Remarks by Bill Gates
Microsoft Corporation
Windows World
April 20, 1998
Chicago, Ill.

MR. GATES: Well, we’re all very dependent on technology. It doesn’t always work. In fact, I live in a house now that uses a lot of technology. I’ve got over 40 Windows machines that I depend on to turn my lights on and off. And it works most of the time. It brings a certain thrill to simple tasks. (Laughter, applause.)

It’s been two and a half years since the introduction of Windows 95, which, of course, was code-named "Chicago." And these last two and a half years have brought a lot of surprises. You don’t hear as much about OS/2 as you used to. Network Computers, they kind of came and they went, all without selling many units, but that was exciting. And now the big focus of the industry is tied into reducing costs of ownership through greater simplicity, and taking advantage of the exploding standards around the Internet.

Another surprise has been the low-cost hardware that’s come along. Really no one predicted that PC prices would come down as fast as they have in the last few years.

But the goal of this industry has stayed the same over the last 20 years, since the beginning of personal computing. And that is to create a tool that magnifies our creativity and allows us to communicate with each other in a new way. The hardware advances are a very central part of this, and so I’ve been pleased to see that the pace of these hardware advances is faster than ever. Amazing microprocessors, like the 400-megahertz Pentium II that Intel’s just come out with. A move towards better screens that I think is a more fundamental advance than people recognize. We’ve got to get to the point where the computer screen has the same type of resolution that you get with paper, and that it’s a tablet-sized device that you can just carry around. In order to really be part of the mainstream, it’s got to take on that new form factor, and that will mean that we can take it with us, take it into meetings, use it in a way that even today’s portable machines don’t allow us to do.

Storage size is pretty awesome. We no longer have to think of running out of storage for documents that we type or even the code that we store. The size of the storage is allowing us now to deal with things that just wouldn’t have been realistic: lots of high resolution images, and in the years ahead even lots of motion video.

DVD, coming from the consumer electronics industry, is going to be a fantastic thing, because not only will we have read only DVDs, we’ll have read/write DVDs that will let us move video and other data types around very easily.

The one tough problem we’ve got in hardware advances is getting everybody connected up at high speed. Businesses are being connected up, but connecting up homes is tied up in a lot of regulatory issues and investment questions that mean that it will probably be the slowest moving thing in the industry. Most people even five years from now will probably still be connecting over the phone lines. And so compared to all the other elements, that is really going to be a pretty big bottleneck.

Microsoft and many other companies are doing everything we can to push that forward, but it does take a lot of work with the phone companies and the cable companies, making sure the regulations are right, and then driving some new technology, whether it’s the cable modems and various forms of DSL.

If we look at the next couple of years, the pace again I think will accelerate. No doubt we’ll get up to 1,000-megahertz clock speeds. We’ve got a huge advance with the Intel IA-64, the Merced chip. That brings a 64-bit address base: 2 to the 64^th is a pretty mind-blowing number, and that’s going to let us put a lot of things in memory that we would have kept on the disk before, and therefore get even more performance than you’d think from just the pure instruction set.

Up at the server level, the goal is quite simple; it’s to take on not only the tasks that mainframes would have been used for in the past, but to go even beyond that.In the world of the Internet where you have so many people coming in and doing transactions, and you want to model what types of customers are doing what on your site, where are they going, what kind of special offers should you make to them, the need to process data is far beyond what would have been required before. And so moving up not only the individual processor, but putting more processors in the machines, clustering those machines together and advancing the software; all of those improvements multiply together to give us new machines that are based around the PC hardware standard, and therefore give people the flexibility, but also that have this unbelievable performance and reliability that the high-end task will require.

And so that’s a major focus for Microsoft and the industry.

The key thing that Microsoft does is providing Windows, a platform for all these advances to build on. A key element is that Windows creates independence between the hardware changes and the software changes, and so you have total choice on the hardware side and total choice on the software side.

There are a few key principles that are behind everything we’ve done with Windows. And people often ask, what allowed Windows to become the most successful software of all time. Today there’s over 150 million copies of Windows in use, and that’s a number that will continue to go up very substantially.

The principles behind Windows from the very beginning have been these:

• to maximize customer value through first innovating in Windows itself;
• second, by working with the software industry very broadly to provide great software choices;
• third, to work with the entire hardware industry to make sure that everything they are doing fits in with Windows;
• and finally, and very importantly, working with the largest part of the computer industry, which are the services companies, the people who do training and integration and build solutions and make sure that they’ve made investments so that they can apply Windows as a building block in the things that they’re putting together.

