Bubble Boys

Feross Aboukhadijeh (YouTube Instant).Photo: Dan Winters

Feross Aboukhadijeh likes to tell the story of how he got famous. It happened last fall, as he was beginning his junior year at Stanford. Google had just unveiled a feature called Google Instant, which shows search results in real time, as you type. “I thought it was kind of gimmicky,” says Feross. But it gave him an idea: If Google could pop out instant search results, why couldn’t YouTube produce instant videos? He bet a friend he could slap something together in an hour. “I lost the bet,” he says. “It took me three hours.”

The result was YouTube Instant, a site that lets you flip through YouTube videos in real time. Say you type in the letter A: The top video that begins with that letter—currently the music video for Adele’s “Rolling in the Deep”—starts playing. Add a B to spell “Ab,” and you see a stop-animation set to the alphabet song. “Abd” gives you the trailer for the Taylor Lautner thriller Abduction. And so on.

YouTube Instant went live at 9:32 p.m. on a Thursday. When Feross woke up at eight the next morning, he had a bunch of missed calls. One of his transcribed voice-mails said, “interview washington post.” “I was like, Nah, that can’t be right,” he says. By the end of the day, YouTube Instant had tens of thousands of views, Feross’s name and grinning face had appeared on dozens of websites and TV shows, and YouTube co-founder Chad Hurley had offered him a job over Twitter.

Feross politely declined. He wanted to continue his schoolwork at Stanford, plus he had other projects gestating. But the experience put him in the crosshairs of Silicon Valley’s heavyweights, if he wasn’t there already. He’d just finished a summer working at Facebook, where he and Mark Zuckerberg had hit it off. (Zuckerberg later came to speak to a Stanford class Feross was T.A.-ing and called him out by name.) After YouTube Instant launched, a Google recruiter made it clear its door was always open. “If there’s anyone more heavily recruited, I’d want to know their name,” says Sean Holbert, course adviser for Stanford’s ­computer-science department last year.

Feross wears his celebrity well. He speaks rapidly but exudes calm, like a presenter at a TED conference. “YouTube ­Instant changed my life,” he says. “People don’t talk to me the same way. It’s like I gained twenty ­badass points. Whether I deserved it or not, I don’t know, but I’ll take it.”

YouTube Instant hasn’t changed the world—it hasn’t even made money. But its story describes the template for Silicon Valley these days, which may be a bubble, but it hasn’t popped yet: If you have an idea for an app, do it now. Throw it up online. Find an audience. Worry about quality later. Best-case scenario, you create the next Facebook. Worst-case, you try again. Even then, chances are you’ll get a job offer you can brag about rejecting. Right this minute, Silicon Valley is America’s opposite: House prices are soaring and demand for young talent far outstrips supply. The ongoing cyberspace race between Facebook, Apple, and Google, among others, means computer engineers enjoy more freedom—and power—than ever before. The barriers to entry for web programming are almost nonexistent. Angel investors are blessing start-ups left and right, and launching a software company is cheaper than ever. Do I take the offer from Google, or take the venture capital to start my own thing? Only in this one little quadrant do people have the luxury to ask such questions. For ­Feross, the son of a schoolteacher and a Syrian-born electrical engineer, the forecast is bright, though indistinct. He may become the next Mark Zuckerberg or Steve Jobs; he may not. But while most of the country is in economic darkness, the American Dream is beaming bright in Palo Alto.

It’s one of the first Saturday nights of spring, and a couple dozen students—overwhelmingly male and unseasonably pale—are packed into the student union at Stanford University, hunched over laptops, chugging Diet Coke and devouring Red Vines. The occasion is the inaugural “Happy Hacky Hour,” an event organized by the computer club (Feross is the president), where programmers are invited to hang out, eat pizza, and do lines (of code). The thump of music and the laughter of females pipe in through the open windows, signals from a far-off universe.

Three kids huddle around the computer of David Fifield, a grad student and former Eagle Scout. Fifield has a window open that displays all the Internet traffic passing through Stanford’s wireless hub. Anytime someone visits a website, the URL—say, http://www.facebook.com—and the person’s IP address pop up on Fifield’s screen. With the right powerful technology, Fifield explains, he could snag their Facebook logins and hijack their accounts. But he chooses not to. “We don’t do that kind of thing at Stanford,” one kid says with a verbal wink.

