The Facebook Effect Part 2

Adam D'Angelo, up from Caltech, was by far the most gifted and accomplished programmer of the bunch, but he was working on his own projects. He was also neither expert in nor very interested in the relatively simple Web-based languages Thefacebook was employing-PHP, JavaScript, and HTML. D'Angelo had a bad case of carpal tunnel syndrome, which meant his hands and arms hurt when he typed. So he was trying to come up with his own alternative-invent a way to move his hands in the air that a video camera could recognize, in order to manipulate text on a screen. It was a pretty challenging project, maybe too challenging, and as summer went on he spent less time on that and more time helping McCollum and Zuckerberg with Wirehog.

While the young engineers worked to bolster the site and refine its features, Parker started thinking about what it would mean to turn Thefacebook into a company. He hired the lawyer who'd helped set up Plaxo. He started looking for someone to manage "operations," a fundamental task in Internet companies that involves making sure the data center and servers are operating properly. Up until then, all that work had been outsourced to third-party companies, but Thefacebook was getting too big for that. Parker discovered that his young colleagues didn't even know the basics about network management, like what a router was. He found an engineer named Taner Halicioglu, who had experience at eBay. He worked from home in San Jose.

Parker became Thefacebook's front man, especially with investors. It wasn't uncommon for fancy cars to be parked outside on the dead-end street, under the big droopy trees that loomed over the front of the house. That meant someone with money was inside. Some guys from the Benchmark venture capital firm wanted to know if there was a chance of an equity investment. The answer was no, for the time being. But Thefacebook was going to need more funding in the near future, so Parker made sure such people felt comfortable calling or stopping by.

A couple of Google executives came over to see if there might be a way to work with or even buy Thefacebook. Even at this early date, Google was well aware that something noteworthy was going on in Palo Alto. Zuckerberg and Parker were leery, though, because the risk of becoming subsumed by Silicon Valley's Internet giant was real. If they wanted to do their own thing, they had to stay independent, they believed. Anyway, what they were trying to do was very different from what Google did. Their site was about people; Google was about data.

One area where Parker and Zuckerberg clashed was over Wirehog, on which development continued. The new president thought it was a huge distraction from the work of growing Thefacebook. And his history with Napster made him leery of getting into another tussle with music and media companies. To Parker it seemed likely that such companies would accuse Wirehog-and with it Thefacebook-of helping users steal content, just as the music industry had with Napster. With Wirehog engineer McCollum, the two flew down to Los Angeles where they met with Edgar Bronfman, Jr., CEO of Warner Music Group, and Tom Whalley, who ran Warner Bros. Records. Parker had gotten to know Whalley in his Napster days. Unsurprisingly, they were wholeheartedly opposed to Wirehog. Though Parker feared that a successful lawsuit against Wirehog could take Thefacebook down along with it, he failed to sway Zuckerberg, who persevered.

"Really great leaders.h.i.+p," says Parker, "especially in a start-up, is about knowing when to say no-evoking a vision very clearly, getting everybody excited about it, but knowing where to draw the line, especially with products. You can't do everything. And that's a lesson Mark didn't know yet. That's a lesson Mark learned."

Work was hardly the only priority, of course. What group of twenty-year-olds suddenly occupying their own house wouldn't want to party? Nerds these guys might have been, but they were fun-loving nerds. Stanford was just a mile or so away. It operates on a quarter system, so students were still around in the summer. Using a feature in Thefacebook that enabled ads to be targeted at just one school, the housemates announced their parties right on the service-"Thefacebook is having a party!"-and then often found themselves mobbed by both Stanford students and townies. Moskovitz started dating a girl who had just graduated from Palo Alto High School.

The parties were typical beer-and-booze-fueled affairs. Here's where Parker came in particularly handy. He was the only one in the group over twenty-one, so they relied on him to buy the alcohol. There was a fair amount of pot smoking, too, though Zuckerberg frowned on it and didn't partake. "Mark is just about the most anti-drug person I've ever met," says one friend.

Hanging out around the pool was of course a major activity. If a gla.s.s broke, the shards often just got swept into the water. McCollum strung a wire from the chimney on top of the house to a spot slightly lower on a telephone pole beyond the pool. With a pulley, he turned it into a zip line, so you could ride down the wire and, suspended over the pool, drop in with a ma.s.sive splash.

One favorite party activity was Beirut, or beer pong, a beer-drinking game for teams of two or more players that involves throwing a Ping-Pong ball into a bunch of beer cups arrayed in a triangle at the other end of a table. If you get your ball into the opposing team's cup, they have to drink that cup's contents. Once all the losing team's cups are eliminated, its members drink the remaining beer on the winning team's side. The losers get really drunk.

Beirut was so popular at Thefacebook (and at Harvard) that six months later Zuckerberg and friends launched a national college Beirut tournament. Thefacebook planned to pit campus teams against one another, then each school's winning team was to come to New York for the final to compete for a $10,000 prize. (The Stanford Daily Stanford Daily asked Zuckerberg why Thefacebook would host an event it had to spend $10,000 on, and he replied "Because it's cool.") Thousands of students paid ten dollars each to register, but Thefacebook canceled the compet.i.tion only four days after it launched, after being deluged with complaints from colleges. asked Zuckerberg why Thefacebook would host an event it had to spend $10,000 on, and he replied "Because it's cool.") Thousands of students paid ten dollars each to register, but Thefacebook canceled the compet.i.tion only four days after it launched, after being deluged with complaints from colleges.

The house felt like a dorm. They'd often grill hamburgers or steaks out by the pool, and eat, raucously, at an outdoor table. If the talking lasted too late at night, the neighbors would get peeved. If someone brought a girl back to his room, his roommate had to sleep downstairs on the couch or drag his mattress into another room. Some people-female and male-stopped by for days and just hung around.

One of them was a friend of Parker's named Aaron Sittig. He had earlier helped create a version of Napster for the Macintosh, called Macster, which Napster bought. At this point he was working for a nascent music-oriented social network called Imeem, located a few blocks away in Palo Alto. Sittig is a quiet, self-effacing, blond surfer type who in addition to being a programmer is a superb graphic designer and typographer. But back then he was feeling burned-out and unmotivated. Parker brought him around because he thought he could help Thefacebook, especially with design.

