(Applause.)

BILL GATES: Thank you. (Applause.) Thank you. Well, thanks for coming to Microsoft. One thing we won't be able to say enough is how much we appreciate all the help you give us, the help you give our users, the feedback you give us, both positive and negative, to help us make better products.

There's nothing more exciting than what software can do to help users and we see that what we've done so far is really just scratching the surface.

So what I wanted to share with you today are some thoughts about what's going to happen over the next 10 years. And you'll find that I'm incredibly optimistic. I think we've laid the foundation with the Internet, with XML, with the new tools that are coming out to actually make more rapid progress in this next decade than ever before.

And that optimism is somewhat in contrast to a general mood in the industry, where people are saying, "Gosh, you know, the dot-com bubble wasn't all real," and we've got a tough economic situation, and even starting to wonder whether big advances in technology are going to drive the kind of productivity advances, and therefore the strong economy that the U.S. in particular saw during the 1990s.

My view is that they're really underestimating the fact that although these things don't happen overnight, the foundation is being laid to do even better than we did during that decade.

And so I talk about this next decade as the Digital Decade. Now, I'm very lucky in terms of making these predictions in that I've got the $5 billion dollars of Microsoft R&D -- (laughter) -- to not only sit and speculate, but to tell those guys, "Okay, you'd better make these things come true." (Laughter, applause.)

So what do I mean by Digital Decade? Some people might say, "Well, aren't we pretty digital already." Well, in a certain limited sense yes we are. E-mail is increasingly popular. Homes of PCs, U.S. is about 56 percent. Internet usage continues to climb. Definitely when people create documents, that's overwhelmingly done in a digital fashion. Word processing, spreadsheet, we've driven that almost to be a common sense thing. When people talk about giving a presentation, they talk about PowerPoint as though it defines the word "presentation."

And so we're very proud of what's come out of that, but in a certain sense it's really not there. For example, take this idea of B2B. People love big numbers where they say, "Okay, this much business is done digitally." Well, what they mean by that is that the purchase order instead of being sent on paper is sent across the network, sent typically through e-mail.

But that whole transaction, the idea of how you find the seller, how you deal with a delay, the product comes, it's not quite right, that is not digital. That's phone calls, faxes, lots of confusion and tons of workers who have to deal with the missed cues that go on in a very analog-driven process.

Most of the U.S. economy are what we call knowledge workers, people who think up products, provide customer service, who are involved in marketing, sales, things where their understanding of information and their effectiveness in seeing trends determines exactly what their productivity is. And even for those workers I would say that we're not even half way to how we can help them get their job done.

If you took away all the computers, people would miss them but the way that buying and selling gets done would be just fine. The way that people use the phone, the way that they organize things isn't fundamentally digital.

And so where can that change? That's a key thing. One idea that we talk about a lot is will there be the energy and the acceptance of these new digital approaches. And so we actually went down to some experts and got them to talk about how they saw the digital decade, what the key things were, and so let's take a look at that video.

(Videotape presentation.)

(Applause.)

BILL GATES: So the digital decade -- (laughter) -- will require a lot of work, but we think we've got the energy to do it. (Laughter.)

There are two ways of looking gat this. One is a technological approach where you build up from the hardware and the software and say what's possible. The other is to take the various user scenarios and say what would be ideal for them, take the knowledge workers, take the enterprises, take the homes and think what kind of convenience will be attractive there. And we try to work the problem from both ends for all three of these groups. We've also seen our software as meeting the needs of all of these three areas.

From the hardware point of view it's amazing what we have to work with. So-called Moore's Law that talks about a doubling in processor power every 18 months or so is destined to keep remaining true for the whole entire next decade.

And so what that means is that the memory size, the computation power will just be phenomenal. So whether it's speech recognition, business intelligence, voice recognition, you name it, we will have the power to perform those tasks.

