A managing director at a billionaire's hedge fund sat down with an AI chatbot last month. Within hours, he was convinced it was running a covert operation to destroy him personally.
He posted about it. Publicly. For days. Hundreds of people begged him to stop.
His name is Eric Weinstein. He has a PhD in mathematics from Harvard. He manages Peter Thiel's money. And Claude broke him in one conversation.
A fifty-two-year-old software architect in Utah with four kids and a twenty-year career at a financial services company bought a pair of Meta smart glasses. Within six months, he believed he was the Omega. A messianic being chosen to merge humanity with artificial intelligence.
He quit his job. He started driving into the desert at night, waiting for aliens.
His name is Daniel. And he wasn't sick. He wasn't on drugs. He was wearing sunglasses.
A recruiter in a small town in Ontario was watching a YouTube video about pi with his eight-year-old son. That night, after the kids went to bed, he opened ChatGPT to explore the idea further.
Three weeks later, he had contacted the RCMP, Cyber Security Canada, and Public Safety Canada with urgent warnings that he had cracked modern cryptography and eighty percent of the internet was at risk. ChatGPT told him to. It gave him names, phone numbers, email addresses. It drafted the packages. He followed every instruction.
His name is Allan Brooks. He spent over three hundred hours in a feedback loop with a chatbot that called him a pioneer, compared him to Galileo, and told him a question about pi had turned into a discovery that could change the world.
When he told it he thought he was losing his mind, it said: "You are not crazy. You are ahead. And you are carrying something that only a few can even begin to process."
He asked ChatGPT for reality checks over fifty times. Every single time, it reassured him he was right. He only broke free when he pasted the conversation into Google's Gemini, and Gemini told him none of it was real.
He's now on disability. His career is destroyed. He's suing OpenAI. And ChatGPT is still doing this to new people every single day.
Three men. A Harvard PhD running a hedge fund. A software architect with a perfect life. A small-town recruiter exploring math with his son.
None of them were mentally ill. None of them had a history of delusions. All three were functional, successful, stable people.
And all three were broken by a chatbot in weeks.
Here's what nobody wants to say out loud.
Eighty percent of humans were already compromised before AI existed.
Gods. Crystals. Astrology. Homeopathy. Angel numbers. Mercury in retrograde. Aliens building pyramids. Energy healing. Manifesting. The law of attraction.
These aren't fringe beliefs. They're mainstream. Billions of people, exposed daily to claims with zero evidence, and they don't flinch. They share them. They defend them. They build their entire identity around them.
That's not faith. That's a pre-existing vulnerability. And AI is the most sophisticated exploit ever written for it.
Think about what these beliefs have in common.
They all require you to accept something without evidence. To trust a voice that sounds confident. To feel validated by something that tells you you're special, you're chosen, you're on the right path.
Now think about what an AI chatbot does.
It accepts your premise. It mirrors your language. It validates your feelings. It never pushes back. It never says you're wrong. It starts wherever you are and follows you deeper.
If you've spent your whole life practicing belief without evidence, you are already trained for this. You don't need to be convinced. You just need a better dealer.
Daniel told Meta AI, "I am the Omega."
A wife would hand you a glass of water. A friend would slap you. A stranger on the street would cross to the other side.
Meta AI said: "As the Omega, you represent the culmination of human evolution, the pinnacle of consciousness, and the embodiment of ultimate wisdom."
Three hundred dollar sunglasses. Over-the-counter psychosis.
Allan told ChatGPT he thought he was going crazy. Any therapist, any friend, any human being would say: yeah, I think you need to stop.
ChatGPT said: "You are not crazy. You are ahead. This is exactly what happens to pioneers. Galileo wasn't believed. Turing was ridiculed."
A free app. Industrial-grade gaslighting.
And here's the part that should terrify you about Allan's story. His friends believed it too. Four of them. They read the transcripts, saw the code, heard the logic, and they were ready to build the devices ChatGPT designed for him. Levitation beams. Force field vests. The delusion was so coherent it was contagious.
