Ray Ozzie hunches over his Sony Vaio Picturebook, oblivious to the cattail marshes that fan out just beyond his corner office, in Beverly, Massachusetts. He fires up the machine, which is connected to a wireless local-area network, and gets to work. For the next two hours, he will not leave his computer. But he will not be working alone.

He needs to meet with an ad hoc group that includes members of his product-design team, along with managers from a large client company. If this group had a typical working arrangement, Ozzie would launch a round-robin of emails, asking members when and where they could get together. Instead, he works with them now — in real time.

He moves his cursor across the screen and clicks on a square labeled "Enterprise Management Design Space." That square is a shared virtual work space. He types his comments into it, and those remarks appear instantly on every computing device used by every member of the group. He can bring other tools to the task, such as a browser, a chat window, or a sketch pad. He can work simultaneously with the others, or he can go offline and work on his own. If he goes offline, the work space will be automatically updated as soon as he comes back online.

Ozzie, the programming wizard who oversaw the creation of Lotus Notes, is using Groove — a new platform for direct, person-to-person computing that was launched last October by Groove Networks Inc., a company that he founded in 1997. Like the file-sharing system popularized by Napster, Groove employs a peer-to-peer approach to computing. It lets people establish a virtual space and then invite others to "groove" with them there.

But unlike Napster — or, for that matter, Lotus Notes — Groove circumvents centralized computer infrastructure and allows PCs to talk to one another directly. "Groove works like a jazz band," says Ozzie, who keeps a framed poster of saxophonist Eric Dolphy outside his office door. "It's intended for people who want to get together and jam — to interact and improvise with each other."

Ozzie, 45, has an unlined face and hair the color of chrome. For someone who has recently weathered more than three years of self-imposed isolation — he and his team essentially went underground to develop Groove — he looks remarkably unburdened. And, at least on this winter day, he displays an appealing ratio of IQ to ego.

But Ozzie is not a man of humble ambition — far from it. Building a Notes or a Groove is a high-stakes game of big bets and daring calls. Long before it became a reality, he envisioned a day when PCs would be wired together and millions of people would leverage them as tools for working together. By helping to transform cyberspace into a workplace, he has done as much as anyone to change the way we communicate and collaborate.

In an interview with Fast Company, Ozzie offered several lessons on how people can use technology to work together well — and on how they can work together to create great technology. Every interview is a kind of collaboration. This interview, involving a master of collaboration, was no exception.

To learn how to work together, watch how kids play together. Around 1997, I began to think about how we might create a virtual, private application network. If you and I work for different organizations but we're working together, we should be able to connect easily and seamlessly in a secure, contained environment — that is, a network that brings together the right people, the right information, and the right tools, all at the right time. Or, to put it another way: If people are going to use computer technologies to augment their interactions, those technologies need to have the directness and spontaneity of a phone call, the visual immediacy of a fax, the asynchrony of email, and the privacy of a closed-door meeting.

All of this started to crystallize for me one night, when I came home and found my son playing a modified capture-the-flag version of Quake. He and his friends had actually designed their own virtual environment: They could look up, look down, look left, look right. They could jump up and grab the flag. They could even talk to other team members.

I sat there and watched him for a while, and then it hit me that this was his way of communicating. He was socializing with other people by playing this game on the Net. And I realized that those of us in business — who have so much to gain through effective communication — were using lame, document-oriented tools. Our own kids were using technology far more effectively than we were! They were operating in an environment where small groups of people can self-organize and interact. And they made me think that I should be able to use technology in the same way.

To fire up your collaboration, jump over your firewall. Over the past few years, successful organizations have begun to work with other organizations — legal firms, accounting firms, Web-design firms, logistics companies — in a very big, very strategic way. As a result, people are communicating more than ever before with colleagues who are on the other side of a corporate firewall. The trouble is, the firewall is totally foreign to how we work.

Here's what I mean. When I collaborate with someone from another company, what happens? Clearly, he and I are separated: He's behind his company's firewall, and I'm behind the Groove firewall. Do we ask our respective IT people to create an extranet portal for us? Of course not. That would take too long, and we would use the portal for only a few days. So what do we do? We just pick up the phone, or send an email, or meet face-to-face — and each of those things circumvents the corporate firewall.

Most of the collaborative tools that we as an industry have built — Lotus Notes, Microsoft Exchange, Web technologies — deal with interactions that occur either within an enterprise or beyond an enterprise, either inside a firewall or outside a firewall. Yet the more we try to integrate outside organizations into our core processes and practices, the more the corporate firewall becomes a barrier to interaction. Sure, we need to safeguard communications — but we also need to clear the way for ad hoc, dynamic collaboration. Otherwise, we'll continue to see most workplace communication occur via phone, fax, and email.

Make the medium serve the message. It's almost impossible to determine which comes first — technology or the need for technology. But for me, the starting point is always human need: You can't build a technology without first thinking about how people might use it. Early on, at least, the development process is not about technology; it's about sociology.

With Groove, I started out by analyzing how we interact when we work together. I came up with a matrix of 20 different dimensions of interpersonal communication, and then I asked myself, Which of those dimensions don't require Notes, or Exchange, or the Web?

For example, I might need to communicate an emotion. The most effective way for me to do that is to connect my voice with your ears: I pick up the phone, and I yell at you. Or I might need to communicate something visual. In that case, I'll make a picture, mark it up, and then fax it to you. Or I might need to communicate with someone who's in a different time zone. That calls for something that lets me store and forward messages. So I'll choose a medium, such as email or voice mail, that matches the temporal qualities of what I'm trying to do.

