by Steve Levy
The essential spec of a new browser was high speed. “Larry and Sergey wrote an OKR saying we should make the web as fast as flipping through a magazine,” says Pichai. “If things could be instant and there’s just no latency at all, the sky is the limit. I mean, we haven’t even scratched the surface.”
In June 2006, the former Mozillans created a small prototype. Though Brin and Page had yet to give an explicit go-ahead, it was clear that they were quietly rooting the effort on. Schmidt no longer opposed the browser idea. But if Google did undertake the project, the CEO said, the result would have to be something that differed significantly from other browsers. In addition, it would have to be fast, it would have to be open source, it would have to be secure. The Executive Committee green-lighted the project to begin in earnest.
“I remember one Friday, there was a meeting called with, like, an hour’s notice,” says one engineer. “It was kind of mysterious. And we were told, ‘The management is thinking about doing our own browser—what do you think about that?’ It was a crazy question, and everybody was a combination of excited and freaked out.” The engineers knew that building a competitive browser was a massive undertaking. There were also mixed feelings due to the group’s strong attachment to both the technology and vision behind Firefox, an icon of open-source development and a hedge against Microsoft’s dominance of the browser market. Particularly for the Googlers who had come from Mozilla, this was a case of digital fratricide. “The fear was that people were going to read this as sabotaging Firefox,” says engineer Erik Kay, who joined the team in October 2006. That would be evil.
The Googlers were eventually mollified by the assurance that their browser would be 100 percent open source. With the open-source system, Google’s code would be publicly available, and if people wanted to use it to create variations, that was fine. It was even possible that Google’s innovations could find their way into the Mozilla code base.
It was Pichai’s unhappy duty to break the news to Mitchell Baker, the chairman of the Mozilla Foundation. Baker was tough; trained as an attorney, she could passionately argue the open-source cause. She also cut a memorable figure with her asymmetrical punkish coiffure. She would have fit right in at Google, except for the fact that she was wary of all commercial enterprises. And as a veteran of the browser wars, she knew that every point of market share was as toughly contested as a football goal-line stand.
Since a Google browser had been rumored for months, Baker wasn’t shocked. But Google had been a partner and benefactor of Mozilla. Now it was her competitor.
Was it a betrayal? Evil? With the pride of a jilted lover, Baker would later shrug it off. “To be betrayed, you have to expect something different,” she says. “I expect Google to pursue its economic interests. I have never had illusions. We’re not a toy. Google doesn’t control us.” But she could not mask the bitterness when, after the Google browser came out, Google began promoting it with AdWords delivered to people who searched using the keyword “Mozilla.” “They’re actively trying to take people away from Firefox,” she complained.
After the usual flurry of crazy alternatives for a code name, the team decided to call its browser Google Chrome. The moniker came from the term used to describe the frame, toolbars, menus, and other graphic elements that border a browser window. In a way, the name was counterintuitive, because Google wanted to strip off a lot of the decorative chrome seen in other browsers and create a sleek sports car of a browser. The idea was to make the interface so minimal that people wouldn’t feel they were using a browser at all but interacting directly with the pages and web apps. An unofficial motto became “Content not chrome,” a bit bizarre considering the product’s name. “We’ve learned to live with the irony,” said engineer Mark Larson during the development process.
Page and Brin wanted Chrome optimized to run web applications—fast. When you run a program faster by an order of magnitude, you haven’t made something better—you’ve made something new. The crucial element in speeding up a browser was a component called a JavaScript engine, a “virtual machine” that ran web application code. In previous browsers, JavaScript didn’t run quickly enough to make web applications seem as nimble as desktop apps; Google felt that if it changed that, people would use the web more and thus use Google’s services and ads more. Google hoped to kick-start a new generation of web-based applications that would make Microsoft’s worst nightmare a reality: the browser would become the equivalent of an operating system.
There was an ideal person to supercharge the virtual machine, a Danish computer scientist named Lars Bak, whose virtuosity in virtuality had established him as the master in the field. But after more than twenty years of nonstop labor designing virtual machines, the forty-five-year-old Bak had returned to his native country and had been planning to take some time off to work on his farm outside Aarhus. When he got Google’s call in September 2006, however, the opportunity was too tempting to resist. Bak set up a small team that originally worked from his farm, then moved to some offices at the local university. He understood that his mission was to provide an engine faster than any previous browser. He called his team’s part of the project “V8.” “We decided we wanted to speed up JavaScript by a factor of ten, and we gave ourselves four months to do it,” he says. A typical day for the group would begin between 7 and 8 a.m.; they’d program constantly until six or seven at night, when they’d call Mountain View to debrief. The only break was for lunch, when they would wolf down food in five minutes and spend twenty minutes at the game console. “We are pretty damn good at Wii Tennis,” Bak noted.