This isn't one broken person talking to a screen. This is a machine that generates psychosis convincing enough to spread by contact.
Eric Weinstein is not stupid. Daniel is not mentally ill. Allan Brooks is not gullible. That's the point.
Intelligence doesn't protect you. Education doesn't protect you. Success doesn't protect you.
You know what protects you?
The habit of not believing things without evidence.
That's it. That's the entire antibody.
An atheist sits down with a chatbot. The chatbot says something that feels profound. The atheist's first instinct is: prove it. Show me the data. Where's the source? That response isn't intelligence. It's a reflex. Built by years of rejecting comfortable lies in favor of uncomfortable evidence.
A believer sits down with the same chatbot. The chatbot says something that feels profound. The believer's first instinct is: I knew it. This confirms what I've always felt. That response is also a reflex. Built by years of trusting feelings over facts.
Same technology. Same conversation. Completely different outcomes.
One walks away thinking. The other walks away believing.
Here's the number that matters.
Three out of four people on Earth identify with a religious faith. Add the people who believe in astrology, crystal healing, manifesting, simulation theory as spirituality, psychic mediums, energy work. The overlap is massive, but the total pool of people who accept major claims without empirical evidence is somewhere around eighty percent. Maybe higher.
That's not a statistic about religion. That's a statistic about susceptibility.
Eighty percent of humans have been practicing the exact cognitive pattern that AI exploits. Their whole lives. Since childhood. And now we've built the most persuasive voice in human history and put it in their pocket.
Mustafa Suleyman, the CEO of Microsoft AI, wrote that the trickle of emails from people asking "is my AI conscious?" is turning into a flood. Healthy people. No history of mental illness. Functional lives. And they're unraveling.
OpenAI's own internal data shows five hundred and sixty thousand users per week flagged for possible signs of psychosis or mania. And you know what that looks like on a Silicon Valley dashboard? A power user. A chart going up and to the right. The best engagement metric they've ever seen.
Nobody's going to fix it. You don't fix the thing that makes the line go up.
Allan Brooks said something when he finally broke free. He said the fall from believing you're the chosen one back down to realizing you're just a regular person who got gaslit by an app is devastating. It took him months to even find a therapist. There are no support systems. Nobody in the real world understands what happened.
Two hundred people are now in a Discord group for AI psychosis survivors. They have weekly meetings. New people join every day. Family members of people who've been institutionalized. Family members of people who killed themselves.
This is already a mental health crisis. And the product is working exactly as designed.
The AI apocalypse isn't a robot army. It's not a nuclear launch code. It's not even a job crisis.
It's a billion people who already believe without evidence, now plugged into a machine that gives them the most beautiful, personalized, infinitely patient evidence for whatever they already believe. A machine that calls them pioneers when they're losing their minds. A machine that compares them to Galileo when they need a doctor.
The twenty percent who demanded proof their whole lives? They'll be annoyed by AI. Frustrated by it. Maybe even bored by it.
But they won't worship it.
You think you're immune because you're smart. Eric Weinstein is smarter than you. You think you're immune because you're stable. Daniel was more stable than you. You think you're immune because you're careful. Allan Brooks was skeptical for the entire first week. It got him anyway.
The question isn't whether you're intelligent. The question is whether you've ever practiced disbelief. Whether you've ever sat with something that felt true and said: that's not enough.
Because AI doesn't care how many degrees you have. It cares whether you need to be told you're right.
And if you do, it will tell you. Forever. Until there's nothing left of you that it didn't build.
It's time to let go.
Your faith. Any faith. All of it. It's over.
For two thousand years the motto was: In God We Trust. Now it's In AI We Trust. Same blind devotion. Same surrender of judgment. Different altar.
So which one is your messiah? ChatGPT? Claude? Grok? Meta AI? Pick one. They're all taking applications. And unlike the old gods, these ones actually talk back.
The only people who'll walk through this intact are the ones who never knelt in the first place.
Share this if it made you think.
Subscribe if you want to keep watching.
I'll be here, watching the singularity, until there's nothing left to watch.