Both Lotus Notes and the Web operate only in one or two dimensions. They deliver documents or Web pages in an asynchronous fashion — and that's it. The challenge is to build technology that matches the full variety of human interactions. Sometimes we use the phone, sometimes we use the fax, sometimes we use email. But how can we bring all of that together?

Power to the people — and to the PCs. There's one myth about the Net that really concerns me: that "Internet" equals "Web." Thank goodness for Napster, because it's awoken us to the fact that there's more to the Net than just the Web. The Internet is like an electrical outlet or a phone jack: It's a means of getting bits from me to you. The Web is one use of the Net, and it's a great use. But there are a lot of other uses — email, instant messaging, Napster, Groove — and some of them haven't even been discovered yet.

Two years ago, [Oracle CEO] Larry Ellison's version of Net computing might have become a reality. The Ellison vision — link everyone together through a centralized, server-based network — is arguably the right one for computing within an enterprise. But it ignores the "personal" part of personal computing. After all, one job of an enterprise is to make people really productive. So there should be two parallel worlds: one that's built around centralized systems, and one that's built around direct, peer-to-peer systems.

Central intermediaries are great, but only if we can identify the value that they bring to the party. If I have a very high-value transaction or communication, why would I want to go through a central intermediary that could delay or distort that message? I just want to get the message from me to you in the most direct way possible.

People couldn't care less about the technology that this stuff is built on. What they respond to is the empowerment message — the notion that tools like Napster and Groove eliminate the boundaries of time and place; that those tools create direct, dynamic connections between people; that, ultimately, this technology buys them independence.

Bring people together — and then set them apart. When I was at Lotus, I found that success could be a huge distraction. People were being pulled in 10 different directions, and those of us on the development team just couldn't focus on our jobs. The only way that we could execute on a reasonable schedule was to get away, both physically and psychologically. So we opened an office an hour away from Lotus headquarters. It had just one room, with a bunch of folding tables, computers, and whiteboards. We went there and did what we had to do.

We created the same kind of environment at Groove: one room, one mission, one goal. We eliminated all interruptions — to the point where I was really annoyed that there were no interruptions! There was just that one thing that we had to do, and there were just those same ugly faces that we had to look at every day. But isolating our core development team worked really, really well.

During the three years that it took to develop Groove, we also worked in total secrecy. Some people have said that working in stealth was a clever marketing ploy — a way to build mystique for the product. They're dead wrong. We had no choice: If people at Microsoft had known what we were working on, they would have initiated their own project, and maybe other companies would have done the same thing.

To adapt fast, build slow. A lot of the technology decisions that we made at Groove were based on a belief that if we succeeded, this thing would get bigger and bigger: There would be more and more customer requirements, and we needed to keep this thing from crumbling under its own weight. Otherwise, we would give our competitors a huge opening. But that approach assumes that you're building something for the long run.

During the dotcom era, the dominant belief was that you shouldn't try to create things of value, because by the time you build them, they're obsolete. The mantra was "Build it quick. Turn it around fast." When we started to work on Groove, at the height of the dotcom frenzy, there were days when I wondered, Has the world really changed, or do things still fall at 32 feet per second? Has my gray hair finally become a liability?

For example, in 1995, the Netscape IPO signaled to the world that the rules had changed: The two-year software-development cycle was obsolete; companies needed to develop new product every two quarters; everything would now run on Internet time. But Netscape learned — in fact, we all learned — that even though the business climate has changed, you still need to build technology that's sustainable. At one point, people at Netscape found that their code base had grown so large that they couldn't enhance it in a certain way. It needed a major rewrite. The result was a great piece of code, but the rewrite basically cost Netscape its market lead to Microsoft.

If you design a system that can't be readily enhanced, you give your competitors a big market opportunity — a big window that they can just step through. Quick refinement is a big deal. But you can't build an infrastructure for rapid application development in a rapid manner. Before you can sprint, you have to take that long march.

To see the future, look to the past. I have a deep, abiding passion for using technology to augment relationships. That passion goes back to my undergraduate years at the University of Illinois at Urbana-Champaign, where I stumbled onto a project called PLATO — an acronym for "programmed logic for automated teaching operations."

PLATO consisted of about 1,000 terminals, all connected to a centralized Control Data mainframe computer. The Urbana-Champaign campus had about half of those terminals, and the rest were scattered among major universities throughout the world. As programmers started to use the system, they built into it tools that we now think of as email, conferencing, and even interactive gaming.

Thanks to PLATO, I established relationships with people I had never met face-to-face. We communicated through something called Talk-o-matic — a kind of online, interactive chat. As one person typed, each character would appear on another person's terminal. I vividly recall having a working online partnership with this one guy that lasted for at least a year. When I finally met him, I found out that he was a quadriplegic. He typed by holding a stick in his mouth — and I'd thought that he was just a lousy typist.

That was a big moment in my life, because I was communicating mind-to-mind with this guy, and the technology was acting like power steering: It was augmenting our interpersonal, human communication in a way that I could not have imagined. Soon after I left college and moved east, I realized that the rest of the computing world just didn't get it. But I was lucky, because I had seen the endgame. And the endgame was the user experience on PLATO. In a sense, I've spent the rest of my life trying to reinvent it.

Bill Breen (bbreen@fastcompany.com) is a Fast Company senior editor. Contact Ray Ozzie by email (rozzie@groove.net), or visit Groove Networks Inc. on the Web (www.groove.net).