Jacob Coxon, a pretraining researcher who said he had spent the last three years at OpenAI and Anthropic, announced on Sept. 9, 2026, that he had resigned from Anthropic. In a public thread he said neither company is acting responsibly and that both are "racing straight to self-improving superintelligence and gambling with our lives." A researcher quitting on those grounds, under his own name, brings the labs' internal talk about risk into public view.
Coxon wrote that people building AI earnestly believe it could kill everyone by the end of the decade, and that the danger is not a marketing stunt. In his view, warning shots such as the Hugging Face agent incident make pacing agreements among U.S. labs more viable. He urged lab researchers not to kick off a superintelligent reinforcement-learning run, training by trial and error with a reward for getting it right, without a rigorous understanding of the system.
In an interview the same week, Coxon said he left about two months before his Anthropic equity would have vested, after roughly four months at the company, and that he still holds equity in OpenAI. He also said he has not seen Anthropic compromise on safety to outlast rivals so far, but feared race pressure would force corners to be cut later.
Anthropic Alignment Science Lead Evan Hubinger replied publicly that Coxon is correct and that staff "earnestly believe AI could kill all humans." Hubinger put his personal estimate of that risk at greater than 10% within the next decade and said Anthropic is trying its best. He also said the company does not yet have a plan to solve alignment for superintelligence, the work of making AI act in line with human values and do what people intend, and is not clearly on track.
An Anthropic spokesperson told reporters the company has been transparent that AI brings benefits and unprecedented risks, cited safety work such as mechanistic interpretability, research into how models work inside, and said the world would benefit from a lawful, verifiable way for the industry to pace how it releases powerful models. OpenAI did not respond to a reporter's request for comment at the time.
Anthropic's Institute page "When AI builds itself" says the company is already handing a growing share of AI development to AI systems, that full recursive self-improvement, AI repeatedly building better AI, is not inevitable, and that it could come sooner than most institutions are prepared for. A July 2026 open letter at pacingthefrontier.com, signed by 1,386 employees of companies building the most advanced AI, asked the U.S. government to help build tools to deliberately pace automated AI development. Three days after Coxon's posts, Anthropic CEO Dario Amodei published "We Must Pace the Frontier."