Remarks by Bill Gates, Chairman and Chief Software Architect, Microsoft Corporation
University of Wisconsin
Madison, Wisconsin
October 14, 2005

BILL GATES: Well, it's fun to be here. This is my first visit, and I'm excited to have a chance to talk to you about the future of software, some of the neat things going on and the opportunities that you all have. Obviously, I think software is the coolest field there is, and I hope by the time I'm done you'll agree with me on that.

I got started in this 30 years ago, and that was when my friend Paul Allen and I left college, I'm not recommending that, and started Microsoft. And it was based on a dream that the combination of hardware breakthroughs and software breakthroughs could create the best tools that man has ever had, a tool for communications, a tool for creativity, a tool for collaboration, something that would really be unlimited in terms of how it could help all the different activities that we engage in. And at the time, that was considered quite crazy; even Intel, who made one of the key building blocks, the microprocessor, and whose Chairman, Gordon Moore, had coined the prediction called Moore's Law that talks about doubling the power of the microprocessor every two years, exponential improvements, it was a key element in the reason that Paul and I saw what we did, even they didn't really understand what personal computing would be.

And the reason is that with all of that great hardware, low-cost MIPS, great disk storage, high bandwidth connections, you have to understand software to think how you can change the device into something that's relevant for everyone. After all, computing at that time was just about large organizations, keeping track of customer databases, doing some scientific problems, but no one had a notion that computers could be out in the masses. In fact, one of the companies I grew up admiring, always loving the work they did, Digital Equipment, their CEO talked about the personal computer as kind of a strange idea, and that's always struck me as bizarre because it was actually Digital Equipment that took us from mainframes down to minicomputers, but the leap of going from minicomputers down to PCs was one that they didn't manage to make.

In fact, a lot of the companies that I admired as I was growing up, Wang Computer that did great things in calculators, word processors, minicomputers, a lot of them have gone out of business. And so as we started Microsoft, we had a fairly humble view that, well, if we're going to start a company, we might as well do it while we're young, we can always go back. I'm actually on leave. I'm not really a dropout at all, I could go finish my courses, although they've changed the course assignments quite a bit, it would be pretty tricky to figure out how to do it at this stage.

As we got going with the very limited machine, the very first machine we programmed to was called the Altair. It was a kit computer. For $360 they sent you a bag of parts, and many people didn't realize that you had to pay extra to get memory. The cheapest memory card was a 256-byte memory card, not kilobytes or mbytes, it was just 256 bytes. We actually figured out how to make the BASIC interpreter run in a 4k byte machine, so the interpreter was about 3.1k that I hand coded, and that left a little less than a k for your programs and your actual data. That was a floating point BASIC.

Today, we're on the verge of having machines where it's not that unusual to have 4 gigabytes of main memory. So that's a mere factor of a million of improvement, a million times as much memory. We already have a million times as much processing speed, we have a million times as much disk storage. Certainly in terms of network connections, we've got a million times the network connectivity that we had back then because we didn't have any network connectivity.

And so all the dreams we had about hardware enabling us to do better software really have come through. And so that takes us to the question, what kind of things can software do? First, we've provided a platform for different kinds of applications, that move to 16-bit computing in the early 1980s with the arrival of the so-called IBM PC and MS-DOS. Later that decade we moved to graphical computing with the arrival of the Macintosh that Microsoft was the company partnered with Apple to do the first applications for the Mac. The day it shipped, we had Mac BASIC, Mac File, Mac Star, we had the only non-Apple applications there. And then a few years later, we came out with Windows, so graphical computing was pervasive, both in the PC base and on the Macintosh.

What we did in terms of office productivity software was quite new. There were some analogs to the past, dedicated word processors, but for certain things like simple to use database, or spreadsheet or presentation software, there was nothing like that before. Today, we find there are even more categories that haven't existed that we can do. For example, take note-taking; we've got a new product called OneNote that's just come out in the last few years, and it's the idea that you can take the audio in, take the ink, take whatever typing you do, and organize and share those in a way that wouldn't have been possible before. So, as hardware gets better, we take and we're able to do more and more things.

So eventually we get to the point where virtually every activity is software driven. The way that a company thinks about analyzing its sales will no longer involve paper. The way that you'll think about photography is already well on its way to not involving film. The way you think about music, you know, this physical storage isn't going to mean much.

My daughter asked me why we call record stores record stores; obviously she's never seen a record, but she has seen a CD. Ten years from now, kids will say, well, what was this whole thing about you have to go open a plastic case up and get something out, and scratch it, why did you have physical media, it's just bits, and the storage that you can carry in your pocket, the storage that you have in the car, in your portable device, you know, it's almost infinite relative to the number of songs you want, so it's obvious that's the way it should be done.

Video is the same way. It's maybe, oh, five or six years behind because, of course, the number of bits that you either need to store or transmit for a full length movie are, say, a thousand times more than a music file. But that, too, will eventually go to the point where we don't use any physical media.

Even more profound is to think about the way organizations will be connected up. Capitalism is about matching buyers and sellers, and actually knowing who might be a seller, who might be a buyer, that's a lot of overhead to it. With the arrival of the Internet, eBay came along, and you could take all the junk in your closet and just stick it out there and maybe get a few bucks for it, and feel like there's a reason to clean out that mess you have, and I'm sure somebody on the other end was thrilled to get whatever those used things were. Likewise, in terms of part-time labor availability, the Internet will allow people who have pieces of time available to establish a reputation and be able to go out and maybe make those things available.

