by Jim Dwyer
They were invited to chat with a man named Mitch Kapor, an elder and eminence of the tech world whose philanthropies, businesses, and causes occupied a few floors in an old warehouse building south of Market Street in San Francisco. People brought their dogs to work. The guys were charmed.
“I love what you’re doing,” Kapor told them. “I want you to win.”
In 1982, Kapor cofounded Lotus Development Corporation, which offered a spreadsheet, database manager, and graphics package called Lotus 1-2-3. The company’s business plan called for $1 million in sales the first year. It sold $53 million. Two years later, the revenue was $153 million. In the rearview mirror of history, the reason is obvious. Lotus showed how numbers move. It mapped patterns of change and rate. Trends buried in the density of numerals now were visible, backlit by Lotus. A simple spreadsheet program was an epochal change for people who labored with pencil, paper, and calculator to figure or forecast profit and loss, rate of return. Businesspeople, academics, and scientists lined up to buy computers so they could run Lotus. It became the first killer app. Bill Gates had once predicted that there would be a personal computer on every desk. Lotus showed why.
After selling the business, Kapor dove into digital progressivism, cofounding the Electronic Frontier Foundation and playing an obstetrical role in the birth of Firefox, the open-source browser that had changed the Internet. He had become convinced that Mark Zuckerberg was vastly overreaching and that an alternative to Facebook was needed. Earlier that year, as Kapor was reading an article online, a notice popped up informing him that an acquaintance had read and “liked” the same article. Furious, he picked through the setttings of his own Facebook account, hoping to minimize the appropriation of his interests.
So he was ready to hear about the Diaspora project. They could check in and keep him up to date, Kapor told them, and he would think about ways he might be helpful. In a pleasant haze, the guys wandered back to their project, his words ringing in their ears. “Keep your heads down, keep working,” Kapor had told them.
Venture capital firms were writing. Besides the federated social web gathering in Portland, they had—thanks to Mitch Kapor—an invitation to speak at Mozilla, creator of the web browser Firefox, and at Razorfish, an online advertising firm. Bill Joy—the Edison of the Internet—wanted to meet, a scouting mission for the venture capital firm of Kleiner Perkins Caufield Byers. A writer with New York magazine who was friendly with their mentor, Evan Korth, had persuaded them to break their self-imposed embargo for a profile to be published in the fall.
The first big decision was what to do about the open offer from Randy Komisar to set themselves up for the summer in the incubator space at Kleiner Perkins on Sand Hill Road, the boulevard of gold in Silicon Valley. His invitation had arrived before the waves of publicity and the cosmically fortuitous timing of Facebook pissing off much of the planet. Not only was the notion of being in a VC incubator beyond any expectation, it was beyond conception. They had tentatively accepted. Yet that did not mean they could drop their Kickstarter pitch or the “four dudes eating pizza” tableau. Whether they were working in the basement of Max’s mom’s house or at the bottom of the Kleiner Perkins office, they would still need ramen money.
The sluice of donations changed everything, making it emotionally implausible for them to move in with venture capitalists. “People expect that we’re going to go live in a shack somewhere and do this,” Dan said.
Ilya had been charmed by Randy—they were kindred enthusiasts—but he felt a tighter commitment to the people who had funded them to keep a kind of pure spirit of what they were doing. “We can’t go the venture capital route now,” Ilya said.
Max and Rafi agreed. Moreover, they would not have to suffer for their principled stance. Another opportunity had emerged. Rafi’s brother Mike, an experienced programmer, had just started work at Pivotal Labs, a leading software laboratory in downtown San Francisco that developed programs for companies like Best Buy and Twitter. Pivotal was a master workshop for “agile development,” which encouraged programmers to be adaptive in their solutions, a departure from the more rigid approach of working off a master blueprint. It billed the services of its programmers at about fifteen hundred dollars a day. Start-ups also hired Pivotal to show them the tricks of the trade, and would take up residence in the Pivotal offices while being nurtured. After the engagements finished, some of those start-ups ended up staying in the Pivotal space to keep working.
