Speech Transcript – Rick Rashid, Future Forum

Rick Rashid Comments
Future Forum
Redmond, Wash., Sept, 5, 2001

RICK RASHID: Hello. Let me bring something up here. So, I was feeling really bad that I wasn’t actually supposed to do any demos today, so I decided I would just pull my laptop out and put up one of the ones that we have. This is actually one you’re going to be able to see later today live, and some of you will hopefully get a chance to get your own faces put into a 3D form. But there I am. I’m waiting for the day when I can actually get somebody like that to actually do the presentation for me and then I can sleep in a little bit later.

Now, see, I can make this guy smile, which is always a good thing. I can make him very sad. I don’t really like to be sad, so we’ll get rid of that.

This is the really valuable thing, especially for executives: You can make yourself look like you’re thinking. (Laughter.) This is an extremely valuable tool, especially I think this would be really great for meetings in the future.

Now, I’m a big science fiction fan, and especially Star Trek , and I’ll mention more about that later, but here is me as a Dr. Spock-like character. And then for those of you who remember E.T. , you know, you can be E.T.

And so I just wanted to put that up. It has nothing to do with my talk, but it is — you know, I feel bad. They didn’t want me to do a demo, so I felt I had to do something.

And as I said, you’re going to be able to — some of you, at least — there won’t be time for everybody to do it, but some of you, at least, should be able to do that, get your faces done live later today.

So let me get into the actual talk itself. You know, if you think back to 1990, Microsoft didn’t have a basic research organization. There was no Microsoft Research. And it’s not surprising, because Microsoft was an awfully small company back then. I mean, we had just been crossing about a billion dollars in sales in that time frame. There were somewhere around 5,000 or 6,000 employees, again a very small organization, I mean roughly a 10th of the size of the company today. And the key products people tended to associate with Microsoft were pretty rudimentary; I mean, DOS, Windows 3.0, Word, Excel, PowerPoint. We were just sort of beginning to think about doing things in an Office framework in those days. So it was a very, very, very different company.

And here is a little — you can see up on the screen there a shot. This is off of the Terra Server, which is something our research group put up a few years ago with images of the earth’s surface, and this is a U.S. Geological Service image. It’s a one-meter resolution. And this is the old campus and there are buildings, you know, 8, 9 and 10. You can see building 6 and I think that’s 6 and 4 there, and 3, I think, off to the side.

And one thing that I notice when I look at this picture is all the trees. They’re mostly gone now. They’ve been replaced by more buildings. So the company has really changed.

I came to Microsoft in 1991 and really I came to Microsoft to create a basic research lab. Starting with basically me, a relatively small group, you know, over the last 10 years we’ve been able to build a pretty significant organization. We now have a staff, if you count all of our interns, visiting professors, post-doctoral candidates that are working with us and our core research staff, there is somewhere on the order of 650 to 680 people working within Microsoft Research, and we have a core staff of 500 researchers working in over 40 different research areas.

We’re now internationally recognized. All of our teams are making enormous contributions. I’ll talk more about that. And we’ve expanded from our original Redmond location out to a number of different labs. We have a lab in San Francisco. That was our second lab that we started, a very small group under Jim Gray. We have a substantial team now in Cambridge, England, and you’ll hear more about that later. We have a very fast growing group in Beijing. They have over a hundred researchers there now. And just recently, two weeks ago, I announced that we would be getting a research lab in Mountain View at our Silicon Valley center in Mountain View. And we’ve already got our first employees there, Roy Levin and Michael Schroeder, Andrew Birrell and Ted Wobber, and that group is growing. We expect it to grow very rapidly over the next few months.

So we’ve been growing, and really now we’re one of the larger research labs in computer science in the world.

You know, as I said, I’m a real science fiction fan. Everybody knows actually I’m a pretty big Star Trek fan. One of the traditions I carried over — I was a professor at Carnegie Mellon University for 12 years, and during the time I was at Carnegie Mellon, every time there was a new Star Trek movie that came out, I would always take everybody that worked for me — back then it was just a few graduate students at the beginning — out to each new Star Trek movie. It gradually got to be more and more people, even at CMU. When I got to Microsoft, I continued that tradition. Now it’s a little complicated; I have to take 500 people and their families out to the movies and it’s a fairly large production where we have to rent a movie theatre and get everything organized, but it’s a fun thing to do and it’s been one of the things that I’ve always enjoyed.