So the Internet has really driven software to this whole new level. Now, what is that actually like? Well, what that is is it's people developing software. What's the process of developing software? Well, people who use software are very passionate. They love to tell us what they don't like about the software, they love to tell us what they do like about the software, and more and more we've been able to use instrumentation inside the software itself to give us a sense of what commands are being used, the order they're being used in, we can see what error messages people are getting, we can see where they pause, what takes time. And analyzing all of that gives a sense of where they'd like to see the product improve.

For example, in our Office software we noticed that a lot of people weren't even finding the advanced functionality, that it was so buried in the menus, the dialog boxes, that they were doing equivalent things, but doing it with a lot of low-level commands. And so that's what gave us the confidence to make a big leap of faith and actually change the user interface in a fairly radical way. So you'll see in the next version, which is called Office 12, coming out next year, the user interface is quite different than it has been, very much a very visual representation where things aren't buried in the way those things were before.

So the life of a software designer is to work in a team, a team that's got marketing people, people writing the specifications, experts on user interface, developers, testers, all really trying to figure out where should they go, what do the customers want.

Another thing that makes the field interesting is that you'll have projects where you're out on your own, you're pushing the frontier. Take what we're doing with tablet computing. That's the idea of having ink on the screen, having that ink be recognized, getting the tablet to be light enough that you no longer feel like you need textbooks or paper at all, so all you're carrying around is this tablet device; when you're in the classroom things can be projected onto that, when you're not, of course, it's still connected up to wireless Internet.

And that should come down in cost eventually to something like $500, so not only in countries like the United States could it become pervasive, but even around the world we can move to much larger numbers and get every student to be able to have one of those, and have access to all of the world's material.

In fact, as I think about the material I had available in the library when I was growing up, it's a small, small fraction of what anybody connecting up to the Internet has today. In fact, today there's more and more lectures about great topics that you can go and find, lectures on physics, or lectures given at this university, or MIT, and so the wealth of material for the motivated student is really quite mind-blowing. And we should want literally every student in the world to be connected up to that.

From the very beginning, the vision of Microsoft has been pervasive computing: drive the cost down, make it as inexpensive as possible, use that high volume to create opportunities so that people developing software, even when they price it at very, very low prices, they're able to get a good return on their investment, because the volume is so strong, it's a very strong opportunity for them.

I want to quickly demonstrate a few products to kind of give you a glimpse of some of these directions I think are pretty neat. This is one called (Max ?) that's just running over here on a PC. What we're doing here is we're taking the idea of photos, and the number of photos you'll have in your library is going to be quite immense. Over your lifetime the number of photos that you'll take will be in the tens, if not hundreds of thousands. And because this software will let you organize them in neat ways, you won't really restrict yourself. You'll take lots and lots of photos, you'll take duplicates and things that are similar and categorize those things together. But, you want to be able to navigate in a very visual way.

Here we have some photos, we can look at this at different sizes. I can take this thing called Italy and decide I want to move some of these things over here. Let me size that down a little bit here. I can click anything in or click it out. Let me actually take all these photos, and so I've got a set there, and use this as a list. Now, once I've got a list, the question is what different ways might I want to view it? I can view it as an album, and so here you can see, let me zoom up a little bit bigger, it's taking the photos and I can reorganize them in these different things, so it's kind of a nice album view, or I can take and put them in what we call a mantle view. So here you see it's a little bit different it's got multiple things, a little bit of a 3D look when we go into play, and this is what you'd see, say, on your TV set, if you want to navigate these different images. You can zoom in on one and see what that is. You'd have various data about it, things that have been recognized automatically in that photo. And so organizing a presentation about your kids or a trip or something can become very easy and it can have a lot of very simple visual impact. So that easy authoring and easy navigation will be a thing that's very important.

Computers will be different sizes. Of course, they'll all be connected to the Internet. The way we can think about them being different is really based on the size of the screen. We'll have on our wrist like a wristwatch computer, kind of like Dick Tracy, if you've ever heard of that cartoon, that may date me.

You'll have in your pocket phone-type devices. These are two that we've been working on that will come out in January. This is a Motorola RAZR-type device and you see it's got the small keyboard there, pretty thin devices. It's a great phone, but of course it's a great messaging device, as well. Another one that we've just announced two weeks ago also comes out early this year is one we're doing with Palm, and this takes our kind of Treo design, the great way they've done the keyboard, and brings it together with rich software.

Now, even devices will go far beyond what they are today. Today they are not your digital wallet, you still carry cash and credit cards around. That's not going to be necessary. Today you still have maps and directions and things like that. Well, this screen it will use the GPS locator, it will know exactly where you are, it will give you a visualization of driving directions. It will know the traffic, and so that will be a lot better than it is today.

We'll also take very rich software and build it in here. If I take this camera and take a picture of a receipt I get at a restaurant, the software should say, hey, that looks like a receipt, when we recognize the numbers there I know the date, I know your calendar, I'll just file that as an expense without your having to even think about it. You take a picture of, say, a product with a bar code, it will come up and tell you what the reviews are of that product, where you can get it maybe for a better price, what the availability is, all very automatically.

If you're in a foreign country and you want to take a picture of a sign to have it translated, it will say, okay, that's not English, send it up to a server that will do a machine translation and come down and make it available for you. So it's very simple and rich software that lets this come together.

