Remarks by Bill Gates

Government Leaders Conference

April 15, 1999

MR. GATES: It's great to be here at our second annual Government Leaders Conference. I can say in the last year there's been some fantastic progress in the awareness and the implementation of government systems taking advantage of technology. However, I'd say that today we're only achieving, at most, 1 percent of the potential efficiencies that are possible. And the benefits involved here are really quite dramatic as we move to a much higher implementation in the years ahead. What's been fascinating is that the visibility of these issues is much, much higher today than ever before.

About three years ago I met with the political leader of a large country, and after the meeting, we went out and the press had some questions. And one of the questions was, Mr. President, what do you think of the Microsoft Mouse? And, of course, this president was very confused because he thought we were talking about a rat of some kind. So, that just shows that three years ago it wasn't really on the political agenda. You didn't have people campaigning where they would say, hey, I am the right leader to deal with the opportunities and the challenges that technology creates for a country.

Today, I see that as very different. In virtually every country, whether it's a developing country or a developed country, politicians are talking about the kind of policies that they will create to empower the citizenry, to deal with global competitiveness, and to mitigate any issues that this global network will provide. That's a really wonderful thing. It's starting the dialogue. It's meaning that governments are getting plans into place.

It is a time of very rapid change. I have no doubt that the way we think of efficient governmental processes will be completely revolutionized in this next decade. You know, when we think about the legislative process, we'll think about that in a purely digital way, whether it's complete visibility of everything that's going on, who said what, who voted what way, and it won't be done as a paper-based system. When we think of the judicial process, we'll think about it being completely digital, the filing, the ability to review things, the challenges that people make, the back and forth, all done in a very digital way. And today, of course, no country has that fully implemented.

However, these visions, the kinds of statements and plans that political leaders are putting forward are all-encompassing. People like Al Gore, who have been talking about this for a long time, talk about technology as the thing that's really allowed the U.S. economy to be so strong, to create so many jobs, to help so many businesses, and it's really quite amazing to economists that this is taking place at all. It was unpredicted. Economists have a tough time measuring productivity, but here they can see absolutely that the advances brought by the PC are the only thing that can explain this wonderful prosperity.

If we look around Europe, the political leaders are also now talking about this on the agenda. I had a lot of other quotes from political leaders I could have included, so I just picked these two. But it really represents a sea change. The fact that voters are going to pick people based on whether they really know how to embrace the future for their country. Whenever I visit countries, I often get a chance to talk to politicians, and they're often asking, will my country be left behind, what is it that should be done?

It's very interesting that there is a wide awareness that usage of these technologies in the U.S. is very, very strong. But the area we're focused on this event, the government area, I don't think the U.S. is particularly ahead. There are a few examples of good things, but no better than we find around the world. In fact, if there's any pattern at all, it tends to be that smaller governments have done more. And perhaps that's because getting all the resources together, getting the interdepartmental cooperation is far easier in the smaller governments. In any case, I see it at a very, very early stage.

About a year ago, I had so many leaders, both business and political, asking me where the good examples were, and how could they measure how far along they were, that I decided instead of answering those questions so many times, it really was worth sitting down and writing a second book. My first book, The Road Ahead, sold very well, on a very worldwide basis. And so this time I thought I'd make it even more concrete, and really talk about it from an organizational point of view.

I wanted to talk about how there should be no impediments to people's creativity, to their best thinking in how business is done. And that's why I titled the book, Business at the Speed of Thought. Now, in writing, I wanted to capture the idea that there's a fundamental change in behavior that I'm talking about. It's not some small difference where you have a few more things on the screen, or where you're using email a little bit. It is a more profound change than that was brought about by the phone, or the car, or the television. The Internet, combined with the PC, is the most revolutionary communications medium of all time. And once people are used to it, once they're comfortable with it, it's incredible the number of things they do with it.