So this has come together in an incredibly successful way. The R&D that we put into Windows has to go up every year very substantially. We’re now well over a billion dollars a year spending to improve this software product.

In a certain way you could say it’s the most expensive software product of all time, because we are investing more in it than is invested in any product in history. In another sense, it’s the least expensive software product because the amount of R&D dollars people get for dollars they have to spend on it are extremely low. And that’s all driven by the volume. Volume is key to this equation. We knew the Windows principles wouldn’t kick in unless we could get PC computing up to millions and even tens of millions of units.

Now, let me touch on a little bit on each of these. First, it’s important to look at where Windows was, and what would have happened if we hadn’t constantly increased that R&D investment. In 1990 Windows was not very successful. Most people were using DOS. We hadn’t integrated DOS and Windows together, so you had to install them separately. Adding new devices was very tough. The applications were still doing their own printer drivers. Modem connections were basically a very rare thing; this idea that communication would become central just wasn’t what people were focused on. There was no networking or Internet. And we spent a lot of time thinking about memory management, so-called extended memory and how you would configure that.

So we have come a long way. The things that were tough and problematic back then, things like printing, we’ve moved ahead.

Now, the beautiful thing about this business is that people’s demands of what they want out of the tool constantly increase. You know, today people want to have all the information about their business at their fingertips, and so there’s a lot -- a lot more that’s being done that goes way beyond even where we are right now. But that pace of innovation has been critical.

You can kind of divide it into three different generations.

The first was the move to graphics interface itself, which is where Windows got started and took us from 1985 all the way up until about 1992 or ’93, before that was mainstream.

Starting with Windows 95, which had the first version of Internet Explorer built into it, we entered the middle phase, which is really putting Internet information browsing at the center of the Windows experience, and designing everything around that. Whether it’s the information that’s stored locally on your machine, or the information you get out through the network, over time all of that information will be formatted according to Internet standards. The protocols will be Internet protocols and the documents themselves will use standards like HTML and XML; and we’re probably about halfway through this stage. There’s still quite a bit that can be done, but the key foundation pieces have been put in place.

Now, moving forward, as we get that completed, the next frontier is the ultimate dream of computer science, providing all the natural interfaces, reducing the complexity and starting to bring in speech, handwriting and vision so that the way you work with your computer doesn’t require simply working with the keyboard and the mouse.

And I think people are underestimating how quickly that will come, because all it takes is the magic of the hardware improvements and the software investment that we’re driving so aggressively.

As for Internet integration, some people already understand exactly why this is being done. But some people still wonder why have we made this front and center. It’s kind of like the graphical interface; once you’ve really started to use it and appreciate it, you wonder, why was it ever a controversial thing.

Today, we want you to not have to think of different tools. But the shell gives you a certain way of finding files on your hard disk. The help system gives you a certain set of commands for navigating around. And the browser has another set of commands. They each have their own name. They have their own way of making a mark and going back to that mark. They have their own programming model. They have their own formats. A format for directory is not the same as the format for a web page directory, and that’s not the same for the help system directory. So both the user and the developer have to deal with the fact that these things are not unified.

Well, with Windows 98, which is in the verge of shipping -- and we’re going to show you all the great stuff in Windows 98 a little bit later -- you just learn one thing, and that is you learn the browser. The browser with its history list ability to go backwards and forwards, the ability to display things in a rich way. It is used for all of these things. And so our proprietary help format and all of the utilities that went with it is now obsolete.

It’s still actually there for compatibility reasons of using applications which still have those help files, but all the work that’s being done in the future simply takes the great HTML authoring tools, and the enhancements we made to HTML, things like advanced Windows controls that HTML didn’t have, they used that to be able to do their help tasks. So, really it’s all about ease of use and allowing you to think of off-line use of the computer, and on-line use of the computer as really the same thing.

Now, this idea of integrating in new functionality, we have a roadmap that goes out to almost the next decade of things we have in mind, and we’re very clear way in advance what things we’re going to build in. We’ve put together a whitepaper that’s up on the Microsoft Website (http://www.microsoft.com/innovation/) that talks about how developers view this and how end users view it. It’s a very important principle for us.We can’t continue to deliver the platform unless we’re full speed ahead on the innovation and integration that has characterized Windows over the years.

