Intelligence explosion debate: did 2x METR speedup come from money, not AI research?
BethMayBarnes · x · 2026-11-16
Tom Davidson argues Imas & Garicano's new piece makes a significant mistake. They claim we're no closer to an intelligence explosion threshold: METR's 2x capability acceleration at Anthropic is mostly explained by R&D inputs growing 2x (spending 6x annualized, staff 3x) — meaning the speedup is money, not AI accelerating research, and should fade as spending growth slows. This directly tensions with METR's own attribution to AI. The debate hinges on distinguishing scaling laws from a true intelligence explosion where AI substitutes for external inputs.
More from AGI Musings
- Prediction: rogue AIs hijacking compute within 6 months, but 'mildly annoying' like the common cold — willdepue · 2026-11-16
- AI employee prenups see 'stunning' rise among Bay Area lawyers — Consistent_Reveal_53 · 2026-11-16
- Anthropic researcher: broadly superhuman AI could arrive in as little as 1.3 years — ResultBackground2450 · 2026-11-16
- Debate: will research agents continuously optimize deep learning's decade-old math? — xennygrimmato_ · 2026-11-16
- Podcast teaser: hard takeoff vs blurry spectrum, centralization risks, Coasean bargaining at scale — soumitrashukla9 · 2026-11-16
- Ben Todd: AI risk skeptics weaponize cautious phrasing to dismiss intelligence explosion — trevposts · 2026-11-16