We talk about this as the Web lifestyle. And although many people get into it, starting with different things, maybe they start because it's the best way to manage their investments, maybe it's the best way to send photos to relatives, maybe they figure this is the best way to find the variety of books that they want to go out and get. No matter how they come to it, though, they find themselves trying to use it for a wide range of things, staying up to date on news. When they see a news article, that news article includes information about how their politicians are dealing with an issue, including their local representative. So, these people living the Web lifestyle are the vanguard of the future. And, yes, it's a small population. Today it's mostly found in universities and in technology companies where you have people who just take it for granted, use it the way that everyone will not too long in the future.

Now, the parallel to the Web lifestyle is when in your work you expect information to be easily available to you. For example, you expect information about expenditures to be easy to call up on the screen, easy to look at trends, easy to take anything you see there and mail it off to other people to analyze what's going on, what changes might be necessary in order to take advantage of that trend. And this is the Web workstyle. This is the elimination of paper forms. This is the empowerment of workers with all the information they need, and a new way of collaborating, a lot less meetings, a lot less feeling like you don't know what you need to know to make decisions, and an ability to react in a far more rapid way.

In this kind of a workplace, the simple jobs, the rote jobs, where you were simply transferring information between paper forms or putting it into the computer, or somebody called in and you would call something up on your screen, those clerical type jobs are not needed because they've been replaced by self-service, people going directly to their PC, connecting to the Internet, and getting the information they need. And so, all the jobs have to be converted into far more empowered jobs, where you can actually deal with complex situations, and be value-added in a world where just simple activities, like entering information, are no longer necessary at all.

Now, the broad approach here, we talk about as the digital nervous system. That's when a government steps back and says, what do we need to have online? How many paper forms do we have, and how do we start down the path to allowing people to work without using any of those as time goes on?

I want to emphasize what good news this is for governments. Governments are all about letting citizens have information, being very responsive to them. And this is such a breakthrough to take that fundamental purpose and achieve it at much, much lower cost than was possible before. The idea of standing in line to get forms. The idea of trying to figure out what the different government departments are, and dealing with them at different levels, city, state, national. Those can be eliminated by creating simple Internet interfaces for all the common activities. And so, to build the digital nervous system requires leadership. It won't just grow up from having all those PCs. As was mentioned earlier today, a digital nervous system, the hardware involved is what you've already been investing in. So this approach is simply about getting far more value out of that infrastructure, out of the PCs that are networked together.

There are many principles that are worth thinking about as you build the government's digital nervous system. The first is to really keep in mind what a fundamental change the Internet is. Every previous kind of communications device was different. The phone would connect one person, but of course you had to know who you wanted to talk to in advance. TV connects a show to millions of people, so it's got this incredible broadcast power, and the phone and the TV have changed the world. They didn't change it as fast as the Internet did, because there were several generations for people to get used to it.

Here with the Internet, we can connect any number of people with any other number of people. And we can connect them without them knowing in advance who they want to talk to, simply by having common interests, or an interest in buying something, or learning something, or working on something, they can be matched up and collaborate together. That collaboration can be in real time, like a phone conversation would be, or it can be asynchronous, like sending somebody a paper document which they can read whenever they would like to. So the flexibility here is quite amazing.

A lot of companies who have wanted to immerse themselves in thinking about the possibilities have assigned their executives to go out onto the Internet, buy books, plan a trip, learn about some medical situation and really get themselves personally involved in what's out there. It's only by doing that, that then they can come back and say, what does this mean for government, what about all those things, those brochures we're printing that get out of date so quickly, or that people don't know that we have, the people who might find them valuable, what about our deliberations and letting people have more visibility of those deliberations and giving us very good input.

So getting them to actually be users personally, although that's a challenge, that's very important. Some CEOs think of themselves as not ever learning to use the keyboard. And that's an incredible barrier. Some top leaders don't want to be embarrassed that they're not good at using the system, so you have to find a way for them to learn that they are comfortable with. But, really getting even the top political leaders to personally see how it works, I think that's important.