Actually, the disk capacity is going up even faster than the processor speed is improving. The optical fiber speeds are going up even faster than that. So an incredibly array of hardware to help us with software applications.

High-resolution screens are an important part of this, getting the fringe right, getting so it's mobile, you can carry it around.

Getting the wireless networks in place: 802.11 we think is a fundamental technology that every business, every home, every convention center is going to be wired up with high capacity 802.11. And that's finally the way that we'll have information wherever we want it. In business take your portable to the meeting. At home have your video, your audio, your photos show up on even very inexpensive little screens where there's Windows Terminal Server projection from the PC to let you get at that information.

So with all the advances, I'd say there's only one thing that's a little disappointing, that's kind of holding things back, and that is that broadband connections going into homes are still quite expensive. The broadband into businesses is quite different. It's expensive, but the value of the information connection is enough that we can think of most businesses as having broadband. Whereas homes even in the next five years there's a question of whether we'll get up to 50 percent penetration.

The key to this is simply that the price has to come down and there are sort of regulatory and competitive dynamics that haven't driven that in the United States.

And the only country that's really done well on this is Korea that's up to 40 percent of the homes with broadband, that is either DSL or a cable modem connection. The pricing there is about $24 a month. So even though we're a lot richer, they've got more broadband.

And so we'd like to see those same things happen, but it's the one area where it's a little hard to predict. Part of that we can mitigate by having a strategy that uses local storage. We always need to do that for offline capabilities, for responsiveness, to deal with network latency. Even as bandwidth goes up, latency does not go down. And so you want to put data and computation as close to the user as possible.

And what that means is intelligent devices like the PC. It even means that the pocket-type devices will have to have a lot of intelligence in order to be able to work offline and have the responsiveness for the richer user interface.

Latency never gets improved, because latency fundamentally is limited by the speed of light. If you're going to send information back and forth, there's no chip miracle that allows that to happen instantaneously.

So there are lots of possibilities coming out of the hardware area.

So what are we going to do with this? Well, first of all our top priority is what we broadly call trustworthy systems, making sure that these systems you can install a new app and it doesn't break things, that if there's a problem that we see that problem, that you can turn to a friend and get sort of screen-driven sharing type support, that the servers or the server goes down, another server takes over that work without the users who are connected up having to know about that. Data is backed up so that if you have any problems you can immediately restore the code and the data and be off and running. And fundamentally these systems have to be viewed as self-repairing.

Now, there's a lot of technology that needs to go into this, and we're starting at a point where today people are very concerned about viruses for good reasons, they're very concerned about the complexity of configuration, the complexity of patching. In a recent incident, the NIMDA, which was really a huge problem for our customers, every single one of the problems exploited by NIMDA there had been a security patch for it for over four months. And so although that doesn't shift the burden off of Microsoft, what it says is the infrastructure for getting those fixes out there is just too complicated. You know, people didn't know if they were up to date. The trouble to do it was simply too great. And so we need to change that quite significantly.

The way I see this working is what we call continuous improvement, that we see through an alerting technology we have called Watson that catches today crashes or enhancing it to take care of hangs. Over time we'll enhance it so even when the user is frustrated they'll be able to explain why the system isn't doing what they expect, that it's too slow, too confusing, whatever, and that will be part of this feedback loop.

We talk about the Watson that we have today as being a sort of red button system, and then we need to enhance that with sort of yellow button feedback, where, "Okay, I'm still running, but you should know I'm not pleased with something that took place."

So that information comes to us. Today we're collecting every day about 7 gigabytes of information. It's been stunning to see where the ecosystem is falling down, where the reliability is not there. This ecosystem is amazing. You know, the fact that you get people who do video drivers and disks and add-ons and tens of thousands of software companies, and part of the beauty of it is there is no permission required. You can go to a bookstore, buy books about Windows, write drivers, write applications and that's a good thing. We want to preserve that openness and flexibility. Nobody has ever had anything like that and it leads to rapid innovation.

