Author: Steven Edwards
Date: 08:10:03 04/26/05
Go up one level in this thread
On April 26, 2005 at 10:33:01, José Antônio Fabiano Mendes wrote: > Harvard Business Review, April 2005 It is a discredit to the HBR that they wold publish this tripe. >Speaking of analytic prowess, what was the significance of your famous matches >with IBM's chess-playing supercomputer, Deep Blue? > >For a start, they were a huge promotion for the game. Nothing made chess more >popular than the match I won against Deep Blue in 1996 and the match I lost in >1997. The official Web site got 72 million hits during the six games of the >second match in New York, which was a higher daily rate than the Atlanta Olympic >Games Web site got in 1996. Wrong. The Fischer era starting around 1970 had a far, far deeper impact on chess popularity. See the USCF annual membership numbers for proof. >But the matches meant a lot more than that to me. Competing with a computer was >first and foremost a scientific experiment for me. I thought it was very >important for society to start communicating with computers, and I knew that >chess was the only field where man and machine could meet. You can't do it with >mathematics or with literature. Chess, however, lies somewhere in between. I >believed that it would be an ideal playing field for comparing human intuition >with the brute force of a machine's calculation. If computer competition was so important to Kasparov, why hasn't he have a long and public history of playing against them? No, the big reason the matches were so important to him was the big money he was guaranteed just for showing up and pushing wood. >The yardstick of victory, I think, should be this: If the best human player--on >his best day, at his peak--can still beat the best machine, then we can say that >the chess master is superior to the machine. And for now, I believe that chess >masters like me still have the upper hand. I can beat the machine unless I make >a fatal unforced error. But when the chess master can no longer defeat the >machine on his best day, then we will have to take a cold, hard look at issues >such as artificial intelligence and the relationship between man and machine. Does Kasparov believe that chess playing programs are an example of AI? Most chess programs have about as much AI as a pocket calculator. >Unfortunately, I don't think everyone shared the same spirit of experiment. The >day after the New York match against Deep Blue, the one I lost in 1997, IBM >stock immediately jumped 2.5% to a ten-year high. It continued to rise >dramatically for weeks. For some reason, Lou Gerstner did not invite me to the >next IBM shareholders' meeting to take a bow! But seriously, I wish that IBM had >accepted my offer for a tiebreaker. To my mind, IBM actually committed a crime >against science. By claiming victory so quickly in the man-versus-machine >contest, the company dissuaded other companies from funding such a complicated >and valuable project again, and that's the real tragedy. Gerstner already wrote Kasparov a couple of fat checks and so likely (and correctly) thought his obligations had been fulfilled. How times change! Some 40+ years ago Arthur Samuel was an engineer (and a very good one) at IBM who developed a checker player that could give masters a rough time. It was a first for any board game. The program also pioneered many techniques that are used in today's chessplaying programs such as bitboards and position learning. What was IBM's stance towards Samuel? It was "We'll let you have machine time, but only for the purposes of testing newly manufactured hardware prior to customer shipping. Also, don't publicize your program too much as we don't want customers to be afraid of our computers." And yet Samuel contributed far more to the computer chess community than the Deep Blue project ever did. > > Did it hurt your pride to be beaten by a computer? > >No, not at all. Let me explain this by telling you a little anecdote. In 1769, >the Hungarian engineer Baron Wolfgang von Kempelen constructed a chess-playing >machine for the amusement of the Austrian empress Maria Theresa. It looked like >a purely mechanical device, shaped like a person. And it played chess very well. >But the machine was a fake. There was a chess master cleverly hidden inside the >device who decided all the moves. > >In some ways, Deep Blue was also a fake. The machine I played with in 1996 and >1997 had no history. Records of its past games were better guarded than >top-secret documents at the Pentagon. And since IBM refused to release printouts >of earlier games, it was impossible to prepare for the match. I couldn't feel >badly about losing because I wasn't playing on a level playing field. Lack of public games does not equate with fakery. Yes, it was not a symmetric relation as the Deep Blue team had all of Kasparov's past games. But Kasparov knew this prior to signing the match contract. Apparently the lack of public games didn't bother him then; his whining and false allegations shouldn't bother us now.
This page took 0.01 seconds to execute
Last modified: Thu, 15 Apr 21 08:11:13 -0700
Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.