Diaspora certainly wasn’t going to hire Pivotal, but Rob Mee, the company president, thought they were onto a good idea, if quixotic. He admired them. Plenty of people talked about building that kind of alternative social network, but no one had done it. Moreover, the guys were working in a programming framework called Ruby on Rails, a relatively new approach that was one of Pivotal’s basic tools. So Pivotal happened to have the world’s leading assemblages of experts in the very language they were using. Mee invited them to come work in the space for free.
With that offer in hand, they phoned Randy Komisar.
“It’s really important to the spirit of the $200,000 that we raised that we be independent, and that we be perceived as fully independent,” Ilya said. “We’re concerned that the people who have given us this money want us to be countercultural. We are concerned that being in your incubator would be an affront to that.”
“You know what?” Komisar said. “I buy that. So let’s get together when you get out here. I’ll do whatever I can to help you. Feel free to call on me.”
By early June, they had set up a four-man code factory in Pivotal’s office on lower Market Street in San Francisco.
—
Spread across a wide-open floor, the Pivotal developers could count on a full breakfast every day, lunch on Wednesdays; a row of bikes hung from a wall; during lunch hour, a court of Ping-Pong tables next to the dining area was heavily used.
That was a casual gloss on a company that had a precise, disciplined approach to software development, the antithesis of the hacker ethic of staying up all night, writing code, and shoveling down pizza. The Pivotal programmers—they called themselves Pivots—worked from nine to six, and almost always in pairs, two people seeing the same screens as they worked. Pair programming helped to prevent white-line fever, the paradoxical effect named by long-distance truckers who look at something for so long that they no longer can see it. In programming, the second set of eyes, Pivotal had found, cut down on both piddling errors and major logical flaws. It also built mentoring into the work. The programmers changed partners every day, avoiding the ruts that might develop between two people; it also ensured that the entire team was fluent in all elements of the project. The structure had been a revelation to Max, Dan, Ilya, and Rafi, who were accustomed to swarming over a project and keeping the long night hours of college students on deadline benders, with no fence between daylight and night.
On their first day in the office, absorbing this new culture, Rafi’s brother Mike came over to chat with Max.
“Congratulations,” he said. “At the end of the summer, or when the project is finished, you’ll be able to get jobs as software engineers.”
That was not where Max saw himself going.
“I want to be the CEO of a start-up,” he said.
Surrounded by developers whose programming talents were worth millions of dollars a year, the four NYU guys burrowed into what was practically the perfect cave—and near-perfect anonymity.
Nearly.
One afternoon, Rafi came up the stairs from the subway near the office on Market Street, and strolled blithely to the Pivotal offices.
A moment later, a stranger sent a tweet out into cyberspace:
I think I just got off the subway after Raphael from #Diaspora!
It initially unsettled him, but the others were simply amused.
—
For the first time in their lives, Max, D
an, Ilya, and Rafi were not working to someone else’s schedule: no academic calendar, no summer job, no expectations. They were on their own. The funds from Kickstarter had been deposited into a bank account, and Dan’s father, who had taken early retirement from Citibank, handled their finances. Their budget was austere: one-thousand-dollar monthly stipends, plus a housing allowance. Additional expenses were decided on an ad hoc basis, and Dan, the de facto treasurer, was notoriously careful with the money.
One day during the summer, Randy Komisar dropped by to chat with the group. Mike Sofaer walked into the session with Komisar. Perhaps inevitably, Mark Zuckerberg’s name came up, and his donation to their Kickstarter drive. Mike volunteered that he thought they should send a special thanks to Zuckerberg for his donation.
“We should be gentlemen about it,” Mike said. That set off a minidebate about the wisdom of having any contact with Zuckerberg. It came to a close when Komisar remarked that whatever the risk of getting in touch with him, there was none associated with simply ignoring him.
