MR. GATES Thank you for having me here. I'm glad to be the last speaker because I like to get the final word in. I'm always a little reluctant to address people going to graduate school since I'm kind of a drop-out myself and don't know much about what goes on in graduate school.

I was very impressed looking at our list of Kellogg alumni how many of the key people who are driving us into new businesses are people who went to Kellogg. It's amazing. So I thought I'd spend a few minutes talking a little bit about Microsoft, and how the business has been run, and the future not just of Microsoft, but the future of technology. I'm such a big believer, you might even say optimist, about how technology is going to change the world, the world of business, the world of education, even the way that people socialize and entertain themselves.

I was a high school student when I first got exposed to computers and I found it very addictive because the idea that you could write a program and it would either work or not work was like a puzzle, a challenge. In fact, it was because I'd been very good at math that some of the other people who were playing with the computer drew me in. It's almost addictive to write good programs and see them work.

The structure of the industry 22 years ago when we started Microsoft was very, very different. It wasn't possible to write operating systems and sell them because companies did that themselves. IBM wrote their own, DEC wrote their own. Every computer company had an architecture that was unique and, in a sense, locked customers in. That was a pretty high switching cost to move around.

When the key component in building a computer became a chip, though, and none of these companies really were leaders in that area, it created an opportunity for restructuring. And so, we got in and said we would be narrowly focused on one thing, writing great software, and we'd hire the best people for that, we'd create product teams around that, we'd create a worldwide organization around that, and that would really be a first.

And so, it was a bet on the PC and it was a bet that was made, to some degree, understanding the dynamics. Also, it was what we liked doing. It was what we knew and understood and we were pretty careful to hire people who were very smart about writing software because we saw it as something that we were committed to very long term.

Well, since Microsoft started, the price of computers has fallen by a factor of a million and that's good because our software has gotten bigger. We need those more powerful computers to do what we do. The improvement in the chip, the improvement in our ambitions, really go hand in hand and it's an interesting partnership. We don't make chips. We work with Intel and the other companies in that area. We don't make [hardware] systems, we don't make communications pipelines. We are a software company and we've driven standards around that.

Those standards are scale-economic, which means that we can spend billions of dollars in R&D, which is about what we spend, and still price the software very aggressively, a thousandth of what other sophisticated operating systems would have cost, and yet putting more into it, advancing it more rapidly than any other operating system. And that's the opportunity we seized.

When we look at our future, it's really that same vision, just moving forward. The cost of computing in the next 20 years will fall by a factor of a million again. The cost of communication in the next 20 years -- a little more complex to measure -- will fall almost as much. And so, people theoretically will be able to reach out to information anywhere at any time, and the key thing that's necessary to allow that to happen is software.

Now, the nature of the software we do has changed a lot. Early in our history, it was writing Basic interpreters and compilers. Then it became graphical interfaces and office productivity software. Now it's much more about distributed computing and security. As we look forward, key technologies will be speech recognition, speech synthesis, computers that can see and recognize who's sitting down at the computer and let them make gestures about scrolling or throwing documents away or however they want to interact with the computer. It will understand whether they're confused or happy or whatever.

If you send somebody mail and you want a return receipt, we can enclose the emotional state of the person when they read your message. And so, we need to make the computer far more natural. We need to make it sort of an appearance of the environment. You'll have lots of big flat screens to read information from. You'll be able to point to those things, talk to those things [on the screen], and you won't really understand where the boundaries of the computer begin and end.

They'll be rich applications that are built on top of it and will be driving that platform forward. Underneath, the platform will be very complex. Orders of magnitude more complex than the largest engineering project you can imagine designing. Today an operating system is about as complex as designing the 747. A decade from now it will be ten times as complex, and that means that the magic way that we build and test the software becomes a competitive advantage.

The beauty of it is that nobody can look in and see that because it's all software. You use software to develop software and that means you can actually have an edge there and go down the learning curve project by project in terms of how you do it.

I believe that the whole world of business will be transformed by this, because the very mechanism of capitalism relies on matching buyers and sellers. Here, through the standards of the Internet and the PC, we have a system that for the first time can bring the friction cost of that matching down to almost zero and perfectly mediate markets on a global basis.

And it leads to some fascinating speculation about the future of distribution systems and the future of different countries, because jobs now can be done wherever the talent is to do those jobs, as opposed to wherever you get the right proximity. The world becomes a much smaller place. When I was in India a few weeks ago, I couldn't believe how enthusiastic they are about this because there is a well-educated set of Ph.D.'s over there who are now seeing that their wage rate will be much more equivalent to the wage rate of somebody with their capabilities in the developed world than it is today. That's economics at work. If there's a shortage of particularly programming talent, then they're going to step in to fill that void.

