As told to Michael Fitzgerald

Second Life is a place where anyone can have just that. It is a richly detailed virtual world where anything a computer programmer can imagine can exist: There are minutely detailed replicas of Rockefeller Center and human-size raccoons; sex and sadism and spiritual retreats; conference calls and a currency exchange. Almost all of it is created by the people who pay to dwell in it. Linden Lab, the San Francisco company that created and owns Second Life, acts as a sort of laissez-faire government. It makes money primarily by selling property, of which it can conjure an infinite amount.

Second Life seems like an overnight sensation--it drew almost a million new residents in the last two months of 2006, doubling its population. In fact, it began in 1999, when Philip Rosedale quit his job as chief technology officer at RealNetworks to realize his lifelong dream of building a virtual-reality environment. Most people he knew thought he was quixotic and certain to fail. He almost did. No wonder the 38-year-old Rosedale describes the recent onslaught of attention as "almost surreal."

What's real is that Second Life is a haven for entrepreneurs, with thousands of businesses selling things ranging from clothes to office buildings to body parts. Business is conducted in Linden dollars, but those can be cashed for cold, hard credit card credits. The in-world economy is now clipping along at $10 million--those are U.S. dollars--a month. Big companies are popping up, too, experimenting with what might be a look at tomorrow's three-dimensional Web.

I was interested in virtual reality in general from when I was a kid. I was making electronics and programming computers from fifth, sixth grade really seriously. I was really into technology. I was pretty good with my hands and still am. We have a machine shop here and I just like making things. But I always thought the best place to invent would be inside the computer, if we could just get in there.

I started a company when I was 17 doing database business systems. I put in little computer systems for car dealerships and an architecture firm, did things like that. My little company put me through college. [Rosedale majored in physics at the University of California at San Diego.]

After college, I moved to the Bay Area and I moved my office right next to these four guys who were building one of the early Internet service providers. They gave me a line over the ceiling, a big Internet connection. This was 1994. I said, Man, you could use the Internet to hook together a lot of computers. You could simulate a world and then we could all go in there.

But then I said, This is a nonstarter right now because for this to be interesting it has to be sexy, it has to be fun, it has to be fast, it has to be within human response times. It has to be like a video game. And in the mid-'90s you couldn't do 3-D on a PC.

I got interested in video compression. I said, I bet I could write an algorithm that would compress video well enough to look at people's faces over those little cameras and do live video. So I built this thing called FreeVue, me and another guy, one of my buddies from physics. You could download it for free on the Internet. You could have six video windows open on the screen at the same time.

Rob Glaser [CEO of RealNetworks] saw it and came to talk to us in late 1996. He convinced us he had a great company, a great idea. I decided that while the networking and graphics for virtual reality matured, I would go be one of the early people at RealNetworks. There were a lot of amazingly smart people there and they were going to give me one hell of an engineering management experience.

I felt that what I was going to build someday, Second Life, was going to be extremely complex from a systems and software standpoint. I needed some experience working with other people and learning how you get people to work together and work on a really big system. I figured I would get to see all that at Real. And I did.

In mid-1999, I came back to San Francisco and I started the office in Hayes Valley, on Linden Alley. That's where we got the name of the company. I hired one guy, Andrew Meadows, who is still here.

We were steadfast in the belief that what we were creating was a complex emergent system driven by an economy and the contributions of a lot of people, just like the Internet but in 3-D and live--you were really there. The other thing was this idea that it ought to be a creative space where people can be entrepreneurial if they want to.

We always believed that such a place, such a platform, would be something that you could reasonably charge money for. If it provided the opportunity for other people to make money, you could reasonably charge some fraction of that.

Second Life was just unfundable. It was just the dumbest idea ever. Mitch Kapor [the founder of Lotus Development] was the only person who got it. Mitch invested in 2001 after I had invested about a million dollars of my own money. I think some of the early angel investors were largely investing in me. They thought I seemed to be a capable, balanced, good-engineering-background entrepreneur, so I could figure out something.

But we could convince absolutely no one that what we were doing made any sense. People said the technology can't possibly be made to work smoothly because there are too many problems with building a simulation combined with broadband, combined with streaming, combined with rendering, talking to many computers at once, the whole idea is just completely impossible.

The second thing they said was, This is not for ordinary people. Even if it is compelling, there will only be a few crazy people that want to do it. And then you guys will be dead.

Oh, and user-created content had never been a fundable idea. Now, everybody's doing it. But in the beginning the idea that random people were going to build a three-dimensional world was just impossible for people to understand. A lot of applications, and all 3-D applications, were top-down designs, where some master designer built all the content and you just wandered around in it. And we were saying, You [the world's residents] are going to build everything. You are going to build these walls. And everybody was like, That's terribly stupid. Nobody got it.