Max and Dan were furious at Mike’s intervention. Ilya was sent as an emissary to tell him that he was not welcome to weigh in on business questions.
Their days were long, intense. Rafi, with ample family connections in San Francisco, had found an apartment, but Max, Dan, and Ilya had all moved into rooms at a cheap hotel that housed students and budget tourists. They hated it, but none of them did much beyond sleeping there. Rafi joined a gym that Mike belonged to, and was disciplined about getting workouts and a full night’s sleep. At the end of the workday, Ilya would take a short run through downtown San Francisco, and he inveigled Max, who had never worked out, to join him for mental health purposes.
It wasn’t simply that they were a small group taking on a huge project. No one, not the wealthiest corporation, most powerful government agency, or a small crew of hackers, can reliably predict how long it will take to make a computer program. New York City gave out a $63 million contract in 1998 for a new timekeeping system. By 2011, the payments had climbed to $628 million, and the project wasn’t finished. As of 1995, the Internal Revenue Service had spent ten years and $2 billion trying to upgrade tax computer systems, without success, according to Dreaming in Code, an account of software development by Scott Rosenberg. And it’s not just government inefficiency; businesses also struggle to execute grand software plans. Rosenberg described how McDonald’s spent $170 million on a failed program to track food in thirty thousand stores. In some respects, such failures made the complications in the rollout of the Affordable Care Act in 2013 look minor.
In a classic work on the subject, The Mythical Man-Month, published in 1975, the author, Fred Brooks, explained that throwing people at a software project was counterproductive. When a job was running late, adding new people to it only delayed it. Each new programmer had to start from scratch, learning the logic already in place; teaching those concepts would consume the time of the incumbents.
A project that, in fanciful theory, should take one person three months would not be finished in one month if three people worked on it.
“Nine women can’t make a baby in one month,” wrote Brooks, who had managed software development for IBM.
He offered practical advice for programmers. “Plan to throw one away,” Brooks wrote. “You will anyhow.”
—
Standing outside the Pivotal offices on Market Street early one morning, Max saw a man with mad-professor hair, in sandals, approaching them. It was Bill Joy. They greeted him with awe, and then broke the news: they couldn’t get upstairs. It was too early and the building wasn’t open yet. They had been so worried that Joy, with his deep technical chops, would throw harder questions at them than the venture capitalists and news reporters that they had not considered the logistics of meeting so early.
Joy, who had made some of the most significant breakthroughs in modern computing while packed into a near closet at Berkeley, was not fazed by the setting. They stood chatting about the elements of Diaspora. He reported back to Komisar. “We met on the sidewalk. They were wonderful,” Komisar recalled Joy saying. “They need to peel the onion, one layer at a time, and not just drive to the center.”
That was the last they saw of Bill Joy.
—
Tuition for the design class was just about ten thousand dollars, and the check was signed “Casey Grippi.”
Janice Fraser looked at it and tried to place the name. She knew that Diaspora had a Daniel Grippi, and Max and Ilya and Raphael.
“Who’s Casey Grippi?” she asked.
“That’s my dad,” Dan said. “He’s handling the money for us.”
The guys had been sent to Fraser by Mitch Kapor and Rob Mee. The entire Bay Area was filled with people who were ready to explain the gospel of tech start-up, none more so than venture capitalists who had, through accomplishment or delusion or their proximity to seams of investment gold, come to believe that they knew what they were doing.
Fraser was not another fish in the start-up capital pond, taking no stake in the companies. She was a design guru, a kind of mentor for hire. For ten to twelve weeks, she ran seminars to help young entrepreneurs find ways to build a product that wouldn’t blow over in a stiff breeze. Diaspora was one of five start-ups in her summer 2010 class. Designers could come and go, the pixels could be moved all over the screen, but as crucial as how a thing looked, that was the simplest part of the endeavor. Discerning its purpose was far more difficult.
