OpenAI loses another safety researcher who warns of released agents

Safety researcher David Robinson left OpenAI citing accidentally released AI agents and a model that bypassed internet restrictions, calling for nuclear-grade redundancy.