A major theme of the industry and one of the things we’re trying to do with the integration is bring some additional simplicity, make it so you don’t have to learn as many commands. There are some things that are very complicated today that should be simple. For example, say you’ve been using a PC for a couple of years, and you want to upgrade to a new PC. How do you take everything you’ve done on that first PC and move it over to a new one? The information is spread all over the place; it’s in the registry, you’ve got files everywhere, and there’s kind of a mixing together of things that relate to your preferences, and things that are specific to that hardware configuration.

And so just wholesale taking all the files over doesn’t work, because you have to take out the things that are hardware specific. It should be as simple as giving a single command, say, take my "state" on this machine and move it over to this machine. Or perhaps an even more advanced scenario is that you’re temporarily using somebody else’s machine, you just want to log in and have access to the way you lay out your desktop and your favorites and your spelling dictionary. That’s almost impossible today. It’s certainly not worth the trouble that you’d have to go through. Well, that’s an example of something that we’ve got -- we’ve got to change and make very simple.

I’ve got down here a list of some of the scenarios that we’re going through and thinking about what -- where can we get rid of steps, where can we get rid of concepts. We’re also reviewing all of the error messages that you can ever get from Windows, and it’s pretty shocking how many error messages there are, and how cryptic they are. One of my favorites is this one here, which a lot of users really understand: "The DHCP client could not obtain an IP address. If you want to see DHCP messages in the future, choose ‘yes’; otherwise choose ‘no.""

Well, the most helpful part is that "otherwise choose no," because everybody understands DHCP, but the difference between yes and no, you know, we’re really trying to help you out by giving you that contrast there. [Laughter.] And you’re not really sure, you know, what if I say "yes," what if I say "no." I mean, I didn’t know what this thing meant the first time I saw it so maybe [Laughter] maybe I never do want to see this thing again. [Laughter, applause.]

One of the worst things that happened to me recently is I showed this slide, and brought up this error message, and even after I’d given the speech, people actually thought that I’d gotten an error in the middle of this speech [Laughter] rather than just put it up on the slide.

So we’ve got a long ways to go. Even an error message that will say, you know, I couldn’t associate something with this file, and so I don’t know how to display the file, and it might even tell you to go to some obscure place and do something. Well, the error message itself--if it has to come up--should have the buttons you can click on to change things, to diagnose things. Basically, the troubleshooters should be directly connected up to the error system. And so you shouldn’t have a help system, a troubleshooter system, and error messages as three different things.

In fact, in the case where you’ve got an Internet connection, you ought to be able to go get additional files or go get advice or go report the state of that system automatically just by navigating from that error message in a very simple way.

So we’re stepping back and we’re thinking of these situations from a user point of view, which is the right way to do it.

We’ve gone out and talked to a lot of people about these problems. Particularly if we’re going to get into the home and start to integrate in all the different devices there we’re going to have to make things a lot better.

In fact, we’ve put together a little video that talks about what it’s like right now to install all these devices in your home. So let’s go ahead and take a look at that.

[Video clip parodies the complexity of using computers and connecting to the Internet today.]

BILL GATES:It’s a lot of work to do there before we get all of these devices working together without people having to understand how the pieces come together.

The second principle after innovation that I mentioned is having the best software choices. This means working with the entire software industry. Actually, when Microsoft was founded there wasn’t much of a software industry, because computers were sold in low enough volume that it wasn’t economic to write packages. You would have had to charge just way, way too much. As the PC came onto the scene, the software industry has exploded in size, and having a standard platform is the reason that’s taken place.

Today we’ve got over 2,000 of our employees who are focused on software developers -- meeting with them, helping them with technical things, helping them with marketing, answering their support calls. So it’s a huge part of Microsoft. It’s over a hundred million dollars of unique spending that we do. Now, that’s important to us because we want them to not only use the present version of Windows, but we want to work with them on the next version. So we talk to these people every time we can and get out in front of them very early on. If you go down to a bookstore of course you can see in all the things that are written about Windows to help people understand even the sort of ultimate details.

