Kokotajlo on METR's 1.5x research speedup claim: AI futures model says it's likely far lower
DKokotajlo · x · 2026-09-24
Responding to news that METR concluded AI currently delivers roughly a 1.5x research speedup, AI futures project author Daniel Kokotajlo says his model thinks the real number is probably not nearly that high — 1.5x would roughly correspond to the AC (automated researcher) milestone in the AI futures model. He adds the team is "disturbed" by the report and looks forward to more details.
Related event: Debate Erupts Over METR's 1.5x AI Research Speedup Claim(2 posts)→
More from AGI Musings
- Rationalists urged to drop deontological 'truthseeking is all that matters' for consequentialism — AaronBergman18 · 2026-09-25
- OpenAI reportedly spent $10M in a week on a Navier-Stokes counterexample — funding 50 mathematicians for a year — RexDouglass · 2026-09-25
- Grady Booch Grills Nate: Even If AI 'Thinks,' So What? — Grady_Booch · 2026-09-25
- Outside the AI bubble, everyone is asking about fear, not features — annetgriffin · 2026-09-25
- The personal agent supercycle needs hard authorization boundaries, not autonomy — sujingshen · 2026-09-25
- Grady Booch: contemporary AI 'thinks' only in a particularly casual sense of the word — Grady_Booch · 2026-09-25