The oft-quoted proof of the rise of machines making human work obsolete is games in which humans lose to computers. This happened in checkers in 1994. It happened in chess 1997. Now computers match humans in Scrabble, backgammon, poker, and even Jeopardy. There is still one exception, “Go”. Why is that? What is so special about Go? The game is similar to Chess in many ways, it is a “deterministic, perfect information game”, meaning a game where no information is hidden from either player, and there are no built-in elements of chance, such as throwing a die. But there are some interesting differences.
For the first move in chess, the player has twenty choices. In typical chess positions there will be around 30-plus possible moves. A typical game lasts about 40 moves before the resignation of one party.
Go players begin with a choice of 55 possible moves. This number rises quickly and soon almost all of the 361 points of the board must be evaluated. Some are much more popular than others, some are almost never played, but all are possible. That makes for 129,960 possible board positions after just the first round of moves. A typical game of Go lasts about 200 moves. As a game of chess progresses, as well as in many other games such as checkers, pieces disappear from the board, simplifying the game. Go begins with an empty board. Each new Go move adds new complexities and possibilities to the situation. The key here is the number of choices available.
The more choices there are, the harder it gets for computers.
The industrial logic was most vividly captured in the idea of the value chain. Value-creating activities were sequential, unidirectional and linear. Those performing the following task must comply with the constraints imposed by the execution of the preceding task. The reverse cannot normally take place. The architecture consists of tightly coupled tasks and predetermined, repeating activities. The output of one task was the input of another. If-this-then-that. Work was algorithmic.
Workers in industrial-age firms were used to the rules that limited choices. The burden of decision making, with the consequent need to communicate and gather costly information, was minimized. Furthermore, by narrowing the scope of choices, the learning requirements for workers were limited. In part, the efficiency-enhancing contribution of mass-production was derived from these lower learning costs.
Work has been designed as a very, very simple game.
Is it then fair to draw the conclusion that the microchip may well replace the human race? Or have we just designed human work plain wrong? Could we, and should we, change the rules of our game?
The most important reason why we need a new concept of work/games is because the players and their contributions in the real world are, at best, too diverse to rank. They are, and should be, too qualitatively different to compare quantitatively as labor. Unlike mechanical systems, human systems thrive on variety and diversity. An exact replication of behavior in nature would be disastrous and seen as neurotic in social life.
The problem we face today is not in the capabilities of humans but in the outdated and limiting conceptualization of work. Work as we know it is mainly designed for machines, not for human beings.
Human life is non-deterministic, full of uncertainty, unknowns and surprises. Creative learning is the fundamental process of socialization and being a human. For a human being, the number of choices or moves in the game of life, in any situation, is unlimited. This is the very hard to copy difference between men and machines.