The Master Switch

by Tim Wu

  Apple’s origins were pure Steve Wozniak, but as everyone knows, it was the other founder, Steve Jobs, whose ideas made Apple what it is today. Jobs maintained the early image that he and Wozniak created, but beginning with the Macintosh in the 1980s, and accelerating through the age of the iPod, iPhone, and iPad, he led Apple computers on a fundamentally different track.

  Jobs is a man who would seem as much at home in Victorian England as behind the counter of a sushi bar: he is an apostle of perfectibility and believes in a single best way of performing any task and presenting the results. As one might expect, his ideas embody an aesthetic philosophy as much as a sense of functionality, which is why Apple’s products look so good while working so well. But those ideas have also long been at odds with the principles of the early computing industry, of the Apple II and of the Internet, sometimes to the detriment of Apple itself.

  As Wozniak told me, the Macintosh, launched in 1984, marked a departure from many of his ideas as realized in the Apple II. To be sure, the Macintosh was radically innovative in its own right, being the first important mass-produced computer to feature a “mouse” and a “desktop”—ideas born in the mind of Douglas Engelbart in the 1950s, ideas that had persisted without fructifying in computer science labs ever since.* Nevertheless the Mac represented an unconditional surrender of Wozniak’s openness, as was obvious from the first glance: gone was the concept of the hood. You could no longer easily open the computer and get at its innards. Generally, only Apple stuff, or stuff that Apple approved, could run on it (as software) or plug into it (as peripherals). Apple now refused to license its operating system, meaning that a company like Dell couldn’t make a Mac-compatible computer. If you wanted a laser printer, software, or virtually any accessory, it was to Apple you had to turn. Apple thus became the final arbiter over what the Macintosh was and was not, rather in the way that AT&T at one time had sole discretion over what could and what could not connect to the telephone network.

  Thus via the Mac, Apple was at once an innovative and a completely retrograde company. Jobs had elected the design principles that had governed the Hollywood studios, Theodore Vail’s AT&T, indeed anyone who ever dreamed of a perfect system. He created an integrated product, installing himself as its prime mover. If the good of getting everything to work together smoothly—perfectly—meant a little less freedom of use, so be it. Likewise, if it required a certain restraint to create and market it, that was fine. Leander Kahney, author of Inside Steve’s Brain, describes Jobs’s modus operandi as one of “unrelenting control over his employees, his image, and even his customers” with the goal of achieving “unrelenting control over his products and how they’re used.”9

  By the time the Macintosh became Apple’s lead product, Wozniak had lost whatever power he had once held over Apple’s institutional ideology and product design. One salient reason had nothing to do with business or philosophy. In 1981 he crashed his Beechcraft Bonanza on takeoff from Scotts Valley, just outside the San Francisco Bay Area. Brain damage resulted in pronounced though temporary cognitive impairment, including retrograde amnesia. He would take a leave of absence, but his return would not alter the outcome of a quiet power struggle that had been building since before the accident. Its resolution would permanently sideline “the other Steve,” leaving the far more ambitious Jobs and his ideas ascendant.

  Like all centralized systems, Jobs’s has its merits: one can easily criticize its principles yet love its products. Computers, it turns out, can indeed benefit in some ways from a centralizing will to perfection, no less than French cuisine, a German automobile, or any number of other elevated aesthetic experiences that depend on strict control of process and the consumer. Respecting functionality, too, Jobs has reason to crow. Since the original Macintosh, his company’s designs have more often than not worked better, as well as more agreeably, than anything offered by the competition.

  But the drawbacks have been obvious, too, not just for the consumer but for Apple. For even if Jobs made beautiful machines, his decision to close the Macintosh contributed significantly to making Bill Gates the richest man on earth. No one would say it was the only reason, but Apple’s long-standing adherence to closed design left the door wide open for the Microsoft Corporation and the many clones of the IBM PC to conquer computing with hardware and software whose chief virtue was combining the best features of the Mac and the Apple II. Even if Windows was never as advanced or well designed as Apple’s operating system, it enjoyed one insuperable advantage: it worked on any computer, supported just about every type of software, and could interface with any printer, modem, or whatever other hardware one could design. After it was launched in the late eighties, early-nineties Windows ran off with the market Apple had pioneered, based mostly on ideas that had been Apple’s to begin with.

  The victory of PCs and Windows over Apple was viewed by many as the defining parable of the decade; its moral was “open beats closed.” It suggested that Wozniak had been right from the beginning. But by then Steve Jobs had been gone for years, having been forced out of Apple in 1985 in a boardroom coup. Yet even in his absence Jobs would never agree about the superiority of openness, maintaining all the while that closed had simply not yet been perfected. A decade after his expulsion, back at the helm of the company he founded, Steve Jobs would try yet again to prove he had been the true prophet.

  JUST WHAT IS GOOGLE?

