AlphaGo's Victory is a Landmark Moment in the History of Artificial Intelligence

Deep Blue didn't win by being smarter than a human; it won by being millions of times faster than a human. Deep Blue had no intuition. An expert human player looks at a board position and immediately sees what areas of play are most likely to be fruitful or dangerous, whereas a computer has no innate sense of what is important and must explore many more options. Deep Blue also had no sense of the history of the game, and didn't know anything about its opponent. It played chess yet didn't understand chess, in the same way a calculator performs arithmetic bud doesn't understand mathematics.

Jeff Hawkins - On Intelligence (2004)

Maybe the only significant difference between a really smart simulation and a human being was the noise they made when you punched them.

Terry Pratchett - The Long Earth (2012)

Human beings are excellent pattern recognition machines. That’s why we see the face of the Virgin Mary in trees and carrots, or can point out galloping horses in the clouds. It’s why your friends console you by telling you bad things always come in threes, and why people say that history repeats itself even when it doesn't. Pattern recognition was an evolutionary adaptation that allowed us to develop rules of thumb in order to navigate in a complex world. When this works well we often give it other names, like intuition. You could even argue that if, as physicists say “the observer creates the reality" then pattern recognition is its own form of consciousness.

That's why last weekend's victory by a computer over the world's best Go player is such a seminal moment. The reason everyone's excited (and why at least 60 million Chinese people tuned in to watch) is because AlphaGo doesn’t work like traditional AI, which relies on brute force computational power and reliable, large memories. Instead, it's based on an ensemble of learning techniques: deep learning, reinforcement and Monte Carlo tree search. These allow the software to explore a very large search space without having to cover every single possibility, and for algorithms which learn and improve through experience. This results in a type of interactive, recursive, pattern recognition that's based in part on the way our own brains work. It's why professional Go players were able to call AlphaGo's 37th move in Match 2 a beautiful act. 

AlphaGo in other words, is the beginning of real progress towards useful AI compared to the false start of chess. In the same way that we automated manual labour, we're automating pattern recognition. We're on the cusp of being able to mimic human intuition, flow and imagination. The potential applications are enormous. Imagine that kind of programming applied to real time routing of traffic in a large city? Or how about diagnosing genetic causes of disease? AI applied to speech recognition gives us universal translation. Imagine a high powered pattern recognition machine set loose on long term weather prediction?

Those are all specialist applications though. The real killer app is AI + the internet. Right now we’re still on the tail end of Web 2.0, an internet that evolved from an information resource into one that allowed us to become part of the process by posting content and connecting to each other. AI makes possible the next round of evolution, Web 3.0. This is an internet that escapes the walls of our devices and becomes a part of our physical environment. It's also an adaptive, intelligent web that provides a far more intuitive user experience based on your own unique way of using it. This stuff is already happening. There’s a website design company called The Grid that has 60,000 beta users (disclosure: we’re one of them) trialling an AI software that changes your website based on how users interact with it.

The move towards Web 3.0, and intuitive computing more generally isn't just about connecting up more devices, it's about a change in the software that runs them. The AlphaGo match has been an amazing demonstration of just how far that software has come. Our machines are starting to be able to recognise shapes in their own digital clouds. And once they get really good at it we're going to have to re-examine everything we thought we used to know about the nature of intelligence.