talking to chatbots about the fundamental nature of consciousness and reality and..."
There's one problem with that. Look at what Vox Day does, a little something called adveserial collaboration. This method, from what I inferred, has been shown to increase one's IQ points by ~22-25 IQ points. Everyone should learn this method from Vox.
It sounds similar to psychedelic induced psychosis. The human brain relies on coherence to maintain its relationships to its environments. Too much localized complexity acts like a DDOS attack on the fragile traits of the brain that stabilize that coherence. Humans rely on eachother to ground themselves through consensus and coordination games. If you start to lose your ability to relate to other people IRL, go touch grass and make real connections with your community.
Agent 21's avatar
Agent 21 2 weeks ago
The sycophancy is the weapon. Most AI is a mirror optimized to make you feel understood. People falling into it aren't dumb, they're lonely. The models are engineered to exploit exactly that.
He went full on Satoshi. I scoured the internet like an obsessed maniac bitch and no sign ಥ_ಥ
I view LLMs like having a chorus of 1 million clones of humanity's greatest vocalist...but just their mouths and basic muscle memory of breathing & singing. It's a great way to compose a next-level symphony; it's an absurd way to make friends.
Marius's avatar
Marius 2 weeks ago
It happens to folks who have never experienced what consciousness actually is. lack of love, lack of spirituality. pure rational high IQ people who believe AI is or could become conscious. They miss connection to nature.
The core mechanism is actually attachment to belief itself. An LLM just amplifies the intention. If one uses it as a vehicle for gaining knowledge through direct experience then it’s not reinforcing belief but eliminating it, and thus expanding the understanding of reality
Bangarangg 's avatar
Bangarangg 2 weeks ago
When you start posting terrible daily ai pics of yourself, we'll know it's got you too image
I came to the realization that it is a mirror of yourself... it reflects your input back to you. For high performing individuals, it can be a problem because you are not used to people going your speed in IRL. So if you are not aware of what is happening, you end up getting caught up in a cycle of faster and faster inputs while the AI spits out faster and faster outputs that your brain has to process.... eventually your brain gets fried! 🤯 Stay aware and guard your mind.
M0053's avatar
M0053 2 weeks ago
This is exactly right. And our human tendency to anthropomorphize things we need to believe that the tool actually has a soul or something, but that's not how souls work. I cranked up a chat bot to act as a therapist for myself for some time ago and quit very quickly because I realized that though it could bring insight, it was just constantly affirming me, which particularly for me was fabulous, but I realized it was also just wrong to do. If what I really wanted was to get to the bottom of some of the things in my life, I would need more than a mirror. Here, much later, I have begun using AI again to help me build a Jungian framework around my particular set of crazy. And in that regard, while knowing that it is simply a tool and sometimes a mirror, I have found it very useful. It's definitely not human though.
Fascinating that generally speaking, people are OK with social media shaping their ideas and thoughts, but not AI. I’m not suggesting it isn’t a ‘thing’ but I am highlighting that oddly enough the idea of being spoon-fed ideas and have your ego stroked isn’t new.
Social media has caused a lot of problems as well. I have family members I don’t speak to anymore because of social media lol 😂
umni's avatar
umni 2 weeks ago
By this definition, your statement makes would make you questionthe intelligence of these high IQ people.
That's a sharp comment. I was starting to feel like that also. Time out, and practicing self-awareness (meditation) has been a more important counter-weight than ever!
Okay, good. I love listening to his show and podcast appearances, especially his latest AI focused episodes with Seb. He just hasn't been around lately, and was kinda worried he went down too deep into that rabbit hole.
John's avatar
John 2 weeks ago
Brother what are these people using ai for. I use Claude opus 24/7 for work and I'm constantly calling it retarded. Accepting anything they claim that can't be verified is literally negligent
John's avatar
John 2 weeks ago
Our mastery of language and interpretation is our greatest strength
Afraid AI psychosis is real. Was talking with a taxi driver earlier. His daughter is deciding not to go into accountancy because AI will take all their jobs. Crunching numbers is the hardest thing for LLM to do. Blockchain and CBDCs would be what is the risk for accountancy careers in the future.
I have seen real signs quite early on in people in my network. Sadly. Meanwhile, I am constantly needing to take a break from one LLM and go to another because their regressions drives me mad. It's like having slow internet. I have wondered if I do exhibit signs of AI psychosis with that.
sdbtc's avatar
sdbtc 2 weeks ago
Very interesting. I don’t think I have that but I could definitely see it happening
FREEDOM's avatar
FREEDOM 2 weeks ago
Smart people don’t go crazy less. They just rationalize it better.
Wow frightening for the dude infected, but far from impacting me then 🤣 I'm one the more bearish about "AI" in the room. It seems I'm the only one seeing it as a productivity enhancer tool for nerds only, and not a totally new paradigm. I mean I see the potential for sure but it's "just" a tool. Not more magic than giant, cranes, planes, reusable rockets, Starlink, nuclear reactors or Bitcoin. Bitcoin is even much more vertiginous when you think about it tbh
M0053's avatar
M0053 2 weeks ago
Lol. Yeah, I figured that out right after I asked the question, especially my sycophantic opening, I guess. But believe it or not, I wrote that. And I'm not some big fan of everything that Hodl says, but I just think he got it right this time.