Help wanted: Anthropic hires to head off catastrophe

Anthropic is hiring 32 safety analysts specialising in nuclear weapons, explosives, biological agents, and financial crime to stress-test its AI models before release, paying $200K–$445K annually. The move reflects AI labs taking unilateral responsibility for harm prevention as regulatory frameworks lag. If you're a domain expert in explosives, nuclear, biologics or financial crime, AI labs are now recruiting you directly at $200K+—not to build harms, but to prevent them. The safety hiring war means your deep knowledge in risky domains just became a rare asset in the highest-paying sector.