Now, I talked about these software teams and the neat work that they do. What that means is that for Microsoft and other companies in this business, getting the best people in, who are really excited about doing software is super important. So one of the things that everybody at Microsoft does is they go out, get involved in recruiting, talk to people about these jobs, and make sure we're finding the right people.

I made a little video of one time I got involved in some recruiting. So let me show you that to get a sense of what it's like.

(Video segment.)

(Applause.)

BILL GATES: Yeah, we had a lot of fun making that. (Laughter.)

Well, one of the places that will change a lot is the living room. Today you think about the set-top box, the videogame, and the PC as three very different things, but, in fact, as you connect the TV with a great screen up to the Internet, you'll have the technology for doing very advanced graphic type capabilities, you think about the kind of interactive games that are going to be offered that are very realistic in nature, you take that broadband connection that lets people socialize, connect up with other people, and really you get a convergence there.

One of the steps in that direction is this next generation videogame that you'll think of as the first high definition generation. I've actually got here an Xbox 360. These don't go on sale until next month, late November, but I just wanted to show a couple things so you get a sense of how much this generation includes and then three or four years later there's another generation and another generation that gets pretty much up to the ultimate experience.

One of the things we talk about is that you can do more than gaming, that you get a digital entertainment experience. So when I turn on my Xbox, I'm not sure if you can see that, you've got these different screens here. You've got your games you can play, you've got connecting up to Xbox Live. Here I've got various media things, so if I say I want to go in and actually use music, I can say, okay, let's hook up a music player that I've got some music on here. So as soon as I get that, it goes in and shows what's available there. So now it sees my iRiver there. And that could be an iPod, anything you want to use -- (laughter) -- we'll connect up to. And so it's going and it's pulling off all the songs, and I can just click and say Play All, and so it's got all my favorites there.

So this is now playing on the living room system that's available. Or if I want to take my camera, again it's just a little USB cable that I've got to connect up and I go in in the same way, just hook it up to that front panel, and then without doing anything special at all, all I'm going to do is back out here a little bit and go back to where I saw the picture interface. Then I go in, and what you can see is it's recognized the camera that's there, I'll select that, and then I can go over and just pick all the different photos that I have on here.

So now I'm just playing these on the TV, music, pictures. Of course, it's a simple command if I want to move these up to my PC or send them off in e-mail. Likewise, if I wanted to bring the music down from the PC, that would be very easy as well.

And so if you compare that to the world that people have had where the devices don't really work together, you have many remote controls, just by having one device that's very visual and a controller that you get to know with the rich connectivity, you've got sort of this vision of digital entertainment, all those things coming together.

I'm going to switch over now to a machine that I've got, it's another Xbox 360. This is one that I have one of the games available on. I've got to wake the controller up. You'll notice these are wireless controllers, which is kind of a nice deal.

I'm not too good at this game, so I'll pick the easy level here. This is Project Gotham Racing, and for any of you who have used it on today's Xbox you'll see a very dramatic difference. What that is is that the graphics processor is over ten times as capable, the memory is over ten times as much, the speed of the different processors is ten times as much. And so we take that and that goes into the richness of the experience and the visualization.

Here I'm accelerating a little bit, I'm going left, I'm going right; boom, I'm running into the wall. You see how realistic this is, it's kind of a nice little Porsche Turbo. (I do own ?) one of those in real life. And so I'm actually racing here in Las Vegas, and so you can see the different hotels that you'd expect there, see everybody's gotten ahead of me here.

A very nice feature on this is that I can use the artificial intelligence and have it go ahead and do the driving for me. (Laughter.) I can learn, oh, you're supposed to brake as you're going through that corner. And I can actually change the viewpoints, so I'm behind the driving wheel, I'm just right here. Also when I'm in any view I can go and look at it and get behind it, get up on the side, get in front of this thing, I can go around and see different viewpoints. And you can see the richness of detail, whether it's the hotels, the parks, and that just comes because you have a dramatically richer visual model of what's going into the environment.

Of course, it's not just the graphics; the idea of being able to connect up over broadband means you can talk to your friends, we'll have a camera so you can see your friends. If you win a contest, you can gloat and they can see you getting excited about that. It's just all built in so it's a far more social experience.

In fact, with that kind of connection, you get a whole new genre of games that wouldn't have made sense without the ability to have lots of people coming in, virtual world type activities as well. And so we're going to see not only things like sports games, racing games, fighting games that have been big, but lots of simple casual games, lots of virtual role playing type environments, so taking gaming to the next level in a very big way.

Now, those demos were things that are pretty close in. Here I want to take something that's more like, oh, four or five years out. What I've got is this is a scenario say I'm in an airport and I've walked in, I've got my phone with me. So my phone has a fairly small screen here, but inside here is a pretty good processor, it's got my e-mail, it's got my information. So I'd like to just take this phone and put it down and have it be recognized.

So, in fact, what there is here, there's a camera here with a little -- actually this is a little light that means it's going to recognize what's going on, and a little projector. So it says, okay, I've noticed this phone belongs to Bill, but I want to make absolutely sure you didn't just pick it up, so I use my fingerprint and I log in. Now I have this large size screen. I could take out a pen and do ink-type things. Here actually in this case I've got a business card that somebody gave me. I'll just put that down and so the camera sees that, actually sees the text on it. I'll go ahead and flip it over because I wrote some notes on there on the back, and then the camera actually notices that, uses handwriting recognition.

And I say, well, I just want to take this and put it into my contacts. So I just turn it over and drag it up there, and then the camera sees, okay, I'll just feed that into my phone, you get these nice little dots that make you know that's going on, so I can pick that up.

