Jacob Coxon, a 27-year-old researcher who worked on pretraining at both OpenAI and Anthropic, has resigned from Anthropic and publicly warned that the AI industry is "gambling with our lives" by racing towards self-improving superintelligence. His seven-part message on X, amplified by an accompanying Wall Street Journal interview, gained nearly 76 million views overnight, with Coxon arguing that competitive pressure between US labs and Chinese rivals is forcing safety trade-offs that could leave AI development "out of control" as soon as next year.

Coxon claimed that many AI executives and senior researchers privately fear their work could prove catastrophic, even while publicly downplaying such concerns, and said colleagues use terms like "crunchtime" and "endgame" to describe the accelerating pace of development. He contrasted OpenAI, where he said staff have not fully grasped the stakes, with Anthropic, where he said the danger is understood but the company feels compelled to press ahead regardless, fearing others will act less responsibly. The Wall Street Journal noted Coxon is not alone in raising alarms, citing OpenAI chief executive Sam Altman's recent warning about looming cybersecurity dangers. Coxon ended his message on a note of cautious optimism about coordination between labs, urging fellow researchers to speak out and consider whether a temporary halt to capability improvements might be necessary.