An Anthropic staffer who led a team researching AI safety departed the company Monday, darkly warning both of a world “in peril” and the difficulty in being able to let “our values govern our actions” ...