The Code

by Margaret O'Mara

  So much of the dot-com boom seemed extraordinary, not least in how quickly its irrational exuberance came crashing to the ground in 2001, taking much of the 1990s prosperity with it. Yet this chapter of the silicon age exhibited similar patterns to those that came before, albeit on a supersized scale. Out of university classrooms and laboratories came young men with novel ideas, met by venture capitalists who funded and advised and connected them into the industry’s networks of expertise. Valley lawyers helped young entrepreneurs cross their t’s and dot their i’s. Real estate moguls plowed under the last generation’s office parks to make way for the next. Evangelists both inside and outside the tech world spread the story of Silicon Valley’s magic, boosting a new generation of boy wonders and convincing America’s leadership class that this really might be growth without end.

  And as there always had been, there were the suits: not just the Wall Street bankers, whose pinstriped swarms grew with every upward tick of the NASDAQ, but the politicians and policy wonks, who turned the well-worn footpath between Washington and the Valley into a roaring superhighway.

  It all started with a browser.

  THE WIZARD

  John Doerr liked thinking five years ahead. When the St. Louis native and Rice University graduate was a first-year business student at Harvard in 1974, he became so intrigued by the promise of Silicon Valley venture capital that he began cold-calling firms to see if any would offer him a job. But the market was at its lowest point. Forget about venture, the VCs told him. Go to Intel and work for Andy Grove. While at Intel, Doerr distinguished himself not only for his affability amid Grove’s mercurial, hard-charging environment, but for being one of the first semiconductor guys to get bullish on microcomputers, pushing the company to make and market its own motherboard to hobbyists even before the two Steves had released the Apple I. As a new Kleiner Perkins partner in the early 1980s, as microcomputer mania seized the Valley, he was hanging out at academic conferences learning about the VLSI and 3-D technologies that would make chips faster and turn cheap desktops into powerful workstations.

  So when Doerr saw his first demo of the Mosaic browser at the start of 1994, he already had been paying attention to the Internet for a long time. Eight years earlier, he had taken a leave of absence to spend several months in Tim Wirth’s Senate office, absorbing all he could about the workings of the ARPANET. Since this stint as Capitol Hill’s highest-net-worth summer intern, Doerr had continued to rack up enormous wins in computer hardware and software. All the while, he was increasingly convinced that the Internet would be the next big thing. And that’s what he wanted to nurture. “We think in terms of building new businesses,” he once explained. “In our more grandiose moments, we think of building industries.” In the 1990s, Doerr and his partners did exactly that.3

  Mosaic was their launching pad. If Tim Berners-Lee’s Web client was the Apple II of the Internet, Mosaic was its Macintosh: the portal that opened up the online world for millions. Created by a group of graduate students at the University of Illinois supercomputer center who were tired of the text-only HTML environment of the earliest Web, Mosaic turned the Internet into an immersive, colorful, point-and-click experience. Within months of its 1993 release, the new browser had seized the imagination of Silicon Valley insiders. Its chief student inventor, Marc Andreessen, headed west to capitalize on the excitement. Andreessen found a first landing spot in the scrubby cubicles of EIT, where Marty Tenenbaum was busily building his e-commerce platform and knew that a good, graphical browser was essential to bringing customers online. But within months, the twenty-three-year-old Andreessen had been lured away from Tenenbaum’s shoestring operation by a plump paycheck waved by Jim Clark, a former Stanford computer scientist who had founded Silicon Graphics.

  Clark had been one of the people teaching John Doerr about the future in the early 1980s. Now, he had become annoyed with his company’s cautious corporate direction (and CEO Ed McCracken’s mooning around at the White House) and was casting around for something new to do. Andreessen was it. After signing on the dotted line, the two new partners promptly flew back to Champaign-Urbana and signed up the entire Mosaic engineering team. Weeks after that, Clark tendered his resignation at SGI and began pitching his old friend Doerr for help with his new venture. Kleiner Perkins put in $5 million, and Mosaic Communications was born.4