But the question is well how do you deal with the downside of that. What if the pieces don't fit well together?

Well, the answer is that this system ought to catch that, and we ought to be able to give people, you know, say to people online, "these are the drivers that aren't working well."

Actually, the first line of defense is to make sure that information gets to those people who are providing the pieces that are having those problems.

It turns out video drivers are 7 of the top 10 of these Watson situations that we have today. So there's a clear action item there. We will work with the right partners and get video drivers out of the top 10.

(Applause.)

Then we'll have another top 10. (Laughter.) But, you know, the scale of these things can be dramatically improved.

In a sense, we had an open loop system. We knew anecdotally that sometimes the pieces weren't fitting together, but we didn't know statistically how to turn that into an action item where we could say to somebody, "Look, you know, we'll go to PC Magazine and they'll rank things according to this data if there isn't quick response."

So the security elements that are involved in this are very important. There's a lot that has to do with moving people up to the NT code base, a lot that has to do with the constant update capabilities. The .NET servers are the first release we'll have where we really take this to a new level. Windows Update is built into XP. .NET Server comes out the middle of next year, and so a lot of advance there.

So with the idea of trustworthy systems we want people to say, "Oh, I'll try out this new app because I know it won't make my machine not work or my speaker do funny things." And so the whole idea of trusting the computer, you know, will people want to put their family photos on the machine; well, if they don't trust the machine they're not going to say, "Oh, well too bad, we'll just have more children and go on those trips again." (Laughter.) They really want those photos to be properly preserved.

So if we don't treat the trustworthy thing as our top priority, a lot of these other dreams we could actually get the pieces in place but people wouldn't be willing to take full advantage of what we come up with there.

People in the office, the productivity scenario: This is probably the furthest along of all the things we've done. And yet if you look at what's missing today, the way that you would do forecasting and sales analysis, the way that if you want to send out a piece of e-mail but limit it to four or five people and say, "Okay, you can't forward this along," or, you know, "This is a message I'd like you to look at," but 30 days from now it sort of evaporates, there's no sort of rights management in this ecosystem.

Today the views of your files is still that simple folder view, and the way you e-mail and their attachments is different than the file system. So we have a lot that the way we're doing replication, the richness of the store simply isn't as effective as it needs to be.

One scenario that we want to go after is making meetings more effective. People might find that surprising, but, of course, ideas like that PowerPoint would be the standard for presentation, in its day that was kind of a wild idea as well.

With meetings, how do you plan them? Is it easy to just essentially have a Web Site where the documents, the attendees, the pre-reading is all there, and that Web Site after the meeting turns into a place people can go and find out what happened and literally see the video and audio of that meeting?

The idea of having digital cameras tracking, videoconferencing, video in meetings, that hasn't been possible partly because if you just have a camera sitting in a fixed position, it can't capture the right person at the right time. Well, you don't want to have to pay to have some film expert in there filming the meeting. That sort of makes the dynamics of the meeting not work well. But it turns out that with some breakthrough software and what we call a ring cam, a 360-degree camera, we can take the video and actually do the direction in software, figure out if the crowd is laughing, okay, pan to the crowd. If somebody is speaking, go and show that person. And we even keep enough information that as you're watching you can literally control your viewpoint but the default viewpoint is right over 95 percent of the time.

It's partly why videoconferencing doesn't work because when that camera is in one position, if the person moves, they're kind of off position, at funny angles, and it just hasn't been effective.

One of the things we've come up with is if you have these meetings digitally and you didn't attend but you wanted to see part of it, we can play it back to you at double the speed of the actual meeting by taking out pauses and that kind of thing, so that you'll be dis-incentived to go to the real meeting. (Laughter, applause.) Let somebody else go. I'll see the subset I care about at twice the speed.

Likewise, remote presence: You know, today that's done largely through audio, but we can do better than that.

