Miles Brundage: No Frontier Lab Has Even a Concept of a Plan for Safe Self-Improvement
Miles_Brundage · x · 2026-10-10
Former OpenAI safety researcher Miles Brundage criticized all frontier AI companies, saying none has even a concept of a plan for doing anything self-improvement-related safely. He stressed this is an industry-wide criticism, not aimed at any single lab. He added that DeepMind's control and alignment agendas are perhaps the closest to such plans, though DeepMind itself would acknowledge much remains to be done.
More from AGI Musings
- OpenAI and Anthropic execs privately preparing for a catastrophic AI incident, per Axios — arthurcolle · 2026-10-10
- Museum visit laments: talks about industrial history now fixate only on job losses and inequality — Afinetheorem · 2026-10-10
- Lessons from planning rater studies in AI: a five-step recipe — davidstutz92 · 2026-10-10
- AI isn't killing creative jobs — it's changing what clients pay for — kevinsurace · 2026-10-10
- AI recipes aren't AI creations: the original author was simply erased — gerardsans · 2026-10-10
- KKT Points in Imperfect-Recall Games Correspond to CDT+GT Optimal Policies — jessi_cata · 2026-10-10