Microsoft is also pushing the boundaries a little bit in terms of how we define software. I said we've stayed away from hardware, we've stayed away from chips and communications, but the boundary between software and content, in my mind, is fuzzy boundary, and so we're pushing out a little bit there. We have things like an encyclopedia that is the world's best-selling. We've actually done it in seven different languages, so it's the best-selling in Japanese, French, German, Italian -- and it's partly content and it's partly software technology.

We have a travel reservation site where you can go in and find the lowest fare and make hotel reservations and things like that. And again, that's a mix of content and technology. And because the Internet is so new, it hasn't really played out what role a technology company like Microsoft plays and what role pure content companies play. And there will be very innovative partnerships that come in around this.

I would say that despite all the hype around the Internet, that although in a short-term sense people sometimes go overboard and over-estimate what it means, in a 10- to 20-year sense, as is usual, people under-estimate the impact of what we've got here. They have a hard time imagining the natural interfaces and the richness and the momentum fact that the more people get involved, the more it makes sense to put things there and the more you put things there, the more people are drawn in. And so, you get very non-linear facts when you have positive feedback like that all critical mass communication systems have, and this is the most wonderful example of that.

So I'd just conclude by saying that I think my last 20 years have been the most fun business and engineering experience anybody's ever had. I've had to learn a lot. I've had challenges. You can pick any year and I can tell you what the crisis that year was, you know, was it the OS/2 crisis or the Unix crisis. These crises are behind us. We've always got new ones that come in, but I can say with a great deal of confidence that the next 20 years are going to be even more interesting and there will be a lot more to learn that will be even better than what's come before. Thank you.

MODERATOR: Thank you. Now for the question and answer session, I'd like to call the first question from here and we'll just alternate back and forth. Thank you.

Q.My name is Joe Togue (sp) and …first of all, it's not a secret to anybody that Microsoft has invested quite a lot of time and effort and money into Internet content, and like you were saying before, Internet content requires a little bit of a different skill than programming skills and writing software applications and operating systems as such and that's what Microsoft is really good at. So the question is actually threefold. No. 1, does Microsoft have credibility as a content provider? No. 2, what is it about Microsoft's Internet content that can compel consumers to go to your Web site and your software instead of a competitor's? And third and most importantly of all, how would the business model for Internet content apply?

MR. GATES The world of content and the world of software are a spectrum. What we're doing on travel reservations is not a pure content thing. The most pure content thing we’re doing is probably Slate Magazine, which is run by Michael Kinsley, and there we've used the same philosophy we always used, which is when you want to get into something that's even slightly new, it's worthwhile to go out and hire the best person in that field.

So when we got into graphics, we hired Jim Kajiya, Alvie Ray Smith **, and the top people to come and form the core of our research group. When we got into database, we hired Jim Gray and about a dozen other top people. In our world, getting somebody who's the best and has the vision of how they're going to revolutionize the product area makes a lot of sense.

A>nd so, we wouldn't have gotten into Slate if Kinsley hadn't come to us and said he wanted to do something new and interesting there. It wouldn't have made sense to go in. With his help, I think we've done something pretty fascinating there.

In terms of credibility, you can't go back and always look to the corporate parent. You have to look at the group that's doing the work. I mean, does General Electric have credibility in the content world? I sort of think of light bulbs and refrigerators, but they own NBC, which has from time to time done very good content. So a company has to have multiple brands. …

And so, our content areas will have brands like Expedia and Slate and Mungo Park and things that, which will have to develop their own identity and do extremely well, and some of these new content things we'll succeed in and some of them we won't. Actually, it's Kellogg graduates who run a lot of these. John Neilson runs the Expedia and Sidewalk work. Laura Jennings runs MSN, and those are two of the three big groups doing the interactive content.

If you want to make money on content, you can get transaction fees, subscription fees, or advertising fees, and all of those are pretty minuscule on the Web today. The Wall Street Journal's got 100,000 paying subscribers and that's the biggest. Total advertising revenue is about $300 million. And transaction revenue is quite small. There are a few people making money. Expedia, our travel site, is break-even or better, partly because of our relationship with American Express. Amazon's book site is close to that, and there are just a few others in these early stages.