But really there are a lot of ways in which I think of what we’re doing as sort of like the Star Trek division of Microsoft. The two little videos that were up on the screen there, these are micro electro-mechanical machines. These are just little machines a few microns across that we’ve built in our research lab. You’ll see a little bit later some of the work we’ve been doing in virtual collaboration. I think of that as sort of the early days of the Holodeck like experience.

We’re working in areas like natural language translation and you’ll hear a little bit later some of the work we’re doing in automatic language translation, not quite the universal translator yet, but we’re working on it.

We’ve actually published papers in quantum state teleportation. I find that kind of bizarre myself. Although my feeling there, the joke I’ve often told about that is that, you know, if Microsoft actually invents the teleporter or the transporter, that the headlines the next day will be, “Microsoft Copies Again.” (Laughter.) You know, we do have to worry a little bit about that.

Scanners, a lot of work in computer vision technologies, you saw some of that in the work on the 3D face modeling.

And a tremendous amount of work on artificial intelligence and we’re really leading the field in a number of areas with artificial intelligence.

We’ve built an incredible staff, and I’ve just put a few names here. You’ll meet many of these people later on today at lunch. It’s just an incredible group. We have three Turing Award winners: Jim Gray, Butlerer Lampson and Tony Hoare. The (Turing Award is basically the Nobel Prize in computer science. We have the inventor of the laser printer Gary Starkweather on our staff. We have Jim Blinn and Jim Kajiya. Both won Academy Awards for their pioneering work in computer graphics algorithms. Michael Freedman and Laci Lovasz are pioneering mathematicians that have won tremendous prizes like the Fields Medal, which is the version of the Nobel Prize for mathematics. Laci Lovasz has won the Wolf Prize and just recently the G
ö
delPrize for the top paper in theoretical computer science. Gordon Bell is the father of the mini computer, the VAX. Many of you remember the VAX. That was before Digital got bought by Compaq and now Compaq is being bought by HP.

And then my own work has been in the area of operating systems. If you’re familiar with the Mach operating system that’s being used now by Apple, that was work that I did. And if you’ve ever heard of the term NUMA in the architecture area in multiprocessors, that’s a term that I coined.

So we have a tremendous staff, and I’ve just mentioned just a few of them.

Now, when I came to Microsoft, my idea was not to just create a traditional computer science research organization in an industrial setting, but really to create something more like a university research department, computer science department and put that in the context of a corporation.

So we’ve really taken this model of a university-style model. It’s a very open organization. We’ve really built on a model that I carried over from Carnegie Mellon of having relatively flat structure, of building around critical mass groups, a very open research environment. We’re very aggressive about publishing the work that we do. We view that as critical. One of the key ways in which we measure what we do is through the peer review process. I think you can always fool yourself when you’re doing research and believe you’re doing better work than you really are. Putting your work out there for peer review is really the way of getting the community’s perspective on what you’re trying to accomplish.

We also have been very aggressive in bringing visitors in to Microsoft. On any given day, we have three to five outside lectures going on, on campus, and we broadcast those on the corporate network and many of those are actually shared out on the Internet through our Web site. We have over 200 outside researchers visiting the organization in any given year. And we work very hard with universities and we consider that to be an important part of what we do. Something like 15 percent of my budget actually goes to universities in one form or the other.

Now, the way I think of the mission, this is what I think we’re trying to accomplish and this is what I tell our researchers, this is what I have consistently pushed over the years is our first goal in research is to expand the state of the art. I mean, if we’re not pushing the state of the art, if we’re not inventing new science, if we’re not getting out there in the peer reviewed literature and making an impact in the conferences and the journals, then we’re really not going to be able to contribute to Microsoft. I mean, Microsoft can always buy technology. We are a big company. We have lots of money. We can do that. The value of having a basic research lab in the context of Microsoft is to be able to do things that are really unique and then be able to bring them into the product organizations as rapidly as possible.

And that’s the second point, which is once you’ve done the work with the research, I mean, once you’ve pushed the state of the art forward in a particular area, then our second goal is to take that work and move it rapidly into products. And I’ll talk a little bit about how we do that.

