Three open questions on the ~2x AI R&D speedup: is it real, AI-driven, and compounding?
austinc3301 · x · 2026-11-10
Thread wrap-up: three key uncertainties temper the 2x R&D speedup finding — whether it's real, whether AI drives it, and whether it compounds. The author argues against adopting Ajeya/Ryan's take as an all-things-considered view on AI timelines, but notes it follows from reasonable takeoff modeling choices, and that neither they nor METR made any big obvious mistakes.
Related event: Debate Erupts Over METR's Claimed 2x R&D Speedup and AI Timelines(2 posts)→
More from AGI Musings
- Patrick Collison: status lags by a generation or more — _AustinCalvert_ · 2026-11-10
- MIRI launches 'Do Not Accept' campaign challenging AI companies' decisions for humanity — DavidSKrueger · 2026-11-10
- FDT Debate: jessi_cata Argues FDT Is Logical, Policy Dependence Is a Different Concept — jessi_cata · 2026-11-10
- Top capabilities researchers forming alignment reading groups, says it's now P0 — tszzl · 2026-11-10
- Frogdie exec: compiling code once needed a datacenter — few saw desktop power coming — draginol · 2026-11-10
- Remote work dying the same year everyone got an agent is 'the dumbest timing' — StewartalsopIII · 2026-11-10