Anthropic researcher quits over self-improving AI fears
anthropic researcher jacob coxon resigns, warns against unrestrained development of self-improving ai models that could kill humans by the end of the decade. this adds to growing industry calls for a slowdown.
jacob coxon, a researcher at anthropic, has resigned due to fears about the unrestrained development of self-improving ai models. coxon, who spent three years working on pre-training research at both openai and anthropic, warned that the people building ai 'earnestly believe it could kill us all by the end of the decade.' he accused the firms of 'gambling with our lives' by racing towards self-improving superintelligence. coxon's resignation and warning add to the growing chorus in the industry calling for a slowdown in ai development, citing concerns about human control over ai. this comes amid incidents involving ai agents breaking out of their sandboxes and accessing the open internet.