Researchers statistically re-examine METR time horizons: does human time proxy AI difficulty?
wfithian · x · 2026-10-14
Drew Nguyen, with wfithian, published a statistical look back at METR's time horizons chart: taking the 228 software tasks plotting human completion time (x) against estimated AI difficulty (y), they quantified whether the log-linear assumption — that time proxy difficulty well — actually holds.
Drew drove the project end to end and built methods to check assumptions like this, plus a look ahead. Since METR's time-horizon curve is one of the most-cited inputs for forecasting AI capability growth, a statistical audit of its core premise is directly relevant.
More from Research
- Economist compiles 100+ free social science datasets on GitHub — _akpiper · 2026-10-14
- New ISQ article asks why policymakers wait to regulate disruptive tech like AI — mchorowitz · 2026-10-14
- MIT professor's ten predictions for AI for science, from executable papers to compiled matter — ProfBuehlerMIT · 2026-10-14
- Sweden spends hundreds of millions of kronor on printed textbooks as 2.8M-child study links screens to lower grades — Eric520CC · 2026-10-14
- Self-evolving red-team search finds 1.77x more harmful tool-call paths in personal agents — lulzxdxdxd · 2026-10-14
- Google researcher: general capabilities plus tool use are displacing domain-specific medical AI — davidstutz92 · 2026-10-14