And now let's say somebody sent me some e-mail, I get the benefit of this large screen, I can see something is coming in here, it's some type of press release, I can go and read that. In fact, I could use the whole screen area here. In this case they want me to again authenticate that I approve this, so again I use this fingerprint reader and then it will go and send that off.

So I essentially have the whole power because of this camera, I have the whole power of this desktop and yet this is something that over time will be incredibly inexpensive, and the idea of having software that does visual recognition will just be commonsense.

So when I'm done I go ahead and pick up my phone and it will realize it's not there, log me off, and so I haven't left behind any information that somebody else might do something with.

So once you get cameras in the environment, once you get displays anywhere, it really changes computing. Some of those displays will be right under your retina, perhaps from your glasses. Some will be in the car, you'll have a heads up type display where information can be just projected up on to the windshield, so it's essentially reality enhancement, so it's telling you things about traffic or where things are nearby that you might be interested in, but doing it in such a way that it's not a distraction for you.

Now, many of these problems I've talked about -- ink recognition, speech recognition -- these are areas that computer science has been working on for a long time, a lot of great work in the universities but then, because of the scale of the problem, get moved into commercial type research that we do in Microsoft Research and eventually find their way into products.

These have been problems that as we've worked on them we've gotten a lot of respect for the way the human brain takes context, takes an understanding of ambiguity and is able to resolve those things. But year after year we're bringing the error rates down, bringing the model that we have of how these things work be better and better and so we'll reach a crossover point some time in the next five years for both handwriting and speech and visual recognition.

So now we're talking about a device that's far more intelligent, far more natural to work with than anything that we have today. And so the frontiers in software are really quite dramatic.

Software is also becoming the tool that gets used in many other fields as well. If somebody wants to understand all the biological data that's coming along, make sense out of that, it's really software and the software tools about networking and communications and state that will help there. So it's people in the software world who will be partnering up, driving breakthroughs in these interdisciplinary ways.

And so this is the field that's changing the world more than any other. This is the one where anybody who works on Office can come in every day and say there are 500 million people who either love what I do or they don't, and if they don't I get to make it better, get that next one out there, and then know exactly how people are going to respond to that.

So it's a fascinating field. It's a field that grows quite a bit over time as the hardware prices come down, as we tackle new capabilities. The fact that the world at large is getting richer means that personal computing will just penetrate more and more. There are some breakthroughs we need like taking software and combining that with wireless networking to do what we call Mesh Networking so we can get the cost of connectivity to be super low cost as well. Getting the hardware and software down in cost is actually easier than the connectivity, but even there we see a path that will let us get from the billion PCs we have today to eventually up to three, four, even five billion PCs so that the kind of empowerment we deliver is very broad.

It's a field that has a great mixture of there are things where a company can get out on its own and really be pushing the frontier like we are with a lot of the natural interface things in the tablet. We can have other areas that are hotly competitive. Take Web search, a very competitive area, we hope to make it even more competitive. And the nice thing about that is that today's search is nowhere near what people want; that is spending five, six, seven minutes traversing through a lot of links is nowhere near as good as having the computer really understand your interests, what you might want to have brought to you, and just have the answer come up. In about 10 percent of searches today, if you type in a sports team, we'll bring you the answer, we'll bring you the latest statistics right there, if it's something in the encyclopedia we'll bring it right to you, but we need to drive that up from 10 percent to 90 percent, and that involves a lot of very deep technology around natural language understanding.

No doubt that will happen and it will even happen in a way that's very beneficial to users, because over time they'll actually get paid for looking at ads. It won't just be that the company takes that money, but there will be an incentive that comes back to the user; so a lot of new ways that that's going to drive forward.

So this is why I talked about software as the field to be in. In fact, I think of biology as the second, not a close second, and then whatever is third is a long ways behind that because working on these projects, seeing what they can do, thinking about the breadth of people all over the world who get to use these things really is something that we only get in this field.

This is a field where we need new thinking, we need people who haven't gotten their minds clogged up by the way things were done in the past, and that's why it's so fantastic for us to bring in a new generation of computer scientists, to hire in the best people and have them join these teams and really take on the very tough problems.

All of you are very lucky, you're exactly the right age to come in in this golden age of computer science and help define a lot of these products. So I'm very excited to see what a lot of you will be able to do with this and take it to a whole new level, and it's going to be fun to be a big part of that.

Thank you. (Applause.)

MODERATOR: So we're going to spend 35 minutes or so asking questions of Bill. So what I'd like people to do is to come up to this mike, these mikes over here and fire away with questions about anything. And I know there are some high school students in the audience and I encourage you to come up, too. Please don't be afraid. Yes?

QUESTION: All right, my question is more about the business of Microsoft. Business is very important when you're getting into anything. I'm just wondering exactly what were the first business steps that you took in creating Microsoft, what was really important that you did?

BILL GATES: Okay. Well, absolutely if you want to hire your friends to work on software, you've got to pay them. And so I thought, well, gee, how am I going to do that. I really wanted to get a team together and be able to write lots and lots of software.

The model that we came up with was a model that was very high volume, low price. And so I remember at the time there really wasn't a software industry, there were maybe a dozen companies who sold products and if you sold a hundred copies of a piece of software that was considered good, and there were only three products that had sold a thousand copies in the history of the industry. And at one point these people called me up and said, "We've heard a lot about your product, we think you may have sold over a thousand copies, we'd like to send you this award," and I said, "Well, that's interesting, we've sold 4 million copies." Because we had the BASIC on the TRS-80 and the Apple II and the Commodore and those things had gotten out in very big numbers.