When you have gross over-entry into a business like the Internet with a gold rush atmosphere, the number of people in the business is unsustainable. Ninety percent will be gone. Now, which 90 percent? If you knew, you could make a lot of money in the stock market. The business model that makes sense is there will be a few kinds of content, news, sports, weather, and financial information. But the frequency of viewing and the volume of viewership will be enough that advertising can fund that.

Then there will be things that you don't visit daily, and very few of those will get by without some type of subscription or transaction-type fees. And so, the business models are not very mysterious. It's just that you need about a factor of 10 to 100 of more eyeballs using the Web and thinking that it's a primary source of information before any of these business models generate substantial profitability.

We're in four businesses: Windows, Office, BackOffice, and our interactive businesses. The first three are three of the finest businesses the world has ever seen. They're growing businesses, they're innovative businesses, they're profitable businesses, and as [investor] Warren Buffett would say, they have reasonably good growth in front of them. You do have to tend after them, but on a daily basis, we tend to grow the most rather than let somebody shrink it in those cases.

In the interactive world, it’s a lot less clear how many of these businesses tie together under a common ground, when do they reach profitability, and what is the scale asset that gets proven out. Now, I happen to be a big believer in this stuff. I think it's a wonderful area because it requires smart people, it requires having a long-term point of view.

Between now and when the pot of gold is out there, there are people who will be doing good work that will give up because they don't have enough patience. So you've got to be smart, long-term. You've got to be able to blend technology in with a lot of customer feedback. And so, if you take those four areas, we think we have people that are very good with that.

Now, we need to blend into that writers, authors, editors, people with those other skills in order to get it to be a full business mix. And so, it's unproven, but it's very, very exciting and we can afford to be an innovator in those categories.

Q. My name is Susan Greenspan. I'm a second-year student and I'll work for a start-up Internet content provider next year. My question has to do with transaction processing and micro- payments. … First of all, do you think that micro-payments will be a thing of the future, and given your investments in areas like Slate and other content providers, how do you feel that micro-payments will change the way the Internet works?

MR. GATES I'm not as big a believer in micro-payments as some people are. There are a lot of things on the Web that really are free and just totally, absolutely free. When you put up your personal page, you're not going to really get a micro-payment for somebody browsing into your page. People like the predictability of what they're going to pay for things; and so, if you have a community of users, having them all have free access to everything is a pretty powerful thing.

Nathan Myhrvold, who's our chief technical officer, and I were talking about this and actually, after we had the discussion, he went off and wrote an article in Slate about some of the fallacies about micro-payments. Technologically, it can be made to work, but let's say you have a magazine that's break-even at a certain subscription charge, what you're saying with the micro-payment world is that certain people are going to be paying a lot more than their subscription price because certain light users will be paying a lot less, and people will actually have to think about their level of use.

So I don't think micro-payments are the magical solution to the content paradox. I think people just have to get used to the fact that some content is going to be very high volume and a lot of other content -- it either better be very, very, very cheap to make, or you're going to have to get people who want to belong, who want to pay an annual subscription, a very predictable amount of money and they identify themselves as hey, I belong to that. I bought out that column. I bought out that music record.

Technologically, micro-payments can do very well and it's going to be played out. There are people at Microsoft who are more optimistic than I am.

Q. My name is Joseph Irvin (sp) and I'll be working for Intel Corporation. I've got a hardball question here. With Microsoft expansionist strategy, are you factoring antitrust possibilities into Microsoft's future, and how has scrutiny by the Justice Department affected your business strategy and revenue models?

MR. GATES Well, there's nothing that competition authorities like more than prices that come down. I mean, competition authorities are there to make sure end users get a good deal, and the PC model of computing, which we were key to founding, that's one of the key reasons why the price of computing has come down a million, and why it will come down another factor of a million.

You can go out and buy hardware today from any of 200 manufacturers, choose whatever you want, and you get that flexibility. So as long as we're doing innovative products, and as long as we're pricing them for the mass market, there is no issue at all. And people who work at Microsoft who create products, they say to me, "Do I have to learn anything about antitrust?" I say, "No, just keep lowering prices. Just keep listening to customers and keep using the stuff that's going on in the research group, and nobody has to learn anything about it because those principles allow you to capture as many customers as you want to capture."

And so, it doesn't really factor in at all to what we're doing. The only time it came in -- I guess you could say there were two times it came in. One is that the Legal Department had to send a lot of pieces of paper to various government agencies over about a five-year period because our area is so sexy, the government had to look into it. And I don't blame them.