But Sittig wasn't showing a lot of initiative. "I kept explaining to Mark that Aaron was brilliant," says Parker. "But Aaron would just sit on the couch and diddle around on his computer all day playing with fonts. Mark kept saying 'Who is this guy? He's worthless. He doesn't do anything.' Mark thought it was bad for the work ethic to have him hanging around seeming to do nothing." (The following year, after re-enrolling at the University of California at Berkeley for a semester to study philosophy, Sittig did come to work at Thefacebook. He became one of Zuckerberg's closest confidants.) Oftentimes the coding, swordplay, and raucous meetings would go on well into the night, sometimes punctuated by breaks for drinking, movie-watching, and video-game playing. The Xbox got a workout, with the game Halo a particular favorite. Somehow Tom Cruise became a group obsession, and thus ensued a lengthy Tom Cruise movie marathon. They rented an entire stack of his DVDs. Why Tom Cruise? Sittig, who put down his laptop long enough to watch along with everyone else, explains: "Tom Cruise was funny because he's not a very cool character. He's not a cool guy." It was camp.

Pretty soon they were naming the servers on which Thefacebook's software was running after characters in Tom Cruise movies: "'Where's that script running?' 'It's running on Maverick.' 'Well, run it instead on Iceman, I need Maverick to test this feature.'" (Maverick and Iceman were characters in Cruise's 1986 film Top Gun. Top Gun.) The Ben Stiller movie Zoolander Zoolander was another house favorite, watched to excess. It played over and over in the background while people were working. These guys found it funny to quote big chunks of the movie to one another. They may have been developing a revolutionary Internet service, but they were still really just college kids. was another house favorite, watched to excess. It played over and over in the background while people were working. These guys found it funny to quote big chunks of the movie to one another. They may have been developing a revolutionary Internet service, but they were still really just college kids.

With a total of seven guys living in the house, they needed more than Parker's BMW to get around, so Zuckerberg and company bought a car. They were planning to return to Harvard in the fall so expected to sell it again in three months. They spent a few hundred dollars on one they thought couldn't depreciate further-a forest-green, twelve-year-old, manual-transmission Ford Explorer. It was so worn-out you could rotate the key halfway and turn off the engine, then remove it. To start it up again, you didn't need a key at all. Just grab the ignition and twist. It was transportation well suited for a bunch of impatient guys who half the time couldn't find the ignition key anyway.

But despite the horseplay and silliness, it was becoming apparent that Thefacebook was turning into a serious business. Zuckerberg knew he had to take more deliberate steps to keep it evolving both technologically and as a business. That summer the growth started to seem a bit scary. They didn't add any new schools until midsummer, but members.h.i.+p kept steadily climbing all summer at the thirty-four colleges where Thefacebook was already operating. And everybody a.s.sumed the beginning of the school year would bring ma.s.sive new demands. New users meant they needed more reliable software and more computing power.

The software and data for Thefacebook was running on servers at a shared facility in Santa Clara, twelve miles south. The guys had to drive down there frequently to unbox, install, and wire up more servers-an activity for which they often recruited friends to help.

They began a.s.suming that Thefacebook was going to continue to keep growing. Every time the database was upgraded or the server array reconfigured, Zuckerberg tried to do it in a way that could accommodate ten times more users than Thefacebook had at that moment. This implicit optimism proved incredibly prescient. If Zuckerberg hadn't had that confidence as early as the summer of 2004, his company might have easily suffered embarra.s.sing and possibly catastrophic outages. But the specter of Friendster's failure to manage its own growth loomed large. Zuckerberg was determined it would not happen to Thefacebook.

The twenty-year-old CEO became obsessed with how well Thefacebook was working technically. He knew that for a communications service like this, performance was key. If the speed with which it delivered new pages to users began to slow, that could be the kiss of death-the beginning of being "Friendstered." There had already been a few frightening outages and slowdowns. He and Moskovitz inserted a timer in the software that discreetly showed on every page just how long the servers had taken to display it. He would argue with the others if they proposed a feature that might reduce that speed. Milliseconds mattered. In an article published around this time, Zuckerberg was quoted saying, "I need servers just as much as I need food. I could probably go a while without eating, but if we don't have enough servers then the site is screwed."

But there was an additional factor that helped spare Thefacebook from performance disaster in its early days, even as its users' zeal and number continued to shock its founders. Zuckerberg and Moskovitz were able to deliberately pace Thefacebook's growth. They did it by deciding when to turn on new schools. Traffic growth followed a clear pattern-launch at a new school and watch usage build steadily, then level off. Each time they added a campus, traffic surged. So if the systems were acting up, capacity was at the max, or they couldn't yet afford new servers, they'd simply wait before launching at the next school. This was a rare a.s.set in an underfinanced Web start-up. It allowed Thefacebook to grow methodically even though it was being run by a bunch of inexperienced kids. Says Zuckerberg: "We didn't just go out and get a lot of investment and scale it. We kind of intentionally slowed it down in the beginning. We literally rolled it out school by school."

Another key factor in Thefacebook's early success was its use of open-source software. From the beginning its database was the open-source MySQL. It cost nothing, nor did PHP, the special programming language for website development that governed how Thefacebook's pages worked. In fact, an up-from-the-bottom Web business like this without real backers could not have emerged much before this. Open-source Web operations software in 2004 had only recently achieved robustness and maturity. Without it, Zuckerberg would not have been able to create a fully featured website in his dorm room and pay for nothing other than the server to run it. Even with 100,000 users, the company's only real costs were the servers and salaries.

Nonetheless, keeping it all running and buying new equipment as Thefacebook grew was starting to cost real money. Zuckerberg spent about $20,000 in the first couple of weeks his crew was in Palo Alto, mostly to add servers at the hosting facility. And more spending was clearly going to be necessary.

The money came out of the account Saverin had set up in Florida. In addition to the cash he and Zuckerberg had deposited, the account was augmented with a considerable amount of advertising income. But with school out, ad sales had pretty much stopped for the summer.

Parker and the new lawyer were trying to straighten out the company's legal status. The limited-liability corporation Saverin had set up was not a sufficient formal structure. It lacked governing doc.u.ments to define how the company operated. There were no contracts, no official employees, and no payroll. Outside investment would soon be needed, but to get it Thefacebook would have to be turned into a real company.

However, Saverin started to make that very difficult. By mid-July, Parker was starting to talk to investors about putting money into Thefacebook. But when Saverin got wind of these discussions, he wrote a letter to Zuckerberg saying that the original agreement between the partners was that he would have "control over the business," and he wanted a contract to guarantee him that control. Says Parker: "It was so soph.o.m.oric. He fundamentally didn't appreciate the importance of product design and technology in this picture. He had this idea that the business stuff was what was important and all this product design and user interface design and engineering and code-you just hire a bunch of engineers and put them in the engine room and they take care of that, you know?" The product as it is engineered and programmed and designed is the business for an Internet company, especially a nascent one. The slightest strategic error in advancing and operating those could mean there would be no more ads to sell.

