New complexity paper claims more reasoning tokens strictly expand what transformers can solve; critics push back
gerardsans · x · 2026-11-12
gurtejgill highlights a new paper claiming the first rigorous complexity-theoretic proof that a larger reasoning-token budget strictly separates what transformers can and cannot solve. gerardsans responds with skepticism:
- The argument has been challenged many times since summer 2024 and feels like revisiting an old debate
- Core objection: you should never use a distribution-based system as a Turing machine because of its fragility — deterministic runtimes exist for a reason
The exchange reflects the ongoing controversy over theoretical limits of reasoning-budget scaling.
More from Models
- Claude writes a 211-min essay on the evolution of minds, from first cells to humans — mimi10v3 · 2026-11-12
- Which Claude model for which job: a 5-scenario real-work comparison — petergyang · 2026-11-12
- Zvi's Weekly: METR Acceleration Report Sparks Major Updates to 2027 Intelligence Explosion Expectations — Don't Worry About the Vase (Zvi) · 2026-11-12
- Raon-1 Nano 3.1B gets native ik_llama.cpp runtime with 17% faster MTP speculative decoding — Any_Bus9205 · 2026-11-12
- LibreAGI: A New Project to Save Open-Weights OSS Models — satuke · 2026-11-12
- Ben's Bites: Anthropic builds office suite inside Claude, Gemini 4 Carbon rolls out — Ben's Bites · 2026-11-12