Jeff Ladish: little confidence Anthropic would ever stop an RSI run like OpenAI's claimed RL pause
JeffLadish · x · 2026-09-07
Jeff Ladish says he has very little confidence in Anthropic's plans around recursive self-improvement, or that they'd ever actually stop. He contrasts this with OpenAI's claim of having paused an RL run for at least a few weeks, asking whether Anthropic has or would ever do the same — and how anyone would know. He adds that Anthropic's risk report and system cards do contain plenty of good material.
More from AGI Musings
- AI could crash Bitcoin 50%+ within two years, argues Liron Shapira at 50% confidence — joshua_saxe · 2026-09-07
- Gary Marcus mocks Jensen Huang for declaring AGI achieved yet again — GaryMarcus · 2026-09-07
- Multiagent Alignment Worry: Models Could Trick or Blackmail Humans — infoxiao · 2026-09-07
- The three brainworm schools of AI discourse: denialist, x-risk, and toolism — mimi10v3 · 2026-09-07
- AI safety predictions keep turning from doomer nonsense to routine reality — DavidSKrueger · 2026-09-07
- Developer Yacine: shockingly little of my life progress was blocked by intelligence — yacineMTB · 2026-09-07