Microsoft

Craig Mundie (Microsoft photo)

Craig Mundie, Microsoft's chief research and strategy officer, has taken over Bill Gates' responsibilities for setting the company's long-term technology vision. In a recent interview, he discussed subjects including the impact of the U.S. economic turmoil on Microsoft, the state of computer-science education, and the status of the company's online services initiatives.

Q: How do you expect the U.S. economic turmoil to affect research and development spending generally, and also at Microsoft?

Mundie: Well, we've always maintained a big focus on our R&D investment, both the "r" part and the "d" part. Even if we expect that the economy globally gets a little soft gets a little soft due to these problems, I don’t think it will have a material effect on our R&D activities. We may moderate certain investments in certain areas across the company, just to deal with the economic situation, but we’re still going to grow substantially this year in both people and our expected business growth. So I don’t think there’s going to be any first-order effect. As to other companies, I think it’s a bigger question. Many of them have not maintained, certainly, the level of research investment that we have. As things get tighter, it will be interesting to see whether they sustain their investments or not.

Q: What about just pure staffing? Will you continue to hire in Microsoft Research, for example?

Mundie: Yes, I think broadly, the company continues to expect to grow by a substantial amount this year, in terms of head count. We think right now, we’ll slow it down a little bit, in terms of rate, but we’re still going to grow. MSR will continue to grow proportionately. … The top management has decided in the last few weeks that we would just slow things down a little bit more in the course of the year until it becomes clear just what’s going to happen in the macro economy globally.

Q: What about government funding?

Mundie: Well, we’ve been broadly concerned for some years that the government has been to some extent reneging -- in my view, at least – on what their commitment to the public should be, which is to fund basic science and research in these key areas. As a percentage of GDP, ever since the collapse of the Soviet Union, in real dollar terms, we’ve seen a decline in government support for basic research. Companies like Microsoft can only take up so much of the slack on that. As corporations have basically diminished their investment in research, it puts that much more squeeze on that. The government shunted a lot of basic research money into particularly the National Institute of Health and other medical-related things, for the last 10 or more years, and you’ve seen a decline, for example, in the support for the National Laboratories. So I think we are in a situation in the United States where we don’t have the level of investment in government-funded basic research that we probably should have.

Q: In your demonstrations to college students, you’re showing how you see PC software on “client” devices combining with online services in the “cloud.” How do you see those trends affecting Microsoft’s business in the coming years?

Mundie: I’m personally super-optimistic about its effect, because I think you’re going to see a strong resurgence of the importance of thinking about how you use this incredible amount of computational power that exists in this diverse set of client devices. As bandwidth costs continue to actually increase, which I think they probably will, and there’s more and more data that’s going to be available, I think the ability for people to get personalized computing out of their client devices, to assist with the digestion of all the information that the world makes available makes available through the cloud services, will be increasingly important.

The other thing I think is going to be dramatic is this change from the old graphical user interface to what you might think of more holistically as a natural user interface. That won’t change the way you might write a Word document, but I think it will dramatically alter the way people deal with the man-machine interface. Not in the individual component ways that I think we’ve talked about for the last few years, where you say, I could add handwriting, or I could add speech recognition, or I could add some machine vision. The real question is how could you bring them together in a way that really does make the man-machine interaction a lot more like person-to-person. I’ve given some demos recently about a robot reception that we’re developing at Microsoft in the research lab. And it’s currently my favorite example of this holistic shift to a natural user interface that clearly, I think, in practice can only be computed by these advanced clients. The amount of ingestion of sensor data and the amount of processing to do that would be non-economic to do in some time-sharing environment. Even if communications costs were zero-latency and free. And they’re not.

Microsoft, I think, is very well-positioned now, because we’ve been certainly remaining focused on building our complete platform for the cloud services. When you look at that, and when you look at how we think the clients will evolve, and our very successful positioning over the last decade to be at least a player in every one of these evolving, intelligent client devices, I think the company is as well-positioned as we could ask to be.

Q: At the same time, you’ve got some pretty stiff competition in the cloud, from people like Google and Amazon, with its Web services. What’s the pitch you foresee Microsoft making to third-party developers to get them to embrace your cloud platform?

Mundie: It comes back to this question of whether you think of them as two things or one thing. My view is that the cloud plus the client, as it evolves, has to be programmed as one thing. Everything today, whether it’s Google or Amazon EC2, they’re all really saying, ‘No, that’s a separate thing. Go program that thing separately.’ So I think that the big step yet to come is when we come up with a programming model or architecture that allows you to think as a developer about the cloud plus the client in a more unified way.

Q: Is Microsoft working on that?

Mundie: That is what we’re doing. It will come gradually. Some of the things that Ray (Ozzie, Microsoft's chief software architect) will show at the (Professional Developers Conference) are a clear step in that direction, but not the end game by any stretch.

Q: One of the common concerns in education is computer-science enrollment. How are you feeling about that these days in the U.S.?

Mundie: I still am very concerned that there’s a fairly precipitous decline in undergraduate enrollments in computer science. I think there’s probably several contributing factors to that. At a time where I think, in the next decade, we’re likely to see the biggest requirement for transformational work in the core computing models themselves, we continue to see the computer science community largely tilted more toward putting computing in science, instead of putting science in computing. We call it computer science, but if you go and look at it, for the most part, we’ve been teaching people proved techniques for how to use computing as we know it to solve other scientific and engineering problems. While that’s clearly important, and a lot of the value is created there, it sort of has left us with a withering level of investment in the core concept of computing itself. And so when you begin to face these things that only come around every 15 or 20 years, where there’s been enough accretive change in the underlying technology base that you ought to think about redefining the computer, then you really start to come up short with the number of people who are really focused on that and the corresponding academic research.

Q: It has been a few months now since Bill Gates left day-to-day duties at the company. How have things gone, and have there been any surprises?

Mundie: No, we had that very carefully planned two-year transition. Things are progressing fine, and along the path we laid out. I think that very first conceptual partitioning between Ray and I, of how we would add Bill’s job to our collective jobs, has actually continued to be a workable arrangement. And so we’re setting here, three or four months into the thing, and so far, there’s been no real hiccups, and things are progressing OK.

Q: Have you yet had that moment, personally, where you were about to dash off an e-mail to Bill Gates, or pick up the phone and call him, and realized that was no longer the proper thing to do?

Mundie: No, I guess I had two years to warm up to the idea, and I haven’t relapsed.

RELATED VIDEO: Microsoft's new Tablet PC prototype.