David Silver Leaves DeepMind: AlphaGo's Creator Seeks Superintelligence

The world of AI has just lost one of its most stable constants. David Silver, a man whose name at DeepMind meant no less than the name of Demis Hassabis, announced his departure. If you remember that legendary 2016 match when AlphaGo crushed Lee Sedol, then know this: it was Silver standing behind it.

He didn't just write code, he spent decades proving that Reinforcement Learning is the only true path to real intelligence. And now this path is taking him away from Google. Why is this happening now?

Context matters more than dry facts here. Recently, DeepMind has transformed from a cozy laboratory for geniuses into a massive corporate workshop, merged with Google Brain. When structures become too large, they start to slow down.

Silver, who always preferred elegant algorithmic solutions over endless "grinding" of data on servers, clearly felt the squeeze. His departure is a clear signal: the most ambitious tasks today are solved not in open offices with thousands of people, but in compact groups of fanatics. Silver's goal sounds modest—superintelligence.

But unlike the marketing promises of most Silicon Valley startups, Silver has something to back it up. His work on AlphaZero and MuZero showed that systems can learn without human input, inventing strategies beyond our imagination. His new project will likely be built around the idea of "pure" learning, where AI doesn't just mimic human text, but learns the logic of the world through interaction with it.

This is fundamentally different from what we see in modern chatbots. This outcome reminds us of recent actions by Ilya Sutskever, who left OpenAI for his SSI project. We are witnessing fragmentation of the "old guard" of AI.

Those who laid the foundation of the industry ten to fifteen years ago are one by one leaving their gilded cages. They take with them not just knowledge, but reputation, which will draw the best talent and billions in venture investment. Google, of course, will survive, but without Silver, the company's research department risks losing that very spark of mad innovation that made them leaders.

What does this mean for us? We will likely see a new player emerge that won't play by the rules of "release a model quickly so stocks go up." Silver always played the long game.

If he decided it's time to step out of Google's shadow, it means he's already put together the puzzle of what the next qualitative leap in technology development should look like. We are entering an era where individual geniuses matter more than the computing power of cloud giants. Key point: the era of DeepMind's dominance in fundamental research has officially come to an end.

Will Silver be able to repeat the success of AlphaGo alone, or will superintelligence prove to be a tough nut to crack even for him?