The Idiot Savant

To purist philosophers, the very thought is galling, even faintly ridiculous. Deep Blue was made, no manufactured, by man and his genius. It is a contraption, an appliance, a contrivance. How could it possibly better its maker? How could an architecture of metal and silicon beat the sophistication of millions of years of evolution it took to arrive at the human brain, the finest machine of all time?

Back in New York, Deep Blue sits in an air-cooled chamber on the 45th floor of a building engagingly named Equitable Towers 8212; in downtown Manhattan, unaware of the tohu-bohu it is causing. It is connected to a terminal by a telephone wire. Across the terminal sits Garry Kasparov, brow sicklied over with the pale cast of thought, as the Bard would have put it. How would he the Bard have described Deep Blue for it has neither brow, nor, as some would claim, 8220;thought8221;? 8220;That wretched appurtenance?8221; Every time Kasparov makes a move, Deep Blue reacts, sifts it with its 32 parallel processors, puts it through a mammoth database, and responds as precisely as its computing power will enable it.

The computing power is enormous 8212; 200 million calculations a second, more than an average club player will attempt in a lifetime. Moreover, Deep Blue does not fret and fume, does not sweat, does not waste time, and does not grimace, get angry or dismayed. Its pulse and heart rate do not change because it has none and it is never distracted. It feels no pressure. But it does not have some of the things the man opposite to it has. It has no intuition, no creativity and no courage. It has been souped up with greater computing power and a bigger database which contains every Grandmaster-level game over the last 100 years which gives it speed and knowledge. But not the subtleties which only the human mind is capable of. Not the experience that comes from knowledge distilled into the subconscious. For instance, in chess terms, Deep Blue is still a sucker for gambits.

A gambit is a strategy in chess where a player loses a pawn intentionally to gain space, mobility and momentum at a later stage. It is a risk. Deep Blue does not take risks. It is a glutton for material and will not sacrifice a piece for a win. Only a human mind is capable of doing that act which constitutes the ultimate beauty in chess.

Deep Blue is an idiot savant.

The computer8217;s strength is its number crunching power and its application largely useless in areas of human endeavour. Deep Blue will recognise truth, but never beauty. Deep Blue also cannot learn from its mistakes. When the first generic Deep Blues an RS-6000 SP computer were made, they were installed in departmental stores to log millions of purchases. They made interesting correlations with the data they gathered: like 70 per cent of the people who came in and bought beer, also bought diapers. Smart, huh? But perfectly useless. Another correlation made by a Deep Blue in a Canadian store makes slightly more sense: a large percentage of people taking out marriage licenses also purchased chainsaws.

Must have been an accident.Of course, Deep Blue has its uses. To the dismay of our own nuclear non-proliferation experts, a Deep Blue prototype will simulate nuclear explosions so that the US can get around the test ban treaty. Transcontinental investors can log into Charles Schwab8217;s web site, run by a Deep Blue, and look up the latest on mutual funds and stock tips. But Deep Blue as a professional chess player? Pshaw! forget it. Leave that to the elegant humans.

Still, Deep Blue could beat Kasparov. But not because it is better than the Grandmaster in the marvelous intricacies of the game. It may win because Kasparov could make an error arising from fatigue or memory loss. Human frailties. In fact, despite being equal in scores at this point to Deep Blue, Kasparov has already exposed it as a silly little fool 8212; a 1.4 tonne number-crunching zombie capable of doing only what humans have instructed it to do. Kasparov himself has hyped the event by saying he saw signs of intelligence in the machine, but that8217;s what it is 8212; hype.

In almost every game, Kasparov has put Deep Blue in a quandary whenever he has gone into uncharted chess terrain, since these moves or openings are not instructed into the computer. In game 2 which Deep Blue won, it turned out that Kasparov resigned a game he could have drawn 8212; a human failing. Deep Blue did not win because it conjured a position Kasparov did not see. Kasparaov failed himself.

By this weekend, we will know if Deep Blue has beaten Kasparov. And well may it have. Big deal. So a Ferrari can cover a distance faster than Carl Lewis. And a 12-foot robot with swift mechanical arms can out-hoop Michael Jordan. Ha. As Kasparov himself put it: 8220;We humans know that there are many animals and machines that are faster, stronger and more agile than we. But none is smarter, more intelligent. In this area, we have enjoyed a monopoly.8221; True. Possibly forever.