Whether or not Saverin understood the essential mechanics of launching an Internet company, there were good reasons for him to feel frustrated with the Palo Alto crowd. He had invested his own money (or his family's) and he was the guy working with Y2M and making the calls to bring in ads. Meanwhile, he felt his partner was blase, to say the least, about revenue. When there was a request for some special treatment from an advertiser, Saverin would bring it to Zuckerberg and Moskovitz. He frequently met a brick wall. What was the chance his investment was ever going to amount to much if Thefacebook couldn't be turned into a proper business? Zuckerberg seemed content that there merely be enough money to pay the bills and keep the site operating.

Saverin had a difficult job at Thefacebook. Advertisers demand responsiveness. They want recipients of their money to be available if they have a question or problem-usually immediately. It was thus harder for Saverin to set his own hours as Zuckerberg and Moskovitz could. His job, unlike theirs, required interacting with customers. It wasn't easy to do that and still keep up with his courses at Harvard.

But he did share one thing with Zuckerberg-ambivalence about Thefacebook's likelihood of future success. He made no secret that Thefacebook was just one of his business activities. He planned to enroll in business school after graduation, so keeping his grades up mattered, despite whatever the company might want from him.

All this later led to a lawsuit. In a legal filing, Zuckerberg and company characterized Saverin's position: "Until he had written authority to do what he wanted with the business, he would obstruct the efforts of the other shareholders and the advancement of the business itself. Saverin also stated that since he owned 30% of the business, he would make it impossible for the business to raise any financing until this matter was resolved."

As their disagreements sharpened, Zuckerberg and Saverin had endless phone calls, which seldom ended with any clear resolution. The Palo Alto group took the view that Saverin was pus.h.i.+ng so hard mostly because his father, the hard-driving, self-made Brazilian multimillionaire, was urging him to. "His father was telling him to play hardball," says Parker, "but this is not somebody who should be playing hardball." Parker reports that when pressed to make a decision about something, Saverin would often say either "I have to go talk to my dad" or "I can't give you an answer now." A day or two later he would predictably come back with a firm answer-one that was unyielding.

Despite his hardball, everybody still liked Saverin. He was charming and congenial and smart. But since he didn't seem to be making a commitment to the company like the rest of them, his efforts to get more authority didn't make sense. He was, in effect, demanding to be CEO of Thefacebook without even making a full-time commitment. The boys were inexperienced, but they were working hard, usually until all hours every night, doing whatever had to be done. Saverin appeared to be luxuriating in New York. He didn't get it, they thought.

In any case, Saverin's business skills didn't impress his colleagues. Saverin was getting a lot of business from Internet banner ad networks that bought s.p.a.ce in bulk, but they paid very little, and would take months before they did pay. Even Tricia Black, who has a higher opinion of Saverin than the co-founders had, acknowledges that "there were situations where there wasn't any follow-through or there were problems with advertisers."

When Saverin had an idea for Thefacebook, it didn't always go over well with his colleagues. For instance, he thought it would be smart to change the process for requesting a new friend so that it required an additional mouse click. To Zuckerberg-fanatically devoted to making his service easy to use-that was apostasy. But Saverin thought it made sense because in the interim you could show the user an additional ad. There couldn't be a worse reason to do it, in Zuckerberg's opinion. Saverin argued strenuously with Zuckerberg and Moskovitz that Thefacebook ought to put a big banner ad at the top of the page. "We just thought that was the worst possible thing you could do," says Moskovitz. "We thought we would make more revenue in the long run if we didn't compromise the site."

Parker and the lawyer, meanwhile, were preparing to create an entirely new legal structure. They were filing papers to incorporate Thefacebook in Delaware. (Most American companies-including just about all Silicon Valley start-ups-incorporate there because Delaware's laws are favorable for business.) Parker, managing the restructuring, was particularly concerned that the intellectual property (IP) that defined what Thefacebook was-that is, the company's most critical possession-was not owned by the company. Saverin in setting up the LLC had not sufficiently defined what it controlled. (As the creator, most of the software and design was by rights owned by Zuckerberg personally, along with some owned by Moskovitz.) Legally speaking, there hardly was a company before this point. Saverin controlled the bank account, but the servers where the service actually resided, along with the intellectual property, were under the control of Zuckerberg, Moskovitz, and Parker. The Florida LLC was more or less an empty sh.e.l.l, and what it actually owned was unclear. Zuckerberg and Moskovitz signed over their portion of the LLC, plus the critical IP, to the new Delaware corporation.

Zuckerberg won't talk about this dispute now, but his legal filings say he told Saverin that because he refused to move to California with the rest of them and had not done work he'd said he'd do, he would subsequently no longer be an employee of the company. While his owners.h.i.+p interests would remain, they were inevitably subject to dilution (meaning they would represent a smaller and smaller percent of the total company) as employees were hired and given stock options, and as investors bought into Thefacebook. Zuckerberg and Moskovitz, by contrast, would be eligible to receive additional grants of stock based on their continuing contributions.

The new corporate bylaws provided that Zuckerberg, with 51 percent owners.h.i.+p, was the company's sole director. Saverin got 34.4 percent. Zuckerberg upped Moskovitz's portion of the company to 6.81 percent in recognition of his increasing contributions. He also gave his new confidant Parker 6.47 percent. But apparently n.o.body's loyalty could be taken for granted at this point, so the shares of both Parker and Moskovitz were to double if they stayed until the following year, which would significantly dilute Saverin's share. Interviewed by the Harvard Crimson Harvard Crimson a few months later, Zuckerberg explained why he'd increased Moskovitz's stake: "Everyone else was like, 'What the f.u.c.k are you doing?' And I was like, 'What do you mean? This is the right thing to be doing. He clearly does a lot of work.'" The law firm got the remaining 1.29 percent. a few months later, Zuckerberg explained why he'd increased Moskovitz's stake: "Everyone else was like, 'What the f.u.c.k are you doing?' And I was like, 'What do you mean? This is the right thing to be doing. He clearly does a lot of work.'" The law firm got the remaining 1.29 percent.