But that idea of being very low price was a radical thing. And, of course, it assumed that the hardware would likewise become low price, and we'd create this virtuous cycle that the cheaper the hardware was the more people who would buy it, the more applications we could write; therefore it would become more relevant and we'd drive that volume further. And that fed on itself, helping to make things like MS-DOS and Windows a big standard.

Yeah, I'd say that a good instinct for business that we really could go out to customers, show them that our technology was the best, that got us bootstrapped, and then the vision of this model, this high volume model really revved up in every way we dreamed about it, and then we were able to even invest in research and a lot of long term things that have been key to our success.

QUESTION: Yeah, with more and more groups coming up with free software like Apache and Open Office, how do you think the future of paid software like commercial software will change?

BILL GATES: Well, free software is, of course, setting new records in terms of volumes and sales every year, and there's an incredible value in terms of the work that goes on in terms of 24 hour support, compatibility. Being able to do deep types of integration, you know, take this tablet work that we're doing, you're not going to see small groups that are disperse taking on something where you have to do what we call integrated innovation, change all the different modules together in lockstep.

Now, of course, once we do that, with some time lag there will be some projects that go do that, but the nice thing is there is always a new frontier. And the fact is that because software is so inexpensive, when I sit down with a corporation, I say, hey, for $100 a year your employees can have the very best tools that collaborate, communicate, produce things; is that worthwhile? When you think about that's probably less than they spend on paperclips, certainly a tenth of what they spend on their phone bill or the desk or any of those things, we can make it so inexpensive that it's very attractive to go with that commercial model.

Free software is nothing new. In fact, the original browser that was done at Illinois NCSA, that was a free thing, the original UNIX work, BSD UNIX was a free thing, so we've always coexisted very well with free software, and people are always going to have that choice. But given the desire to have these systems be self-managing, extremely secure, very high level, the frontiers that we'll tackle on the commercial side will keep us in a very good position in terms of being able to offer a better value.

QUESTION: I went to a Microsoft certified technical school before I came here, I got my associates in computer information systems, but now I'm here in computer science. And I was wondering what you think is the major distinction between technical schools and liberal arts schools computer science graduates.

BILL GATES: Well, I may not be -- you probably know better than I do, to be honest. There's, of course, a whole spectrum of where you get up to the most in-depth research at the graduate level, and then all the way down to the kind of exposure to computers that even students who aren't going to be writing computer programs, they're thinking of the tools they have. And within that range there is so much breadth to this field.

I think it's very important for people to get a sense of programming and data structures and algorithms; the more people that can know that the better. And so I think as you move into the more liberal arts type format, the willingness to create abstract concepts that then later when you're in the workplace it's a generalized skill that you map to the particulars, that's more the approach, whereas in the technical schools they're literally teaching you how to do network administration with Netware or Windows Server or something like that. So I think they both have a very important role.

When Microsoft is hiring developers, we're less interested in the specifics of what they already know and more about their conceptual knowledge, the idea that they're very good at doing design type tasks, whereas our customers when they hire in people to help with their systems, they just want somebody who already knows exactly how to work with those systems.

QUESTION: I was wondering, for you to lead the software industry obviously you have to know like keep yourself updated about the software programming techniques and stuff. I was wondering approximately how many hours you spend like reading or thinking about software programming things per week. When you were the CEO still doing like management work and now that you're the Chief Software Architect, how has it changed? And are you still interested in keeping yourself updated?

BILL GATES: Well, I think it's fun to stay up to date. This is a fast-moving field, and it's hard to know all the things that go on. Of course, the Internet is now this amazing tool in terms of the RSS feeds you sign up to, the search requests you do, the e-mail you have connecting you up with lots of other smart people who are tracking the field.

And so within Microsoft I know who the experts are in various areas. We have lots of reviews where the research people come in and talk to me about the work they're doing. Several times a year I go off and do what I call a Think Week where people will send me papers they think I ought to read, and that's where I get to look a bit far a-field, for example, understand satellite technology to say will that give us data connectivity in Africa, or look at something like Lumileds to see how that will change projectors to see how quickly that will give us projection type capabilities on every flat surface that we might have. And so people are sending me things and there's a lot of deep expertise.

Microsoft Research is going out and working with the top universities around the world, and so as there are breakthroughs in, say, various machine learning techniques or, say, using statistical techniques to do machine translation, they're bringing that back and making me aware of that, because I'm in the role of making sure the 5 billion in R&D we spend a year, that we're making the right bets, that we're putting that money on exactly those things.

So I love staying up to date. As you said, dropping out of being a CEO and just being the Chief Software Architect was a big help for me. Microsoft got complex enough that I was spending less than 50 percent of my time on technical strategy. And so in the year 2000 Steve Ballmer took that over for me and that's been a great thing. Now I'm up to about 80 percent of my time is on product strategy.

QUESTION: You showed us a lot of new and upcoming things like your Xbox 360 and your MAT software and then some things that are coming a lot farther in the future like the -- (source audio break) --

BILL GATES: The new version of Windows that's late next year, it's about a year from now, and the new version of Office that we're calling Office 12 that goes with it, and that's the one that has the new user interface that I mentioned.