  The rapid rise of the company that soon was renamed Netscape showed the tightly knit, decades-in-the-making Silicon Valley network in prime form. The leadership Doerr assembled for Netscape was a seasoned mix from tech and other venture-backed new industries, all grade-A men for a grade-A idea. Doerr recruited AT&T Wireless CEO Jim Barksdale, formerly of FedEx; his chosen VP, Mike Homer, was a veteran of Apple. Wilson Sonsini provided outside counsel. Kleiner Perkins itself was bursting at the seams in those days with high-tech venture partners: Sun’s Vinod Khosla, Fairchild and Apple’s Floyd Kvamme, and Regis McKenna, who at Doerr’s urging had joined in 1986 as a general partner. Even Netscape’s real estate had history: its first offices were sublet from HP, and it soon scaled up to grand headquarters on the newly-remediated site of Fairchild Semiconductor’s Mountain View manufacturing plant. The Wagon Wheel was just around the corner.5

  There was no better product around with which to create an entirely new industry than Navigator, the browser Netscape released in early 1995. Built to run on any type of computer—Mac, PC, Unix workstation—Navigator decisively left the desktop wars of the 1980s behind. Designed to run on high-speed broadband networks (like the NSF-funded one Andreessen had enjoyed in his U of I lab) and secure enough to host financial transactions, Navigator anticipated a dial-up–free future where nearly anything could be bought with a single click. The company further disrupted the usual order by giving its browser away for free. It was a startlingly efficient way to build a user base. One year into its existence, Netscape had six million customers.6

  Netscape was just one star in the growing Internet constellation financed by John Doerr and Kleiner Perkins during those heady days of the mid-1990s. The firm put money in scrappy Northern Virginia–based America Online (AOL), the successor company to The Source, which was elbowing into CompuServe’s and Prodigy’s market share by inundating American households with free CD-ROMs pitching its wares. And KP incubated entirely new companies, like @Home, which promised to bring high-speed broadband into homes across America. (The venture later became part of one of the dot-com era’s most celebrated failures. At the height of the late-’90s froth, @Home merged with the Internet portal Excite in a deal valued at $7.2 billion; two years later, the stock had lost 90 percent of its value and the company was bankrupt.)

  And an earlier generation of Kleiner Perkins hits helped the online world build out and scale up. In March 1995, making good on its promises that “the network is the computer,” Sun Microsystems introduced Java, the first programming language written with the Web in mind. Java product manager Kim Polese sensed that the new language was going to be a big deal after her team pitched a story about the launch of what they then called “Hot Java” to Silicon Valley’s paper of record, the San Jose Mercury News. When the big day came, Polese frantically flipped through the business pages to see if they’d picked up the story. Not finding it, she cast down the paper in frustration. Then she saw the front page. There it was: “Why Sun Thinks Hot Java Will Give You a Lift.”7

  Java wasn’t just a lift; it was a game changer. It brought the Web alive, allowing programmers to build applications with graphics and animations as robust as any found on current desktop software. And Java ran through a browser, not an operating system. The Web was not going to be balkanized like the PC platform. Programmers could build applications on Java and have them run on any kind of machine. Websites went from static and clunky to animated and agile. As Sun CTO Eric Schmidt put it, “It’s the difference between a telegraph and a telephone.” Java made Sun a jazzy Internet-era brand and turned Polese into a Web-era superstar (s
he soon left to start Web multimedia company Marimba). By 1996, Java so ruled the world that KP opened a $100 million fund solely to invest in companies building Java-based software.8

  Doerr, who had emerged from the 1980s chip wars as a faithful disciple of Japanese management principles, called all of this “the Kleiner keiretsu.” It was a collection of different firms sharing resources and expertise, operating symbiotically to create an entire market ecosystem—from the networks and routers to the browsers and portals to the programming languages and software applications. It was how Japan’s electronics consortia had ruled the world in the 1970s and 1980s. But it also fairly characterized the web of personal relationships and specialized expertise and tacit knowledge that made Silicon Valley soar. Other Valley VCs, both veteran and new, had giant Internet wins, of course. But Kleiner Perkins came to symbolize the dot-com generation more than any other. The firm was the MITI of the Internet era, and John Doerr was the wizard behind the curtain.9

  FARM TO FACTORY

  Gordon Moore liked to observe that the most important thing Fred Terman’s university did for Silicon Valley was to graduate 800 masters and PhD students per year, replenishing the region’s intellectual pool. The other Moore’s Law certainly held true in the case of the young people who came out of Stanford’s computer science and engineering programs in the 1990s. Four decades of innovation in AI, software design, and networking had turned The Farm into the home for many of the world’s best computer scientists, and a magnet for many of the very best graduate students. Stanford’s encouragement of tech transfer meant that the very best of them already had well-formed dreams of commercializing whatever they built. If you wanted to study computing, you went somewhere else. If you wanted to study computing and found a company, you came to Stanford.10

