Talks about MIPS, Cisco & Rambus and the role of Universities.
Universities can play vital role in the 'real world'
As the dean of the Stanford University's School of Engineering for the last three years, John L. Hennessy has been a vigorous proponent of allowing faculty and students to work in Silicon Valley companies. He speaks from experience. During a sabbatical leave from Stanford in 1984-85, he co-founded MIPS Computer Systems (now called MIPS Technologies Inc.), which specializes in the production of chips based on his research at Stanford. This is an edited transcript of his conversation with Mercury News Staff Writer Scott Herhold.
You encourage students and faculty to become involved in the private sector. A lot of traditional academics might distance themselves from such close cooperation. Why do you do this?
We do it because we see it as a way for researchers at Stanford to bring their research out to the real world. Technology thrown over the wall in the form of publications often hits a piece of barren earth and dies away.
We've also seen over the years that having faculty go out and spend a few years in industry -- and then come back to be faculty members -- makes them much better teachers and researchers. They bring a set of experiences back to the classroom about what's it's like in the real world.
Everyone knows the example of Andreas Bechtolsheim, who helped found Sun Microsystems, and Sandra Lerner and Leonard Bosack, who co-founded Cisco Systems. What are some lesser-known but equally significant contributions from Stanford faculty or students?
Well, a couple of others that are probably lesser known: Professor Mark Horowitz's involvement in Rambus. Rambus was a company that actually took a long time to build, because the fundamental ideas were quite revolutionary. Changing the DRAM industry is something that takes a long time to occur. But those ideas go back to research concepts that Mark had quite early. A similarly revolutionary change was Professor John Cioffi's work with Amati, which really pushed ADSL (asymmetric digital subscriber line, which moves data at high speed over ordinary copper phone lines)
What are the kind of mechanisms that Stanford uses to allow faculty to go out and do this?
We generally let the faculty take a leave from their position. They'll generally disappear for a period of one to two years. And they'll probably be around enough at Stanford that they can supervise their graduate students. But they'll primarily be involved in whatever endeavor they're in, often a start-up company. And then they usually come back. Occasionally, faculty members get enthralled enough with the external world that they stay there.
How does the sheer pace of the Internet age change the endeavor we've been talking about? Does it force faculty members to quicken the pace of their ideas, to shortchange basic research?
In one way, it's created lots of opportunities. We probably have more faculty on leave at start-ups than ever in our history. And I think that's because there's such an explosion of opportunity with the Internet, and a technology vacuum being created, where industry is almost pulling things out of the university as fast as they're being created. I actually feel that the normal entrepreneurial mechanisms of the valley -- entrepreneurs leaving from established companies -- do quite well in building what I would call the shorter-term focus companies, the companies that are figuring out what Cisco is going to want to buy three or four years from now.
I think the university is more important for generating paradigm shift companies. Even in our applied work, our research focus is probably 5 to 10 years out.
Without giving away any trade secrets, can you give me an example?
We've had a lot of discussion with the networking companies -- 3Com, Cisco, Sun, AT&T, Lucent -- focusing on questions like how the Internet is going to work when there are a billion users. How the protocol is going to work, how you're going to scale, how you're going to deploy a system that must be incrementally deployed, must be incrementally updated? That's a great set of problems for the university to be thinking of. And clearly, it's far beyond the horizon of what most networking companies are currently worried about.
Is it difficult to keep faculty or student attention on those long-term problems when there may be short-term opportunities that can turn people very wealthy?
We've certainly seen the impact of the Dot-com phenomenon, particularly among our students. Some students have decided to take advantage of short-term opportunities. We're a large graduate institution. A lot of our population is Master's students who are going to come and go in a year anyway. Some of the Ph.D. students have become infatuated with the technology. So those things will happen from time to time. But as long as there's some reasonable balance, and people try to focus on these long-term things, I think it will work out.
Are there any potential conflicts of interest that need to be addressed?
Certainly there is potential for conflict of interest. Probably the biggest problem we worry about is how to insure that faculty are choosing their avenue of research in a way that's not overly biased by short-term potential commercial opportunities. We try to cover those bases. Every faculty member does conflict-of-interest disclosure statements every year. We try to be especially careful where there's funding from an outside interest in which the faculty member may have some monetary or financial connection. We try to encourage people to keep these things as separate as they can.
We've been talking about successes. What about failures? What have your faculty learned from failures?
What I learned -- and I think a lot of faculty have found this -- is that you can start a small company with a great technical idea. You can recruit the best technical team in the world, and that's not enough. You have to build the rest of the company -- marketing, management, all those components of the company. And I think a lot of faculty are surprised how difficult that aspect turns out to be.
Who owns the technology? Suppose that a faculty member comes up with a good idea, presumably on university time. Does the university own the rights to the idea? How is that split?
The current university rules say that if the invention involves a significant amount of university resources, or if it falls under something like government funding, then the university owns the technology. So there are some things that people do on their own. I think Yahoo is the best example. Basically, it was a couple of graduate students hacking on something they found interesting. In the case of MIPS, or Silicon Graphics, the university owned the technology, and issued licenses to the companies.
Has that ever wound up a matter of dispute?
Sometimes there are disputes over what the licensing fee should be, whether it should be an exclusive license, or a non-exclusive license. In some cases, if there's a potential for conflict of interest, we'll sometimes opt for a non-exclusive license, to say, ''Look, this technology is available to all comers.'' In general, the university's Office of Technology Licensing wants to avoid exclusive licensing if it looks like there will be many potential customers for the technology.
By and large, it's worked because the Office of Technology Licensing believes that job one is getting the technology out, job two is getting an appropriate return to the university. If you switch those around, then the word ''appropriate'' becomes the subject of much debate, and can damage the ability to transfer technology. Then nobody's won.
www7.mercurycenter.com |