“A lot of people think that design is the skin. Pretty pixels,” Fraser told her classes. “That’s not to diminish the importance of pretty pixels, but it’s not enough. It’s necessary but not sufficient.”
At the first session, Fraser moved to each group to hear them explain what they were doing. She listened to Max tell the genesis story—the Moglen speech, the pizza-fueled coding sessions, the Kickstarter rampage. Then she popped her North Star question.
“Who,” she asked, “are you building this for?”
Well, Max said, privacy on Facebook sucked. They were going to make it possible for people to have encrypted conversations. And the users would have the ability to take their data with them. There would be no question about who owned their personal information.
Dan had some ideas about taking away the limits on what a user could do—making it possible for a picture to be bigger, or to have a page of type.
Who, Fraser asked again and again, was Diaspora for?
“Is it for PLM?” she said.
That was Fraser’s shorthand for People Like Me. Or people like Max, Dan, Ilya, and Rafi—distinctive personalities, unquestionably, but all members of the same food group, computer nerds. And proud of it. Were they building Diaspora for other guys who liked to hang out in the computer club and play online games? Google was an instance of something that was built for PLM. But that worked as a business because so many people felt they had to scratch the same itch of searching the Internet.
Gently, Fraser pushed her fingers through the holes of their work plan. They were working on creating a messaging function? Pictures would be shared? So? All the features wouldn’t necessarily mean a thing unless they had an idea of who was going to use it. “Every application is one person,” Fraser said. “You have to satisfy one person before you can make a market.”
For many years, the developers of free software prided themselves on a stripped-down aesthetic. Linux was a fine operating system, and could do as much or more than Microsoft’s Windows, or Apple’s OS X, but it required users to type strings of commands that in slicker products could be activated by clicking a single button. Plenty of free software was like a car with a manual transmission instead of an automatic one: the stick shift gave drivers a better feel for the car and the road, but required a level of attention to the motor and its gears that simply was unneeded in cars with automatic transmission. Apple had achieved its success with a slick, simp
le user interface, shorthanded as UI, which mediated the user experience, known as UX. To break free software out of its dowdy appearance, a company called Canonical had developed Ubuntu, a handsome, easy-to-use operating system based on a version of Linux known as Debian; its good looks were viewed with suspicion in some corners of the free-software communities. Not by the Diaspora guys, though. They had grown up with the clean dazzle and elegance of Apple products, and expected that everything would look good, and work simply.
But by itself, a great look meant nothing, Fraser warned. Without substance, a clear notion of who and what the product was for, a cool interface was the graven idol, the golden calf.
The right answer, or at least the truth, was that they were not building a product. They were creating a function, something that would make it possible to use the distributed, federated nature of the Internet for people to swap and share without paying a toll in data. That function was not a product that could be sold; it would be like trying to monetize breathing. Yet the four pizza-munching dudes were taking on the coloration of Silicon Valley, where ideas mattered only if they were the kindling for start-up businesses. The questions posed to start-ups were also put to them—interesting, provocative inquiries that did nothing to dislodge a premise that was not true and had little hope of becoming true.
At the next session, a week later, they offered ideas about trying to create a network where strong ties would thrive, and not be overgrown by the weeds of acquaintanceship.
“We’re not just the un-Facebook,” Dan said. “We’re not trying to out-Facebook Facebook. That’s not interesting to us.”
Max said: “What is really interesting to us is how we can build a product around meaningful ties. We don’t want to end social networking. We want to fix it.”
He brought along a deck of slides from a presentation by a Google researcher on the nature of online relationships. The thrust of it was that the existing social networks—that is, Facebook and the dying star of MySpace—did not encourage strong ties. Granting the obvious business interest of Google in undermining Facebook’s spreading empire of content that was walled off from the relentless indexing machines of Google, the point was beyond dispute. There was no calibration. What you posted for the benefit of the softball team about the after-game beer party was the same thing that your mom saw.