Gates goes to college

CAMBRIDGE, Mass.--Despite growing competition, slimmer technology budgets and an ongoing antitrust battle, Microsoft continues to dominate the software industry.

But Bill Gates, the company's chairman and chief software architect, sees a new challenge on the horizon: a dwindling pool of computer science students.

Last week, Gates embarked on a recruiting tour of sorts. He visited five colleges steeped in computing history to deliver hour-long talks: the University of Illinois at Urbana-Champaign, Carnegie Mellon University, Cornell University, the Massachusetts Institute of Technology and Harvard University. His message was in part a history lesson for those too young to remember the industry's roots and in part a call to action for the nation's best and brightest students.

Gates has a lot at stake. His company, with a research and development budget of some $6.8 billion a year, increasingly relies on innovation to keep it ahead of growing commoditization in the PC software market. His talks were designed to rekindle interest in computer science programs, potentially boosting the number of qualified graduates headed to Microsoft and other tech giants.

Gates sat down with CNET News.com on the campus of MIT to discuss why computer science programs have fallen out of favor, how Microsoft views the offshore-outsourcing phenomenon and what Microsoft--and the industry at large--is doing to attract talented people

Q: Why you are doing this campus tour now? What's the urgency?
A: Computer science is at a stage at which it can do more for the economy, more for the way business is done and more to solve the great problems that have been discussed for a long time, like speech understanding, vision and solving the very tough problems related to security and reliability that are important for us to get the full potential out of the work we have done.

We're at the time when those problems are on the verge of being solved, and we just need lots and lots of great people.

I think the United States has got to rededicate itself to letting the best and the brightest not only to be developed here but also to come here.

The buzz, the expectation level, is lower than when there was all of this hype, and people are thinking that there aren't such tough problems to be solved--that they'll just get resolved overnight. And virtually all the things that happened there in the late '90s pointed out this need to get to this next level.

So as I am speaking to these computer science students, I'm showing them my excitement, my optimism, and I am explaining how their work is going to be critical to the advance in other fields, including sciences like astronomy or biology. I'm also discussing how computer science plays into how businesses understand and prepare for what is going on and how they make decisions. Or how it helps people at home get more flexibility to organize their memories or stay in touch with people or be creative.

The results that we have in Microsoft Research are, you know, sort of exhibit A of all this. Patience and the quality of people we have been able to attract are making a big difference. The collaborations we have with the top universities are what make me so optimistic about the coming breakthroughs.

Why do you think that fewer students are enrolling in computer science classes and programs in general?
I think that it's a combination of things. I don't think we are as good as we should be at pointing out that these jobs are very social jobs, for which you are working with lots of people, and that collaborative skills are very key. A lot of these jobs, like program management, don't necessarily involve just sitting there, working on hard-core code--though there are plenty of jobs that are of that nature.

I think that kids want to have an impact. We need to articulate that computer science, other than perhaps biology, is the area where the greatest impact and change in the world will come, and that between this and biology, about 90 percent of the way the world changes is right there.

We need to have heroes, people who make computers usable by handicapped people, people who advance the security mission. When I have had talks with college faculties, I ask if people know the Turing Award winners and whether they think there is enough visibility for the advances that are taking place.

I'm also trying to cast great people in universities or great people in Microsoft Research as role models. Certainly, some of the computer science drop-off is related to the end of the dot-com boom notion that free money was available--just compare today with '98, when we saw a bulge in response to that former notion. But if you look at China and India, they're still on the increase.

The top work force and the great research is still very much in the United States, but if you look 20 years ahead, there will be some challenge to that. But we are going to be doing the lion's share of our development work in the United States, because we have the best work force here, and the job opportunities for people in computer science will continue to be phenomenal.

You know, there is a worldwide shortage of great computer scientists because of the seminal role their skills will play across so many activities and sciences. When the United States asks, "Wow, how do we stay ahead, making sure that we are taking these great universities we have and the few companies that do invest in long-term research? How do we make sure that we are doing even better at that?"--it should understand that those questions are key to the path forward; the rededication to that uniqueness.