Joseph Huang (WiFiSlam) and Ivan Lee (Geomon).Photo: Dan Winters

Everyone here has a project. Two students are working on Habut, a web ­service that helps you create good habits by sending you periodic reminders (e.g., “Go outside”). Another student is starting a Q&A site for medical topics. A pair of undergrads, Dan Thompson and Grant Mathews, are working on StanfordHub, a discussion forum pegged to the student-government elections. “This guy is a total frickin’ genius,” says Thompson, nodding at Mathews. Mathews doesn’t react.

Thompson, a chatty, Bieber-haired then-sophomore, came to Stanford from the aggressively traditional St. Paul’s School in Concord, New Hampshire, planning to study psychology. “I got into CS because here it’s cool,” he says. When you understand the pleasure of hacking—the application of creativity and problem-solving—spending a Saturday night writing code doesn’t seem that crazy.

Pleasure comes with pain. You code for hours, you run the program, it fails, you debug it, run it again, pass out on your keyboard, wake up, code some more. But when it finally works, it’s a rush. Thompson, who has dabbled with cocaine in the past, compares it to the drug. “Writing code to me is the same experience,” he says. “It’s misery, misery, misery, misery, euphoria.”

Hackers don’t always sit down at their computers with a goal, says Feross: “They just want to see what they can make a computer do, even if no one uses what you build.” For him, tinkering started at an early age. Whenever his dad brought home anything requiring assembly, ­Feross would put it together. Reading the manuals gave him power over his parents. When his mother grounded him, he’d get revenge by setting the child lock on the microwave. When he got his first computer, driven by an old Intel 386, he installed a bunch of viruses and then fixed it himself. He started programming in middle school and created an AP-study-notes website that still gets 10,000 hits a day. “My dad called me Computer Guy, and he called my brother Sports Guy,” says ­Feross. “I hated that. I wanted to be Sports Guy too.” Feross plays club basketball now—at six-foot-five, how can he not?—but he’s come to terms with his skill set. “Now I’m proud to say I’m a computer nerd.”

When I met Feross at his Stanford dorm, his lanky frame was draped in a Facebook T-shirt. It was almost like the company had called dibs on him. Last summer, Feross interned at Facebook, where he worked on a small team that was rebuilding the Facebook Groups application. As the launch date approached, Mark Zuckerberg started working out of the team’s office.

Zuckerberg doesn’t code much for Facebook anymore, the same way that Steve Jobs never hand-coded software for the iPhone. But, as the Groups team was adding the finishing touches to its product, Zuckerberg said he wanted to write a few lines. “Everybody was like, Ohhhh, Zuck’s gonna write code,” says Feross. Someone set up an easy bug for him to fix—adding a link to a picture, or something—and he went to work. Five minutes passed. Twenty minutes. An hour. “It took him like two hours to do something that would take one of us who’s an engineer like five minutes,” says Feross. It was like a retired slugger coming back for one last at-bat, for old time’s sake, and finding he’d lost more of his game than he’d reckoned. Still, he got props from Feross & Co. for getting his hands dirty.

It’s moments like these that growing tech companies struggle to hang on to. When Zuckerberg gets in the trenches with the grunts, it sends a message: Program­ming is the core of their world. Coding isn’t about making money or scratching some OCD itch. It’s about doing what you love and, yes, changing the world. Engineers shopping their talents talk about impact; they say they want to work wherever their contribution will make the most people happiest. You can have the best day of your life working at a big firm like Yahoo, and you still won’t affect the company’s value. At a small company, you can triple it.

The result is that in order to recruit young talent, companies try to seem smaller while getting bigger. In January, when Google announced that Larry Page would step in as CEO of Google in April, the move was sold as a return to the company’s start-up roots. Since then, every­thing has been squeezed into this narrative, from prioritizing mobile to tying the size of bonuses to success in social-media efforts to the campus workshop where Googlers can build stuff out of wood or metal or Legos. Google also seduces free spirits by famously letting employees spend 20 percent of their time on projects they are passionate about—some of which turn into major Google products, such as Google News and Gmail.