  In 1902, the New York Telephone Company opened the world’s first school for “telephone girls.” It was an exclusive institution of sorts. As the historian H. N. Casson described the qualifications for admission in 1910: “Every girl shall be in good health, quick-handed, clear-voiced, and with a certain poise and alertness of manner.” There were almost seventeen thousand applicants every year for the school’s two thousand places.10

  Acquiring this credential was scarcely the hardest part of being a telephone girl. According to a 1912 New York Times story, 75 percent were fired after six months for “mental inefficiency.” The job also required great manual dexterity to connect dozens of callers per minute to their desired parties. During the 1907 financial panic in New York, one exchange put through fifteen thousand phone calls in the space of an hour. “A few girls lost their heads. One fainted and was carried to the rest-room.”11

  People often wonder, “What exactly is Google?” Here is a simple answer: Like its harbinger the telephone girl, Google offers a fast, accurate, and polite way to reach your party. In other words, Google is the Internet’s switch. In fact, it’s the world’s most popular Internet switch, and as such, it might even be described as the current custodian of the Master Switch.12

  Every network needs a way to connect the parties who use it. In the early days of the telephone, before direct dial, you’d ask the telephone girl for your party by name (“Connect me with Ford Motors, please”). Later on, you’d directly dial the phone number, from either memory or the telephone directory, which seems rather a decline in service. Today, Google upholds the earlier standard, but on the Internet. Needing no address, you ask for your party by name (typing in “Ford Motor Company,” for instance), and Google shows you the way to connect with them over the World Wide Web.

  The comparison with Bell’s telephone switchboard girls might sound a little anticlimactic to describe a firm with ambitions as grand as Google’s, but this reaction betrays a lack of awareness of the lofty import the switch has in the information world. For it is the switch that transforms mere communications into networking—that ultimately decides who reaches what or whom. And at the superintending level, which most networks eventually do develop at some point, it is the Master Switch, as Fred Friendly reminds us, that will decide who is to be heard. However many good things the Internet has to offer—services, information resources, retail outlets—it hardly matters if you can’t get to them.

  There are, of course, some differences between Google and the switch monopolists of yesteryear, including the firm that is arguably its truest forerunner
, Vail’s AT&T. For one thing, Google is not a switch of necessity, such as the telephone company was, but rather a switch of choice. This is a somewhat technical point, but suffice it to say that Google’s search engine is not the only Internet switch. There are other means by which to reach people or places on the Internet, as well as other points that might be described as “switches,” like the physical routers that direct the flow of Internet traffic on the data packet level. There are plenty of ways around Google: you can use domain names to navigate the Internet, or use one of Google’s competitors (Yahoo!, Bing, and the like), or for the truly hard-core, simply remember the IP addresses (e.g., 98.130.232.209), the way people once used to remember phone numbers. In fact, unlike AT&T, Google could be replaced at any time. And yet if by 2010 Google wasn’t the only game in town, it was clearly the most popular Internet switch; by its market share of the search business (over 65 percent) it clearly qualifies as a monopoly.

  In some ways, Google nevertheless enjoys a much broader control over switching than the old AT&T ever did. For it is not just the way that people reach one another to talk, but the way most people find all forms of information, across all media platforms, at a time when information is a far more prominent commodity in our national life and economy than it has ever been. Siva Vaidhyanathan’s aptly titled book Googlization of Everything points out how much power this gives Google over global culture. As he writes, “for a growing (but not universal) portion of the Web and the world, we allow Google to determine what is important, relevant, and true.”13 As he suggests—and how many would disagree?—whatever shows up on the first page of a Google search is what matters in forming our sense of any reality; the rest doesn’t.

  To understand this unusual level of consumer preference and trust in a market with other real choices is to understand the source of Google’s singular power. But is this a stroke of cold luck, or is Google something special? Quite enough has been written about Google’s corporate culture, whether one looks to the cafeterias that serve free food, the beach volleyball, or the fact that its engineers like to attend the Burning Man festival in the Nevada desert.14 Not that such things aren’t useful inducements to productivity or the exception in corporate America, but they are more nearly adaptations of a general Silicon Valley corporate ethos than one particular to Google, a point the company readily admits.

  Boiled down, the Google difference amounts to two qualities, rather than any metaphysical uniqueness. The first, as we’ve already remarked, is its highly specialized control of the Internet switch. We shall describe the nature of that specialization in more detail presently, but for now let us say it accords Google a dominance in search befitting an engine whose name has become a verb and synonymous with function. (No one Apples or AT&Ts a potential new boyfriend.) While the firm does have dozens of other projects, it is obvious (certainly from their individual direct contributions to cash flow, which are minuscule) that most, including the maps, the lavishly capacious Gmail accounts, even the hugely popular YouTube, are ultimately trial balloons, experiments of a kind, or a way of enhancing the primacy of the core business of search, whether by creating complementary information resources or simply engendering the goodwill that comes of offering cool stuff for free. The second Google difference is in its corporate structure. The firm, while having as many ventures as it has engineers, eschews vertical integration of these efforts to a degree virtually unprecedented for a communications giant. This structural distinction may be hard to grasp, so let’s explain it more carefully.