There were six very quiet years. It wasn't discouraging the first couple of years because we were just having fun. When we got to be 20, 25 employees it got pretty stressful. What we didn't really think about was if the content alone is what compels other people to come and join, then you have a pure word-of-mouth exponential growth model, and that means you are going to have to wait a long time for the plane to take off.

It was discouraging as we grew and we started to make some progress, to the point where residents were making stuff. And the investors would still just say, You have got to turn this thing into some sort of video game. It needs to have a purpose. They wanted us to make it like something that had come before it so that they could value it.

When we couldn't grow it as quickly as it needed to, we had one round of layoffs. There were 31 of us and 11 of us left. That was in late 2003, when we pretty well thought we were dead.

And then we did one discontinuous thing: We recognized that there was a core of people who were really starting to want to build the content and invest in it and really value it. And we said, What you have in Second Life is real and it is yours. It doesn't belong to us. We have no claim to it. Whatever you do with Second Life is your own intellectual property. You can claim copyright on it. You can make money.

We said the same thing about land: Land is yours to own and resell. We had been reading Hernando De Soto's The Mystery of Capital and Jane Jacobs and all these books about innovation and ownership and why great places are great places. And we said, Let's just make this a real world. Let's let it have a real economy and let's make property have real value. There was a lot of buzz around that. The investors could see this thing starting to go. In early 2004 we got a couple million bucks more.

I think I felt like we were going to make it maybe in early 2005. I felt like the fundamental network economics of buying and selling stuff was irreversible.

Our approach to engineering was this: Tell everybody in an e-mail every week what you are doing, then make some progress of some kind and tell everybody in an e-mail how you did it. That was our organizational scheme. We said, Everybody is smart here. Identify what you are going to do today and get it done.

That evolved over time into the work system that we have today. We have this huge database of stuff to do. You choose your own work from it. So groups are formed more organically. I am pretty critical of traditional business styles. The biggest way you avoid that is you continue to make everybody entrepreneurial, which is easy to say--everybody says garbage like that in big companies. But the way you really are entrepreneurial is that you have to set your own strategic direction. That's what entrepreneurs do. You have to take risks and you have to expect to be held accountable.

We have this thing we built called the Love Machine. The Love Machine allows anyone who works here as a Linden employee to send anyone else a brief note that says "Thank you for doing this for me." There is a little webpage where you can go to send an e-mail, and then you get a little e-mail that says "Love From Philip" in the subject and it's got text in it. Now, you think, what's the big deal about that? Well, all of that stuff goes into a database. Your review carries that. Everybody is sending love to each other. It creates a positive collaborative environment.

Most businesspeople communicate in a mostly negative way. If people are encouraged to be entrepreneurs and take risks, they can also become combative and competitive. You have to balance that. So we built the Love Machine for balance. We joke that some day we will be more famous for the Love Machine than for Second Life.

We use a lot of the ideas from The Wisdom of Crowds. We vote internally on tasks. And when you get something done you can say, "Oh, I got 17 votes on this." And again, you use that as part of your review.

We also use anonymous spot surveys for a lot of stuff. So I send out surveys saying, like, "Should we get rid of me as CEO?" Or I send out several options: "We should get a new CEO: now; when we have 200 people; when we have 500 people; never."

There were some nevers. I think people didn't figure I was good for a thousand-person company. I actually think I am, but I'd be fine not doing it, either.

We don't even have a concept of budgeting here, really. For example, we don't have a travel budget. If you travel you have to send an e-mail to everybody that says how much you spent and why it was worth it.

As an entrepreneur in high school, I thought starting a company was about a process where you do all these official things to create a company. I was so proud of my business license. I hung it on the wall. Now I've realized that a company is a culture and a model and a business surrounding an idea or ideas.

People are starting to do business meetings in Second Life. When I looked at that originally, I was like, Why would anybody do a business meeting or a conference call in a virtual world? What a dumb idea. Actually, it is not. You are more likely to speak the truth to me in Second Life than you are in a real business meeting.

An enormous amount of intellectual energy is going to move into this world, and some of what we are doing in the real world will therefore be displaced. You can imagine New York City being kind of like a museum. Still an incredibly cool place to go, but with no one working in those towers because work, creative work, where you are engaging with other people face-to-face, you are going to do in a virtual world. It's going to leave these cities [gestures toward downtown San Francisco] and move into digital worlds. It is easier to do things there.

For a full archive of past How I Did It features, visit