David Silver

In the early 2000s, artificial intelligence researchers used the game of Go as the ultimate benchmark: a 19x19 board game with more possible combinations than there are atoms in the observable universe. The consensus was clear: no computer would beat a human champion for at least 20 years. David Silver, a British researcher who had studied mathematics at Christ's College Cambridge and later earned a PhD in reinforcement learning at the University of Alberta, decided to take it as a challenge. Not because it seemed easy, but because he believed that artificial intelligence had to learn to reason, not just to memorize.

On January 28, 2016, Silver's team published the AlphaGo paper on the cover of Nature. It was the first AI to master Go by combining deep learning and reinforcement learning. Two months later, in March 2016, AlphaGo defeated world champion Lee Sedol in a five-game match (4 to 1), watched by millions around the world. The turning point came on Move 37 of the second game: a move so unexpected that the commentator took more than a minute to understand what the machine had done. Lee Sedol said he had never seen anything like it in his twenty-year career.

But Silver didn't stop there. In 2017 he published AlphaGo Zero, which learned to play Go from scratch without seeing a single human game, purely by playing against itself. In 2018 came AlphaZero, which in under 24 hours taught itself to master chess, shogi, and Go simultaneously. That work earned him the 2019 ACM Prize in Computing, the most prestigious award in the field, and in 2021 he was elected Fellow of the Royal Society, the world's oldest scientific academy. In 2026, after more than a decade at DeepMind, Silver founded Ineffable Intelligence with the goal of building AI that learns entirely on its own, without human data.

The lesson I take from David Silver is this: sometimes the greatest contribution is not solving the problem everyone is trying to solve, but reframing it. He didn't ask "how do we get a computer to memorize good moves?" He asked "how do we get it to reason from scratch?" That shift in perspective changed history. Today, every time you see an AI model learn something new without anyone teaching it step by step, you're seeing the echo of that question.