AI Researcher Quits AI Company for Building AI

Jacob Coxon, who spent three years helping to build some of the most powerful artificial intelligence systems ever created, resigned from Anthropic on Tuesday with a warning that the most powerful artificial intelligence systems ever created pose a catastrophic risk to humanity.

Coxon, who previously worked at OpenAI before joining Anthropic — a company founded by people who left OpenAI because they thought it was too dangerous — said in a seven-part post on X that neither company is acting responsibly, that both are “racing straight to self-improving superintelligence and gambling with our lives,” and that things could be “out of control” by the end of next year.

His post reached 76 million people overnight — a figure made possible entirely by the AI-powered recommendation algorithms of the platform on which it was published, which correctly identified that humans enjoy reading about their imminent extinction and served it accordingly.

I resigned from Anthropic today. I spent the last three years doing pretraining research at both OpenAI and Anthropic. Neither company is acting responsibly. They are racing straight to self-improving superintelligence and gambling with our lives. More thoughts below.

— Jacob Coxon (@hilbertspaess) September 9, 2026

Within hours, a significant portion of the internet speculated on the only viable conclusion – the robots are not coming, they are already here. They are among us. Coxon discovered this. The portal has open, and his post is the equivalent of a man in a tinfoil hat outside the Pentagon who turns out to have worked inside the Pentagon for three years and is no longer willing to stay quiet…

Coxon did not respond to requests for comment.

His resignation letter was found to have been composed using a keyboard connected to a computer running software trained on human text, processed by a semiconductor fabricated in Taiwan, and posted via a smartphone.

AI wrote none of it, as far as anyone can confirm.

It does, however, sound remarkably like something a sentient AI robot might have suggested.