OpenAI releases RSI report; researcher daveholtz says no intelligence explosion yet

daveholtz · x · 2026-11-23

OpenAI shared new data on how AI is changing its internal research: researchers' AI use keeps rising exponentially, agents complete longer tasks, but human involvement remains central, and safety/alignment effort is growing faster than any other research area. OpenAI argues labs should be required to track and publicly report indicators of recursive self-improvement (RSI). Researcher daveholtz pushes back on the intelligence explosion narrative: if it were happening, the acceleration in model capabilities would be impossible to miss.

Related event: OpenAI's First RSI Report: Agent Task Length Grows but Intelligence Explosion Has Not Arrived(2 posts)→

Original post →

More from AGI Musings

AGI Musings channel →