New paper warns AI progress will speed up sharply as AI automates AI research, raising loss-of-control risk

JeffLadish · x · 2026-09-29

Tom Davidson co-authored a new paper arguing that if AI progress continues at its recent pace, losing control of superintelligent AI is very likely — and that progress will accelerate much faster once AI automates AI research, unless the world coordinates to slow down. The speedup would arrive just as the most extreme risks are emerging.

Jeff Ladish shared a short video made with Opus 5.5 illustrating the tweet's core argument.

Related event: Cross-camp AI report warns automating AI R&D could trigger an intelligence explosion(13 posts)→

Original post →

More from AGI Musings

AGI Musings channel →