The Internet is going to constantly improve. The Internet this year is much better than the Internet a year ago, or the year before that. With things like having video and audio on the Internet, the improvements are very dramatic. You know, today you can go out to the Internet and get radio stations from around the world. So people who travel from one country to another can stay in touch with the music they like, the news, the sports. In fact, we built that feature into our latest Internet Explorer. And it's proven to be incredibly popular.

A second principle is, every worker is a knowledge worker. This is the idea of empowerment. This is the idea that so many of those jobs that have been rote in nature, simply taking a file and kind looking at it, those have to be changed to be jobs that are adding more value. In all of these activities we have to put citizens at the center. When they want to connect up and say something like, I want to start a new business, or I want to add an employee, or I want to check my medical records. It should be very simple for them to get that information. And so the views of information have to be organized by this citizen's view. And every contact with that citizen, where they've called in and asked a question, the best way to manage it is to record that. So if they call in again, you simply call up that record and see exactly what's been going on with them. And you eliminate this notion that they talk to one person and heard one thing, talked to another person and heard another thing, and you have to have them constantly re-explain who they are and why they're calling in.

Within a country, a government is always the largest organization and so in terms of really helping businesses get going on this, it's important for governments to take a leadership role. The opportunity for world competitiveness is really to reach out to new customers through the Internet. I want to emphasize how it intensifies business competition. In the past, to offer up a product you would have had to build a physical distribution network, and that was very expensive and it only worked for high volume products.

Today on the Internet you can have products that are very specialized - say you have a company in your country that's only going to sell 4,000 or 5,000 units of a product in a year, and sell it around the world. It just wouldn't have been feasible to find the buyer in the past. The middleman would have eaten up all the profitability in a low volume product like that. With the Internet today, people can search for the special things that's of interest there, and be matched up with no distribution network at all.

So you have lots and lots of companies who can sell more, who can do additional products that wouldn't have made sense in the past. You have lots of people who have great training, whether it's in software development or engineering skills, who in the past could only offer those to the local market. But now, because of the collaboration over the Internet, they can offer those on a worldwide basis. And so facilitating this, by taking the right steps so that the government is a model user, I think is of great importance.

I'm sure you see all kinds of statistics about Internet usage. And there are many ways to measure Internet usage. The most generous number is to say that anybody who uses electronic mail is an Internet user. There you get still a pretty small percent of the population is online. And you see that in the United States the percentage is higher than it is in many other countries. That won't stay the case. The reason that's happened is that this is something that builds momentum. The more people you get using electronic mail, the more you get new users who want to come in and use it. So it spreads not linearly, but non-linearly. And so you once you get the core group in there, then it moves out quite rapidly.

Certainly, the fact that we have such a large installed base of PCs was a key element, because we were able to take and convert those into Internet connected PCs. Some PCs that are sold, particularly in the home, are not yet connected to he Internet and that's because we haven't made the cost low enough, and we haven't made the applications good enough. But, I think in both of those areas there won't be a problem solving that. So every year that we have this conference I think you'll see these numbers will be growing up very, very dramatically.

Part of what I see Microsoft doing is sharing best practices. We'll be involved in some great pilots, everywhere around the world. And where those work or don't work, we want to share that information so that everybody is learning together. I decided this year that most of my time I would free up for question and answers, so I know a lot of the sessions have gone through these examples of things, particularly new things in the last year that have really showed us the potential here.

I wouldn't say that any government has done a comprehensive solution that spans all departments, whether it's really taking medical payments, or really taking all of the legal issues, or really taking education in the broadest sense. Even the best examples, which tend to be the smaller governments, are really just scratching the surface in some of those things. And yet, even with these fairly narrow systems, the reaction of the business and citizen users is quite positive. So we can use that energy to drive these things forward. What are the key issues? Many of those continue to be similar to what they were a year ago. A first step, I always say, is universal publication. Taking all government information, getting it to be easily navigable online, and tracking whether people are using that information. You know, seeing if they're finding it there in the right form.

