In 1972, Bobby Fisher played Boris Spassky for the title of world chess champion. Game 6 of that series is arguably the greatest chess match of all time – Fisher played an unusual opening and later made a series of brilliant tactical moves that left Spassky with no choice but to resign. At the end of the game, the reigning champion Spassky joined the live audience (yes, there was a live audience for a chess match) in standing and applauding Fisher.
Beyond the Cold War backdrop, what made this match so extraordinary was the element of surprise. Fisher practiced for years and years against himself, playing both sides of the chess board, developing strategies in his mind. He didn’t test his tactics in other tournaments because that would have tipped his hand, so he had to calculate every possible counterattack by himself before risking his plan on the world stage.
And that is why Game 6 will never be replicated. For the last 10 years, computers have regularly defeated the best humans in chess. Grandmasters spend most of their time now in front of a computer, running simulations, testing theories, and devoting hours to rote memorization. Humans can’t see 40 moves ahead, but computers can. And when you look that far ahead the game of chess becomes a deterministic puzzle.
Some say that the number of potential moves in a game of chess is greater than the number of atoms in the galaxy. It is something like 10 to the 50th power – that is roughly a million, billion, billion, billion, billion, billions. It is a number so large that it might as well be infinite. But computers don’t need to test *every* move to figure out how to win. They just solve for the *best* move.
If a computer can reduce the near infinite game of chess down to one move, isn’t it only a matter of time before almost everything can be reduced down to the best move?
Take for example finding a soul mate.
In 2010, 17% of couples who married met on an online dating site. Today that number is 33%. At the current growth rate almost every married couple will have met online in 20 years. And to the extent married couples have kids, well then, in a strange way computers are breeding humans.
What about picking a home to buy? Or whether you should cook chicken for dinner? These feel like choices, but perhaps the best outcomes are knowable if you have the right inputs.
Or maybe all the inputs aren’t necessary, and all you need is to believe that computers are right.
For instance, whenever my wife and I drive we always have a debate about whether to use Google or take my “shortcut”. It is a quick debate and Google always wins. There is a certain satisfaction in knowing that every outcome was examined and the computer picked the best. And since it is impossible to know the alternative, Google’s best outcome becomes a self-fulfilling prophecy.
I am a little troubled that what makes us smarter also makes us predictable. Assuming we rely on a computer to make choices for us, a faster computer could arguably figure out what we are going to do before we do it. That faster computer would seem omnipotent and able to predict the future. This is already happening on Wall Street, as computers trade stocks against other computers, trying to outwit one another.
My hope is that all this technology will make us better humans. But perhaps we will regret losing the magic of making mistakes, and wish for a simpler time when we got stuck in traffic, married our high school sweetheart, and accidentally bought a house downwind from the sewage treatment plant. Of course, this assumes serendipity existed in the first place.