Dan Thompson (StanfordHub).Photo: Dan Winters

Facebook, while considerably smaller than Google, with over 2,000 employees, is growing fast. But it cultivates an image of start-uppiness and agility. Zuckerberg has an ethos: “Move fast and break things.” Several times a year, the company holds 24-hour hackathons. Facebook also brags to prospective hires about its enviable engineer-to-user ratio, which as of two years ago was about one to 1.26 million—the kind of leverage most successful start-ups can’t match.

Anti-bureaucracy also means giving people at the bottom access to people at the top. Every week, Google hosts ­“TGIFs,” giant companywide meetings where employees can ask execs questions about what’s going on in the company. “That’s very much part of the Google culture, not to take everything for granted, to question authority,” says Michael Brandt, a recent Stanford graduate who has a job lined up at the company. Facebook prides itself on its flatness and Zuckerberg on his accessibility. “I’d be there at 8 or 9 p.m., and we’d talk for a half-hour about whatever was on his mind,” says Feross. “It was really personal.”

Then there were the perks. It’s a cliché by now that every big tech company’s campus is a mini-resort. Apple’s proposed new site looks like the set of Close ­Encounters of the Third Kind. Facebook boasts Ping-Pong tables and Xbox 360s. Employees at the “Googleplex” in Mountain View can take after-work dance classes and then get massages before hopping a free shuttle home. Zynga, the gaming company in San Francisco, lets employees bring their pets to work. And, of course, free food every­where—food for your dog can’t be far off. Even AOL, that nineties relic, has a fancy new Palo Alto building, complete with a pinball machine, a Razor-scooter dock, and a room with a drum set.

“The shoe thing is overrated,” says ­Joseph Huang, bouncing on a roller chair. Huang has been homeless the last few days, crashing on couches while he figures out his living situation. Skinny and barefoot, with spiky hair and an oversize T-shirt from the Montreal Jazzfest, he comes across as a tall Asian rabbit.

“Ideas are a dime a dozen.”

Huang is sitting amid a cluster of tables in AOL’s Palo Alto offices, a space dedicated to StartX, an accelerator for start-ups involving Stanford students. Huang’s start-up is one of ten fledgling companies selected in the spring session by the accelerator, which helps them with everything from legal issues to accounting to office space to connecting with financiers. His company, WiFiSlam, analyzes local wireless signals to create a kind of indoor GPS. Fire up their Android app, and a little dot on a map tells you where you are. Commercial uses could include museum walking tours, private-security coordination, or navigations systems for malls.

“I worked at Google, but it’s not hungry enough,” he says. The advantage of working at an accelerator is you’re surrounded by entrepreneurs with similar brains (large, hyperactive, mildly Asperger-y). After an internship at Microsoft, then-Stanford undergrad Ivan Lee decided to try something different. He and a few friends created Geomon, a sort of augmented-reality Pokémon game for your iPhone that integrates elements from the world around you. If you’re at the beach on a sunny day, for example, you’ll fight different creatures than if you’re in a rainy city. Even if Geomon doesn’t work out, he won’t go back to Microsoft. “I think I’d rather stay and do more start-ups,” he says.

“Taking a job is in some ways like a ­second option,” says Akshay Kothari, who with his friend Ankit Gupta converted a Palo Alto garage into office space for their iPad-news-app company, Pulse. “If you say you’ve taken a job, it’s like, ‘Oh, you haven’t figured it out?’” Last year, the two Stanford classmates started building the news app as part of a class at the Institute of Design. Five weeks later, it launched on the iTunes App Store, where it sold for $4. It now has 5.5 million users.

Feross buys into the Silicon Valley meme of entrepreneurs as pirates: They know the risks, but they do it out of love. “If you’re an entrepreneur, you have a kind of delusion,” he says. “You’re willing to do something that’s kind of ridiculous, which is probably gonna fail 99 percent of the time, because the adventure of doing the thing itself—the journey—is so enjoyable that you don’t give a crap about the end.”

The start-up life has its downsides. For every Pulse, there are a hundred non-Pulses. Entrepreneurship is all-­consuming. Starting a start-up isn’t work—it’s parenthood. “If your company could be worth half a billion dollars by the end of the year or zero dollars, the decision of whether to work this weekend is no longer a decision,” says Brandt, the Stanford graduate. “To me, that’s terrifying,” says another Googler. “No salary, tallying living ­expenses. I could not do that.”

