Ten years of AlphaGo Zero: the tabula-rasa Go AI that beat its predecessor 100-0
1a3orn · x · 2026-10-23
Researcher 1a3orn marks the 10th anniversary this month of DeepMind's Nature paper Mastering the game of Go without human knowledge — AlphaGo Zero, which reached superhuman Go purely via self-play reinforcement learning and beat the champion-defeating AlphaGo 100-0. His observation: notably less happened in the first five years after the paper than in the second five, underscoring how AI progress has accelerated.
More from AGI Musings
- Sentdex mocks AI race: 'building a new atomic bomb, someone should act — not us' — Sentdex · 2026-10-23
- MicroCapClub's Ian Cassel: AI erases the info edge, meeting management is the last moat — rickasaurus · 2026-10-23
- Prediction: personal agents will replace apps as the primary interface — HarveenChadha · 2026-10-23
- Geoffrey Hinton: 'I'm not a doomer,' puts AI extinction risk at about 1% — Logibenq · 2026-10-23
- Reddit Post Claims It's Near-Certain a Backdoor Already Sits in Popular LLM Harnesses or Inference Engines — Ori_553 · 2026-10-23
- waitbutwhy's classic AI explosion chart: from 'monkey tricks' to 'The fuck??' — pickover · 2026-10-23