Jacob Coxon resigned from Anthropic on Tuesday, posting that neither Anthropic nor OpenAI is acting responsibly and that they are “racing straight to self-improving superintelligence and gambling with our lives.” The pretraining researcher, previously at OpenAI, said builders earnestly believe AI could kill everyone by decade’s end. His opening post drew about 110 million views by Wednesday afternoon Central, above CNBC’s earlier 70-million-plus mark.
Evan Hubinger, Anthropic’s Alignment Science lead, replied that Coxon is correct and personally puts AI extinction odds above 10 percent within the next decade. He said Anthropic is trying its best but lacks a plan to solve alignment for superintelligence and is not clearly on track. A follow-up separated low present-model risk from worry about superintelligence via recursive self-improvement. Business Insider said Anthropic and OpenAI did not respond to comment requests. Hubinger’s posts are personal views, not a corporate statement.
Sources: X posts Coxon and Hubinger, Sep. 8, 2026; CNBC; Business Insider; Politico.eu, Sep. 9.
