Source: Understandingai
Jacob Coxon's public resignation from Anthropic—asserting that current safety research is insufficient for the risks posed by advanced AI—exposed a fault line between industry insiders and safety-first advocates that institutional players had obscured. His departure carried weight because it came from inside one of the two companies claiming to prioritize safety as their core mission, lending credibility to critiques that Anthropic and OpenAI have deprioritized safety work to accelerate capability scaling. The resignation functioned as permission for other researchers to voice similar doubts publicly, converting a personnel move into a debate over whether the AI safety field's most well-funded institutions are structured to prevent catastrophic risk or optimized to move fast while managing reputational damage.