Communications: You know, it's great that people have cell phones and faxes and e-mail, but it's really a mess because the really scarce resource, which is the user's time, can be used up by junk mail or your phone ringing when you don't want it. And even if you're just writing a memo, you often get new mail, and it's so tempting to say, "Well, it could be an important piece of mail." Well, 90 percent of the time it's not but you interrupt your thought process and go off and read that.

So the user has no control. You have many phone numbers, many e-mail addresses. Instant messaging is just an addition to this. We don't really help you state what are your criteria of when you'd like to know things, not just communications but a stock price change, a scheduling change. The computer should be an information agent and work on your behalf and screen things according to criteria you care about. So if you're in a meeting, the threshold is very high. When you go back to your office it's somewhat lower. When you're at home messing around it's even lower and the different messages come to you when that is appropriate.

Scheduling things so people can organize things digitally through schedules: We've really not given people enough power to fundamentally use the schedule for things like groups and families and make that the primary way those things are done.

So there's a ton we can do to just make communication better, to have real time communication. For example, if you have two people calling each other in their offices, why doesn't the screen of their computers allow them to exchange lists and documents and work together? When you make a phone call, that should happen automatically. You know, that other person, you have their name in your buddy list, your name in your contacts and you can see the phone number you called, see their IP address and make that connection.

A Holy Grail for many people in the computing world and Microsoft in particular has been the idea of reading off the screen. You know, we've sort of made an advance on this with Encarta. We took the encyclopedia from being a print-based document to something that overwhelmingly is now electronic, so it's more up to date, richer and interactive. But for most long documents people don't like to read off the screen.

There are two big reasons for that. One is just the resolution and fonts, the appearance. Well, that will get solved with good software, including ClearType, and much better screens.

The other is -- and this is a tougher problem -- is if the screen is in a fixed position you're just not going to read it for long periods of time. It's just too fatiguing for your neck muscles and everything to go look in that one direction.

And so what you need is something you take with you. Well, a portable PC, you could use it that way, but it's still kind of strange if you're reading to adjust your viewpoint. Subconsciously when you read a book or a long memo you're doing that but you're not having to think about it. A portable still doesn't work very well for that.

And so we've been working on this idea of this tablet device. You know, why is that interesting? Well, you go to a meeting, take it with you. Take your notes. You're in the meeting talking about a spreadsheet and something looks unusual. Just point at it and expand it, expand it by time, geography, which person, which product. Dig into that information. It's live information. You want to tell somebody that they ought to do something related to that, scribble a little note on it and point at their name and your name list and off it goes. If you have follow-up items you want to write down, those would be recorded there and you can share them with your workgroup. So we really need this tablet-like device.

Now, there was an attempt to do this six or seven years ago where the hardware and software just weren't ready. Today, we think it is ready. We've been working with hardware partners. It's a year from now that we expect these to ship.

I actually brought one of the two form factors with me. One form factor is just the thinnest possible device and it literally will be tablet like. That one you'll have a keyboard, but it's detachable, either wireless or USB connected. And so when you take it you can sit in bed and hit the Next page or scroll. You know, there are a few buttons on the thin little tablet. It is a full PC though. It's running all the software.

Another form factor is what we call the convertible, where you have a normal portable like this one and you use it with a keyboard like you normally would, say you're in your office. Somebody says, "Come into this meeting, let's talk these things over, let's look at these numbers, we've got a problem," you take this thing and you rotate the screen around and then you have this device here.

And so this is a surface and we've insisted that the digitizer here be extremely high quality. It's 133 samples per second. So you take your pen and you're just taking notes and doing things. The handwriting recognition software has advanced incredibly. Some things you'll leave as ink, some you recognize, and it's all there. And, of course, when you read you just have a few buttons here so you can sit and page down, you know, take this thing, carry it wherever you want to go.

And so that is in a sense you can call it an evolution of the portable PC, but you can also call it revolutionary because once you've got all that information in digital form the opportunity of software to add value and do a better job is quite incredible.

