How Deep Learning Took Over
In late 2012, Geoff Hinton arrived at a computer science conference in Lake Tahoe to sell a tiny startup with no products, no customers, and only three employees. He was sixty-four, dealing with a severe back injury, and spent the negotiations standing over a laptop balanced on an upside-down trash can. Yet the companies circling him were among the most powerful in the world. Google, Microsoft, Baidu, and DeepMind all wanted the same thing: the ideas behind deep learning.
Hinton had spent decades defending neural networks, computer systems that learn patterns from data instead of following detailed human instructions. For much of his career, that choice made him an outsider. Most of the field favored symbolic AI, where engineers wrote rules by hand. Hinton kept betting that bigger data, better math, and faster computers would eventually make learning machines far more useful than rule-based systems.
The auction unfolded over email, with bids jumping by at least $1 million at a time. The price climbed past $40 million, and the scene grew stranger by the hour. Microsoft worried Google might be intercepting messages. Hinton and his students watched offers arrive in a Gmail inbox while sitting in a hotel room that looked nothing like the birthplace of a technological shift. He finally stopped the bidding at $44 million and chose Google, not because it offered the highest possible price, but because it gave him the best place to continue the work.
That sale marked more than a business deal. It announced that the center of artificial intelligence was moving away from systems built by hand and toward systems trained by experience. The shift quickly spread into speech recognition, image search, translation, self-driving cars, medicine, surveillance, and military tools. It also brought new dangers. These systems absorbed the strengths and flaws of the data used to train them, and the scientists who once worked quietly in universities were suddenly pulled into the middle of a global struggle over money, power, and control.
Full summary available in the Readsome app
