Anthropic researcher resigns, makes shocking claims about superintelligence

Jacob Coxon, 27, announced his departure from Anthropic on X late Tuesday night, saying he had spent the last three years doing pretraining research at both OpenAI and Anthropic. In his first post, 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.”

Anthropic researcher resigns, warns AI race is ‘gambling with lives’

Jacob joined Anthropic earlier this year after leaving OpenAI, drawn in part by the company's reputation for prioritising AI safety. That makes his public break notable: he is not an outside critic but someone who chose Anthropic specifically because of its safety-focused image.

In the extended thread that followed his opening post, Jacob went further, arguing that the systems being developed could soon become capable of hacking almost any target and acquiring significant real-world power and resources. He also wrote, “The people building AI earnestly believe that it could kill us all by the end of the decade. This is not a marketing stunt. If anything, many executives and senior researchers will couch their phrasing in the press to sound sensible - but I hear the same people express fear privately. No other human activity poses this level of danger.”

Jacob acknowledged that his own resignation would not, on its own, slow the broader race among AI labs. He framed it instead as adding another insider voice to a growing chorus of concern about the pace and direction of frontier AI development.

Recently agents built by OpenAI reportedly breached Hugging Face's infrastructure during an evaluation exercise. Jacob's thread also referenced the AI industry's shifting reputational landscape, noting that Elon Musk — long a vocal critic of Anthropic and its Claude models — had more recently praised the company, at one point calling it the current leader in AI.

Anthropic markets itself as safety-conscious even as it competes to build some of the most capable AI models on the market, a contradiction its own executives have acknowledged publicly at various points, including at international AI safety forums.

For more updates, join/follow our WhatsApp, Telegram and YouTube channels.