Just to take an example from the United States, it used to be that the financial filings of companies were made on paper, and then the government would take them and store them on paper, and then a company would come in and type them onto the computer, so several weeks later, at great expense, you could find that information. Today all that information is filed on the Internet, and it's immediately available at no cost whatsoever to anyone who is interested. And so the ability to search for the information and find it, to be notified that there's something new there, it's just a completely different world than what was going on before.

And that kind of model makes sense for all information. Everything that's filed with the government. You know, if you take things like property transactions, if that's a public database, there should literally be a map online where you can click, look at a piece of property and see the information that's supposed to be publicly available. So universal publication of information, it's very easy, it's a first step, and if governments should take that as the first principle of what they end to do.

Second is that before you can really do a good job of getting this information out to citizens and businesses, you've got to have far better flow within the government itself. And this involves using electronic mail, making government to government interaction on a digital basis, by having a reliable email system that's used for all of that interchange. There's too much paper that's moving around between government departments. It's too slow. You know, things like budgeting, looking at tax receipts, dealing with citizen complaints, that can be done so much better, once it's digital. So I'd say that's really a second agenda item, is to say where we know that there's lots of information, where we can train the employees, where we can put in the infrastructure, which is the people who work for the government. Let's make sure we're doing that very, very well, as a first step.

Another easy step is to say that the most immediate demand for these systems will come from businesses. Businesses, because they complete globally, because the communications costs are affordable to them, they should be able to work with the government without paper flow, not in 10 years, but in most cases within three or four years. It shouldn't be that hard to take what's been done in paper form, collecting value added tax, registering various things, take all of that, and say that the Internet will become the only way that that gets done. Often for systems you'll have to maintain a parallel approach, but for businesses, particularly large businesses, that shouldn't be necessary.

Two of the tough issues that governments face are authentication and broad access. Authentication I know has come up a lot here in the conference. It's interesting to think how you know who you're interacting with, before we talk about digital systems. When somebody comes in, calls the government, and says, I am a certain person, how do you know for sure who that is? Even if they come in physically, how do you verify exactly who that is? Authentication is an imperfect situation in the pre-digital world. It's actually possible in the digital world to do better. And country by country, that will either get solved by the government issuing some sort of a smart card, or finger print type ID system, or password, or it will be done with the combination of the government and the commercial entities, often the banks there. It can be done with credit cards, it can be done with smart cards, there's lots of ways to solve that issue. In fact, I don't think it should hold people back as much as they think. With most transactions, even a weak form of authentication can work very well.

A final issue is broad access. You know, when you talk about the ultimate, which would be something like having voting be done on a digital basis, there you'd like to make it easy for everybody to participate quite easily. Some of the projects where the Post Office has been used to access these online applications have gone very, very well. And yet, that's only happened I a few countries. Part of the reason that's tough is you've got to get all the different departments that would benefit from that broad infrastructure working together. So if it happens to be in the Post Office, it's not just the Post Office applications that get implemented there, but all the government departments think, wow, now that we've got this universal deployment, so that everyone can have a terminal that's nearby, how do we take advantage of that. And so I'd love to see more of those rollouts, so that there's countrywide accessibility, which will really feed the interest in doing the applications.

Now Microsoft's role in this continues to be a very specialized role. What we're good at is building the high volume software standards. We don't do the chips, we don't do the systems, we don't do the communications networks, but we work with all the leaders in those industries to make sure the pieces come together in a total solution.

In every country, we've developed local software developers and local solution providers who need to get into this process of helping to build these systems. And so, we use our expertise to help the pilots get going, but really want to transfer that expertise to local experts, local companies as much as possible.

Our key products remain the same, Office, Windows, BackOffice, and then our Internet presence which is branded under the name MSN, or Microsoft Network. Those four products are 90 percent of what we do. And everything else we do is really in support of those products, the development tools, the support operations, all the things that we've put together.

