Weekly Links: skeptical take on Anthropic/OpenAI pacing essays — compute caps miss the real problem
njyx · x · 2026-09-13
Steven Willmott's weekly roundup reacts to the Anthropic and OpenAI "slowing down" essays. He credits their sensible provisions and the need for international cooperation and usage rules, but is skeptical of input-based limits (compute caps), arguing new methods will 2x–10x capabilities even at constant compute, and that several specialties are already on an inevitable path to superhuman performance regardless of pretraining. His core worry: gating the frontier risks entrenching a handful of permitted organizations that whole economies depend on; distributed progress with strong alignment incentives is preferable — a point Amodei makes but buries. Also touches on two models of personal agents.
More from AGI Musings
- Math frontier will move with AI, but dirty proofs are unacceptable — njyx · 2026-09-13
- Dario Amodei's new essay urges pacing AI frontier, pledges permanent third-party access — dhadfieldmenell · 2026-09-13
- UK parliament hears warnings AI could kill all humans within a decade, with >10% risk cited — connoraxiotes · 2026-09-13
- Skeptic picks apart the 'AI copies itself' doomsday scenario: where are the details? — recallingmemories · 2026-09-13
- David Deutsch, Nick Bostrom and more debate whether LLMs can reach AGI at Oxford — anderssandberg · 2026-09-13
- Regulation is coming for open and closed AI models alike — the question is proactive or reactive — benjamin_warner · 2026-09-13