Closed AI evals face criticism: opaque, single point of failure, and potential conflicts of interest
evijit · x · 2026-10-18
Researcher evijit argues that the proliferation of closed AI evals the public cannot analyze or fix is bad for science and unsustainable, given a single point of failure hosted by one org. He raises key questions: how can third-party researchers measure eval quality, who pays for the compute (and do well-resourced labs have conflicts of interest), and what happens to small OSS labs and community developers?
More from AGI Musings
- 'Junior hiring isn't cooked' crowd mocked as WSJ advises job seekers to update LinkedIn — felpix_ · 2026-10-18
- Researcher Mariya Ivasileva warns AI research culture is eroding intellectual honesty — mariyaivasileva · 2026-10-18
- MIT lets a Claude-powered agent freely "play" on a simulated island for 30 hours — and finds play expands its capabilities — LuizaJarovsky · 2026-10-18
- Using AI well demands more skill than most assume, argues researcher — davidmanheim · 2026-10-18
- Epic CEO Tim Sweeney urges devs to openly embrace AI: most already use and love it — lucasmeijer · 2026-10-18
- Text-only AGI skeptics: all modalities are needed for general intelligence — Angaisb_ · 2026-10-18