This is a very broad release because what people expect out of Windows in terms of performance and security and dealing with new hardware and wireless networking, people expect a lot out of Windows. And there's the end user's point of view where you want things like rapid search to find all your information and great peer-to-peer networking, great voice handling so you can just do voice without any quality problems. And then there's the whole point of view of the IT department where they want to be able to deploy things very easily, they want to be able to secure things very easily. And then there's the point of view of the developer where they want a graphic subsystem that makes it just a few hours to write a little application like MAT. In fact, that was the whole reason we did MAT was to show that as the operating system has moved up to this new level, that it's easy to write applications like that, not that that is hard like it would have been in the past.

So Vista really is big news for the developer, it's big news for the IT department, it's big news for the end user.

Now, Vista has taken about, oh, four and a half years for us to get done. We in the future would like to update parts of the system every year, like the browser every year, the media player every year, and so we're doing more layering in Vista to give us that agility that some pieces can be on a very quick cycle whereas other things like the scheduler or the file system would be on more like a three-year cycle.

QUESTION: I'd like to know what kinds of improvements would you like to see in the education system in this country, and in all countries.

BILL GATES: Well, the United States does quite a good job of taking its top 10 percent or so of students and giving them a reasonable high school education and then having universities like this one where they get a great university education. If you looked at the top ten universities in the world in biology or computer science, somewhere between eight and ten of them would be in the United States, probably nine I would say at this stage, and that's an unbelievable thing.

So we're doing well on that part, and it really helps the country a lot, that's why we have Intel and Microsoft and Google and so many companies that really got started here and rely, the vast bulk of their talent will always be here in the United States.

It's the other 90 percent that this country is not doing that good a job on; in fact, if you looked at math, science scores where you compare the U.S. to other nations that are well off, our fourth graders are about third or fourth in a ranking of the top 20, and then we fall to almost 18th by the time you're a senior in high school. There's a lot of reasons for that in terms of the size of the high school, the curriculum, the incentives for the teachers.

So we really need to fix that in order to have this broad strength that we need, both competitively and because of the nature of the new jobs. The nature of the new jobs require far more skills than in the past, not because those jobs are going overseas but because automation is eliminating those jobs. People talk about China as this great country doing manufacturing and that's true, but the number of manufacturing jobs in China has not been growing because they're also engaged in this automation process. So all the jobs are moving up to require essentially a college education.

MODERATOR: I'm going to ask several questions. I so wish I was sitting on the other side right now. Well, tell me what a typical career path, if I do a degree in computer science and join your company, tell me what a typical career path would look like.

BILL GATES: Well, there are a lot of career paths. Typically when somebody comes in, they'll come in as a developer or a tester or a marketer or a program manager, and they'll come in often and work on a fairly small product or a specific feature of a large product, like you'll take, say, OneNote that's part of Office and work on that team. And that team is about 20 people in total, great spirit about how they're going to change the world there and do neat things. And so you get to know do you like the user interface design piece, do you like the coding design piece, do you like the business issues around marketing, do you like going out and spending time with customers, and that sort of determines your career path.

A lot of the people who do dev are the ones who will work in development for a period, work in programming management for a period, and even then work in marketing and then move up and take on management responsibility.

Then again we have amazing people. Peter Spiro is a person that came from this university and he's a database expert, I mean super expert, and he's always worked on database. He is changing the state of the art in database, he's been given a really big charter to do that, gets to hire people in. And so he started with a small development group and now he's got that whole group that's under him. And so that's moving up a purely technical ladder and he does some management things but he lets other people take the parts he doesn't like of that.

So it's a pretty diverse set of activities that most people when they start out at Microsoft don't even have a clear sense of exactly where it's going to go, how many different products will they work on, which of the activities they find are the most fun. But if you can get in and even just at the summer and then get a sense of it, then you get on the track that's a fit for you.

QUESTION: I was thinking a lot about people in my generation typically are pretty good with computers, but I look at my dad who runs a business and he calls me all the time, "I can't check my e-mail" and I explain the same thing over again, and the same with my grandparents.

And I was looking at a lot of these devices and even I think of people in my generation that you put them in front of a computer and they really don't know what to do, they're kind of out of it. And I was wondering what kind of progress is being made in the software field as far as making these more accessible for people that maybe don't have the natural ability to just like be computer savvy.

BILL GATES: Yeah, well, consider people are doing more and more with computers. The problems that we had in the past, it used to be getting the right printer driver was very complicated. Well, most of the time that's not a problem now. Getting connected up to the Internet, that's still not always as simple, debugging what's wrong with your Internet connection, but over the next several years -- that's a big thing in Vista is that it automatically figures out for you at which level your Internet connection is breaking down so you don't have to sit and do the typical ping test, the set of things that you would go through it's going to do for you in a very automatic way.

So as we take something and make that easy, take e-mail and make that very reliable, then people want to do group collaboration. And so, wow, that's kind of hard, complex, flaky; once we get that right, then they want Digital Rights Management and workflow, and so we have to conquer that next frontier.

There's a lot to be done, I'd say that the living room today is a mess, even without a computer in it, all those different remote controls, and having devices where because they don't display things visually you don't really know what state is it in, why isn't the video showing up. Here when you have something that can be giving you instructions on the screen and everything can interface through this, it can be made a lot easier.

So I think the living room is some place that software technology actually can be a big plus there.

There will always be people who use these systems in very light ways, and knowing what those profiles are, that they just want to do e-mail, just want to do photos, and then tailoring the user interface to those scenarios, that's one of the technologies we're working hard on is you just pick your scenario and you're not presented with much that pushes you outside of that.