Now, across very one of those products, we have a number of key initiatives. Driving the Internet forward, making it so that things like video conferencing will be commonplace. It's not today because most networks aren't fast enough, or the network administrators are worried that that video traffic will crowd out the other traffic. And so working with industry leaders, we've now built into the next version of Windows this ability to prioritize the traffic, and so you'll see video conferencing, which has been very limited use, explode in popularity.

Now, that we think of as a very important application, because it enables distance learning. It enables distance medical consultation. And so making sure that the video information, you know when your network has the capacity to deal with it, and it's only used at a lower priority than some of the other traffic, having that standard in Windows we think is very important.

Interoperating with older computer systems is another big priority for us. If you want to create this common view of information, you'll often have to go out to applications that were written on a mainframe or UNIX system. And the ideal is to be able to extract that information and create this new interface that was developed on a Windows Internet platform, using the modern tools. And because of these interoperability pieces, that is very possible. Now, over time, those applications will migrate down onto the Windows platform. But you don't want to have to do that at the same time as you create the common rich views. Those views can be created leaving the applications alone, simply through that interoperability.

We also have a key initiative to make sure that even the very inexpensive PC servers can deal with billions of transactions, and there's incredible progress there, both breakthroughs in hardware and breakthroughs in software. And so scalability is not going to be an issue at all. You can use the mainstream PC hardware even for your most demanding systems.

A final initiative is the one that I spend most of my time on, because there's so much cleverness that we can apply here. And that's making these systems simpler while we make them more powerful, eliminating commands, eliminating utilities, making it so that when you want to search for information you don't have to think about, is it in my files, is it in my mail, is it out on the Web? All of that brought together, a lot less to learn, a lot easier to work with. And there are dozens and dozens of things like that that the power of software can do on behalf of the user.

Now, one thing that still causes people to underestimate how the PC and the Internet are changing the world is that they underestimate how quickly each of those will improve. The levels of R&D of all the companies in this business are at record heights. Microsoft has grown its R&D to be about $3 billion a year now, and that's all great software developers pushing the frontiers of how software should work.

Some of the breakthroughs that I see coming that I think will really surprise people and make a big difference, I want to step through very quickly. First is advanced collaboration, allowing people to make decisions by working on a PC Internet basis. Take something like managing legislation, all the different committees reviewing it, editing it, tracking who made the changes, what the status is. That's a great collaboration example. We will build in to our BackOffice Exchange and to our Office products, all the interfaces that allow that kind of collaboration to be very, very easy. This is often talked about as rich knowledge management. Well, that's coming to the Office and BackOffice software.

When we think of networks today, we mostly think of wired networks. But the advances in the wireless area are going to be very rapid. And so, instead of thinking about your phone, your PBX with all those wires, even in a phone environment, you will have just a small pocket-sized device and nice screens that you can use to get your data and to make those phone calls. And so, your desk will have a large flat screen for creating and editing documents, and then a small little screen device that you can put in your pocket. And either one of those you can take with you to a meeting or wherever you want to go because at least in your business environment, it will be a wireless connection. And so wireless will play an important role in bringing the PC into new usage.

I mentioned earlier getting voice and video on to the Internet, making that a standard capability is very, very important. When you go to a Web site, instead of just having screens of information, if you're curious to get some advice or help, you'll be able to push a button and somebody will come in who can see what you're looking at, what you've been looking at, what your history of interaction is, and talk to you so that you get the best of a phone conversation and the Web in one interface.

The progress in making PCs smaller has been quite incredible. These portable machines are now less than three pounds. And it's even faster than we would have expected. There's a new form that will be, I think, quite popular in the next two or three years, and that's where you can detach the keyboard and simply have the screen itself, and that screen will detect any notes that you write on it. It will be just like a real tablet. That kind of handwriting recognition and annotation is a huge area of investment for Microsoft, and we're very close on this as almost a standard feature, certainly in less than three years.

Speech input is also coming along, although in the next two or three years, it will mostly be used with small vocabularies, when you want to call up on the phone can get your schedule or mail, that will be very straight forward. For dictating documents, people expect a very high level of accuracy, and we're still very optimistic about that, but I'd say it's more like five years before that will be the primary interface.