The key point here is that our job in research is to make sure there is still going to be a Microsoft 10 years from now. You know, this field moves very rapidly. As I said, you know, 1990, 1991 wasn’t that long ago and yet the world was completely different. Ten years from now who knows what the whole landscape is going to be. In fact, we’ve seen whole trends come and go in three-year, four-year periods. So our job is to make sure that there is going to continue to be a Microsoft and to a larger extent a computer science community in the future.

Now, I said one of the things we’re trying to do is to expand into the state of the art. One of the ways we measure ourselves is what is our impact in the conferences, the key conferences and the journals that are out there. These are just a few statistics from some of the top conferences, but if you look at like the CHI Conference, the Computer-Human Interface Conference that was held here in March in Seattle, over 13 percent of the papers came from Microsoft. At SIGGRAPH we published more papers since 1996 than any two organizations have. In PLDI, which is one of the major programming language conferences, that was just held here — I think it was in July was when it was held — we had over 30 percent of the papers this last year. It was 25 percent a couple of years ago.

So we are making a huge contribution. We’re one of the major publishers in computer science research today, and we really take seriously our responsibility to work within the professional societies, to work with the journals and conferences, and many of our researchers are editors of journals. They’re chairmen in conferences. They participate in the program committees and are part of that process.

We also take seriously our responsibility to the academic community and working with it. We held a very successful faculty summit here just about a month and a half ago, bringing in 300 faculty from all over the world and really talking to them and letting them talk to us about what’s going on in the academic research community and ways in which we can help and participate.

We contribute more than $75 million a year in direct cash donations and in software to academic institutions around the world. And one of the key things that we’ve been trying to do is really make software available for teaching and research. We have a fellowship program that we support as we’re supporting graduate students, again around the world, both in China, other parts of Asia, Europe and the United States. And we’ve really been working with a number of universities now as we move into this new .NET era, where we’re really thinking about building a true distributed computing environment and working with universities to make available courseware, and we just announced just about a month or so ago an academic version of VisualStudio.NET as a way of making it easy for professors and universities to use these technologies in the classroom.

We’re also, as I mentioned, not just trying to push the state of the art forward, but we really are trying to drive Microsoft products. Basically, anything that you’re using from us today that was in the video — I think I made the comment in the video too — anything you’re using of Microsoft today has technology in it that came out of Microsoft Research. You know, over this 10-year period we’ve had a big impact on what Microsoft is like as a company. Some of the kinds of technologies, I mean the reason we were able to sim ship Office 95 and Windows 95 was because of optimization technologies developed in our programming languages group. And this is critical. We would not have been able to do that. It was an enormous advantage for us in the marketplace of being able to bring those two important products together, through each of those products together.

In Windows 2000 one of the critical things that we were able to do for Windows 2000, one of the reasons why it’s been basically one of the most stable big releases we’ve done in recent years is a huge emphasis on automatic bug detection. We were able to bring that technology in at the latter stage of the Windows 2000 release and we’ve now been pushing that out to the developers all across the company.

A number of the key technologies that drive Microsoft products, things like the Windows media/audio technology, ClearType, the collaborative technologies that we’re using, the intelligent searches in Windows XP and in a number of our server products, IntelliMirror, these were all things driven out of the research organization working with the product groups.

And we’ve been an incubator for major product parts for the company. When you think of what is the Microsoft digital media division today, well that was a group I started as a research group back in 1994. We spun that out into a product organization in ’96 and it’s been growing ever since.

The Tablet PC group grew out of a group that came out of research. The e-book is the same. Windows CE came out of our advance technology group.

I was the leader of our first e-commerce group at Microsoft. The data mining facilities in SQL Server 2000 came from the research group and then we worked closely with the product groups to make those technologies happen.

And all the new technologies we’re beginning to bear now in terms of natural language processing, the speech recognition support, not just in English, but in Chinese and Japanese and a number of other languages are coming from our research organization.

