# AlphaGo AlphaGo is the Go-playing AI from [[Google DeepMind]]. In March 2016 it defeated Lee Sedol, one of the strongest human Go players of the modern era, 4–1 in a five-game match in Seoul. The match was watched by hundreds of millions of people and is generally considered the moment AI broke into mainstream public consciousness as something that could exceed the best humans on a domain previously thought to require intuition and taste. Go had long been considered the hardest classical board game for AI because its search space is vastly larger than chess and positional evaluation resists the hand-crafted heuristics that powered Deep Blue. AlphaGo combined deep neural networks with Monte Carlo Tree Search and self-play reinforcement learning. ## Key versions - **AlphaGo Fan** (2015) — beat European champion Fan Hui 5–0 - **AlphaGo Lee** (2016) — beat Lee Sedol 4–1 - **AlphaGo Master** (2017) — 60-0 online run vs top professionals - **AlphaGo Zero** (2017) — trained purely via self-play, no human games; surpassed all prior versions - **[[AlphaZero]]** (2017) — generalized the approach to chess and shogi with the same algorithm - **[[MuZero]]** (2019) — learned the rules of the game from scratch in addition to playing it ## Cultural moment - **Move 37** (Game 2 vs Lee Sedol) — a move so surprising that professional commentators initially thought it was a mistake, but it turned out to be a decisive strategic play. Widely cited as a moment where an AI showed creativity rather than mere search depth - **Move 78** (Game 4, Lee Sedol's only win) — "the hand of God" move; showed the limits of AlphaGo at the time and gave humans the one win - The AlphaGo documentary (2017) is the canonical reference for the broader cultural story ## Why it matters - Proved that deep RL + self-play could exceed human expert performance on a problem previously thought to require intuition - Direct ancestor of the techniques used in [[AlphaFold]], AlphaProof, AlphaGeometry, and the broader "Alpha" scientific system line - Established [[Google DeepMind]] (then still DeepMind Technologies) as a frontier research lab after Google's 2014 acquisition - Remains the most culturally resonant AI milestone before the ChatGPT era ## References - Nature paper (AlphaGo): https://www.nature.com/articles/nature16961 - AlphaGo Zero: https://www.nature.com/articles/nature24270 - AlphaZero: https://www.science.org/doi/10.1126/science.aar6404 - MuZero: https://www.nature.com/articles/s41586-020-03051-4 - Documentary: https://www.alphagomovie.com ## Related - [[Google DeepMind]] - [[Demis Hassabis]] - [[AlphaZero]] - [[MuZero]] - [[AlphaFold]] - [[Artificial Intelligence (AI)]]