Another scenario is e-business. As I said earlier, this notion that e-business is done or even that we're 10 percent of the way there is not right. And yet all the benefits of e-business -- efficiency, matching different buyers and sellers -- that is phenomenal and that can be seized.

It's very complicated to do, because we're talking about computer systems across the Internet where the developers at the two ends have never met, exchanging rich information and not having security problems and dealing with the whole richness of the business relationship. It's very tough to do. XML is key to that. This new generation of tools, VisualStudio.net is fundamental to that. Getting businesses to think about how they get information between their knowledge workers and their back-end systems to use XML is very fundamental to that.

For us it's a big change, a big change. SQL needs to change to embrace XML. Office needs to change to embrace XML at its heart, and our development tools are already far along in doing that.

So the enterprise and the knowledge workers are two domains that we've done very well in. The home I'd say we're even less far along. Yes, there are PCs in the home doing knowledge worker type things, but if you go to the living room and that's less the case.

What we want between the Xbox and the PC and an 802.11 wireless network to really fulfill this dream of your music is digital, your family photos are digital, you're notified of the things you care about. Entertainment is brought to a whole new level. You watch TV when you want to watch a show, that it's available to you.

And so making that easy and cheap enough, there's a lot of work there but the pieces are really falling into place.

One scenario is with the 802.11 network if you get a little screen that's got an 802.11 receiver, just a dumb screen, it doesn't have a disk or software, much software, it's just dedicated, take that little screen and put it on your refrigerator. Well, your PC will notice that that device is there and say to you, "Do you want to have your music selection, photos, weather, news, whatever you want, play Reversi, on that refrigerator thing. You know, I've got Windows Terminal Server here, so I project out onto it and just tell me what you want the default homepage to look like." So you edit that up, you get that.

That's not limiting it in any way, because you can have buttons on that homepage that take you to anything that matters. That homepage can know your schedule and alert you five minutes before you're supposed to leave automatically, or show you various traffic for things that you're going to go to, so it's intelligent because it's got the full power of the PC.

And you don't have to think about will I put my music on this thing and my photos on this thing and move them all around because it's projection out from a single unified store.

So music, photos, video, this stuff is going to go digital. Sharing it, organizing it, annotating it, the cameras are great now. The printers are great now. One amazing thing about this tablet PC is we built 400 of these and sent these out and looked at how people used them. They actually use photos twice as much as you do on a normal PC. Even though we did nothing with photo related software, but the idea that you could just take this thing and share it and look at it together and click through the photos, you know, it's so much more natural to have this in your hand than say, "Well, come up to my PC" and there's one chair there and let's try to gather around something that's not designed for that. This is a much more social thing for selecting music or photos. And so all these things really come together in a synergistic way.

Games: You know, we're at an exciting juncture here for this. Xbox just released about a week ago and it's off to a great start. When we introduced Xbox the games were clear: We couldn't say it's a multi-function device. We had to say this is the best game machine ever. You know, the notion that, "Oh, it plays games and it does something else," there was no open-mindedness to that. (Laughter.)

Now, this has a 733 megahertz Pentium processor, 64 meg of memory, has the best NVIDIA graphics chip that they've ever made, 8 gig hard disk, Ethernet connection. It's kind of a very general-purpose machine. (Laughter.) And so we might be able to do something else with it. (Laughter, applause.) But the first and foremost thing is to get a huge install base based on playing the best games, and then we can take that broadband capability and the fact it just runs essentially PC software and show you how either this device on its own or working together with a PC in the house will let you do neat things.

So we're really sitting at a pretty interesting juncture. You know, handwriting recognition, speech recognition, business-to-business, easy interoperability, decent development tools: All of these things we've dreamed about for the last 20 years, but it's only now that the right level of power in the systems and the right kind of R&D to take on these tough problems is really in place.

So these next 10 years will be even more fun than the last 20. Thank you.

(Applause.)