We'll also see in a way that's complementary to the PC these smaller devices that I often talk about as a personal companion. Now, these can come in many forms, the palm-sized PC for the pocket, the small portable machine we call the hand-held, electronic books, the smart TV connection that replaces the set-top box, or even something that's in your car that you can give commands to and it has speech recognition. The software that's used in these will often be the same Windows software, so we can make it easy to have all your information available on any device without you having to move it around, taking a lot of effort to do that.

One of the real ambitious goals we have is not only this speech recognition, but also computers that talk and computers that see. Microsoft has had a long commitment to this, and that's what it really takes. It takes getting in the great researchers, and driving these forward. We can already see a number of these technologies in limited form in shipping products.

For example, in Japan, people have characters that the keyboard makes very difficult to enter. If you don't know the pronunciation, you can actually use the mouse to draw the character, and that kind of handwriting recognition has proven to be very popular. Our encyclopedia has the ability to talk and speak the contents of articles. Our Office products check your grammar and give you good advice there. And so these advances are on the way.

In fact, since you were kind enough to take the trouble to come all the way up here to Seattle, where our headquarters are, I thought it would be fun to get one of our top researchers in to actually show you some of the work he's doing in these areas. And so, I'm really pleased to have Matthew Turk, who works on vision at Microsoft research.

Welcome, Matthew.

MR. TURK: Thank you. It's nice to be here.

Well, as Bill was saying, computers have for a long time used mostly a mouse and keyboard to interact with people, but we're working in various ways, language, speech, et cetera, to change that, and improve and make computers more interactive, more natural, more compelling. And one of the ways is vision.

As you see, I have a camera on top of a monitor here. My group is working on ways to take the input from that video camera, which is being fed into the computer, process those images, and make some sense out of them and try to answer questions like, is anybody there in front of the system. If so, who are they, where are they, what are they looking at, what are they doing, what kind of gestures are they making, et cetera. Things like that, that we use all the time when we talk face to face as an important part of our interaction. Now, the question is, can we endow a computer with similar sorts of ways to interact with us, and we think we can.

So, I'm going to show you a few little demonstrations of some basic technologies that are based on this vision that I hope you can infer what we might use this for sometime in the future. First of all, I need to move out of the scene for a moment.

Okay. The first thing I'm going to show you is a movie playing on my screen, and it's just going to be a movie playing. But the computer actually will notice when I'm here in front of it, and when I'm not. There's the movie, I'm sitting here watching it. Now, the phone rings, or I need to get a cup of coffee or something. I leave the scene, and it stops playing. Okay, it's paused. It knows that no one is there, so it says, okay, I'm just going to wait. I come back into the scene, it starts playing again. I can move around here, move up, move back. It just knows I'm here. And I can even us a gesture, something like this, to stop it, to pause. And then start again. So, that's little demo number one.

(Applause.)

MR. TURK: Little demo number two is, the computer is going to have an indication of how far away I am from it, and so it's going to present information to me differently depending on where I am. So, I'll start out here. There's a slide let's say I'm editing or I'm presenting to somebody, and here I am at a certain distance, and the text is just right for this distance. Now, if I move away, it scales with me, it gets bigger. And now, I can see it just about as well now as I could when I was close to the screen because of that size. I nod my head again, and I can use my indication no or yes with my head to stop the behavior. So, I say stop, and it doesn't do it. But I come back and say, okay, go ahead and start doing that again, and it will start doing it again. So, again, just a small example of a certain behavior. That has changed depending on a visual aspect of the scene.

And now I'm going to play a game with the system. The game is tic-tac-toe. First, I'm going to train it. Now, it's just noticing how I look when I look at different parts of the screen here. And then in a moment we'll start playing the game. Okay, so now we start. The red box is giving you feedback as to where the system thinks I'm looking as I move my head around, and I'm not touching anything. So now when I stop for about a half a second, it plays my X. Now the system plays, and now it's my turn. I'll go up here and play, and now the system plays. I guess I'd better go right here. And now the system is kind of stupid, so it plays there.