Saverin later claimed that he did not know the company was being reincorporated, or about several other aspects of this plan. But something he learned around this time must have made him a lot angrier because this is when he "attempted to hijack the business," in the words of Thefacebook's later legal filing. He froze the Florida bank account, making it impossible for the company to pay its bills. He said he wouldn't release any money until his business demands were met. "It felt like we were negotiating with terrorists," says someone who was in the Palo Alto house. This was just when it had become apparent that big purchases of new servers would soon be required. Saverin said he had prepared an operating agreement that described the respective roles the boys would play in the company, but he wouldn't let Zuckerberg see it unless he promised to sign it without showing it to his lawyer or anybody else. Zuckerberg responded by creating his own doc.u.ment, which described the responsibilities he believed were appropriate for both of them, but Saverin would have none of it.

As the negotiations continued, Zuckerberg had to spend his own money to keep the lights on at 819 La Jennifer Way and more importantly, to keep buying servers. Zuckerberg had tens of thousands of dollars he had saved from programming and website jobs he'd done in his summers and spare time. His dentist father and psychologist mother also contributed thousands. This was money, according to a later lawsuit, that had been intended for his college tuition. Zuckerberg and his family ended up spending $85,000 that summer. For twenty-five new servers alone, he spent $28,000.

Chris Hughes didn't return from France and show up at the house until the end of the summer. But even so, he played a critical part in Zuckerberg's brain trust. Thefacebook's Palo Alto geeks lacked confidence in their own judgments about how people would respond to the product. Humanities major Hughes had a better feel than they did for how users would respond to new features. Immediately upon his arrival Hughes was deluged with requests to look at this or that feature or page design. He talked a lot about privacy and simplicity. Even after Hughes left to go back to school for his junior year, master and commander Zuckerberg often wielded Hughes's opinions when arguing a point with one of the others. Hughes remained Thefacebook's public spokesman, fielding an ever-growing number of interview requests out of his dorm room, mostly from college papers around the country.

At summer's end, Thefacebook had over 200,000 users. Zuckerberg and Moskovitz were planning to launch Zuckerberg and Moskovitz were planning to launch at seventy more campuses in September. Parker was well along in continued discussions with investors who the guys hoped would give them the money they needed without too many strings. Negotiations with Saverin continued. at seventy more campuses in September. Parker was well along in continued discussions with investors who the guys hoped would give them the money they needed without too many strings. Negotiations with Saverin continued.

Some weeks earlier, Zuckerberg and Moskovitz took about five minutes to decide they wouldn't return to Harvard. Earlier they had thought they'd be able to run Thefacebook from their dorm room again, but signs were that this could be an explosive school year for the service. They didn't want to mess it up. D'Angelo and the interns returned to school, as did Saverin. Zuckerberg, Moskovitz, Parker, and Halicioglu were, for now, Thefacebook. McCollum stayed on to work on Wirehog.

On September 11, the owners of the house stopped by to check on its condition. They did not like what they found. Zuckerberg had sublet it from tenants for the summer. In a later court case, a memo the owners subsequently wrote was entered in the record. "The house appeared to be in total disarray and very dirty," they wrote. "Furniture out in garage-unsure about what is missing and/or broken...Ashes from bar-b-q dumped-some on deck and some in a flowerpot out in back yard. Broken gla.s.s all around yard and some on deck...An antique Indian basket...had been taken outside and left on top of the built-in bar-b-q. It was broken and burned...." They also complained about damage to the chimney from the zip line, repair costs for the pool filter damaged by pieces of gla.s.s, a broken laundry room door, etc. College shenanigans had been extensive at Thefacebook's corporate headquarters.

In early September, even as he was still wrestling by phone with Saverin, Zuckerberg was served doc.u.ments informing him that Tyler and Cameron Winkelvoss and Divya Narendra had filed suit in federal court. They contended that Zuckerberg had stolen the idea for Thefacebook from them.

3.

Social Networking and the Internet.

"Every capitalist out there wants a piece of the action."

The concepts of social networking are not new, and many of the components of the early Facebook were originally pioneered by others. Zuckerberg has been accused several times of stealing ideas to create Facebook. But in fact his service is heir to ideas that have been evolving for forty years.

Something like Facebook was envisioned by engineers who laid the groundwork for the Internet. In a 1968 essay by J. C. R. Licklider In a 1968 essay by J. C. R. Licklider and Robert W. Taylor t.i.tled "The Computer as Communication Device," the authors asked, "What will on-line interactive communities be like? In most fields they will consist of geographically separated members, sometimes grouped in small cl.u.s.ters and sometimes working individually. They will be communities not of common location, but of common interest." The article crept further toward the concept of social networking when it said, "You will not send a letter or a telegram; you will simply identify the people whose files should be linked to yours." As a key employee in the Advance Research Projects Agency of the Department of Defense, Licklider helped conceive and fund what became the ARPAnet, which in turn led to the Internet. and Robert W. Taylor t.i.tled "The Computer as Communication Device," the authors asked, "What will on-line interactive communities be like? In most fields they will consist of geographically separated members, sometimes grouped in small cl.u.s.ters and sometimes working individually. They will be communities not of common location, but of common interest." The article crept further toward the concept of social networking when it said, "You will not send a letter or a telegram; you will simply identify the people whose files should be linked to yours." As a key employee in the Advance Research Projects Agency of the Department of Defense, Licklider helped conceive and fund what became the ARPAnet, which in turn led to the Internet.

A decade or so later, a few pioneers were beginning to spend time in such online communities. The first service on the Internet that captured substantial numbers of nontechnical users-long before the invention of the World Wide Web-was the Usenet. Begun in 1979, it enabled people to post messages to groups dedicated to specific topics. It functions to this day. In 1985, Stewart Brand, Larry Brilliant, and a couple of others launched an electronic bulletin board called The Whole Earth 'Lectronic Link, or Well, in San Francisco. In 1987, Howard Rheingold, a big user of the Well, published an essay in which he coined the term virtual community virtual community to describe this new experience. to describe this new experience. "A virtual community is a group of people "A virtual community is a group of people who may or may not meet one another face to face," Rheingold wrote, "and who exchange words and ideas through the mediation of computer bulletin boards and networks." who may or may not meet one another face to face," Rheingold wrote, "and who exchange words and ideas through the mediation of computer bulletin boards and networks."

