Arguably the most influential invention of the century, the integrated circuit infiltrates every aspect of daily life and has become indispensible to the industrialized world. No foreseeable invention appears ready to displace it. And yet the IC is young enough that one of its creators-Jack Kilby of Texas Instruments Inc.-is still with us.

At that time, Kilby was granted a freedom unusual in the late '50s and even rarer today. For the rest of his career, he would be allowed to choose the path of his work, even dabbling in solar-cell research at one point. Officially retired since 1983, Kilby has an office at the TI center that bears his name, where he can be found on most Fridays. We caught up with him on a recent midweek day and asked him to share his thoughts on the 40th anniversary of his landmark invention.

The industry's very exciting to watch, and I'm very glad to have had a part to play in it. But I think I know too much about how it's come about to think that any single individual is entitled to very much credit for it. What we see today is the work of tens of thousands of the world's best engineers working for 30 years.

I'm not in a position to judge it, but the changes are tremendous. There was a time when most of the semiconductors were made by a few large companies with very broad lines, each attempting to cover as much product area as possible. Today we no longer have what I would consider a broad-line semiconductor company. If Motorola redeploys the way they seem to be doing, that'll be the last of them.

What we have in its place are groups of specialists-even a company as big as Intel is in that category-and a tremendous number of small groups that have managed to split off areas of the business that I would have called impossible: CAD specialists, photomask specialists, package specialists. Much of the process development has shifted to the semiconductor-equipment companies, which is a major change. Originally, the semiconductor companies had to, or at least chose to, build their own equipment, because they felt that they gained a proprietary advantage that way.

The most interesting thing to me is to see the cost reduction in electronic functions. In '58, one silicon transistor cost $10, and you had to put a few dollars of other components with it to make a gate. Today that cost has been reduced by well over a million, and this has opened up many new fields that certainly weren't considered 40 years ago: replacing simple mechanical functions, for instance, such as timers and things of that sort, in addition to the more calculator- and computer-like functions.

Prices are going down faster than the costs, but that's kind of been the history of the industry. I'm sure there were people who panicked when the cost of a single transistor went from $10 to $1. Or from $1 to 10 cents.

For the system to keep working, it depends on a little magic. These cost reductions have to trigger enough increase in volume to offset the reductions. When that is slow to happen, it gets pretty sticky. People making DRAMs have seen that in the last few years. Those costs are now extremely low, and while the amount of memory in a computer is increasing, it hasn't gone up an order of magnitude, which is what it would take to offset that.

The business has been built on continuous change right from the start. It's hard to remember, but we went through six completely different types of transistors from 1950 to 1960. Change has always been a part of it.

Pox on singing neckties
There are a number of things I could do without. Musical Christmas cards. Neckties that play a tune. And I guess that if I really knew all of the places where microelectronics were being used, I could name some more. But you don't know these. It's so pervasive that there are all kinds of applications that aren't even visible anymore. I can understand there might be 30 microprocessors in my car.

I don't have the foggiest idea what most of those do.

I think the cost of the IC will continue to decrease, and [use] will continue to expand. At the moment there's no replacement in sight, and so I would assume that people will continue to look for new applications.

Actually, the personal computer is just the latest of things. We were fortunate that there was a whole series of products that came along at about the right time-the calculator, games, what TI called the home computer and then the personal computer. Each of these was about an order of magnitude bigger than before, or at least an order of magnitude, so it gave the appearance of almost continuous growth. It really wasn't. It was a series of discrete products.

It's getting increasingly hard to find a product that will expand the market by an order of magnitude. That wasn't too hard when the calculator or the wristwatch came along, because the markets were small. Today they're gigantic, and it's probably not reasonable to expect that a single product will have that kind of impact again. It may change from these blockbuster new products to a steady succession of a multiplicity of products. I suspect that won't stop.

As for the PC industry slowing down, that's Mr. Gates' fault. For a time, the number of new programs being developed for new applications produced a real need for a more powerful PC, more memory, faster disk, more sophisticated logic. In the last few years, that's sort of tapered off. I suspect the time will come when people come up with new functions for PCs to do that will expand the market in a couple of ways. One is increase the number of people who think they need PCs. The other is to increase the need for more sophisticated PCs.

Probably the most crucial invention was the invention of the transistor, which made all of this possible. But I've been very pleased to see integrated circuits grow, progress and spread as the cost has been reduced.

I guess I think I started something worthwhile, but the changes have been so great that it's almost hard to recognize. I don't know that it's possible for a single individual to keep up anymore. If it is, I'm certainly not that one.

On Robert Noyce
TI and Fairchild battled for years to claim the IC as their own invention. In the end, a compromise deemed Kilby and the late Robert Noyce co-inventors. EE Times asked Kilby for his thoughts on the controversy.

[Noyce] was a very remarkable man. I considered Gordon Moore a close friend for many years and still do. I didn't know Noyce as well, but they've done a tremendous job [with Intel]. Companies have a vested interest in advancing their position and tend to take an interpretation of the facts that favors them-usually with enough truth that it's very hard to separate these things out. You shouldn't look at those things, I don't think.

Return to: IC @ 40