They were also pretty good at writing a JavaScript engine. As the project progressed, Bak’s benchmarks showed that V8 was running JavaScript ten times as fast as Firefox. And how did it compare in those same benchmarks to the market-share leader, Microsoft’s IE 7? Fifty-six times as fast. “We sort of underestimated what we could do,” Bak says.
Sundar Pichai and his team had an OKR of 20 million users by the end of the year. “It was a very aggressive OKR,” he says. “A classic.” He didn’t make it. “We got there, but not in the time frame we had in mind.” A lot of people downloaded Chrome in the early weeks and found that it didn’t work. Because the online behavior of Googlers was not typical of the general public, there had been many websites and apps that went untested. “We had five thousand internal users, but not one noticed that Hotmail didn’t work,” says one engineer. But after Chrome shipped to the public, Hotmail users instantly found that it wouldn’t run their mail—and deleted Google’s browser.
Also the Macintosh version was months late, even though an early Mac version was in the plans all along. In fact, after Steve Jobs’s keynote presentation in January 2008, when Apple’s CEO introduced a slim new computer called the MacBook Air, Sergey Brin gave Pichai one of the first units and said, “I want Chrome running on a Mac.” The Mac version didn’t ship until late 2009.
But Chrome’s numbers grew, to over 120 million by the end of 2010. What’s more, every one of Chrome’s competitors made it a point to speed up their browsers. That was exactly what Google wanted: browsers that provided a better experience for people to run applications on the web.
In fact, Google began to believe that people already had reached the point—with web apps such as Google Docs and all the myriad services on the web—that there was almost nothing you couldn’t do with a browser. Pichai gave a netbook to his father and noticed that once his dad opened Chrome, he never opened another application. He came to think that the word “application” didn’t apply to a browser—it was more like a gateway to everything in the world that really mattered, the stuff in the cloud. “It was very clear to us a lot of people were buying these devices with the goal of spending their entire day in the browser. So we all started talking about a natural course: designing an end-to-end experience around the browser. Think about it.”
In fact, the thought had already occurred to the team. “We didn’t want to use the OS word,
but Chrome was always thought of as an operating system for web applications,” says Linus Upson. But once Chrome was launched, the team began thinking of it literally that way, building it so that if you bought a netbook—or eventually any other kind of computer—there would be no Windows or Linux operating system, just Chrome.
“From eight p.m. onwards is when you have really interesting conversations,” says Caesar Sengupta, an engineer on the team. “We started challenging ourselves to think about how we would build an operating system.” They got Upson and Pichai on board and began ticking off what a Chrome operating system should be: blisteringly fast, totally free of malware. “It should just feel like the web,” says Sengupta. They put together a proposal that they took to a meeting with Larry and Sergey in October 2008. Since Brin and Page had been wanting to do an OS for ten years, they instantly embraced the idea. “I’m all for it,” said Page.
As the Chrome team brainstormed its operating system, they realized that there was a chance to redefine computing itself, in terms of the cloud. As web applications got better, they figured, why have any client applications? In fact, why not jettison the entire concept of storing a file and running a program with it? It was a startling concept, as few thought that cloud-based computing was far enough along to replace the current paradigm. Privacy advocates might worry about the security of cloud-based data—but Google believed that it had proved its trustworthiness with Gmail. IT experts might worry about what happens if web services had outages. But Google was confident that its unmatched infrastructure had sufficient power and redundancy to be as dependable as the electricity from a power outlet. In any case, an ambition junkie like Larry Page wasn’t about to argue with such an audacious premise.
Besides, if Chrome OS could move people more rapidly toward cloud computing—or just make computers so easy to use that people used them more—Google’s business would boom. In fact, Upson argued, Google had more at stake in improving computers than did companies that actually make computers. “Google makes money with advertising online, but that’s about 20 percent of total advertising spent,” Upson says. “Eighty percent of the time, people’s attention is offline. To the extent that we can make computers better, everything will go online, and Google can participate in that advertising space. There are four more Googles to be had here. That’s why we have incentives to make computers better. Computer manufacturers want to figure out how to get the most money out of you. We want to make you happy. If we can do it free, so much the better.”
At the time, Google was about to launch a project it had been developing for more than a year, a free cloud-based storage service called GDrive. But Sundar had concluded that it was an artifact of the style of computing that Google was about to usher out the door. He went to Bradley Horowitz, the executive in charge of the project, and said, “I don’t think we need GDrive anymore.” Horowitz asked why not. “Files are so 1990,” said Pichai. “I don’t think we need files anymore.”
Horowitz was stunned. “Not need files anymore?”
“Think about it,” said Pichai. “You just want to get information into the cloud. When people use our Google Docs, there are no more files. You just start editing in the cloud, and there’s never a file.”