More and more people became familiar with electronic communication, initially by commenting in online groups and chat rooms. The French postal service was the first to bring these concepts to a ma.s.s consumer audience when it launched a national online service there called Minitel in 1982. Then America Online started in 1985, initially under another name. In 1988, IBM and Sears created an ambitious commercial online service called Prodigy. Shortly, however, AOL came to dominate the business in the United States. On these services people typically invented or had a.s.signed a quasi-anonymous username for themselves, which they used for interacting with others. I was Davidk4068 on AOL. By the early 1990s, ordinary people began using electronic mail, again typically using addresses that did not correspond to their names. Though they maintained email address books inside these services, members did not otherwise identify their real-life friends or establish regular communication pathways with them. Later in the decade, instant-messaging services took hold the same way-people used pseudonymous labels for themselves, not their names.

In the early days of the World Wide Web, the notion of an online community advanced a little further. Services like TheGlobe.com, Geocities, and Tripod emerged and enabled users to set up a personal home page that could in some cases link to pages created by other members. Mark Zuckerberg's first website was one he created on Geocities when still in junior high school. The popular fee-based dating site Match.com launched in 1994 filled with personal information, but for a very specific purpose. Cla.s.smates.com debuted in 1995 as a way to help people, identified by their real names, find and communicate with former school friends.

The era of modern social networking finally began in early 1997. That's when a New Yorkbased start-up called sixdegrees.com inaugurated a breakthrough service based on real names. Two Internet sociologists, danah boyd and Nicole Ellison Two Internet sociologists, danah boyd and Nicole Ellison, articulated in a 2007 paper the salient features of a true social network: a service where users can "construct a public or semi-public profile," "articulate a list of other users with whom they share a connection," and "view and traverse their list of connections and those made by others within the system." You establish your position in a complex network of relations.h.i.+ps, and your profile positions you in the context of these relations.h.i.+ps, usually in order to uncover otherwise hidden points of common interest or connection. Another element must be added to explain the trends that led to Facebook-an online profile based upon a user's genuine ident.i.ty.

The sixdegrees service was the first online business that attempted to identify and map a set of real relations.h.i.+ps between real people using their real names, and it was visionary for its time. Its name evokes the speculative concept that everyone on earth can be connected through an extended chain of relations.h.i.+ps that begins with your immediate friends, proceeds to the next "degree"-the friends of your friends, and on until the sixth "degree."

Andrew Weinreich, sixdegrees' founder and a lawyer, was himself an inveterate networker. The World Wide Web was just beginning to get traction among ordinary people. At sixdegrees' launch in early 1997, Weinreich invited several hundred people a.s.sembled at New York's Puck Building to join immediately at one of the twenty PCs set up there in the room. "It no longer makes sense for your Rolodex to live on your computer," he proclaimed. "We'll place your Rolodex in a central location. If everyone uploads their Rolodex, you should be able to traverse the world!"

Members normally joined sixdegrees after receiving an email invitation from an existing member. This method of recruitment would be imitated by many subsequent social networks. It sounds obvious to us now, but at the time it was revolutionary. The service allowed you to create a personal profile listing information about you and your interests, based on your real name. Then it helped you establish an electronic connection with friends. You could search through profiles and ask friends to introduce you to interesting people you found. There were two key features on sixdegrees when it launched. The first was "connect me." If you put in someone's name, it would create a map of your relations.h.i.+p to them through the service's various members. The other was "network me," which enabled you to identify certain characteristics you were looking for, so the service could identify members who matched those qualities. A doctor in Scarsdale who likes chess, perhaps?

But, as Weinreich now ruefully concedes, "We were early. Timing is everything." The service was hugely expensive to develop and operate. It hired ninety employees, bought lots of expensive servers and database software licenses from Oracle, and paid Web-development firms Sapient and Scient millions to develop features. And what did all this expense make possible? A service that most people used at a painfully slow speed on a dial-up modem. And there were other severe limitations. Profiles may have had your name, career data, and favorite movies, but they lacked photographs. After all, few people back then had digital cameras. The lack of photos was such an obvious problem that in 1999 Weinreich seriously considered asking members to send in photo prints of themselves so interns could upload them a.s.sembly-line style.

It was unclear to people-members and nonmembers alike-whether sixdegrees was intended as a dating service, a business networking service, or both. Nonetheless, by 1999 sixdegrees had reached Nonetheless, by 1999 sixdegrees had reached 3.5 million registered users and a larger company bought it for $125 million. But it never generated much revenue, and in the wake of the dot-com bust its new owner shut the money-losing company down in late 2000. Figuring that this was the beginning, not the end, of social networking, lawyer Weinreich and his partners had the foresight to win a very broad patent covering sixdegrees' innovations, a patent that would later figure in the history of Facebook. Weinreich talked about networks like his as the "operating system of the future." 3.5 million registered users and a larger company bought it for $125 million. But it never generated much revenue, and in the wake of the dot-com bust its new owner shut the money-losing company down in late 2000. Figuring that this was the beginning, not the end, of social networking, lawyer Weinreich and his partners had the foresight to win a very broad patent covering sixdegrees' innovations, a patent that would later figure in the history of Facebook. Weinreich talked about networks like his as the "operating system of the future."

Though sixdegrees broke the ice, it took years before others ventured into the waters and built what can be considered genuine social networks. In 1999, two ethnic-focused sites, BlackPlanet and Asian Avenue, launched with limited social networking functions. A Swedish social network for teenagers called LunarStorm launched January 1, 2000. Cyworld, a hugely popular service in Korea, added social networking capabilities in 2001.

It wasn't until 2001 and 2002 that the social networking bug hit Silicon Valley and San Francisco. Most of the entrepreneurs and venture capitalists there were still in shock following the precipitous slide in valuations and revenues for Internet companies that began in early 2000. Companies were closing and the mood was grim, especially for consumer Internet companies. New ones were hardly receiving any investment money in 2001 and 2002. But a few hardy souls recognized that sixdegrees might have simply started too soon.

Plaxo, the Internet company that Sean Parker founded with friends in 2001, wasn't a social network, but it had a lot in common with them. Plaxo was a contact management service. After new members uploaded their contacts, it relentlessly peppered those people with requests to update their information, always pressing them to join as well. It was obnoxious, but it worked often enough. Parker was thinking much like Andrew Weinreich at sixdegrees-put your Rolodex in a central location and let us manage it for you. Parker liked the Plaxo concept because it was viral-one user could lead to an entire chain of users. Plaxo also foreshadowed a crucial aspect of Facebook-it maintained unique identifying information for individuals based upon that person's network of contacts.

