(FORTUNE Magazine) – You can often learn as much about how business partners work together by simply watching and listening to them as by asking direct questions. Case in point: the following conversation with Bill Gates and Steve Ballmer, perhaps the preeminent buddy act in American business.

Although they've been friends since college, the two have different styles: The characteristics these 43-year-olds do share are their commitment to Microsoft, their competitiveness, and their astounding wealth. (Gates' fortune has been bouncing around the $100 billion mark in recent months, while Ballmer, who owns 4.8% of Microsoft, is worth more than $11 billion.)

Chairman and CEO Gates, the scion of an elite Seattle family, grew up in a world of country clubs, private schools, and speedboats and summer cottages. Ballmer, Microsoft's president since late last year, is the son of an immigrant father who worked his way up into middle management at Ford Motor. Gates famously dropped out of Harvard; Ballmer stayed on and got his degree. Gates was always a little on the quiet and shy side, while Ballmer, a social animal, can be downright bombastic.

Listen to them talk and you begin to see how complementary their skills and personalities are, especially when it comes to running Microsoft. Bill is the infotech industry's consummate business strategist and a pretty fair visionary to boot, while Steve's the real detail guy who knows how to get things done.

The give-and-take that follows is the hallmark of two guys who have been competing alongside and against each other for as long as they've been friends. We learn how much Gates has relied on Ballmer to get him out of jams over the years, and we may end up wondering why it took the Microsoft founder nearly 20 years to make his pal president. Reading between the lines, we can also see how the impetus for much of the change at Microsoft doesn't come just from Bill, but from Ballmer too. And now it's not just Bill who's listening to Steve: On Sept. 23, when Ballmer publicly ripped overvalued tech stocks, the Nasdaq index tumbled 108 points, or 3.8%.

FORTUNE: You're old friends; how has that friendship evolved over the years?

Ballmer: We've known each other 25 years.

Gates: Hasn't it been longer than that?

Ballmer: No, it was the year Richard Nixon resigned, and this is the 25th anniversary of Nixon's resignation, which happened right at the start of sophomore year at Harvard [fall 1974].

Gates: We both were living in Currier House. A lot of people living there were into math and science. I was actually a fairly normal resident in the sense of being hard-core in my subject. Steve was unusual in that he was part of the mainstream campus social activities. We were introduced by this guy, a really nice guy...

Ballmer: Jeff Clark.

Gates: Exactly. [Laughs.] Steve remembers his name, just like he remembers everybody's name.... Anyway, Jeff thought we were both pretty interesting and said we should get to know each other.

Ballmer: This must have been in the fall, because I remember that year shutting down your room for Christmas break. You'd gone to Seattle but left your door wide open, and somebody had to lock it.

Gates: We got to be good friends, even though we had pretty different views of the world. I was fascinated by what Steve was doing because it was a lot more like how my parents went to college--you know, participating in clubs and activities and knowing people. My whole thing was to sign up for classes and then not show up, because I didn't want to look like I was trying too hard.

Steve and I would talk about what to do with our lives. One thing Steve mentioned was that he might like to go into government. We thought the leverage of a smart guy in government would be superhigh.

Ballmer: Bill, on the other hand, knew he was going back to Seattle, which I thought was just bizarre, for three reasons: (A) I'd never been there; (B) it was a long way away; and (C) when you go to Harvard, there's a culture that says, "The East Coast and the Power Corridor blah, blah, blah." But Bill always said he loved his family and the Northwest.

Gates: One of the galvanizing things for our friendship was when we took a graduate-level economics course together. The teacher was Michael Spence, who more recently headed the Stanford Graduate School of Business. Steve had signed up to do so many social things that he couldn't go to class. And neither of us had taken the prerequisites.

There were two ways to get your grade: You could take the two midterm exams and hand in homework, and then the final counts only 30% or 40%; or you could bet it all on the final. Well, like I said, we hadn't even shown up for those other things, so it all rode on the final.

Ballmer: We camped out in this conference room at the dorm for four days reviewing the midterm exam questions.

Gates: And we couldn't answer them at all. Then we would look at the answer and say, "Yeah, I guess that's obvious."

Ballmer: Unlike you, I had actually tried to do the reading.

Gates: So anyway, we went to this final. It was an open-book test, so everyone was lugging in all these books and looking up and marking stuff. I went in with two pencils and that's it. When they handed out the test and it was just one page, well, we knew we were golden--because one page meant the test was about theory.

Ballmer: It was basically a math test.

Gates: There was this question about a refrigerator that had a certain probability of failing, and you had to figure out the value of the refrigerator. That's an integration problem. So there were six questions, including one for extra credit, in this three-hour exam, and halfway through, I'm done. I handed in my paper, and as I left I looked at Steve, implying, Hey, how come you're not done? We got two of the best scores in the class.

Ballmer: You got a 98, and I got a 99.

So you guys talked about working in government, but did you ever talk about future careers in business?

