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.
More from AGI Musings
- Malthus Saw It in 1798: Wealthy Avoiding Kids, Delayed Marriage Repeats Today — SydSteyerhart · 2026-11-16
- Runway CEO Predicts Fully AI-Generated Cultural Phenomenon by 2027 — c_valenzuelab · 2026-11-16
- Reddit debates the AI scaling curve: everyone says sigmoid, but where's the ceiling? — Successful-War-9050 · 2026-11-16
- '5% worse and 10% cheaper': the button with no vote is already everywhere — eigenrobot · 2026-11-16
- Effective Altruism is top-down and hierarchical, not a grassroots movement — ctjlewis · 2026-11-16
- Lawyers say clients' AI use creates unrealistic expectations and fake case law in settlements — hoofnagle · 2026-11-16