Still, there has never been a better time to be a geek with a dream. Launching a new product is all but free. “The incremental cost of starting a start-up beyond hanging out and doing nothing is basically zero,” says Paul Graham, the Viaweb co-founder turned tech sage who now runs Y Combinator, a start-up incubator in Mountain View. Server space is vanishingly cheap. Open-source software is readily available. While California faces one of the worst budget crises in its history and national employment numbers remain stagnant, venture capitalists stalk the campuses of Stanford and Caltech, looking for fertile ideas on which to sprinkle dollars.

And thanks to social networks and mobile devices, your audience is right there: If you build it, they will download. When Pulse went live on the iTunes App Store, it was clunky and slow, with an ugly 3-D graphic pasted across the background. Still, it quickly became the top-selling paid iPad application, with more than 15,000 downloads in the first few weeks. “I was like, Why are people buying this shit?” says Kothari. Its second iteration was smoother, with no 3-D monstrosity and soon a price tag of zero. But the decision to launch immediately—rather than sit around perfecting the app—was crucial.

The most important shift for programmers isn’t the economy or even social networking. It’s a concept called abstraction. Abstraction basically means automating low-level tasks (like designing a button for a website from scratch) so that you, the creator, can focus on high-level problems (like how the website looks and feels). Over the last decade, coding has become increasingly abstract. Thanks to the open-source software movement, engineers have access to vast online troves of free code they’d otherwise have to write themselves. At the same time, companies like Apple and Google have built tool kits for programmers who want to design apps. A lot of iPhone apps use the same scroll feature, for example. That’s because Apple offers it as a free “black box” tool—you can use it, you just can’t open it up to see how it works. In a sense, coders aren’t builders so much as D.J.’s, pulling a few lines of code here and a black box there and smushing them together.

YouTube Instant is, in essence, a mash-up of two existing technologies: Google’s “suggest” algorithm, which lets you search the most common terms, and YouTube’s embedding capability. “I didn’t do these things myself,” says Feross. “I just combined them together in a way that hadn’t been done before.”

Abstraction helps engineers build new things faster. Pulse came to life in only five weeks because many of the components—scroll functions, buttons, fonts, layouts—already existed. Other start-up logistics get abstracted, too: Buying server space, once an ordeal on par with renting an apartment, now takes twenty minutes online, thanks to companies like Amazon that handle it for you. Advertising your product no longer requires a sales team, thanks to ad networks. Companies like StartX and Y Combinator facilitate the legal and accounting ­mumbo-jumbo as well as the investment meetings.

All this would seem to mean that anyone with an idea could bring it to market, regardless of programming talent. In fact, it means the opposite.

At a picnic table outside the Stanford Coffee House, Feross flips open his ­MacBook Pro and shows me his latest invention. It’s a site called Instant.fm that allows users to create playlists online. It sounds obvious—anyone who knows ­Spotify or Grooveshark gets the point. The difference is Instant.fm’s simple ­interface, video capability, and easy sharing. Feross’s key insight was pulling the music instantly from YouTube—sound familiar?—thereby dancing around copyright issues, and cataloguing tunes using data from another music site, Last.fm. Like YouTube Instant, the genius of ­Instant.fm isn’t its component parts but their clever combination.

When I point out to Feross that the whole playlist thing has been done before, he smiles. “Ideas are a dime a dozen,” he says. “Execution is what matters.”

I hear this over and over. “I’m not a big believer in ideas,” says Panagiotis Ipeirotis, a professor at the Stern School of Business at New York University who’s also a programmer. “I believe in execution much more than the idea.” “It’s one thing to have a great idea,” says Dan Thompson. “It’s another to be an engineer and to have a great idea.” For example, Thompson was recently working on an app for the iPhone called Recollect, which would use Bluetooth to help people find stuff they tend to lose, like their keys. Thompson wasn’t the first to have the idea—Phone Halo and Cobra Tag already exist. The difference, Thompson says, is he’ll do it better.

Even the great American fable of our time, The Social Network, is about the triumph of execution over concept. “If you guys were the inventors of Facebook, you’d have invented Facebook,” Jesse Eisenberg’s Mark Zuckerberg tells the Winklevoss twins in the film.

The idea that there are no new ideas is not a new idea. But as the barriers to entry for tech entrepreneurs drop, it’s truer than ever. If ten people have the same idea for a piece of software, the one who succeeds will be whoever launches it fastest, then iterates as quickly as possible. “You have to be embarrassed by your first product,” says Lee, who helped create Geomon. “If you’re not embarrassed, you’re taking too long to get it out there.” As Feross puts it, “Done is better than perfect.”