Eventually I think speech interaction is another way that we'll be able to draw in a broader set of users to use richer things, because then it's not our command structure, it's their normal way of talking.

MODERATOR: We've got several young women in the audience. I'd like to see several of them step up to the mike. We only have two right now.

MODERATOR: My name is Paul Cekas.

I was wondering, there's a lot of talk these days about Google becoming really massive and competing with Microsoft directly, and I was wondering if you contribute that mostly to media hype or if there are any plans in the future or what the plans are to compete with them or even collaborate with them on like projects at a technical level.

BILL GATES: Well, mostly we'll compete with them. (Laughter.) There are certain industry standards and things like that we can work together on, and there are certain policies about advertising and things that make sense to tackle together.

But by and large today when you use a search engine, they make a lot of money and you don't get any of it, it just all goes to them. And that's because people view them as having the best search engine. As we make it clear to people that we have a search engine that's as good or better, then we can start to have some competition where we take and say, hey, your time is valuable, we'll give you either free content for doing search in our environment or we'll even just monetarily pay you for the fact that you do that, and you'll have basically the kind of competition you always see in the software industry. And that's very healthy where you're both in trying to improve your product, coming up with a business model that's more attractive for the users.

We see that in this game space. I mean, at this point it's largely ourselves and Sony. They're going to do the PlayStation 3, we've got Xbox 360, users are going to compare those things, there's room for both to succeed and so we'll both be funding R&D out into the future.

Google is in this kind of honeymoon period, which is great in a way. Every day people tell them they know everything, they do everything perfectly, and the longer the world tells them that the better it is for us. (Laughter.) We like to be underestimated. And, frankly, most of our history people haven't underestimated us, they've thought of us as kind of this miracle company, and so it's kind of fun to be in this period where people are thinking, yeah, will they really do a better search, will they move ahead on that, and the reality will strike fairly soon. (Laughter.)

QUESTION: I had a question about the application of software to more professional scientists, to say maybe the medical scientists, and how they could influence human decision-making in seeing, for example, patients or research.

BILL GATES: Well, certainly the idea of using software to take this massive data that we're gaining about medical conditions, that is the state of the art. There's no way to take genomics, proteonomics type information and make heads or tails out of it without far better software tools than we have today. And so we're very excited about that and very involved with that. The whole approach called Web Services can help with that.

One of the neatest things we did actually wasn't in biology, it was in astronomy where we took Jim Gray, one of our researchers took the idea of seeing there's all these disparate databases, how could you create a virtual layer that would let people essentially query against the union of those databases instead of having to go to each of them one at a time.

Now, astronomy was easier because it's a very complex field, but not nearly as complex as medicine, and you don't have the privacy issues, stars don't worry about that, you don't have this proprietary data of who controls the data issue in astronomy, and so we did that as kind of a proof of how you could stitch data together and let people see patterns across many, many sources of information.

We need to do the same in the biological sciences, and in the Microsoft Research group we have a lot of people who are MD PhDs and really got into computer science because they saw they needed those tools, in particular machine learning techniques like Bayesian approaches they've pioneered because they thought those would be helpful.

Now, moving that up to the level of diagnosis, helping with diagnosis, that's probably one of the trickiest areas, because you always want to be very conservative about putting the judgment of the machine in the place of a human doctor. For some period of time it will be much more making sure the records are easily available and helping maybe to suggest things rather than that kind of final decision about what's going on.

Now, if we get out far enough in time, we'll be dealing with real intelligence, but for the foreseeable future it will just become a much better tool.

QUESTION: My question is, as you've stated, technology is everywhere, especially in the home. Do you feel, do you foresee or fear that society might reach a saturation point, and if you don't foresee that, do you feel that technically society could be essentially dumbed down because everything has become so easy for them that they're not thinking?

BILL GATES: Well, certainly take agriculture, most people here probably don't -- including me, me most of all, I don't know much about agriculture, and a hundred years ago that would have been considered a travesty, I don't know how to ride a horse, I don't know how to plant, I don't know anything about those things and yet I eat very, very well. (Laughter.) So something that used to be considered so crucial, now we can count on about 1 percent of the workforce to create enough food that actually we're one of the greatest agricultural exporters in the world.

And so you will have areas of endeavor like that where the efficiency becomes pretty unbelievable. Manufacturing, certainly the idea over time that assembly gets done on a lower scale and that lower scale stuff is very driven by software-type devices, many manufacturing jobs will disappear because of that kind of capability.

So a lot of the things we automate out of existence I don't think we have to feel that bad about. Administrative jobs used to be where the boss would scribble things on paper and then the secretary would type things in or they'd take dictation and type things in. If that still exists today, it's in archaic organizations. Today, the job of an administrative assistant is a far higher level job having to do with scheduling and resource organization that is more stimulating, more interesting, more value-added than just that rote thing of essentially data entry that was a big part of that job in the past.

Now, I can't say that it's totally without dangers, there's certainly the danger of having to do with privacy issues that the data is there and who's actually getting access to it. We're certainly seeing problems where bad people get into these environments and try and, say, do identity theft in some way. We need to put in safeguards for that.

I certainly think that people still should understand a little bit about computer programming, a little bit about the concept of mathematics, even though these computers, they're very good calculators, you want to have those concepts in mind because as you get more advanced, you're kind of building on those notions.

You know, when I see a kid whose not very good at handwriting, the fact that at a young age they can use typing, that's probably a positive thing, it's just defined muscle skills that are developing later and causing the problems for them.