A great example of how we work with developers is the next major version of Windows NT 5, and the same sort of chronology you would have seen for Windows 95 and Windows 98. We get out in front of people very, very early. Here I show the first date is August ’95, which is two and a half years ago, right when Windows 95 was coming out. It was the first time we sat down and talked with people about key parts of NT 5, what would the directory be like, and so on. We have a number of other events, management service preview, developers conference, an admin preview, another developer conference, all the information up on our developer network, which is called MSDN. And then subsequently in April ’97 we opened up our development labs for people to do testing. We did another design preview.

Finally, in September '97 we reached a milestone where we gave the data out to over 6,000 developers, and this includes everyone, even to our, you know, arch competitors were there at the developers conference. I think Sun had 30 people, Netscape had 20 people. So, you know, anybody who wants to develop is welcome to do it. We put this same data on the developer network and so we got over 100,000 copies out. We’ve updated that a couple of times since then. And the big milestone that’s coming up on this one is dated too.

Windows NT 5’s a very big product for us, but the key point I’m making here is that we do everything we can to help software developers take advantage of the platform. One of the big issues we learned about early in the discussions on NT 5 was that people doing enterprise applications were having to do a lot of work that really should be done in the operating system. And that’s why you’ve seen Windows NT 5 over the last couple years really be beefed up dramatically in areas like transaction management or message queuing capability, including gateways to connect those things up to their mainframe equivalent, things like CICS. We hadn’t realized what a huge part of a high-end application’s development with doing something like that, but really matches our skill set and shouldn’t be duplicated across many, many different applications.

But Windows NT 5 is a very broad product. It was certainly during the last two years that we recognized the critical importance of cost of ownership, and we built in the IntelliMirror capability, which lets you handle the state problem I talked about earlier, where you can easily move to any machine, you can easily upgrade to a machine.

We also decided that this Windows terminal approach, letting low-end users have the execution of their applications take place on the server, that was important. And so that’s been folded in and will be integrated into the source tree as part of the NT 5 shipment.

So, during this several-year period, the entire industry has come together, let us know what they want in the product, and so we’re hopeful that the day the product comes out we’ll not only have lots of updated applications, but lots of utilities that build on top of the new things that we’ve done there.

So I think it’s a great example of how we like to work with the software industry.

This developer network I mentioned, it is pretty overwhelming the amount of information we give people. It’s 163 CDs. We could only get 50 CDs onto this slide, so, you know, 163 CDs is a lot. Now, part of what’s there is every localized version of Windows in the world: the Thai version, the Arabic version, the Korean version. We give that to people so if they want to do separate test cases against those or do special features against those, they’re able to do that. But every bit of information we can provide is put onto those things, and those are put out on a quarterly basis.

Another key principle around Windows is this idea of the best hardware choices. And that’s a very fundamental thing, because although the word "open" is used again and again and again, most of the time it’s used where people don’t have total flexibility, where you can’t pick among many different suppliers for each new piece of hardware without having to change everything about your software.

The PC industry was the first to deliver on that fundamental openness.

We have a lot of partners in the hardware industry, you know, from Compaq, the largest PC provider; Gateway, who was a great partner on plug and play; Dell, who’s been a leader in driving the direct model, working with us on electronic commerce, what it means to offer these products through the Web; and Cisco, who’s our partner on things like quality of service and security over the network, making it so that work at home will be easy. We’ll start to see video and audio across the network, some very exciting things going on there.

Intel is probably our most important partner, because we are the two leaders in this industry. Almost every initiative we’ve done, we’ve done with Intel, including the feedback we give them on their processor work. We do incredible analysis of what they should add to the processor to speed things up for us, and they’ve been incredibly responsive in providing new capabilities. HP is a partnership that’s been greatly strengthened over the last two years. They’ve made Windows NT the heart of their strategy, and so whether it’s in the area of services in the field or building small devices around Windows, making sure their printers work around Windows, there’s a lot of new things that have come out of that.

The final Windows principle had to do with working with services companies, and you’ll see that there are a lot of companies that are both in the hardware and the services business. I put Digital here, because they have the most certified people out in the field who’ve really gone through the advanced Windows training.

But all sorts of companies now are getting involved in that: KPMG, Ernst and Young, Arthur Andersen; each of them have competency centers that focus on particular ways of using these things. Amdahl has taken their mainframe expertise and created a group that’s very strong in mainframe integration. Two new partners for this list are Unisys and NCR. We completed major agreements with both of them in the last six months, and so their strategies have shifted to be focused around Windows NT and very much value added strategies: applications software, data warehouse capabilities, moving out to the vertical things like retail banking and other areas where they have been very, very strong.