This launch-now-iterate-later approach puts power in the hands of the doers rather than the thinkers. Business students with an idea but no tech background are often left twisting. The listserv of the Business Association of Stanford Entrepreneurial Students gets a dozen e-mails every week from someone looking to hire a programmer. Very few hear back. “If you don’t have the skills, I don’t know what to say to you,” says Geoff Woo, the group’s former president.

Ideas still matter. But the ability to tweak and hone and hack and crack—to move fast and break things—is essential to the idea-formation process. Intellectual turnover is therefore high. “Fail fast,” the mantra of tech entrepreneurship, means everyone moves onto the next idea sooner, which in turn means there are a lot of ideas out there—many of them bad.

There’s a profound disdain among programmers for what one student called “Facebook for your pet”—that is, a lazy rip-off that doesn’t address a real need. Huang calls it the “cat versus zebra” phenomenon. Back in March, it was announced that a company called Color raised $41 million in venture capital from Sequoia Capital and Bain Capital Ventures. The idea sounds cool: It’s an iPhone app that lets you share pictures with everyone within 100 feet, which could eventually create a truly massive photo database. But, everyone wondered, is it really $41 million cool? Over-the-top rhetoric from investors—one said it could “change every­thing”—deepened the skepticism, and Color became shorthand for “Bubble 2.0.”

The Color buy fueled the sense that the race is on to create the next Facebook or to convince investors that you will. Yet among the companies I saw, many of them were pushing back against Facebook or at least trying to rein in its influence. For example, Path, a photo- and video-sharing service based in San Francisco, limits users to 150 friends. Most people don’t have meaningful interactions with more than 150 people in the world, Path figures. Why not focus your sharing on them? Path isn’t “pushing back” against Facebook, its founders say. (That, despite the fact that Path co-founder Dave Morin came from Facebook, where he co-invented Facebook Connect.) It’s just the next step in the evolution of social networking.

Some engineers are trying to throw off the shackles of Facebook altogether. “Facebook is now the issuer of driver’s licenses on the Internet,” says Michael Fischer, a Stanford Ph.D. student who bridles at the site’s power. He’s building a program called DeepSocial, which is basically a LIKE button with more control. Instead of sharing your likes with all your Facebook friends, it shares that information with a select group of people. And instead of tracking your likes centrally, as on Facebook, it communicates the information through your e-mail account, decentralizing the data. Of course, the odds that the next Facebook wins by being the anti-Facebook are pretty much zero. The whole point of paradigm shifts is you can’t see them coming.

One Saturday night, Feross and a few friends gathered in an unlocked lecture hall and screened the anime feature Howl’s Moving Castle. “In high school, I saw kids who watched anime, and I was like, I don’t want to be one of those kids,” says Feross. “They’d bring swords to school and fight each other during study break.” But recently he’s gotten into the genre, particularly the films of Hayao Miyazaki. He admires ­Miyazaki’s process: Instead of plotting everything out in advance, the director has to see what the animation looks like before he can finish the story. “Playing it by ear and just going with the flow—it’s kind of like what you do when you’re hacking on stuff,” says Feross.

That also describes Feross’s recent approach to life. This summer, he didn’t go to work at Facebook, or Google, or ­YouTube. He had an internship at Quora, a question-and-answer site started by former Facebook employees. He figured it would be useful to see how a start-up ­operates, in case he ever wants to launch his own.

Ever since YouTube Instant, Feross says, he’s felt the urge to create more—to wield his powers, in anime-speak. “It was a big wake-up call,” he says. “I now realize I have the power to make something people will use. I had the power all along.

“Successful people aren’t any different from you and me,” he continues. “They’re not inherently more brilliant. The difference is they had the wisdom to get their hands dirty and be part of the game instead of just observing it.”

Failure is an option—and it’s not even a bad one. “I’m not as afraid of uncertainty,” says Feross. “I don’t think Mark Zuckerberg knew what he was doing when he built Facebook.”

People like Zuckerberg and Bill Gates and Steve Jobs surrendered to the process and changed the world before they really knew what they were doing. It can happen. A lot of kids in Silicon Valley are counting on it.

Bubble Boys