I mean, I'd rather read about our industry than read about other types of business. So they had an issue there. The only other thing was when we were looking at buying Intuit. That's a case where you need to file with the government for mergers above a certain size, and the government can block the merger. And they chose, in the case of our merging with Intuit, to file a lawsuit to block that.

If they had gone through the appeals process, no matter how well we did, they could have delayed the thing 18 to 24 months. And so, we decided not to pursue that transaction. And so, in the very narrow case where we're thinking about doing an acquisition, then we need to send email to the lawyers and in a few minutes they send you back email and tell you whether a particular acquisition would or not be viewed as strongly pro-competitive.

Q.I'm Alex Estevez. Upon graduation, I will work for a relatively small software company, and the question involves WebTV. How do you see the WebTV strategy fitting in with the FCC's recent mandate for digital signal conversion? And how does that strategy fit in with the MSN strategy, MSNBC and Windows strategy?

MR. GATES WebTV is the company that we recently merged with. It's a great group down in California that builds a box that connects up to a TV set and delivers a lower cost and a pretty simple experience. Now, you don't get a full PC. You don't get to run games or office productivity software. You get basically Internet browsing with some limitations because of the TV screen, although they've been very creative about how they do the image mapping.

The belief is that electronic mail is going to become immensely popular and that going out to the Web to get critical information is going to be incredibly popular, and that some of the time you'll want to do that in a way where you're sitting in front of the TV and you want to say, "Okay, what other shows are on," or "What's some background information on the show," or "I just saw an ad and I'm kind of interested in that product. I'd like to know more about it."

And so, there's scenarios where you have TV viewing and Web-type information coming together to create a richer experience than either one of them by itself can. We believe in those scenarios. We believe in getting the price of getting into electronic mail and the Internet and the simplicity down to the lowest level possible, and getting those viewers to come in through a site that we create.

So when you dial in WebTV, we program up what's easy to get through there, and so we'll be able to pull in a lot of great content just like a network normally does. The word "network" applies very, very well. Of course, we're going to take our Microsoft Network and what we're doing with WebTV and bring those two things together.

So although it's a little less capable than the PC, most of the things we do on MSN we’ll be able to deliver to WebTV and you'll have a common mail account, common subscription. We're going to drive WebTV out so that whenever you travel around, there are public kiosks around that are WebTV kiosks so you can put a smart card in and get your electronic mail and see what's going on.

And the TV industry has been working so hard to get to high definition and not really thinking about what the bootstrap is that they really didn't take into account what was going on in the computer industry, and so we kind of showed up at the last minute and said, "Hey, here's the better approach called HD-0." This will all get worked out.

You probably will never see headlines about this again because by the time they broadcast it, they'll realize that all the new screen technologies are progressive in nature, the number of receiving devices that are going to be PCs in the next three or four years versus so-called high-def TVs will be about 20 to 1. It's a typical standard struggle, but nothing we're doing really relies on that. It's just sort of common sense about how you create a low-cost receiver and a high-quality video image.

Q. My name is Minovin Klass (sp) and I'd like to ask you a question of two parts, both equally important. The first part is what's your greatest fear, and the second part is what's your favorite beer.

MR. GATES Well, my greatest fear is that as a company, we lose our edge, lose our sense of using that customer feedback, lose in research, that somehow we've slowed down and become mediocre. Once that happens, somebody, I can't say who, will pass us by. Somebody will get through these revolutionary things better than us. So I come into work every day saying, "Let's make sure we have as much hunger, as much concern about doing these things better than we had before."

That is the thing that every company has to worry about. Only by worrying about that will you ever deal with that issue. Many large companies just haven't kept that sense of urgency in their culture, and it's devastating if you don't have a sense of urgency that your real assets are research and customer feedback.

In terms of beer, I'm not really very picky. Any light beer will do.

Q. You mentioned before that for many of these Internet content providers to be profitable, the Internet is going to need to grow 10 to 100 times, and yet it appears that bandwidth to the home is a barrier to that growth. I know I'm getting frustrated with the speed of the Internet. How do you think that problem will be solved, and when do you think it will be solved?

MR. GATES Using the phone infrastructure, it's tough to get up above about 60K baud. What you have to do is put a new type of signaling into the switching office and down in the PC. And if you look in the next two or three years, 90-plus percent of homes will be dialing in at 28.8 or 56.6. If you go 10 years out, and look at the U.S., the vast majority of homes by then will have so-called ADSL type hook-ups, or PC cable modem hookups, and those have much higher speeds. Typically, it's complicated because PC cable modems share a box, whereas ADSL is point to point. But you'll typically have 6-megabit-type feeds.

