TEXT ONE

A MACHINE THAT THINKS*

Computers, the electronic brains behind intelligent metals, miraculous foods and replaceable organs, are growing smarter by the minute. According to Moore's Law (propounded by Gordon Moore, cofounder of Intel), they get twice as smart every 18 to 24 months. They are already 3,300 times smarter than when microprocessors were introduced in 1971. Paul Horn, senior vice president of research at IBM, says Moore's Law will apply for at least another 15 years. This means that when country singer LeAnn Rimes is 31, computers will be nearly 200 times more powerful than those we use today—or roughly 660,000 times more powerful than the first microprocessor.

If that doesn't boggle your imagination, consider this: We're fast approaching a time when computers will design themselves, with no help from us lowly, jabbering primates. As computer scientist Carl Feynman puts it, "We've discovered that the gray mush in our heads is probably not the best material for thinking." Computers will soon be using their superior "thinking material" to make themselves even smarter. Humans and computers are already collaborating on the design of new generations of computers, but, Horn says, sometime down the road, people will be phased out of the process.
At that moment, our relationship to our world will change more profoundly than it has at any time since Homo sapiens made friends with fire. For eons, human beings have been far and away the most intelligent entities on earth. But this era is coming to an end. If you are a human and not a computer, you may be wondering: Is this good? No one has any idea. We are crossing a Rubicon—perhaps into paradise, perhaps into hell, definitely into a life our pitiful brains are incapable of anticipating.

Should we put the brakes on while we still have the chance? Forget about it, unless you're willing to use a typewriter while everyone else uses Microsoft Word, or make your kids amuse themselves with tic-tac-toe while their friends play Nintendo. "The competitive advantage—economic, military, even artistic—of every advance in automation is so compelling that passing laws, or having customs, that forbid such things merely assures that someone else will get them first," writes Vernor Vinge, a computer science professor at San Diego State University. Every day, computers produce some new improvement in our lives that we find impossible to resist. And every day we fail to resist, they come closer to cutting themselves loose from us.

An IBM computer called RS/6000 SP is a compelling symbol of what's to come.

Better known as Deep Blue, it beat the world's greatest human chess player, Garry Kasparov, in 1997. And that's only a fraction of what it can do. Its talents have also been used in such pursuits as engineering, air traffic control and financial analysis.

Deep Blue doesn't approach any of its tasks the way people do, but make no mistake: It's a machine that thinks. To deny this, writes Yale computer science professor Drew McDermott, "is like saying an airplane does not really fly because it does not flap its wings".

Perhaps there's nothing to worry about. The potential upside of living with Ubercomputers is fantastic to contemplate. Computers will warn us about threats to our health long before doctors can, enable us to exchange business documents by simply shaking hands, help us understand foreign languages without learning them.

And how they will entertain us! Forget about movies on demand; we'll have experiences on demand. Want to talk philosophy with Plato, match wits with W.C. Fields, play a game of Battleship against Lord Nelson? Entirely possible. Horn, a jazz enthusiast, relishes the thought of someday jamming with Charlie Parker. "The computer will react to my playing in the way Charlie Parker would have done," he says. Sounds great, until some spoilsport asks if it's not possible that Bird-the- Computer might be so revolted by his playing, it will simply cover its cyber-ears and order him to get lost. Horn laughs. "That's exactly the point," he says. "Something just like that might happen."

That's the downside: These new computers might not like us very much. In fact, they may well regard us as we regard poorly trained dogs. Scientists like Feynman are searching for ways to increase the chances that computers of the future "don't have malevolence, like their creators." But, he cautions, it may not be possible.

Here's another caution: Predicting the future is a fool's game. Countless thinkers of the past have tried, only to miss the mark by a billion gigabytes. Consider this effort by H.G. Wells in 1939: "After the second world war [there will remain only] a patchwork of staggering governments ruling over a welter of steadily increasing social disorganization. It will be the Dark Ages over again. There will be a return to primitive homemade weapons, nonmechanical transport, a new age, if not of innocence, yet of illiteracy, and slow, feeble and less lethal mischiefs will return to the world."

On the other hand, it's also possible that all these predictions underestimate the changes that lie ahead. Lew Platt, chairman and CEO of Hewlett-Packard, tells a story about ENIAC, the first general-purpose computer, built in 1946. After it was introduced, Popular Mechanics reported that "ENIAC is equipped with 18,000 vacuum tubes and weighs 30 tons." And then the magazine took a wild flight of fancy, daring to suggest that on some extraordinary day in the future, a computer of equal power could be-constructed with "only 1,000 vacuum tubes and perhaps only weigh one and a half tons." Today, says Platt "the computer power in those thirty tons of ENIAC can fit in your pocket ... in fact, your pocket calculator is more powerful." (4 620)