Craig Barrett, Intel's CEO, says the U.S. educational system does not teach students enough science and mathematics and is geared more toward producing lawyers and consultants. What's your opinion?
Relative to Asia, he is right. And you walk the halls of Intel--not Microsoft as much--the same trend is there.

We need to have heroes, people who make computers usable by handicapped people, people who advance the security mission.

A lot of tech's talent pool is comprised of people who are educated, at least in part, outside this country. And even if people come in as immigrants, there's a notion that the sciences are more objective; that your connections (or lack thereof) and accent may not hold you back like they might in law, politics, sales and other career tracks.

Immigrants have always been disproportionately represented in the sciences. The proportion of the great U.S. science that was done by people like John von Neumann and Alan Turing is pretty unbelievable.

What is Microsoft's position on offshore outsourcing? How do you decide what development goes on within this country and what development goes on outside the country?
It is interesting. We found it easier to do research in multiple locations than to do product development in multiple locations. And we will push some product development projects to India and some to China, but the lion's share will stay where it is, because we think that the best work force is here.

I think that the United States has got to rededicate itself to letting the best and the brightest not only be developed here but also to come here. Essentially, you not long ago could predict somebody's job opportunities more by what country they were in than by anything else. And it literally is the PC and the Internet around it that has gotten people to say, "No, these communications networks and collaborations software mean that the greatest determinant of your job opportunity should be your education level."

It's not that unskilled workers in India are getting high-tech jobs.

The United States will have to say, "What it is that we are going to continue to be the best at?"

But college graduates in India are getting jobs, not just in software development but in call centers and you name it--architecture, even some types of legal work. Any job can now be done with modest driven efficiency at a distance.

The world is therefore tapping into more human skills and innovation, and the lower prices that come out of that will be great for everyone. The world is just better off. Some of these impoverished countries will now participate in the positive development cycle, and that is great for everyone.

The United States will have to ask, "What it is that we are going to continue to be the best at?" We have to rededicate ourselves to the importance of intellectual property and a better education system. In the 1980s, we had this angst about Japan--that it would just go after industry upon industry and take over. We can look back at that and say, "That was wrong." But it actually created a level of humility and a rededication to not just matching the Japanese model but emphasizing what was different and better in the United States, including our research universities.

It created this sort of healthy paranoia.
Right. We really questioned whether our cost structure and our focus and efficiencies were correct. It was actually during the 1980s when we had this angst that the foundation was laid for the great advances in chips and personal computing. Communications in the 1990s benefited the entire world--the United States disproportionately--which is impressive, given that it's already the richest country in the world.

Today, some of these global competition issues surrounding the many educated people in other countries should make us think about our role, the key principles around research funding and the great education that will let us keep our unique position.

Does that put responsibility on the government to put in place favorable legislation?
Certainly, because this issue's big. Politicians and voters will have to think hard about which way to go. Some would say, "Let us get out of this world trade system.

I dare Wall Street to come up with a demo with which they can say, "Hey, we will change your world and make it a better place, and you can help handicapped people."

Companies are giving us products that are too inexpensive while too high-quality. Consumers then want to buy them to improve their lifestyle. Let's not let people buy those things. Let's go backward on free trade," versus others who'd say, "OK, how do we raise our level of competitiveness? And everybody in the world is going to get rich and that is OK."

It isn't like war, for which you have a winner and a loser. In this case, everybody can win, but you need to have some specialization. We are always going to have a higher cost structure than those of other countries. We therefore must have to have some excess; some productivity to make up for that. Some people have been far too concerned, not recognizing our edge, including the immigration piece, the education piece and the intellectual-property piece.

What technology and innovations is Microsoft thinking about that will draw in college students and people who may be thinking about leaving the computer industry? What will Microsoft do to get them excited? Will that have the same effect as open source and Linux in generating enthusiasm?
Had what effect?

Getting people interested in this technology and this movement; feeling that they are changing something and that they can have a part in developing something new.
That is a very narrow phenomenon. Seriously. Go mingle with these computer science students, and just think of the numbers involved here. There are hundreds of people who do that, yes. But computer science has moved on from operating systems and kernels--let's not kid ourselves. You don't even learn machine language anymore. Don't think that there is some correlation between people working on that stuff and what is going on with these students here. That's more and more overseas. It has nothing to do with the United States.