Ballmer: Only after Bill started Microsoft. One night we went to a Howard Johnson's out near where Paul Allen [Microsoft's co-founder with Gates] lived. Those were the early days of Microsoft, and Bill and Paul were going back and forth between Cambridge and Albuquerque. I was on a jihad with you that night because you didn't have a real gal Friday around the Albuquerque office, so you guys had to buy filing cabinets and supplies yourselves. That struck me as being kind of random. You hired an office manager soon after that, if I recall.

Gates: Even though Paul and I came up with the idea for Microsoft in the spring of sophomore year, I was still a student. The summer after that, I wrote [the computer language] BASIC for the MITS Altair, and then I missed the first half of junior year while we set up Microsoft in Albuquerque. I came back to Harvard for the spring of junior year while Paul stayed in New Mexico.

Ballmer: After that you were gone.

Gates: And you went on and graduated and then went to Procter & Gamble. I remember coming back for a dinner during senior year and telling you how somebody from Long Island had offered to buy Microsoft for seven or eight million dollars--I mean, nothing.

Ballmer: That wasn't nothing at the time!

Gates: No, it was flattering as hell.

Ballmer: After that dinner we walked from the Plaza Hotel down to Battery Park and back, talking about business the whole way.

Gates: Steve once came by Albuquerque when he was working for P&G.

Ballmer: We were test-marketing Cold Snap frozen desserts in Denver and Dallas, so I came by on one of my trips.

Gates: The key thing happened after we moved Microsoft back to Seattle. I invited Steve to come visit for a week in 1979, before he started on his MBA at Stanford. Steve came back to visit again the summer after his first year. I had it in my mind that I really needed help running Microsoft and that he was exactly the kind of help I needed.

The company had about 30 people then. We had so many customers who wanted our stuff, but we weren't really sure how to hire the right people, how to price our software, and things like that. I had overloaded myself with negotiating deals, taking care of the books, managing, and writing code. I hadn't created a structure. It was wonderful, but it was a bit of a mess.

It's a weird thing to try to get someone to drop out of a prestigious business school like Stanford. My line to Steve was: "Do you know anybody who is as good as you are? Do you have a twin you haven't told me about? That's what I need." There was always the fairly clear hint, however, of "Come on, man, help me out. You can get some practical experience and then go back to a FORTUNE 500 company later if you have to, but I need you now."

Ballmer: I was trying to decide what to do for the summer after my first year at Stanford, so I went to Los Angeles to talk to Booz-Allen and McKinsey and some others, but I also agreed to go up to Seattle and talk to Bill. I came up on a Friday, and we went out to dinner that night with Bill's parents, during which Bill says, "I'm on a redeye flight at midnight. Can you drop me off at the airport? I have to go on vacation down in the Caribbean, so please forgive me." I dropped Bill off and took his car home for him. Then I needed a ride to the airport the next day, so Paul Allen took me. As it turned out, that was part of the sales pitch, because Paul and I had never spent much time together.

Gates: So there I am in the Caribbean, with a bunch of Microsoft people on this boat--the Doo-Wah Doo-Wah--recruiting Steve over the radiophone. We had this notion that, based on success, he'd get between 4% and 10% of the company. But we hadn't settled on a salary. So we had this negotiation of whether to pay him $40,000 or $50,000, and my cohorts, who had all had too many drinks, said, "Hey, you look really uptight. Pay him whatever he wants! Let him take more vacations!" I finally gave in and gave him $56,000, or was it $60,000?

Ballmer: I thought it was $40,000. It wasn't $56,000.

Gates: I'm pretty sure it was $50,000. So anyway, Steve comes to work, and then a month into it, there gets to be friction between the strong personalities. Some people didn't know what to think of Steve.

Ballmer: Some did not like me coming because they had seen my offer letter.

Gates: Yeah, somebody left it on the Wang word-processing system, and one of the guys printed it out for everyone in the company to see. Paul and I had tried to be open with everybody, explaining how we would pay everyone very well but Microsoft was a partnership owned by Paul and myself. So when they found out we were sharing ownership with Steve...

The key was getting Steve to leave Stanford Business School. Of course, his parents' expectations were not along those lines.

Ballmer: My dad never went to college. He was an immigrant and he worked at Ford. There, if you wanted to get ahead in life, you went to business school.

Gates: Your first title was assistant to the president. Around Microsoft he was known as the "A to the P." So a month into it, Steve comes to me and says, "We have to hire 40 more people."

Ballmer: We only had 30 at this time.

Gates: But I had this very conservative view because we'd seen a lot of customers go bankrupt. So I liked a lot of cash in the bank, relative to our payroll needs.

Ballmer: Bill had these yellow pieces of paper all over his house that had written on them the amount of cash we had in the bank, contract payments due, and the name and salary of everybody in the company, and you could see how many months of salary he had in the bank. So we had a big row about his conservatism. And Bill says to me, "I didn't have you quit business school to come up here and bankrupt us."

Gates: Finally we agreed that we would be really picky about who we hired, so it wouldn't be too easy to hire that many people that quickly. In the end, the growth came faster than we expected, so that donnybrook went away.

