Anthropic alignment researcher Joe Benton leaves for METR to work on AI risk assessment
EthanJPerez · x · 2026-09-03
Joe Benton has left Anthropic after about 1.5 years to join METR, where he'll work on embedded assessment of AI risks. He called Anthropic great but said he'd always leave for higher-impact work. Former colleague @sprice354 praised the move and joked about Benton now "holding us accountable from the other side" — a notable case of safety talent flowing from a frontier lab to an independent evals org.
Related event: Anthropic researcher Joe Benton departs to join AI evaluator METR(6 posts)→
More from Companies & People
- Why Finance Is the Best Beachhead for AI Agents: 25-Minute Tasks That Take 11 Days — inductionheads · 2026-10-31
- Cisco President Gives 27-Year-Old Engineer Unlimited AI Tokens to Rewrite 98% of Legacy Code — inductionheads · 2026-10-31
- OpenAI to train on Daybreak Blue security-research sessions from Nov 29, with no opt-out — ccerrato147 · 2026-10-31
- Engineer with OpenAI and Anthropic offers shares prep resources, including a surprising one — StasBekman · 2026-10-31
- New exposé claims one AI company's private reality is uglier than its public drama — AIandDesign · 2026-10-31
- OpenAI kills products fast: Sora lasted 6 months, Atlas 9, Agent Builder 8 — emollick · 2026-10-31