What bridged the gap from R1-style short reasoning to today's long chains?
yoavgo · x · 2026-09-25
Yoav Goldberg adds that the known training stack roughly explains DeepSeek R1-era 'short and okay-ish' reasoning, but current models show far more impressive abilities. What bridged the gap between short reasoning then and the long, complex traces now remains unexplained — an open jab at the opacity of frontier reasoning-model training.
More from AGI Musings
- AI risk is a governance problem before it is an existential one, argues exec — ingliguori · 2026-09-25
- BlackRock sees AI driving digital asset demand as agent-to-agent payments loom — tallmetommy · 2026-09-25
- Dan Faggella pushes back on 'AGI will naturally be caring': an alien GPU god is not a parent — danfaggella · 2026-09-25
- Why Anthropic's Biology Bets May Win: Moore's Law and the Hardware Lottery — IgorCarron · 2026-09-25
- Robert Miles: only people who never talk about AI would say 'never anthropomorphize' — aran_nayebi · 2026-09-25
- Jensen Huang accidentally calls for shutting down OpenAI, per Zvi's podcast breakdown — Don't Worry About the Vase (Zvi) · 2026-09-25