About 3,000 years ago in China, the board game Go was created, and today it’s one of the most complex games in existence. Two players take turns placing black and white stones on a grid, aiming to capture the most territory by surrounding their opponent’s stones.
Early Go computer programs aimed to make the best moves by storing vast amounts of prior games and expert playing strategies. But when developers created AlphaGo in 2014, they used artificial intelligence to take a different approach. While earlier programs relied heavily on strategies derived by humans, AlphaGo stored only the game’s basic rules and improved by repeatedly playing against itself. Sounds interesting, but how good is AlphaGo?
In March 2016, AlphaGo played Lee Sedol, one of the world’s best Go players at the time. During the second match, on move thirty-seven, commentators of the game felt there were only two plausible moves — one being offensive and the other defensive. But AlphaGo instead played a stunning third move. One commentator said, “Not a single human player would choose that move.” It seemed there was a glitch in the computer program. But AlphaGo won that game, and it turned out that move thirty-seven was not a glitch. It was a move so brilliant that it surpassed the present understanding of all Go players.
AlphaGo played this move precisely because it was not trained by humans. It became so good from playing itself, without human intervention, without being told the “proper way to play.” By merely observing patterns, it devised a strategy that was never contemplated during the game’s entire 3,000-year history. AlphaGo ended up defeating one of the world’s best grandmasters, not by knowing more human strategies, but by being unconstrained by them. In other words, AlphaGo won because it knew less than the grandmaster and was therefore able to see more.
AlphaGo teaches us that not all human advice is good advice. This chapter is about realizing that you will receive a lot of bad advice in your lifetime. You will be told, “This is how it’s done.” People will say, “I expect you to do this.” Sometimes the advice given is helpful. Other times, there’s a lot of flashiness but not much depth, which means it is well-packaged advice. It sounds good, but if you interrogate this advice, if you break it down, it turns out to be a terrible idea.
Your job is to develop a compass to distinguish between what makes sense and moves you forward from what is nonsense and drags you backward. This is not something you can delegate. You can’t ask someone to think for you. Remember, if you seek only to please others, you will end up making many bad decisions.
This article is an excerpt from Chapter 5 of Choose Your Work
Footnote
One commentator said, “Not a single human player would choose that move.” AlphaGo. Directed by Greg Kohs. Moxie Pictures and Reel As Dirt, 2017.
Dan Dobos writes about decision making, personal growth, human potential, fulfillment and helping people choose the work that they are meant to do. He is the author of Choose Your Work.