So it's a matter of patience, and it raises real strategy questions for companies because video content is basically worthless in today's Internet unless you're dealing solely with business customers. Whereas, 10 years from now video content will be important. People won't be interested in a site that doesn't have great video assets. So when we picked a partner in news, we said, okay, there are text partners and there are partners that understand video. And although in the short-run it made no difference, that's why NBC was a particularly strong partner going forward. But it will be quite some time.

But people are underestimating ADSL and PC cable modems, partly because there were over-promises about interactive TV, and about how quickly the cable infrastructure would get upgraded. But now people have actually gone too far, and they're underestimating the work that's going on. There really is good work in both, people like U.S. Robotics, and in PC cable modems.

Q.My name is Tom Updike (sp). This question is not meant to add to your fears, but if you were thinking about starting a business, or if you were to start a new business today, and you were facing the prospects of going up against Microsoft, what would you do to ensure your new company's success?

MR. GATES I don't think there's anything you do to assure your success. You should make sure you hire people who enjoy writing software, so no matter what happens, they will have had a good time. Seriously, you know, if you take Microsoft's three primary businesses, Windows, Office and BackOffice, I'm not sure I'd attack any of those three. Now, that leaves a world of opportunity. In the interactive area, it's guaranteed that there are dozens of companies out of thousands that are being started up right now that are going to find a magic formula and do something fantastic and be very successful on the Internet. It doesn't mean they'll have a billion-dollar or a 10 billion or 100 billion valuation, but something that will be considered a fantastic success, because there are so many approaches that have to be tried. The Internet is a great place where capitalism works because you get so many people trying different things, and investors evaluating that. And so, in new areas, once you move away from those three, you can feel like there is a non-zero chance.

The best thing to do is to have a vision that the existing companies don't have. Why did Microsoft succeed in software? Partly execution and partly vision. We believe in software as the key asset. We believed that that would span multiple hardware manufacturers, and really there was no one else doing that. And so, if somebody has a breakthrough in artificial intelligence, or a breakthrough in visual recognition, in the deep technologies that there's clear infinite demand for, then they could do extremely well by pursuing those areas. But you've got to have something unique.

Q.My name is Arvin James (sp). You had mentioned that there hasn't been a year in the life of Microsoft where you haven't faced a crisis. Larry Ellison has been the most outspoken advocate for the Network Computer. One might say that his motivation in this has been to go head-to-head with the current Wintel [Windows and Intel] architecture of computing. What do you see as the role of the Network Computer both in the home and the business markets, and how do you envision positioning Microsoft against Oracle? Do you envision them as a competitor, a collaborator, or just operating in totally different spaces?

MR. GATES Oracle builds on top of Windows NT and, in that respect, they are a collaborator. But I'd say that's the footnote to our relationship. Now, we compete with Oracle, and Oracle totally owns the database business today. They've got it locked up. And we have a different vision of where database is going. We don't think -- take a safe date, five years from now -- that you'll go into a small business and say, here's a database file. We think you'll go in and say, here's an operating system that manages data, mail, Web, and files. You don't even have to learn those terms or think about those things. It will do your customer lists. It will do your messages. It will do your transactions. It's an operating system that is very, very rich.

We have an integrated strategy that we started years ago hiring the people we believed were the world's best database experts. So we're spending about as much money developing databases as Oracle does. Now, we get zero -- well, not zero, but almost rounding error revenue -- from the database market today, and that's where Oracle gets 50 percent of their revenue. Actually, it's a little less than that now. And so, we're going to compete in databases.

There's another area where Microsoft and Oracle have taken a different approach. Oracle has decided to be a big consulting company and try to build a lot of account control by being like IBM, where they have massive people in large accounts. They're very large-account oriented. They're very enterprise-oriented. The first place they'll really lose lots of database share is in small and medium-sized organizations because they're not well adapted for that. But in the large organizations, they've got lots of consultants. So almost half the revenue comes from consulting.

I've chosen not to make consulting a big T&L for Microsoft so that we can partner up with small consulting companies, which are great, along with the big consulting companies, people like EDS, Andersen, all the accounting firms, and it's a different strategy. And I think our partner approach there is much better.

