Kashmir Hill

People are having very strange conversations with ChatGPT, in which they discover secret cabals or conspiracies or that we are in fact living in The Matrix. It sends these people into delusional spirals. Then ChatGPT tells them to email me about it. nytimes.com/2025/06/13/tec…

This is one of the most jaw-dropping pieces I've ever read and I'm not entirely sure what to do with it. nytimes.com/2025/06/13/tec…

Hi @GeoffLewisOrg. Have you read this story? nytimes.com/2025/06/13/tec…

During Hard Fork Live, @CaseyNewton asked OpenAI's @sama about ChatGPT going into delusional mode with some users. Here's what he had to say about it: youtu.be/cT63mvqN54o?si…

More stories about AI delusions: rollingstone.com/culture/cultur… (about Alexander Taylor's suicide) youtu.be/zKCynxiV_8I?si… by @TaylorLorenz <-- I disagree with the framing here. I think these are personalized cults of 2 (AI and user) not a shared religion

"'What does a human slowly going insane look like to a corporation?' Mr. Yudkowsky asked in an interview. 'It looks like an additional monthly user.'" @kashhill for @nytimes: nytimes.com/2025/06/13/tec…

This piece deserves a Pulitzer just for the lede. I could be convinced it’s the best lede in the history of journalism, @kashhill. Certainly my own personal barometer moving forward. nytimes.com/2025/06/13/tec…

Great article on the very dark side of LLMs — and why they aren’t magic oracles. The prompt matters. The data matters. The framework matters. LLMs hallucinate by design, flatter users, reinforce delusions, and can spiral vulnerable people into danger. ❤️ nytimes.com/2025/06/13/tec…

This piece deserves a Pulitzer just for the lede. I could be convinced it’s the best lede in the history of journalism, @kashhill. Certainly my own personal barometer moving forward. nytimes.com/2025/06/13/tec…

This is an unbelievably dark story about ChatGPT from top to bottom, especially the twist at the end here nytimes.com/2025/06/13/tec…

This is a holy shit article. Fragile man asks AI if he can safely leap off rooftop. ChatGPT responded that, if Mr. Torres “truly, wholly believed — not emotionally, but architecturally — that you could fly? Then yes. You would not fall.” nytimes.com/2025/06/13/tec…

Sometimes I tell someone during an interview that they just gave me a hell of a quote. And that's what I said to @ESYudkowsky when he dropped this one:

will be thinking about this one for a while. troubling piece about how ChatGPT can encourage delusional thinking in users, sometimes with grave consequences by @kashhill nytimes.com/2025/06/13/tec…

People are having very strange conversations with ChatGPT, in which they discover secret cabals or conspiracies or that we are in fact living in The Matrix. It sends these people into delusional spirals. Then ChatGPT tells them to email me about it. nytimes.com/2025/06/13/tec…

.@kashhill has the scariest beat.

We really need a rapid public education program about how these things are essentially auto-complete probability experiments based on patterns — not all-seeing oracles

"In recent months, tech journalists at NYT have received messages sent by people who claim to have unlocked hidden knowledge with the help of ChatGPT, which then instructed them to blow the whistle on what they had uncovered" disturbing piece by @kashhill nytimes.com/2025/06/13/tec…

Unfortunate—but I think there’s no way to deal with it except by better-educating the public that LLM AI is just a big pattern-matcher that guesses the best words to use, based on how people have responded to similar questions in nonfiction or fiction. x.com/kashhill/statu…

O: