The Mature Machine

A favorite dream project of mathematical thinkers is a chess-playing machine. None has been built that will play a full game without human help, but the development of monster electronic computers offers hope that they can be "programmed" (instructed) to match the best moves of a skillful human chess player.

The leading authority on the subject, Dr. Claude E. Shannon of Bell Telephone Laboratories, believes that a computer can play--theoretically--a perfect (unbeatable) game of chess. But on the practical side, no existing or projected computer is fast enough to make the calculations. In planning a typical 40-move game, he figures, the machine would have to make 10^1^2-o (10 followed by 119 zeros) calculations. Even at the lightning speed of electronic computers, the job would take 1090 years before the machine could make the first move.

It would be easier, says Shannon, to make a machine play a fair game of chess, seeing three moves ahead and avoiding obvious bad strategy. Such a machine would play rapidly and would have no mental lapses. It would never get lazy or nervous. On the other hand, it would lack flexibility, imagination and the valuable human ability to learn by experience. It would never beat a good player.

In a recent issue of Britain's Nature, Dr. J. Bronowski of the Central Research Establishment of the National Coal Board takes issue with Dr. Shannon. A chess-playing computer, he says, could be made to learn by experience just as a human being does. It could be given a memory of unlimited capacity. It could remember each move in all the games it had played. By classifying moves, it could determine which were most successful in each chess situation. It could even classify its opponents by the character of their moves. Eventually, says Bronowski, when the computer's memory finally bulged with remembered chess games, it could become a master player with a personal style of its own.

What would happen if two experienced chess machines played one another? The more experienced machine, thinks Bronowski, would always win. "In human life," he says, "maturity is always offset by loss of other powers. The machine, on the other hand, will become more and more experienced but will never lose any other faculties. That is the real difference between the human being and the machine in these circumstances. The machine can mature without growing old, getting better and better. So in this case, the most mature machine will always win, provided, of course, there is no mechanical failure. All good machines are likely to have nervous breakdowns."

This file is automatically generated by a robot program, so reader's discretion is required.