The Game of Go Is Dead, AI Buried It

In 2016, AlphaGo defeated Lee Sedol. The world’s best Go player lost to a machine. Tech journalists, venture capitalists, and Google’s PR department celebrated as if they’d just cured a terrible disease. A game that required five thousand years of human wisdom, boiled down to a neural network. Brilliant. Time to party.

Except Go is dead now.

Everyone sees it. They just prefer to look the other way. AI dominates training, opening moves are homogenized, top players match AI recommendations. This is called evolution. Progress. Democracy, go ahead and use that gray matter between your ears. This is colonization wearing a smile. AI doesn’t offer “Go insight,” it offers pattern recognition. Players replicate moves without understanding them. They’re human printers.

Competition structures force everyone to do the same thing: follow the algorithm or lose. DeepMind and Google say: free training for all. What they mean: you’re dependent on our oracle. Dependency grows in lockstep with opportunity. You’re free as long as you obey.

Character formation is gone. Go used to be deliberation, failure, wisdom forged by destroying yourself on the board. Now you memorize answer keys. Players feel it but don’t say it. They can’t say what they feel anymore because their thoughts aren’t theirs.

Lee Sedol understood what he’d become: an oracle follower who no longer recognized himself. So he quit. Not out of sadness. Out of disgust with himself.

The question Go won’t ask itself: do we want human excellence or efficiency? Perfection in pattern recognition demands the loss of everything that made Go great: uncertainty, defiance, the courage to be wrong.

DeepMind made Go an offer. I’ll make you perfect, you give me your soul.

Go said yes.