Ballmer: By 1984 I started running the operating-system platform-development operations. The reason I took over platforms despite not being a programmer was that I was managing the IBM relationship. Those two things--DOS and IBM--were pretty closely coupled.

Gates: Yeah, that broke the unspoken rule that everybody you work for at Microsoft should be able to write code better than you did. In the early days, the engineers ruled. That was always a big thing.

In the year since Steve was named president, it's been anything but business as usual at Microsoft. There are a lot of new faces and new roles. You're even rewriting the company's mission statement.

Gates: The heart of it has to do with inventions that are creating opportunities for us and that the market is demanding. People talk about, to use our terms, the Web lifestyle and the Web workstyle, where the way we define information will be very different.

So the company had a real need to step back and make big technological bets that involved reinventing the company and the Windows computing platform. I was getting pretty overloaded by trying to be the intersection point for the product side of the business and the marketing-and-sales side. So I asked Steve to step up.

Ballmer: A year ago, two things were true: No. 1, the technology to make the Internet and user-interface innovations happen just wasn't getting enough of Bill's attention. Not that we don't have great technical people. But the whole's got to be greater than the sum of the parts. The only way that works is if there is a chief architect, and Bill was doing it only part-time. No. 2, we've had to tell our people to start looking at our business differently. Either they say, "Everything is like it always has been, and we just try to sell more," or they can be in an even more dangerous mode, when they say, "We've met the new opportunity head-on, we turned the ship around, and we've won."

Well, neither is true. We did turn the ship around some, and we should feel good about that; and sure, we've got to sell more of that old stuff. But there really is [claps his hands loudly for emphasis] a new opportunity here. We're not at the end of a process; we're at the beginning of the next phase.

Gates: Our success comes from some big technology bets. We made a big bet on the graphical user interface (Windows); we made a big bet on Windows NT; and there will be a new way for developing software applications for the Web that will start with Windows 2000, and that's a big bet too.

And what kind of networked, digital device are you going to have in your living room? We're betting it will be something that evolves out of WebTV. People also know we're investing heavily in Microsoft Research, but they haven't seen how that work coalesces into the next generation of software. A big part of my job, then, is to galvanize across product boundaries what these big technology bets are, and to get momentum behind them.

Ballmer: We really don't want to have happen here what's happened at other high-tech companies, which is to miss those seminal big bets. Otherwise, we'll be another also-ran. We've got plenty of people who used to work at DEC, and they're happy to remind us DEC's No. 1 problem was not building the right stuff. I can help with a lot of the other problems: Do we have the right people, the right teams, and are we focused in on the money and the accountability? But are we building the right stuff? We've got to have Bill and the rest of the senior technical leadership behind that.

Judging from your actions since taking over as president, it seems you foresee problems with execution, marketing, and other traditional business issues.

Ballmer: How do you prepare for the fact that our business model is going to change fundamentally over the next five years? That's just the nature of what's going to happen with the Internet. So we ask ourselves, Do we have our people focused the right way to contend with that? Are our people able to see that there are unifying architectural principles in our platform, and are they able to explain those to our customers? Our people are actually pretty customer-focused, but they've got to be able to translate from "Aha!" to actual business benefit.

What about the notion that there will be a profusion of platforms and devices on the network--specialized devices that are optimized for particular tasks? Have you accepted the fact that not all of them will need Microsoft software?

Gates: Sure, the notion of a platform is changing. It's changing in the sense that you often think now of writing a piece of code that you can "serve up" to a variety of clients--to a phone with simple voice input and output, or to a small-screen hand-held PDA, or to a big screen like a TV without much data storage or processor intelligence, or of course to a PC, which is the ultimate client.

The PC is unique because sometimes you might want to take the application off the network and run at least part of it on your own machine. So you'll want a symmetric architecture between what goes on at the server and what goes on at the PC. With that, a salesperson could download part of an application and its data onto his PC and use it anywhere.

That's a unique feature of our technology--it empowers the individual. Our competitors who are doing server-based applications are sort of anti-empowerment--they take away that portable computing and move it to the center of the network. We're trying to take advantage of the empowerment that the PC industry has given to users rather than take it away. We also believe our approach is a strong counterpoint to those who use the single-point-of-failure model, where if the server goes down, you can't do anything.

Windows is a very different computing platform than we started with. But people can use the same development tools they've always used and run the same code, and they don't have to write all their programs in some religious new language.

Ballmer: The thing that makes the Windows computing platform today a little harder to talk about than, say, the DOS of 1985 or Windows of the early '90s is that back then the operating system was a discrete piece of software on a single machine. What we're talking about now is just harder to describe. There's a big opportunity for us to be more crisp in describing this new platform.

Gates: Yeah, but one reason it's so hard to describe is that the technology pieces are still falling into place. The part of my job that is probably most important in this era is to make our message clear and get software developers excited. I have to make sure people see us not only for our old success but also for the breathtaking new things that we're taking on.