The Grandmasters Who Never Read a Chess Book

What happens when a generation learns differently? The lessons from the first field AI dominated...

The Pass of Killiecrankie…

AI is a normal technology. That means we can learn from past experience. Not to predict the future but to the right questions.

One field where AI has run way ahead is chess. Deep Blue beat Garry Kasparov, the reigning world champion, in 1997. Progress since has been so rapid that a chess programme on your phone now beats the world’s best players.

Yet there are now 605 million casual chess players and 1.64 million people have an active rating with FIDE, the world governing body. The number of grandmasters, those playing at the highest human level, has more than trebled since Kasparov lost that match.

The first lesson is simple: people don’t stop doing something they value just because AI does it better.

But how has that growth happened? What else has changed?

Kasparov signalled the future immediately after his defeat. He told us Deep Blue had been playing moves that no human grandmaster would even have considered.

From different moves, the new generation has built a completely different way of learning and playing the game. Javokhir Sindarov, winner of this year’s candidates tournament, spoke about it in El Pais:

“It’s very rare that I read books. My coach insists on it, but I prefer that he teach me things at the board. For example, I’ve never looked up games of great champions of the past on my own, like [Cuban player José Raúl] Capablanca [who was champion a hundred years ago] or [Soviet player Mikhail] Botvinnik [champion in the mid-20th century], although I have seen some when my coach shows them to me.”

The article concluded:

“The logical deduction from that fact is that deep strategy, medium- or long-term thinking in a game, positional considerations… are now secondary, because computers have taught them to play in a different way.”

Learning differently means training differently. ChessBase ran a long interview with Sindarov’s trainer, Roman Vidonyak. The regime is high-intensity training. 8 hours a day for 10-14 days at a time. Sitting at the board working on 17 key skills. Without the support of a computer or any other device.

Contrast this with traditional chess. Strong players would spend hours studying a single position or working through opening variations. Traditional chess demanded deep understanding; AI has shifted the focus to calculation.

Sindarov will challenge for the chess world championship against Gukesh Dommaraju. Barely 20, the reigning champion is 6 months younger than the challenger.

Both are exceptionally strong players. They prepare and play differently but their absolute standard is not higher than previous world champions. Yet they are confidently beating older players who learned in previous eras. They have reached the highest level faster than ever before.

AI is not making us dumber. It is not replacing thinking, it is changing the way we think. Modern players can use AI to apply centuries of accumulated knowledge without spending years working through it piece by piece.

Learning and training differently has made it possible to use calculation in areas that previously relied entirely on judgement. It has changed the way the most advanced skills are developed, not just the basics. This is the mechanism by which a younger generation, brought up on AI, is defeating their elders.

What happens when the younger generation stops worrying about old style entry-level jobs and starts building AI-native skills and businesses instead?