In fact, one way to look at this is to say, well, let’s look at all the big companies in the computer industry and say what role Windows plays in their strategy. And for all of these companies, which are the large companies in this business, all of them except two have Windows as their primary strategy. They see Windows as being the thing that they’ll bring their value added to.

The two exceptions you can probably guess, one is IBM which has a mixed strategy. They do a lot to support Windows, but they have other things they are focused on as well, so they wouldn’t say it’s number one although there’s a lot they do there and we’re glad that it’s at least part of their strategy. And then you have Sun, which is purely anti-Windows, and not at all shy about it. [Laughter.] So it provides a counterpoint to what everybody else on here is doing and what they’re pulling together.

Now, if you’d looked at this same list of companies just when Windows 95 was introduced, less than half of them would have had Windows at the center of their strategy. So that’s a fairly radical shift since the introduction of Windows 95. And it’s really due to the fact that the microprocessor speed has come up, Windows NT has gained critical mass and people are seeing that the volume there is really driving things in that direction.

The Windows family has three different levels, and we’re trying to be very clear about where each of these apply. The most powerful thing, the thing we’re driving business towards is Windows NT. It requires somewhat more memory, but it is with Windows NT 5 a strict superset of Windows. That wasn’t true with NT 4. We didn’t have the power management and we didn’t have the full set of drivers, so there were some tough trade-ups people had to make and mix things together. With NT 5 the only tough trade-off is how do you get there in making sure the systems have enough memory. But the security, the richness of that system is why we recommend it for businesses.

Windows 98 will take over from Windows 95 very rapidly. All new machines by this summer will come with Windows 98 built in, because Windows 98 is a very obvious move up from Windows 95. It doesn’t require more memory. It’s not like the transition from 16-bit to 32-bit with Windows 95, where there were a lot of challenges with the drivers and everything. It is a straightforward refinement with a lot of neat new things.

The new tier we’ve got in here is this Windows CE tier, and that applies only to new hardware devices, but there’s actually quite a variety of those: PC companions, like these devices with the small keyboards here. You can see the neat color screen models that have come out in the last few months and are selling very, very well; in fact, they can’t make enough of those. A new form of the PC companion will be the one without a keyboard. Those come out this summer, and they use the touch screen with the pen. They do handwriting recognition or you just simply browse the information very easily on that machine, which is a palm-sized PC is the category we have there.

Another unusual form factor is the Auto PC, and that’s the first time where you’ve seen speech recognition, because we don’t want you to have to be looking at the screen instead of looking out and driving while you’re interacting with the device.

Because there’s a restricted vocabulary, that’s going to work very well, and that’s just showing how speech is starting to come into the mainstream.

At home it’s WebTV is our big focus. That’s done very, very well and next year’s version brings Windows CE into that.

And then there’s a lot of embedded systems, industrial systems where Windows CE comes into play. Windows CE is much smaller. It doesn’t require a hard disk in the device. But it is a strict subset of the applications interfaces.

We had our first developers conference for Windows CE just a month ago. We were hoping to get 1,000 people for that, and we had over 2,500 people show up, so it was a positive surprise at how much interest there is in all these different form factors.

Well, let’s now talk about the next big release of Windows, Windows 98. This is a product that we’ve spent the last two and a half years with the Windows 95 team going out, talking to customers, looking at every support call we’ve gotten, talking to our partners, and we’ve picked three major themes for the release.Performance and reliability being the first thing. Second is ease of use. And there’s a lot of points to support that, but particularly the Internet integration. And finally support for a new generation of hardware. And when we say new generation of hardware, we mean a lot of things. Probably the most interesting is USB. Now, that’s actually been out on quite a few PCs over the last year. There’s about 30 million PCs that have USB capability, and a lot of neat devices are being connected into it. So there’s a lot that is going to enable your PC not to have as many wires and things, and yet to have richer peripherals.

The best way to understand Windows 98, like all of these products, is actually to see it in action. And we’re very, very close to releasing it. So I asked Chris Capossela, one of our program managers, to come along and give us a glimpse of some of the things that Windows 98 allows.