So overall I'd say it's a good thing, I think we need to guard against some of those problems. I don't see any saturation, to be honest. As we take something like photos and make those digital, that just grows how much people want to use that. As we take music and make that digital, people will be able to have their friends share with them more music and so their tastes will be broader and it creates new opportunities.

So by and large, I think the path we're on is pretty good, as long as we take the security and privacy issues and pay a lot of attention to those.

QUESTION: I was wondering, there's a lot of companies, Mandrake, Red Hat to start off, for example, that have been able to turn profits using Open Source software. I was wondering why Microsoft has stuck to the proprietary and shared source software even though it has millions of programmers at its disposal.

BILL GATES: Well, we have millions of programmers at our disposal. I mean, the whole goal in the Microsoft world is to get people to build software on top of Windows, and that includes people who don't want to be paid for their work and it actually includes people who want to send their kids to school and buy a house and all these awful things. So we embrace both commercial development and non-commercial development on our platform.

What's happened on the Linux platform in general is that the people who work on that just basically have an attitude they're not going to pay for anything and so you end up with a whole stack of noncommercial software. But that's fine for some users, but it means that they're not attracting in the people who want to do the commercial work in that environment.

So when you want to fund the great research, you've got to have a business model that works with that. You can look out there and see what the success is of people who don't create some sort of payment for their services. I wish them all the luck in the world, but I do think the numbers kind of speak for themselves.

QUESTION: I was lucky enough to find a copy of Microsoft Bob and get it to run on Windows 2000 back in the day, but what I've been thinking about over the last year is about user interfaces. And I was just wondering like if you guys are ever planning on making it so that as far as customizability comes out where like I could make motion graphics and have it so that everything I'm looking at is something that's made by me, because I'm kind of an artist kind of guy. So I'm just wondering if like there's ever going to be a moment in time where I can create an entire user interface, whether it's 3D, 2D or whatever, that would be completely customizable to the core.

BILL GATES: Well, the degree of customizability you always have quite a range because at some point you get fairly complex semantics in terms of how the different user interface elements relate to each other. And so the ability to reskin something, that is have different colors and different drawings and things like that, that's always very straightforward. Then you get another level, which is you can have sort of styles where a single set of changes applies across a wide range of things. Then you can have things where you take some of the things like how windows appear and disappear and customize those.

What we're trying to do is use an animation tool, we actually introduced this at our PDC, use that animation tool ourselves to develop all the Windows user interface and then when we ship Windows it actually ships with those animation files, and so you can get in with our cool animation tool and go in and make changes.

Now, there will be some changes where you'll violate some of the constraints that the software is expecting, where it expects something to be visible at some point in time, so you could get yourself a tiny bit in trouble, but it's fine, we'll let people go out and have that capability, so it's our desire to give you that kind of richness.

This version of Windows does not go all the way in doing it, it goes a little bit further than we have in the past. It will actually be the next version where we have everything built on top of Avalon that we'll give you that kind of flexibility.

MODERATOR: I think we just have time for a couple more questions.

QUESTION: And I have a question, you've mentioned research and development fairly frequently, and especially in the development platform as being a .NET developer. So I'm wondering how does Microsoft in its developer, idea or vision for developers treat research and development, and I guess like aspect oriented programming, intentional programming, database research, OO databases, stuff like that. As a developer sitting here, how should I be positioning myself? Like I really did research, but should I be pushing myself into research or should I trust that others will develop those research entities?

BILL GATES: Well, that's pretty much up to you. It's no doubt, to take the specific topics you talked about, the idea of having something that has all the benefits of object oriented databases and yet it has the same capability as traditional databases, that is something we and others are working super hard on, and it's a very necessary element because we have this world of XML data that is more heterogeneous than the classic table type structures, and we want to be able to process that in a rich way. And so there's been an explosion of funny little query languages and things that is being rationalized around a rich query model.

The whole way that queries relate to programming languages, there's a need for deep innovation there, and we showed at our PDC, talked about this iterator concept that we think is the way that we get it right into the language and so it's kind of a useful thing. So object oriented databases, (Peter Spiro's ?) job is to make that built into the system, just a standard way that you get things done.

Intentional programming is something we believe in but that's more vague. It's not like there's an existence proof of that. We want to write less code, we want the code to often be declarative in nature. Take the idea of managing a bunch of systems. What you'd like to do is say here's the services we want to run, here's the pool of hardware, we want these things to be prioritized with this kind of responsiveness, and, hey, Microsoft, your infrastructure software, you deploy this, if a machine breaks you redeploy it, we don't want to have any humans thinking about this at all, we just write some declarative constraints and then you go and solve that.

That idea is very important, rule-based programming, declarative programming has got to be part of how we solve the management problem, what some people call autonomic computing, how we solve business intelligence, and so those higher level abstractions we are investing in that more than any other company.

But unlike the OO database, I can't tell you exactly what year we'll have that solved. It will probably be rolled out over an entire decade as we make more and more progress on that.

So it's a Holy Grail, we love it, it's right up there with parallel programming. In fact, it fits in, in that some of those higher level abstractions lend themselves to automatically being broken down into features that can be run in parallel, and that's a very urgent thing because the new chips that are coming along really force us to do more parallel programming.

MODERATOR: Well, I'm sorry we are out of time for questions, but Mr. Gates was at Michigan this morning, and I do want to give him a memento of the winning team. So thank you very much.

BILL GATES: Super. (Applause.)

END