AI safety drama: zero detected misalignment incidents from Chinese labs despite Western lab turmoil

basedjensen · x · 2026-09-29

Reacting to drama around OpenAI's evals lead, the poster says they don't fully buy that observed incidents were helped by EA saboteurs making an AI safety point — but notes there still seem to have been zero detected incidents from Chinese labs. The quoted tweet questions how OpenAI expects "misalignment" to stop with this person running evals, noting he announced joining Twitter to "sound the alarm" on September 6, two days before Coxton's incident.

Original post →

More from AGI Musings

AGI Musings channel →