Ten Years After AlphaGo: The Game That Changed Everything

Ten years ago, a moment of machine creativity changed the world's perception of artificial intelligence forever.

On March 9-15, 2016, DeepMind's AlphaGo defeated world Go champion Lee Sae-dol in Seoul—a feat most AI experts expected was at least a decade away. The match, watched by over 200 million people, wasn't just a victory; it was a paradigm shift. When AlphaGo played the now-legendary "Move 37" in Game 2, a play so unconventional that professional commentators initially thought it was a mistake, it demonstrated something that stuck: AI could do more than mimic human intelligence. It could innovate.

"Move 37 was a definitive preview of the AI era—proving it wasn't some distant, vague future, but a reality arriving on our doorstep," DeepMind CEO Demis Hassabis wrote in a blog post marking the anniversary.

Go has 10^170 possible board positions—more than atoms in the observable universe. To crack it, AlphaGo combined deep neural networks with Monte Carlo tree search and reinforcement learning, an approach DeepMind published in Nature in 2016. After beating Lee, the system evolved: AlphaGo Zero learned entirely from random play and became arguably the strongest player in history. AlphaZero generalized further, mastering chess and shogi from scratch in hours, even beating specialized programs like Stockfish.

But the game was always a means to an end. "The technology was ready to be applied to our real goal of accelerating scientific breakthroughs," Hassabis wrote.

That goal materialized fastest in biology. In 2020, AlphaFold 2 solved the 50-year protein folding problem—predicting 3D protein structures from amino acid sequences. The team folded all 200 million known proteins and released them in an open database. Today, over 3 million researchers use it. In 2024, Hassabis and AlphaFold lead developer John Jumper won the Nobel Prize in Chemistry.

The AlphaGo playbook proved portable. AlphaProof, combining language models with AlphaZero's architecture, became the first AI system to achieve silver-medal level at the 2024 International Mathematical Olympiad, solving four of six problems. An advanced version of Gemini with "Deep Think" mode went further in 2025, earning a gold-medal score—35 out of 42 points—at the 2025 IMO.

AlphaEvolve, a coding agent, found novel matrix multiplication algorithms—a fundamental operation powering neural networks. In validation studies with Imperial College London, DeepMind's AI co-scientist independently arrived at the same hypothesis about antimicrobial resistance that researchers had spent years developing.

"Ten years after AlphaGo's legendary victory, our ultimate goal is on the horizon," Hassabis wrote. "The creative spark first seen in Move 37 catalyzed breakthroughs that are now converging to pave the path towards AGI."

The game that started it all is still being played—and the stakes have never been higher.