Sam Altman confirms OpenAI uses recursive self-improvement to build its next models
rohanpaul_ai · x · 2026-11-18
In a Semafor interview, Sam Altman confirmed OpenAI uses recursive self-improvement (RSI) to build its next models: 'Are we developing new models faster than we would without using our technology? Yes.' But he denied any fully autonomous loop — no button that lets the model design its successor without human oversight — and admitted OpenAI could push further along the RSI spectrum than it does, but 'I don't think we could safely do that right now.'
More from AGI Musings
- "No SaaS, no apps": devs debate agents rewriting the entire software stack — alvelda · 2026-11-18
- Ex-OpenAI researcher So8res: labs should collude on safety and dare DOJ to sue — trevposts · 2026-11-18
- Yudkowsky thinks doom is wickedness, Davidad thinks it's weakness — and that decides whether we live — repligate · 2026-11-18
- Professor: non-native students write better with Grammarly than with ChatGPT or Claude — arvindsatya1 · 2026-11-18
- signulll: attention converts to money easily, but money no longer buys clout — signulll · 2026-11-18
- "Supervised vs self-supervised" is a false dichotomy, argues UT RL lecturer — KyleMorgenstein · 2026-11-18