Jacob Coxon quits Anthropic with chilling warning

Jacob Coxon, a researcher with experience in pre-training at Anthropic and OpenAI, has stepped down publicly, delivering a dire warning about the trajectory of artificial intelligence. He argues that leading AI companies, in their pursuit of self-improving superintelligence, are gambling with humanity's safety. Coxon's concerns have been echoed by Evan Hubinger, a team lead at Anthropic, who concedes a 10% chance that AI could eradicate the human race within the next decade.

Anthropic and OpenAI, in their rush to develop increasingly powerful systems, may be jeopardizing our existence, Coxon contends, and urges the industry to adopt more cautious development practices.