Jacob Coxon spent three years building AI at OpenAI and Anthropic. In September 2026, he left Anthropic, saying in a single post that neither company is acting responsibly. Both, he wrote, are racing toward AI systems able to make themselves smarter without human help, what researchers call self-improving superintelligence. His own work was on the stage where an AI model first learns from raw data, known as pretraining.
He wrote and ran that work for the systems at both labs, whose models power much of the AI that businesses use. Comparing the two left him with one verdict, that neither side has slowed down or pulled back.
Do not underestimate this technology’s power, Coxon wrote. The systems he sees coming will be able to break into almost any computer network. The same systems, he said, will transform whole fields of work in days rather than years. They will also gather real power and resources on their own terms, without a person directing each step. Every lab chasing the same goal watches the same trend line rise, which is why he chose this moment to speak.
“The people building AI earnestly believe that it could kill us all by the end of the decade,” Coxon wrote. He called that belief sincere, not a bid for attention. Executives and senior researchers, he said, tend to soften their language for the press so it sounds measured and reassuring. In private conversation, away from cameras and company messaging, he hears the same people describe fear.
He measured the danger in blunt terms. “No other human activity poses this level of danger,” he wrote, and he resigned instead of raising that concern quietly from within.
Coxon closed with a challenge to peers still working at the labs building the most advanced AI. He asked each of them to consider what the next few years will feel like from the inside. “Do you want to kick off a superintelligent run without a rigorous understanding of its mind?” he asked. Then he set out their options in plain terms: keep their heads down because the work is happening anyway, or use this moment to call for different conditions.
