InformationWeek: After a career in semiconductors, you're on the board of ZettaCore, which recently added a new management team from the semiconductor field. How did you become interested in this company?
Vadasz: ZettaCore is developing technology in molecular memory. Molecular electronics isn't a field on its own, but it can augment the semiconductor industry in certain areas. We're adding to that industry to overcome certain limits, and it's part and parcel of the future of the semiconductor industry. Nanotechnology isn't about silicon replacement--at least not for a long time. It's about augmenting what we have in semiconductors, and also in areas like textiles. To get data in and out of these nanostructures, we'll have to work with silicon.
InformationWeek: Intel doesn't invest as heavily in long-term, basic research under its own roof as some other technology companies. Why is that?
Vadasz: There is basic research at Intel, but you can't do everything on your own. The all-encompassing, centralized research model is gone. Long-term research is best done in a university environment, and advanced development is best done in a corporate environment.
InformationWeek: That sounds like what Intel chief technology officer Pat Gelsinger would say.
Vadasz: Pat is a good student. You need to make smooth transitions from research to product development. The best long-term researchers are at universities, and funding that research allows you to understand multiple ideas at the same time. It's more economical. [Intel VP] David Tennenhouse runs a research council with hundreds of university programs. That kind of work is much more prevalent today than 10 years ago.
InformationWeek: There are some large changes in the offing in the way supercomputing will get funded by government, as the National Science Foundation brings out a new budget agreement with its supercomputing research centers. What effect will that have on the push toward developing a so-called petaflop computer, capable of processing a quadrillion math operations a second?
Vadasz: The [federal] budget is way too tight for all early IT research. We always have pet projects, and now they're obviously security related. Once, it was supercomputers for atomic research. We're still living too much off research that was done 10 to 15 years ago. As far as supercomputing, the life sciences will drive that--genomics and so-called drug design. That's absolutely the driver.
InformationWeek: Intel's had some success selling its Itanium chip in scientific- and technical-computing markets. For example, the second-fastest supercomputer in the world, an SGI machine at NASA, uses Itanium. Will the chip ever be a success in business data-processing markets?
Vadasz: There is some success with high-end transaction processing. But technologies take time to take hold. Look at computer video. That's starting to happen now, but it's taken 10 years.
InformationWeek: Or voice over IP. I remember seeing those demos in the mid-'90s.
Vadasz: Yes. I give about five lectures a year at Stanford in areas where there's new technology and a regulatory mess. I've talked about voice over IP, broadband, and open-source software. With voice over IP, there's a coming debate about what constitutes a link to you, versus a link that was a telephone line. You can put all kinds of services on top of that new link. But we're trying to straightjacket everything into a 100-year-old definition of universal service.