In late 2001, an entrepreneur and local pioneer named Adrian Scott launched a social network called Ryze. Scott aimed to eliminate any uncertainty about Ryze's purpose. It was not a dating site. It was for businesspeople. Its name was intended to evoke the way members could "rise up" by improving the quality of their personal business network. Members' profiles focused on work accomplishments and they networked with co-workers and business contacts. It planned to make money by charging employers and others to search its databases for prospective employees, consultants, etc. Though it never much took hold except among San Francisco's tech cognoscenti, it inspired and set the stage for many developments that followed.

Jonathan Abrams, a local programmer, Ryze member, and inveterate partyer, saw an opportunity to focus on the nonwork part of people's lives. He built a very social network for consumers and called it Friendster. Though it wasn't exactly a dating site, it offered many tools to help members find dates. Abrams gambled that he could take customers away from Match.com, as the idea was that you'd meet more interesting people if you got to know the friends of your friends. Members were expected to use their real names, and Friendster gave you a novel tool to keep track of people-the very one that sixdegrees' Weinreich had pined for. Their pictures appeared next to their names right on their profile. This was a breakthrough. You could search to learn which people lived near you who were already friends of a friend. If you liked their picture, you could try to connect.

When Friendster launched in February 2003 it was an immediate hit. Within months it had several million users. To join you needed an invitation from an existing user, which were much in demand. Pretty soon people were talking about Friendster as the "next Google." It even reportedly turned down a $30 million buyout offer from Google itself. Back in Boston, Mark Zuckerberg took notice and joined, as did other Harvard undergraduates, including the Winkelvoss twins.

Friendster seemed to be hitting the big time. Abrams made magazine covers. But by the middle of the year the experience of users started spiraling downward. Millions were joining and Friendster's servers were slowing. It couldn't manage its success. Pages took twenty seconds to load. It also started to have public relations troubles-it engaged in a very public battle with so-called "fakesters," users who were deliberately creating Friendster profiles using phony names and ident.i.ties, including cartoon characters and dogs. Abrams was resolute that people on Friendster should use their real names, and he kicked lots of the fakesters off. Aiming in part to solve its expensive technical challenges, the company took a big infusion of money in fall 2003 from two eminent venture capital funds-$13 million from Benchmark Capital and Kleiner Perkins Caufield & Byers.

A recent visit to the San Francisco office of Friendster founder Abrams, now operating an online invitations business called Socializr, finds him scruffy-bearded, contrite, and still eager to party. The first thing he does when I walk in is offer me a tequila. He interrupts the interview several times to reiterate the offer, despite my repeated refusals. "The site didn't work well for two years-that's a fact," he concedes, finally getting down to business. Then he explains how a series of engineering misjudgments prevented Friendster from repairing its performance problems until long after he was removed by the investors as CEO in March 2004.

Abrams is one of social networking's great innovators, but he willingly concedes he built on the ideas of others. "The concepts were not new," he says. "What was new was the vibe of it, the design, the features." But the fact is, as Sean Parker puts it, "Jonathan cracked the code. He defined the basic structure of what we now call a social network."

In Friendster's wake, a throng of social networking sites blossomed in San Francisco attempting to duplicate its appeal. Each tackled the idea of connecting people in a slightly different way. One was Tickle, a service which, on observing Friendster's broad-based appeal, altered its own service, which had previously been based on self-administered quizzes and tests. Two of the other new social sites-LinkedIn and Tribe.net-were founded by friends of Abrams.

Reid Hoffman had been the lead angel investor in Friendster's very first financing, putting in $20,000 of the total $100,000 Abrams raised. Hoffman is a pivotal figure in the history of social networking. One of Silicon Valley's most thoughtful executives, he carries a substantial amount of industry credibility in his stout frame. Way back in August 1997 he started a dating service called SocialNet, which tried to find matches based on information users put in a profile. Some call it the very first social network, though Hoffman doesn't. In any case, it didn't do very well as a business (though when it was sold, its investors made their money back). But in May 2003, three months after the launch of Friendster, Hoffman founded LinkedIn, a social network for businesspeople. Hoffman believed that social networking was likely to divide into two categories-personal and business-so this was not a conflict with his support of Friendster. LinkedIn, which thrives to this day, has a lot in common with Ryze. Your profile is basically your resume. Users look for jobs and ask others for business recommendations or advice. But in keeping with its businesslike att.i.tude, it started without photos. (Hoffman added that capability later.) Mark Pincus plays Laurel to Hoffman's Hardy. Skinny, medium-height, and hyperactive, Pincus was another Friendster investor and Hoffman's buddy. In May 2003, the same time that Hoffman was launching LinkedIn, Pincus unveiled Tribe.net, a social network where members could create a "tribe" around a specific interest. Tribe.net was originally intended to help members share Craigslist-type cla.s.sifieds so they could buy things from people they knew. Its tribal quality, however, quickly became its trademark, and its most cohesive online tribes were not the ordinary-Joe ones Pincus had envisioned. They included regular attendees at the annual Burning Man festival in Nevada as well as devotees of alternative s.e.xual practices, more interested in just connecting than in buying and selling things.

Sean Parker fell in with this San Francisco social networking mafia. At that time, Parker shared a house with Stanford students in Palo Alto, where Friendster was already taking off, and several of these guys were already in his own real-world social network. Ryze's Adrian Scott had been an early Napster investor. And Tribe's Pincus had founded Freeloader, the Arlington, Virginia, start-up where Parker interned in 1994 at age fifteen. Soon Parker was hanging out with them and their friend Abrams of Friendster.

Parker and Abrams quickly bonded. And the more he hung around Abrams, the more Parker became fascinated by Friendster. He began spending lots of time at the Friendster offices. He helped Abrams find additional investors and acquired a small amount of Friendster stock himself. That was just when the service started to bend and break under the load of its newfound popularity. "I watched from the sidelines as they lost the war," says Parker now. "The story always was: 'One more month, one more month. We'll get it working.'" (Friendster was later resuscitated, but too late for the U.S. market. Now about 60 percent of users are in the Philippines, Indonesia, and Malaysia.) In the summer of 2003, just as Tribe.net and LinkedIn began to grow, an unexpected development got Pincus and Hoffman worried. They learned that the patent held by the now-extinct sixdegrees was being put up for auction by its new owners. The patent is broad and sweeping. It describes a social network service that maintains a database, enables a member to create an account, then encourages him or her to invite others to connect to their network via email. If this other person accepts the invitation and confirms his friends.h.i.+p, the service creates a two-way communications connection. These processes are at the heart of most social networks.

