METR researcher joining Anthropic sparks concerns over eval lab independence
ShakeelHashim · x · 2026-10-28
The news that Jasmine Dhaliwal has joined Anthropic drew criticism: commentators argue such moves make it hard to convince the public that METR — which evaluates frontier models, including Anthropic's — is truly independent of the labs it assesses. The episode highlights a structural tension in AI safety: eval organizations' credibility depends on independence, while talent circulates between evaluators and the evaluated.
More from Companies & People
- Ashley Vance visits University of Waterloo to decode its cracked-engineer reputation — kyliebytes · 2026-10-28
- Law firm AI debate: local-only Kimi/DeepSeek vs Harvey's portfolio strategy — ivan_bezdomny · 2026-10-28
- Gamma Is Now Worth Over $2B, But Its First Idea Confused Everyone — dunkhippo33 · 2026-10-28
- The 1931 P&G memo by 26-year-old Neil McElroy that invented the product manager role — aakashgupta · 2026-10-28
- LinkedIn's APM becomes "Product Builder": PM-to-engineer ratio hits 1:1 — aakashgupta · 2026-10-28
- "If AI can do it, humans shouldn't": team lets AI agent review PRs — Yuchenj_UW · 2026-10-28