OK, so what is what Microsoft is doing in terms of innovation to get people interested?
If people sit down and see what we are doing in accessibility to let handicapped people such as the visually impaired use computers, they see the impact. We're advancing how computing can be used to make learning easier or more engaging. We're changing how graphics work both for serious work and pure entertainment--those graphics breakthroughs are so much dramatically better than what came before. We will be able to translate text so that anybody in the world can get at the information largely in English today.

We can tell students about how speech recognition can be used or show them digitalization approaches that let businesses better understand their data. Just demo this stuff, and people really get excited about it.

So there is a disconnect between all of that and what's really drawing more people in. There is a lot of very deep stuff that is critically important for the people who really want to work on the hard platform stuff, the invention required on software verification, new security algorithms, new compiler algorithms, etc. The opportunities are there, and people do connect to them; people do want to make an impact. I dare Wall Street to come up with a demo for which they can say, "Hey, we will change your world and make it a better place, and you can help handicapped people." I dare a law firm or a bread company--these can be wonderful companies, don't get me wrong--to improve the world. Doing so today involves some mix of computer science applied to many domains and to biology.

Will the computer languages we have now keep pace in the next five years, or is there some advance needed on that front?
The languages we have today are not good enough. The type systems aren't high-level enough, the way they bind to XML (Extensible Markup Language) is not direct enough, and the way they are oriented around message passing is not strong enough. In the academic arena, there are actually suggestions of how to solve every one of those things. And I am not saying that Java or C or Visual Basic or even COBOL (Common Business Oriented Language) are going to go away. These will continue to be super important languages.

COBOL keeps ticking along.
Yeah, absolutely. Because people have applications written in it. But there will be a lot of languages that respond to Web services and XML and high-level verification. Perhaps the security angle is the strongest.

		Get Up to Speed on... Web services Get the latest headlines and company-specific news in our expanded GUTS section.

But the data model for XML is not well matched with today's languages. Currently, we are doing some extensions ourselves. We and others will both extend existing languages, and new languages will come along. But it is super critical when we think of these environments that they not be oriented around one particular language; that they let the academic community or commercial people do rich language extensions and yet work with existing code and tools.

Several years will pass before commercial customers see Longhorn and a new version of Office. What does Microsoft do between now and then to drive revenue and keep corporate customers coming back?
The issue is doing exactly what customers want, and you know what they want is very focused on reducing their information technology cost and increasing security and reliability. Even in the next six months, we will have Windows XP SP2, and in the next year or so, we will have the Whidbey release, which is about a big advance in Web services and a big advance in letting people improve the properties of the applications they write.

We're taking all these things we have done of source code analysis and tracking tools and actually building those into Visual Studio so that developers get that same thing. A ton of the Whidbey release of Visual Studio is taking things that we have to invent for our own very demanding high-scale process and offering this for small-scale projects, where the same imperatives exist. We will also have a release of SQL Server. We will have an update of OneNote, we will have an update of InfoPath. And we will have a new server release that will likely come up before Longhorn.

So we have a lot of updates, but each update is tied to what the big customer issues are. We'll see increased uptake of Office 2003, and we will do more templates for Office 2003. We will see the uptake of XP in general and SP2 in particular getting out there.

Has the consumer part of Microsoft's business become more important, as IT spending has flattened in recent years?
No. We will always get the bulk of our revenues from business customers, particularly large enterprise customers, who are very demanding on making systems very manageable. The amazing thing is that eventually, when consumers have many machines in their home--which isn't now--they will effectively demand zero management. So everything we are doing to make these corporate customers delighted now--driving the management cost down and down and down--will ultimately take the form of what the consumer will demand: no management.

Having many servers, many different applications, is a high-bar goal today. But everything our primarily customer base--what is today and what always will be--drives us toward something super important for all the customer bases. Because the large customers have IT departments, they are more articulate about how they want the pieces to fit together. But security and management are not just unique to them.