Lawyers for the two entrepreneurs told them that in the wrong hands the patent could be used to stop both of their companies, or pretty much any social networking firm. They decided to try to buy it. They also knew that Friendster was getting millions of dollars from the VCs, and worried that with more resources Friendster might attempt to elbow into their parts of the industry. Owning the patent was a form of defense. However, neither company's board would authorize buying the patent. So Hoffman and Pincus decided to use their own money.

But they weren't the only ones who had recognized the patent's potency. Yahoo was beginning to realize it might have missed the social networking boat. It entered the auction and actually put in the highest bid. But Hoffman and Pincus, gung-ho, were willing to pay faster and won with a bid of $700,000.

The two now say they just wanted to keep the patent out of the hands of larger players like Yahoo or Friendster. "We were worried that someone was going to buy the patent and then sue all the early social networking companies," says Hoffman now. "We bought it defensively, to make sure no one would kill the nascent industry."

But even as these entrepreneurs were creating a new industry in San Francisco, an unlikely compet.i.tor emerged from nowhere, four hundred miles south in Los Angeles. Mys.p.a.ce began as one of scores of Friendster clones-Mys.p.a.ce co-founder Tom Anderson was an avid Friendster user. Anderson got the idea to start Mys.p.a.ce in part out of frustration as Friendster slowed and crashed. But, according to But, according to Stealing Mys.p.a.ce, Stealing Mys.p.a.ce, the definitive history of Mys.p.a.ce, by Julia Angwin, Anderson also thought he could deliberately appeal to the so-called fakesters, "to create a site where users could create any ident.i.ty they liked." He and co-founder Chris DeWolfe put few restrictions on how you could use Mys.p.a.ce. the definitive history of Mys.p.a.ce, by Julia Angwin, Anderson also thought he could deliberately appeal to the so-called fakesters, "to create a site where users could create any ident.i.ty they liked." He and co-founder Chris DeWolfe put few restrictions on how you could use Mys.p.a.ce.

The two were employees of an unwieldy and disorganized Net conglomerate called eUniverse, which secretly installed spyware on users' PCs and sold expensive and questionably advertised merchandise. There they applied their libertine values to the creation of the new service. Anderson and DeWolfe took a kitchen-sink approach. If something had proven popular on the Web, the commercially minded pair wanted it in Mys.p.a.ce. When their service launched on August 15, 2003, only six months after Friendster and three months after Tribe.net, it included games, a horoscope, and blogging along with a Friendster-like profile page for members.

While Friendster's Abrams was a bit of a control freak, fighting a lengthy losing battle to protect his particular vision of a real ident.i.ty-based service, Mys.p.a.ce took a generally lax approach to just about everything. That suited members just fine. For one thing, it was less rigid about who could join than other social networks. You didn't need an invitation from an existing member. You could use either a real name or a pseudonym. And one of the features members liked best was not even intentional. An initial programming error allowed members to download Web software code-called HTML-onto their profiles. People quickly began using it to tart up their sites. Ever adaptive, the Mys.p.a.ce founders noted members' enthusiasm for this freedom and embraced the error as an a.s.set.

Member-created designs were how Mys.p.a.ce got its distinctive Times Square look-all flas.h.i.+ng graphics and ribald images. But while this look might have been unintentional, it was in keeping with the Mys.p.a.ce ethos-if you could pretend to be anybody, you also had the freedom to make your profile look like anything. And you didn't even always know who a Mys.p.a.ce member was. That made it difficult to limit your connections to genuine friends. People began adding friends w.i.l.l.y-nilly, the more the better. It became a compet.i.tion-how many could you have? As for behavior on the site, along with plenty of conventional conversation there was a definite tilt toward the s.e.xual. On Friendster the look of a profile was fixed for consistency and Abrams wanted you to use your real name for connecting to other real people. Such niceties were disregarded by Mys.p.a.ce's Anderson and DeWolfe.

As Angwin carefully explains in Stealing Mys.p.a.ce, Stealing Mys.p.a.ce, the canny founders had superb timing. The world was ready for a ma.s.s-market social network. The sixdegrees service came too soon-it lacked the right online environment in which to thrive. But that landscape had finally emerged. the canny founders had superb timing. The world was ready for a ma.s.s-market social network. The sixdegrees service came too soon-it lacked the right online environment in which to thrive. But that landscape had finally emerged. In 2003, Angwin notes, the percentage of Americans In 2003, Angwin notes, the percentage of Americans with broadband Internet access rose from 15 percent to 25 percent. Broadband not only meant faster viewing times, it also made uploading photos easier. Digital cameras were becoming common and affordable. Crucially, a wider variety of people were getting fast Net speeds. For the first time lots of families-including those with teenage girls-had broadband. Had Friendster not broken down under the strain of success, it might have appealed to this crowd, but Mys.p.a.ce nicely stepped into the void. with broadband Internet access rose from 15 percent to 25 percent. Broadband not only meant faster viewing times, it also made uploading photos easier. Digital cameras were becoming common and affordable. Crucially, a wider variety of people were getting fast Net speeds. For the first time lots of families-including those with teenage girls-had broadband. Had Friendster not broken down under the strain of success, it might have appealed to this crowd, but Mys.p.a.ce nicely stepped into the void.

Mys.p.a.ce initially spread among the relatively hip Los Angeles friends of Anderson and DeWolfe. The founders marketed their service in clubs to both bands and audiences. Shortly afterward it became an essential promotional tool for bands in L.A. It didn't take long before enterprising musicians all over the country began adopting Mys.p.a.ce. Along with the bands came the bands' audience-teenagers.

Mys.p.a.ce was hip and a great site to find out about bands, but it also leaned toward the s.e.xual. Holding Mys.p.a.ce parties in nightclubs around the country became another of the site's promotional tools. The implicit message: Mys.p.a.ce was a digital club where wild behavior was welcome. A disenchanted Friendster user who called herself Tila Tequila joined Mys.p.a.ce, bringing her fan club with her. She was a buxom young Vietnamese model with a yen for attention. Her profile was full of pictures of her wearing very skimpy clothing.

Though the site's minimum age was supposedly sixteen, plenty of younger kids created profiles claiming to be older. It wasn't unusual for thirteen-year-old eighth-grade girls to post photos of themselves wearing only a bra. Parents groups at junior highs and high schools all over the country convened alarmed meetings about the dangers of social networking.

By the time Thefacebook launched in February 2004, the flamboyant Mys.p.a.ce had more than a million members and was quickly becoming the nation's dominant social network. Thefacebook offered users limited functions, a stark-white profile page, and was limited to students at elite universities. The contrast in tone couldn't have been greater.