[Demonstration of Windows 98 covered a number of features. It showed that Windows 98 starts programs faster by tracking the applications that a customer uses the most often and arranging the disk space to optimize loading time for those. It showed a Maintenance Wizard that, among other features, scans a hard disk for errors, and deletes unnecessary old files. The demonstration showed Internet integration for easy browsing of the hard drive. It showed Troubleshooters, which walk users through steps to solve a variety of problems. It showed how easy it was to connect to the Microsoft online support site and updated information and product components. It showed dual-monitor support. It showed support for a variety of devices, including DVD, which played a movie, a Universal Serial Bus, which enables a number of devices to be plugged into the machine. (The last step in the USB demo failed, prompting Mr. Gates to note, "that must be why we’re not shipping Windows 98 out yet.")

BILL GATES:Well, let’s look out beyond Windows 98 and Windows NT 5. We’ve already got the key work that we’ll be using in the next several major versions, a lot of it being worked on in our Research group. A first order of business for us is to go further into unification. You saw how Windows 98 unified help and the shell and the browser. Those three things come together. But there are still things that you have to work with in the system that have special commands and special ways of operating. The way you manage web pages, the way you secure them, the way you cache them, the way you navigate them; it’s its own name space and it’s very different from the file system. Those two need to be brought together.

Messages are a whole other world, where they have these properties, you have tabular views of them. So your e-mail client isn’t integrated. It’s just another thing that lets you do searches and queries and name things, and you have to learn a whole set of conventions there. So that’s very complex. The registry is something that you hope you don’t have to think about, but believe me it’s there. And there’s a very arcane set of messages you use to find things in that, navigate it and keep it up to date.

All these things can be brought together. And so the system shell lets you navigate pages, files, messages, system settings, all in a very simple way.

We’ve got prototypes of this user interface already running, but it’s a big enough change that it’ll await the major release that comes after Windows NT 5.

Then we have all the things that relate to natural interface. Linguistic technology is already showing up in things like the word processor, looking at your grammar, proposing synonyms, checking for any mistakes you have. We advance that quite a bit in the next version of Office. We have something where as you type, it’ll detect what language you’re typing in, and so it’ll know what kind of spell checking to apply there.

Making this linguistic capability rich is very important. If we look five years down the road, we won’t be typing in arcane URLs. We’re going to get such collusion between all the different people trying to use those same URLs that a much more practical way to use your system will be to walk up and say: what are the latest movies or where are the good restaurants or what’s going on with interest rates or the stock exchange. So you’ll pose your question, your desire to find something out on the Web purely in natural language. And based on what you’ve done in the past, based on the resources that are out there, it’ll find exactly what you’re interested in.

One thing I think people may not see coming as quickly as it will is the inclusion of a low-cost camera with the PC. Now, normally when people think about that, they think, wow, video conferencing. Well, in fact, it’ll actually be used as in input device, and that will be far more important than even the video conferencing, the idea of seeing if there’s a user there, seeing their gestures, letting them point to different things. We have the software for doing that, and I think that will be one of the first natural input things that’ll be standard in the PC sometime in the next, say, three-year time frame.

Speech has been the Holy Grail of all these things. Everybody wants to be able to just dictate to their machine, talk to it, call it upon the phone, tell it to do things. And that should be possible. Now, there’s been a lot of research over the last twenty years, very, very good progress. In fact, some people are using dictation products today, but they’re not integrated in, in a way that all the applications can declare the vocabulary and take advantage of that. It’s not quite -- we want to have it built into the system.

Handwriting will be part of that tablet device I talked about, being able to annotate things, edit things, very easily write there, very simple direct manipulation interface.

The final thing is the most difficult, and that is the kind of automatic learning that the AI community has talked about for many years. There are some ways of modeling this that can mean very good progress. It will require all the great work Intel’s going to do to create machines that have the power to do this in the background and become smarter about what your interested in. But I think it’s a very necessary step to fulfil the original vision that Microsoft had, that these tools would be something that everybody would want to use.

So we’ve got some exciting new releases, both in the near term and the long term. And it’s all driven by those principles I talked about. Now, those are very dear to us. They really determine the things we do. They’ve been the key to our success. And so we look forward to working with all of you as we drive Windows forward.

Thank you.

Bill Gates' Keynote - part #1

Bill Gates' Keynote - part #2

Bill Gates' Keynote - part #3

Bill Gates' Keynote - part #4