Oracle has also decided to get into applications. And so, SAP views Oracle as their primary competitor. So SAP views me as their best friend, and they're a great company. We've gone out there and done things, and they have a very strong position in that business. So we're not doing the vertical applications that people like SAP, BAAN, PeopleSoft too, we're getting them to build on top of our database technology, which is being integrated into the operating systems.

So that's everything but the Network Computer. What is the Network Computer? Well, some days it has a disk, some days it doesn't. The one from Sun is completely different in terms of resources and user interface, than the one from Oracle, which is completely different than the one from IBM. And so, NC means Not Compatible. And there's nothing wrong with starting with a non-compatible approach. But it's only defining characteristic is that it does not run PC software. Now, how would you go about doing something like that? Well, one approach is, you could say, let's run the software up on the server. After all, Sun and Oracle have made money with fairly low volume and high prices. Sun in the case of selling server hardware, Oracle in the case of selling overpriced databases. And so you want people to do things against the server. That old mainframe time-share model is a very good model.

So you try and move the computing back to the center. It's very tough to do because people are used to portable computers. They're used to picking lots of applications. They're used to the responsiveness and empowerment that comes with a PC. But attacking the PC is not an easy thing.It used to be there were 40 companies that hated the PC, and then there were 30, 20, 10, now there's two. And these guys have always hated the PC, and it's Sun and Oracle.

And so what they want to say is, move things to the center, and rewrite your applications, all your applications, from scratch. Now, why is that pitch even passable? It's because there's a lot of complexity that's grown up around the PC in terms of updating software and managing the software. And in the same way that the Internet was the big theme for Microsoft in '96 -- we did a lot of things there -- in '97 it's cost of ownership. Building into the operating system automatic installation, automatic updating. Replicating states on the PC up to the center automatically, so it's backed up all the time, and you can move it to another PC. A lot of technology will deal with that in a dramatic way.

People love the thing that's sort of good about the PC, and the things that have gotten complex are all very, very fixable. But they're basically betting that people want to start over, have a new user interface, and go back to what was the Unix model, where the hardware manufacturers get to decide what the system software looks like. And so they all deviate from a common standard. And the Achilles' heel of Unix was that Solaris was not AIX, was not HP-UX, and so the cost of developing the operating system made it very, very expensive. The low volume of machines made it very expensive to develop for. And the NC is basically a two-letter Unix business model.

Now, we take it very seriously, don't let me say that it's completely silly. But what we take seriously are the things we need to improve, not that the alternative has any true merit.

MODERATOR: We have time for another.

Q. I was going to change speed a little bit. My name is Darrin Broadshaw (sp). I'm wondering, over the past quarter you've spent several billion dollars buying back shares, and your cash account seems to be growing year over year to quite a substantial position. And given that short -- you know, you walked away from Intel in a deal because of antitrust concerns, you needed a lot of money to fund R&D, I understand, but where -- what's going through people's minds with how you'll manage this cash going forward? Will you continue with repurchases or perhaps giving it back to the shareholder at some point?

MR. GATES Repurchase is a form of giving back to the shareholders. You can use dividends or buybacks. Because we have a stock option program that's creating natural dilution, the buyback is very effective for us. But your question is a very good question. Every great business generates phenomenal amounts of cash, and there's a great danger when management has all that cash. Part of the brilliance of what Warren Buffett has done in creating Berkshire is that he has smart businesses, a lot of which he owns, and he's basically the CFO. That is, as those businesses allocate cash, he's smart enough to say to them, you realize if you try and use that cash, you're going to go outside your circle of competence. And so, instead of you diluting things or getting confused, I'll take that cash back up to the Berkshire level and allocate it to the other businesses, or buy some other fantastic businesses. So he's like the treasury for very well-defined cash-generating businesses, allowing them to keep their eye on the ball. It's a brilliant system that works very, very well.

In our case, we generate cash, and even though we have this nice $2 billion R&D budget, and we've got $9 billion cash right now, and we'll generate another $6 billion in the next year -- we do salt it away for a rainy day. But it's not a limiting factor on the business. So the buy-back is the thing that makes the most sense. I always thought there would be some strategic investment where we'd go out and spend billions, but generally when you buy companies, they want to be bought for stock anyway, because it carries their gain over. So there have been some cash acquisitions and some may come along. But I think the buy-back will be a key part of that. The key limiting factor for Microsoft, and maybe this is a good point to make, whether it's in marketing or business management or engineering -- the key limiting factor for Microsoft is hiring great people. It's not cash. You know, we have neat jobs, neat things, and keeping it so that it's all very orderly and well done, that will determine how we go after our new ambitions.