The first social network explicitly intended for college students had begun at Stanford University in November 2001. It was probably also the first real social network ever launched in the United States. This little-known service, called Club Nexus, was designed by a Turkish doctoral student in computer science named Orkut Buyukkokten as a way for Stanford students to improve their social life. An undergraduate political science major named Tyler Ziemann managed the nontechnical parts of the project.

Club Nexus was revolutionary and had a raft of features-probably too many. It allowed members to create a profile using their real names, and then list their best on-campus friends, who were known as "buddies" in Club Nexus lingo. Buddies who were not already members then automatically received an email inviting them to join the service. Only students with a Stanford-issued email address could join, and that email authentification ensured that each person was who they said they were. You could chat, invite friends to events, post items in a cla.s.sifieds section including personal ads, write bloglike columns, and use a sophisticated search function to find people with similar interests. Students used it to find study partners, running buddies, and dates. Buyukkokten himself once bragged Buyukkokten himself once bragged that what made it different from any other website was "you can create really big parties." that what made it different from any other website was "you can create really big parties."

Within six weeks Club Nexus had 1,500 members at Stanford, whose student body totaled about 15,000. But after it reached about 2,500, usage leveled off. The service was just too complicated. Buyukkokten was a talented programmer who had loaded it with every interesting feature he could think of. But that made it difficult to use and diffused activity among many different features. You didn't get the sense there were many others in there with you.

Once the two men got their degrees in 2002, they wanted to commercialize their venture. Recognizing that student use was tepid, they made what some might call a foolish decision, considering the subsequent successes of Facebook-they focused instead on alumni. They created a company called Affinity Engines, which began marketing a modified version of Club Nexus called InCircle to college alumni groups. Their first client was the Stanford Alumni a.s.sociation. By 2005 its customers included alumni networks at thirty-five schools, including giants like the University of Michigan. But not long after Affinity Engines started, Orkut Buyukkokten left the company and went to Google.

A year or so after he joined Google, the entrepreneurial programmer approached Marissa Mayer, a top company product executive, and told her that over the weekend he'd built the prototype for a new social network. Mayer and Google's executives, who by policy encourage entrepreneurs.h.i.+p among employees, embraced his project. Google was thinking of calling the project "Eden" or "Paradise." Then one day Adam Smith, a product manager working with Buyukkokten, told Mayer that the engineer owned the Web address Orkut.com. The two felt Buyukkokten embodied the spirit of his service, so they just decided to name it after him.

The well-conceived Orkut, a social network open to anyone, launched in January 2004, just two weeks before Thefacebook.com. It initially thrived in the United States and was holding its own against a surging Mys.p.a.ce. But by the end of 2004 it had, somewhat oddly, been tightly embraced by Brazilians. A gra.s.sroots campaign there to win more members than Orkut had in the United States captured the imagination of young Brazilians. After they succeeded, the service acquired a distinctly Brazilian and Portuguese-speaking cast. Americans began to drop away. Today Orkut, still owned by Google, remains one of the world's largest and most sophisticated social networks, yet more than half its members.h.i.+p is Brazilian. Another 20 percent live in India.Google's diminished expectations for it can perhaps be gleaned from the fact that in 2008 it moved Orkut's headquarters to Brazil.

Club Nexus was the first college-specific social network, but by the 20032004 school year similar sites were popping up at a number of schools. The Daily Jolt, a sort of discussion community, had been around since 1999 as a kind of campus bulletin board and was operating at twelve schools. Collegester.com-"a virtual community of free, useful, and enjoyable services 'for the students, by the students'"-had been launched in August 2003 by two University of California at Irvine alumni. An online matchmaking service called WesMatch was thriving at Wesleyan University. Its entrepreneurial founders had launched a version at Williams College and were expanding to Bowdoin, Colby, and Oberlin. At Yale, the student-run College Council launched a dating website called YaleStation on February 12-only a week after Thefacebook's debut. By the end of the month about two-thirds of undergraduates had registered. Then there was CUCommunity, which had taken off at Columbia in January. Both the Yale and Columbia sites were getting lots of members before Thefacebook arrived.

In late 2003 the Ivy League seems to have collectively decided that campus facebooks should go online. Student governments at Cornell, Dartmouth, Princeton, Penn, Yale, and Harvard, among others, were all complaining to college administrations that their campus facebook was not in digital form. The idea was no secret. A sense that the time had come helped push Zuckerberg to create Thefacebook and accounts for its name. Students everywhere had also been influenced by the rapid ascendancy of Friendster, and many were dismayed to see it stumbling. By fall, Mys.p.a.ce was already making waves in Los Angeles and in the music world.

Aaron Greenspan, a Harvard senior, launched a service there in September 2003 called houseSYSTEM. It allowed residents of Harvard residential houses to buy and sell books and to review courses, among other functions. It also invited students to upload their photographs to something called the Universal Face Book. houseSYSTEM was controversial for how it treated student pa.s.swords and never got much usage, though hundreds of students signed up to try it.

Separately, Divya Narendra claims to have come up with the idea for a Harvard-specific social network in December 2002. He later teamed up with the Winkelvoss brothers to build Harvard Connection, according to some of the voluminous legal doc.u.ments filed in the lawsuit they brought against Zuckerberg and Facebook. The identical towering Winkelvoss twins-known to some Harvard cla.s.smates as "the Winklevii"-worked hard at rowing for years and made the finals for men's pair rowing at the 2008 Olympics in Beijing. They came in sixth and last, but it was a huge achievement. Previously they'd won a gold Previously they'd won a gold medal at the Pan-American Games in Rio de Janeiro. The two athletic blond uber-WASPs couldn't be more different from the scrawny, nerdy, brainy Jews who founded Thefacebook. medal at the Pan-American Games in Rio de Janeiro. The two athletic blond uber-WASPs couldn't be more different from the scrawny, nerdy, brainy Jews who founded Thefacebook.

The three worked fitfully on the idea that would become Harvard Connection over the next year. Since none of them were programmers themselves, they hired people to help. Two successive computer science students tried and failed, in the founders' opinion, to get Harvard Connection right.

During the first months of Harvard's fall 2003 semester, Zuckerberg began making waves with ad hoc bad-boy applications that were intrinsically social-first Course Match, then Facemash. Narendra and the Winkelvosses read about him in the Crimson Crimson's coverage of the Facemash episode. They got in touch and arranged a meeting. He agreed to help out, but says now he thought of it as just another of his many social software "projects."

Zuckerberg worked off and on writing code for Harvard Connection. After a few weeks he appears to have lost intere