One way to “teach” a computer to play chess is to simply program it to know all the legal moves and give it a goal (in this case, to take the king). If you have even mildly decent chess skills, you’ll most likely be able to beat that computer if you play a game with it.
Another way to teach a computer to play is to program it to play itself over and over and instruct it to track the likelihood that each move leads to a win or a loss. A few days and many games later you’ll have a computer that even a chess master would have a hard time beating.
In the first scenario, the computer follows instructions. In the second, the computer learns.
Our greatest successes often aren’t products of our ability to follow instructions. Our greatest contributions much more often come from what we’ve learned.
And while we humans are incredible learning machines, a computer has two distinct advantages over us:
First, it has a flawless and instantly searchable memory. Our brains can’t do that. Fortunately, we don’t need to—that’s what we have the computer for.
The second disadvantage is one that the computer can’t help us with: we’ve got egos. Unlike most of us, the computer is not concerned about short-term win/loss records. Instead, the computer “cares” about learning.
Just like us, the computer only learns by looking backward. It can’t know if each new move will lead to a win until the end of the game. But at the end of every game, it’s more prepared for the next one. As far as the machine is concerned, a loss is equally valuable to a win. The end result is the same: new learning. The computer doesn’t need wins—it needs more games.
Unlike computers, we’ve got a bias towards winning, and of course—in the long run—we should all want to win. But in the short-run, we’d be much better off looking for more games.
Too often, both in the short-run and the long, we don’t take the time to look back and mine our experiences for learning. We’re too distracted by the pain of the loss or the self-congratulation of the win.
Even worse, too often we don’t play the game at all—avoiding a loss seems safer than risking a win.
If you’re playing the game using only the rules you know and playing with a mindset of loss-aversion, you’ll still get some wins mixed in with your losses. But you’ll never become a master.