  • • •

  A medieval architect looking at the skyline of New York City or Hong Kong would be astonished that the buildings manage to stand without flying buttresses, thick walls, or other visible supports. A nineteenth- or twentieth-century industrialist would feel much the same bewildered awe regarding a major Internet firm like Google. You might call Google a media company, but it doesn’t own content. It is a communications company, but it doesn’t own the wires or airwaves over which packets reach people. You might accept my characterization that Google is simply the switch, but a switch alone has never before constituted a freestanding company—what basis is that for value? Many credit Google with ambitions to take over the world, but an industrial theorist might well ask how such a radically disintegrated firm could long endure, let alone achieve global domination.

  Compared with other giants, like Time Warner circa 2000, or Paramount Pictures circa 1927, or AT&T’s original and resurrected incarnations, Google is underintegrated and undefended. The business rests on a set of ideas—or more precisely, a set of open protocols designed by government researchers. But that is the point: it is the structure of the Internet, much less than anything particular to the firm itself, that keeps Google standing. It traffics in content originated by billions of people, none of them on salary, who build the websites or make the home videos. It reaches its customers on wires and over airwaves owned by other firms. This may seem an improbably shaky foundation to build a firm on, but perhaps that is the genius of it.

  If that seems a bit abstract, it is well to remember that Google is an unusually academic company in origins and sensibility. Larry Page, one of the two founders, described his personal ambitions this way: “I decided I was either going to be a professor or start a company.” Just as Columbia University effectively financed FM radio in the 1930s, Stanford got Google started. With its original Web address http://google.stanford.edu/, the operation relied on university hardware and software and the efforts of graduate students. “At one point,” as John Battelle writes in The Search, the early Google “consumed nearly half of Stanford’s entire network bandwidth.”15

  Google’s corporate design remains both its greatest strength and its most serious vulnerability. It is what makes the firm so remarkably well adapted to the Internet environment, as a native species, so to speak. Unlike AOL, Google never tried to resist or subdue the Internet’s essential structure. It is a creature perfectly suited to the network as its framers intended it. In this sense, it is the antithesis of AOL.

  Google’s chief advantage, as we have suggested, can be summarized in a single word: specialization. Companies like AT&T or the big entertainment conglomerates succeed by being big and integrated—doing everything, and owning everything. A company like Google, in contrast, succeeds by doing one (well-chosen) thing, but doing it better than anyone else. It’s the trait that makes Google the hedgehog to so many others’ fox. The firm harvests the best of the Internet, organizing the worldwide chaos in a useful way, and asks its users to navigate this order via their own connections; by relying on the sweat of others for content and carriage, Google can focus on its central mission: search. From its founding, the firm was dedicated to performing that function with clear superiority; it famously pioneered an algorithm called PageRank, which arranged search hits by importance rather than sheer numerical incidence, thereby making search more intelligent. The company resolved to stand or fall on the strength of that competitive edge. As Google’s CEO, Eric Schmidt, explained to me once, firms like the old AT&T or Western Union “had to build the entire supply chain. We are specialized. We understand that infrastructure is not the same thing as content. And we do infrastructure better than anyone else.”

  Google, between content and transport

  Unlike AOL Time Warner, Google doesn’t need to try to steer users anywhere in particular. They need only focus their resources on helping you get wherever you want to go, whether you know where that is or not. Needless to say, it is a great plus not to be involved in trying to persuade anyone to consume, say, Warner Bros. content. Such was the inherently corrupting project of AOL when Steve Case joined his company to Time Warner. Case had assumed that any Internet company would need control of both wires and content to succeed in the 2000s. He was wrong.

  That’s the advantage. On the other hand, Google’s lack of vertical integration leaves it vulnerable, rather like a medieval city without a wall.* He who controls the wires or airwaves
can potentially destroy Google, for it is only via these means that Google reaches its customers. To use the search engine and other utilities, you need Internet access, not a service Google now provides (with trivial exceptions). To have such access, you need to pay an Internet Service Provider—typically your telephone or cable company. Meanwhile, Google itself must also pay for Internet service, a fact that, conceptually at least, puts the firm and its customers on an equal footing: both are subscription users of the Internet. And so whoever controls those connection services can potentially block Google—or any other site or content, as well as the individual user, for that matter.

  Nor is this matter of infrastructure the firm’s only weakness. A concerted boycott among content owners—website operators or other sources—could achieve the same choking effect. Under long-established protocols, any website can tell Google that it doesn’t want to be indexed.† In theory, Wikipedia, The New York Times, CNN, and dozens of other websites could begin telling Google, “Thanks, but no thanks,” or conceivably strike an exclusive deal with one of Google’s rivals.

  How Google reaches customers

  Both of these vulnerabilities are a direct consequence of Google’s corporate design, of the fact it owns no connections and no content. As we shall see, the firm’s most determined enemies have begun to understand and exploit these frailties.

  In Chicago in 2005, AT&T’s Ed Whitacre took a break during a typical day of empire building to grant BusinessWeek’s Roger Crockett an interview.16 In the midst of his campaign to reunify the Bell company, the CEO was refreshingly clear about his strategy. “It’s about scale and scope,” he told Crockett a few times, “scale and scope.”

  Crockett asked, “How concerned are you about Internet upstarts like Google, MSN, Vonage, and others?”