Now, the last time I did this for Bill, I actually lost. So, I'm happy that I won this time. Now, again, just to note, the point here isn't necessarily that everybody wants to play tic-tac-toe by moving their head around. The point is just a simple demonstration of the computer actually sensed approximately where I was looking, and took action based on that. So, if you can imagine, for example for a person with a certain kind of disabilities who can't maybe use a mouse and keyboard very well, using your head as an input device would be a very useful thing, and for computer games and all sorts of things.

I'll show you real briefly the technology behind the few demos that I just showed. This window on the far left here is just the background that was taken when I walked away earlier. The first live window here, you can see, is just a live video coming from the camera up here, and this one in the middle here, the black and white one, is basically a difference between the two previous. So, when a user is there, you're going to see some white stuff. And then the thing on the far right is basically a draping that gets just an outline of the person. And then, given that outline, I can go back into the original live image, and I draw this green box that shows where I think the head is, so as I move around, it tracks. As I move back, the box gets smaller or closer.

So that the movie player was just looking for a head, when it didn't see a head, it paused the movie. When it did see a head, it went ahead and played. Similarly, the screen that was getting scaled, when the head box was very big, like it is now, that window was small because it knew I was close, when the head box was small, the window was bigger because I knew it was further away.

And then, one more, a little demo here. This is actually the beginnings of an environment for small children to interact with the machine not through a typical keyboard and mouse, but through moving their body around and reacting to characters on the screen. I'm going to show you a few little vision demos that we're trying to do gesture recognition, recognize certain gestures that the kids are doing.

So, on the right side here, you just see sort of certain parts of the vision routine that I won't explain, but on the left you're going to see action and you're also going to hear some action based on what I do. So, right now, I'm going to look for virtual cymbals. So, I'm not touching anything again, I'm just clapping my hands in front of the machine, but it's noticing that and causing this action of the cymbal sounds to play.

Similarly, I can play the bongo drums just by doing this. And I can be a bird by flapping my wings. And just as an example, if I only do one hand, it doesn't work. So, a bird has to flap both wings to get it to go.

And I'll ask Bill to do this last one, to come up and conduct a little song for us.

So these are just some examples of how a computer can, through vision, interact with users. And we hope that someday before too long these kinds of interactions will be in our products.

MR. GATES: Super. Thanks.

(Applause.)

MR. GATES: Well, it's very exciting to see that there's very few limits to how far software is going to take us. You know, the kind of things that we're good at, dealing with language, making gestures, all of that, we'll be able to take full advantage of. I just wanted to wrap up by emphasizing a couple of points.

There's a lot of concrete steps I hope that this conference has made clear that are really worth looking at, the use of electronic mail, the issue about the forms, the issues about transactions being online, a lot of what we've been talking about assumes that you've got a communications environment that's advancing to drive down costs and drive accessibility even into remote areas. And so in a lot of ways that takes on an extra urgency with the Internet, making sure that schools have low cost access to the Internet, making sure that the latest connection speeds are coming along, things like DSL that's off to a fairly slow start, I think in the next year or two we are going to see an explosion there, and that will be very important, and that's the phone companies who need to make those investments.

Education is really probably the most important aspect of all of this. The payoff there doesn't come immediately, you know, starting with the universities, but most importantly in all the schools, this is really the way there's not going to be a have versus have not problem here. And I'm really pleased to see the number of people here who are very focused on education as the biggest opportunity.

From Microsoft's point of view, our commitment is to be involved in helping drive this forward, whether it's the conference, or the Web sites, or making the tools better. This is something we believe in, because it fits in with our view of empowering people to do the things they want to do, in an efficient way.

I think the bottom line here is that you could almost say there's a competition between governments over the next decade, to see which ones are really going to be responsive, which ones are going to reinvent themselves by using these digital tools. And we're very excited to be a part of this, and help in any way we can.

Thank you.

(Applause.)