CeBIT 98: Gerstner Keynote P1 - Deep Computing

IBM Chairman and CEO Louis V. Gerstner Jr. delivered the following remarks during the keynote address for CeBIT 98 last night.

Here are some of the highlights from Gerstner's speech.

I've been looking forward to this evening for a long time, because I've known for many years how important CeBIT is to the global information technology industry. So before I go any further, I want to thank you all for inviting me to participate in this important forum.

So, tonight, while I'll talk about the power and potential of information technology, I hope to temper my remarks with the perspective I had when I came to IBM five years ago, the perspective of a customer. It's certainly easy to see why raw technology dominates these events. It's seductive. It's breathtaking. And it's penetrating every aspect of our lives.

As we stand here today at the opening of CeBIT, I believe we're at the threshold of a very important change in the evolution of this industry. This young, and in many ways, very immature industry, is about to play out in its most important dimension. That's because the technology has become so powerful and so pervasive that its future impact on people, businesses and governments will dwarf all that has happened to date. I believe there are two trends that are most significant, and bear the closest watching.

The first is what we call "Deep Computing." The term is inspired by our chess-playing supercomputer Deep Blue. Deep Blue is an amazing machine capable of calculating 200 million moves per second. But speed, while essential, wasn't enough. After all, Deep Blue's predecessor was also quite fast, but it lost to Kasparov two years ago. The difference the second time around was an infusion of knowledge, human chess knowledge, thousands and thousands of chess moves, games and outcomes captured as mathematical algorithms. This is what let Deep Blue mimic the workings of the human mind race through millions of possible chess positions and extract the best one. And it worked rather well. Deep Blue is emblematic of a whole class of emerging computer systems that combine ultrafast processing with sophisticated analytical software. Today, we're applying these systems to challenges far more vital than chess.

Let me talk about two important application areas, starting with simulation. Simulation is about replacing physical things with digital things; recreating reality inside these powerful computer systems. In the pharmaceutical industry, the ability to simulate the interaction of chemicals and do it with computers rather than test tubes and petri dishes, can speed up by years the discovery and testing of new pharmaceuticals.

Computer simulation saves time and money. It gives customers a competitive advantage. And it can do more than that. Recently, the U.S. Department of Energy asked IBM to build a gigantic supercomputer to simulate nuclear weapons so that they will never have to be exploded for test purposes again.

The second type of "Deep Computing" is what's known as data mining. Some people call it business intelligence or the ability to extract insight from mountains of information and see relationships and trends that were previously invisible. Banks are looking at spending patterns and other demographics to tell which customers will be the most profitable over the long haul. Healthcare companies are analysing millions of patient records to find hidden indicators of disease. These tools are helping slash the staggering cost of insurance fraud in healthcare, a 100-billion problem annually in the U.S. alone.

Insurance companies can now spot aberrant billing practices. One company in the U.S. has saved $38 million, after an initial investment in this technology of only $400,000. In one instance, they found a doctor sending in a bill once a week for a procedure typically performed once or twice in a lifetime.

We believe "Deep Computing" is one trend that will have a profound effect on commerce and society. Of course, in concept, throwing tough problems at big computers is nothing new. Its roots can be traced to the origins of this industry. The difference today is that these systems are far more powerful, and affordable enough to be used by businesses, governments and institutions of all sizes.

Let me cite a few examples, drawing on what we at IBM have learned from helping thousands of customers embrace the Net in the past year.

One of the most contentious, fast-moving and bare-knuckled battles is being waged today, believe it or not, in book selling. The leader in this online race is If you haven't heard of them, don't feel bad. Three years ago, nobody had. They didn't exist. Their customers still don't know where they're physically located, and they don't care. exists only in cyberspace. But with 2.5 million titles, it's nearly 15 times larger than the world's largest physical book store. Open 24-hours-a-day, every day of the year, and they recently served their 1-millionth customer, in Japan, one of 160 countries to which Amazon ships books. Until recently, they had the market all to themselves. Now the traditional book sellers like Barnes & Noble in the U.S. and media firms like Bertelsmann in Europe are jumping in. Can virtual companies like Amazon battle against and beat the entrenched brands?

The same kind of transformation is happening in retail banking, car sales, music and entertainment, and insurance. And it's not limited to the commercial world. Public sector institutions are being buffeted by the same powerful forces.

In higher education: There's a university in Canada: Athabasca University, that delivers 100 percent of its courses by what's called "distance learning." No students on campus. All instruction is delivered online. And they've captured nearly 30 percent of all MBA students in Canada.

All of us must realise this isn't a spectator sport. We aren't just watching Every enterprise and institution will grapple with these issues and at the highest management levels.

The toughest, most jugular decisions that need to be made aren't which browser? which server? They're core management and policy issues. This only escalates as the networked world marches on.