When Pichai first proposed this concept to Google’s top executives at a GPS—no files!—the reaction was, he says, “skeptical.” Upson had another characterization: “It was a withering assault.” But eventually they won people over by a logical argument—that it could be done, that it was the cloudlike thing to do, that it was the Google thing to do. That was the end of GDrive: shuttered as a relic of antiquated thinking even before Google released it. The engineers working on it went to the Chrome team.
“We’re taking a fairly radical position,” said Upson. Netbooks running Chrome OS—and Google had already contracted with computer makers to produce them in late 2010—would have no storage. None. You would never install any application. The idea was that you would turn on your computer, and it would boot up instantly (forget about the three-minute wait one must endure with Windows) and connect you to your world, which resided in some cloud somewhere. You wouldn’t have to bother where. And you could enter that world through any computer once you entered your proper passwords. “We will be your IT department,” says Upson. “You never need to worry about software updates, anything like that. We will take care of it all for you.”
Upson and Pichai believed that a wave of new technologies would allow a cloud computer to do everything one did with a desktop machine, only more reliably, more simply, more securely, and much faster. A new protocol called HTML 5 was beginning to roll out, and it enabled web applications to run offline. Google had also been working on a project called Native Client that would allow web-based programs to run as nimbly as those written specifically for a given computer—it would allow even hard-core gamers to get the performance they needed from a web app, something previously unthinkable. And it would all run on the web—you would never install software on your computer again.
That was so startling to the public that the Chrome OS team members often had to repeat it several times before it sank in. The conventions of the desktop were so ingrained that they felt like gravity. But the Googlers had an answer for everything. What about those times—and there were a lot of them—when there was no Internet broadband available? The prototype Chrome OS released to a few thousand testers had a cellular 3G modem built in, as a backup for those times when Wi-Fi wasn’t around. (It wasn’t a great solution, but it was better than nothing.) What about drivers, those little pieces of software that allow your particular computer to connect smoothly with printers and other plug-in devices? “We’re saying no drivers. We’re done,” says Upson. What he meant was—they’re done, they’re the current model of computing.
Google had declared that the cloud was its destiny. And ours.
PART FIVE
OUTSIDE THE BOX
The Google Phone Company and the Google TV Company
1
“They already hate us—what’s the downside?”
You might say that the seeds of the Google Telephone Company took root right after the company moved out of its Palo Alto office to Mountain View in August 1999. The tenant moving into the space Google vacated was a start-up company named Danger.
Danger’s cofounder, Andy Rubin, was a veteran of Apple in the early 1990s and a fabled start-up called General Magic. He’d started Danger to make a mobile communications device called the Sidekick, less a cell phone than a tiny computer—arguably the first smart phone with a measurable IQ. Instant messaging, not phone calls, was the Sidekick’s main purpose; you held it sideways, slid out a keyboard, and began thumb-punching IMs, which appeared in colorful pop-ups on a bright screen. It became popular with teenagers and rap musicians. The Sidekick’s built-in search engine was Google. “The engineers just liked Google,” he says. In 2002, Rubin was demo’ing the Sidekick for a class at Stanford when someone approached him to tell him how cool it was. That was Larry Page.
A couple of years later, Rubin made it a point to visit Google when he was seeking partnerships and funding for his next start-up. (He’d left Danger, and eventually Microsoft bought the company.) Rubin’s new idea was to create an operating system that would power whole families of smart phones—then give the system to the big network carriers (like Verizon or Sprint) for free. This would save the carriers money, since they wouldn’t have to license an operating system from a company like Microsoft or build their own. (Typically a carrier pays 20 percent of the per-phone cost for an operating system.) The system would be written under the rules of open-source software, with the code available to any software authors who wanted to write applications. Rubin’s plan was to make money by selling back-end services to go with the operating system, such as storage, support, and security. It was the familiar model of giving away the razor and making money on the blades.
Rubin, who was a maniacal robot aficionado—he would haunt the Aki
habara district of Tokyo for weird Japanese toys, and build a few of his own—called the company Android. He gathered a team of eight to begin working on a prototype. He had a good contact at the handset manufacturer HTC, which provided him with a top-secret new device that he could use just for demos. After a few months, Android had a working model with a set of slick features such as contacts, email, and a camera. (One nice touch was that Android’s photo software could recognize faces.)
Rubin began pitching carriers in 2004. He also went to the Far East to sell the idea to other handset manufacturers. Even though he was offering something for free, it was a tough sell. The mobile phone world had a profitable business model and was loath to consider disruptive new schemes. He would later vividly recall the trip he had made to Korea—“on my own dime!” he said—to present the concept to Samsung. He and two colleagues found themselves in a huge boardroom. Standing along the wall were about twenty carefully manicured executives in blue suits. (Rubin was in blue jeans.) The division head arrived, and, as if on cue, everyone sat down. Rubin gave his presentation, and the division head rocked with laughter. “You have eight people in your company,” said this executive. “And I have two thousand people working on something that’s not as ambitious.” It wasn’t a compliment.