And so the key thing — and I mentioned this in the video, but I think the key thing, and you’ll see this as you listen to the talks today, is what we’re really trying to do in research is we’re just really trying to break down barriers. We’re really trying to make computers really usable in ways that they haven’t been before. We’re trying to extend human capabilities, breaking down the barriers between people and computers, really allowing computers to augment what people can do to enhance their lives in a more direct way, to have interfaces that are more natural, where people can interact better together, breaking down barriers between people. Right now it’s too hard to work with colleagues in remote locations. It’s too hard to work on a shared project, and we want to make that easier, make that simpler to do.

Break down the barriers between different parts of people’s lives, and you’ll see more of this later, but the notion here is again there are so many different things that you do in your life and they’re not very well connected today. We can help to bring them together, to let you be able to get the information you need about your family when you’re at work, the information that you need about work when you’re at home, when you’re in your car, when you’re in the airport or wherever you happen to be, when you’re here at this meeting.

Another area we’re trying to push is advancing digital media, really advancing the state of the art, breaking the reality barrier. You saw a little bit of that with the face modeling. But if you think about what computers are going to be capable of doing over the next few years, you know, it’s only probably in the next three to five years that we’ll be able to represent on a computer screen the level of reality that you see when you look out a picture window. That’s where we’re going and we need to build the software technology to make that real.

And finally really building that distributed computing infrastructure that science fiction writers have talked about I think for the last 20 or 30 years, you know, building that infrastructure that lets programs run in the network, that lets them interact with computers no matter where they are and break down the barriers that make that difficult to do today.

Some of the product technologies you’ll be seeing during the day and you’ll be seeing during the demos, some of the work we’re doing in natural language processing, speech, information agents, we’re putting a huge effort in privacy and security, data mining and data analysis, and high-performance computing is a critical component of the things that we do. We’ve pioneered the high-speed networking support, for example, that was in Windows 2000 and in Windows XP that has allowed Windows to be the fastest networking platform on the planet and currently holds the Internet 2 LAN speed record.

Finally, breaking down the barriers, again it’s only going to be a few years before we’re going to be able to have software that really creates the same level of visual reality that exists in the physical world, that we can really talk about human interfaces, not the old fashioned computer oriented user interfaces, that we build the software that can see and speak and listen and learn. We really build facilities to let people collaborate together in a realistic way, that we really provide a credible notion of what a mobile office is and really do make computing just part of the environment that we live in and that we work in.

So that’s kind of an introduction to the organization, the things that we’ve been doing. It’s been an incredible 10 years and you’ll hear a lot about the different things that we’re doing now. But it’s just been amazing. I mean, when I first started at Microsoft Research, I got a huge amount of advice. I think whenever you begin to start something, you know, you’re the target of a tremendous amount of “advice,” and I use term “advice” in quotes.

One element of that advice was that it was really going to be hard, you know, there were very few examples of new basic research organizations that had been very successful in the computer science area. During the 10 years in which Microsoft Research was started there have been a number of research labs started and have not had the level of success that we’ve had.

So I’ve been tremendously lucky as a person, and I’ve really enjoyed the time that we’ve had, and I think we’ve built a tremendous organization and have a chance to meet a lot of people.

What you’re going to be hearing later today, Bill is going to come on stage and talk about his work and the importance of Microsoft Research in making Microsoft successful.

You’re going to hear from our groups in Cambridge, in China and here in Redmond and each of these presentations will give you a much more in depth view of the technologies we’re developing, the direction that we’re going, the work that we’re doing and you’ll have a chance to see a lot more of the actual technologies in action.

And Amitabh Srivastava is going to be talking about the impact that research is having in the creation of software and the creation of new software products, and that will be part of today’s presentations as well, and then, of course, lunch and the other things that you’ll be seeing.

So that’s what’s going to be happening today, and now it’s 10:00 and I’m going to be very pleased to introduce Microsoft’s Chairman and Chief Software Architect. Bill was really, with Nathan Myhrvold, he was the one that convinced me to come. I was a very difficult person to convince to come to Microsoft. Nathan used to refer to me as Dr. No because I think I told him no so many times when he tried to convince me to come. But in my meetings with Bill, with Nathan, I was really convinced that Microsoft was going to be a fun place to be and that this was a unique opportunity to create a basic research lab in the context of a software company.

So let me introduce Bill and thank him once again for the great 10 years that we’ve had.

BILL GATES: Great work, Rick.

(Applause.)