  The grad student dubbed “Doogie” became the Internet era’s first brilliant case in point. After steaming through his first two Stanford degrees, Jerry Yang had found himself stalled as he worked on his doctorate. When he entered the program, he’d opted for computer-aided design as a focus, which in 1990 seemed like a great bet at a moment when Silicon Graphics was becoming the darling of the Valley. A few years later, the landscape looked a little different. Thanks to being at Stanford, he and his closest friend in the PhD program, David Filo, had cottoned on early to the wonders of Mosaic. When their advisor went on sabbatical, the two hunkered down in a pizza-box-strewn trailer at the fringe of the engineering quad and started spending all their time building websites: a homepage for a fantasy basketball league, a tribute to sumo wrestling.

  The two spent so much time Web surfing that Filo put together a list of his favorite sites so that he could better navigate the swelling sea of online information. Yang—whose undergraduate work-study job shelving books in the Stanford library had introduced him to the joys of the Dewey decimal system—realized they could go one better by turning the list into HTML and putting it online for others to enjoy. The attention-shy Filo didn’t really want his name on the thing (they were supposed to be writing dissertations, after all), so Yang simply dubbed it “Jerry’s Guide to the World Wide Web.” It was late 1993. Within a matter of months, the site had ballooned into a real business, as more people downloaded Mosaic, piled online, and tried to navigate through the Internet’s dizzying bounty.11

  By the start of 1995, the guide had gotten a million hits, Web-related start-ups were brewing across the Valley, and the two grad students realized that their time playing hooky in that trailer had generated a serious business opportunity. Yang and Filo left Stanford, taking with them a number of their campus friends, who now became full-time “Web surfers”—human beings who read the web and attempted to thoughtfully sort its contents into various categories. In contrast to the overwhelmingly male engineering teams that populated the new generation of start-ups, a good number of these early surfers were women.12

  Now courted by the Valley’s leading investors, the grad-school dropouts opted for $3 million from journalist-turned-VC Mike Moritz (now at Don Valentine’s Sequoia Capital) and named their new venture Yahoo! “Their greatest strength,” elder statesman Valentine observed sensibly about Yang and Filo, “was the recognition of their weaknesses and their lack of experience.” Once again, the Silicon Valley mentorship network, new and old, got to work. Moritz and Valentine connected Yahoo! to lawyers, PR experts, and Internet providers. They found a seasoned CEO, Tim Koogle, a Stanford graduate who’d been working in tech companies big and small for two decades. Marc Andreessen offered Yang and Filo server space at Netscape so they could move off Stanford’s overburdened network.13

  The corporate exclamation point and goofball job titles—Yang was “Chief Yahoo,” and it got more fanciful from there—weren’t the only ways in which the operation reflected new trends in the Internet-age Valley. Yahoo! ran on programming smarts, but it wasn’t software or hardware that could be sold at a retail store or bundled into a desktop. Its product was content: a distinctive classification system for the Web’s vast information landscape, a job first performed by the human Web surfers, then later—as Yahoo! scaled up to undreamed-of heights—by an algorithm. Intel sold microchips; Microsoft sold packaged software. Yahoo! and its dot-com brethren gave away their product for free. The only way to make money was through advertising. It was a profoundly new model for Silicon Valley.

  After four decades of perfecting its art, the Valley’s business ecosystem knew exactly how to nurture funny little companies that built chips or wrote code. It had little inkling of what might happen when those companies became information platforms.

  THE COMMERCIAL SUPERHIGHWAY

  The viral spread of the Netscape browser and the ease with which Yahoo! opened up the Web to surfing brought more and more people onto the Internet by the middle of the 1990s. They were browsing, searching, chatting, and e-mailing. But they were only doing a minimal amount of buying and selling. The act that later became a mundane part of Americans’ everyday consumer life—go to a site, point and click, enter your credit card number, and hit “buy”—was an alien, anxiety-inducing idea in the early dot-com era. There was no guarantee that a credit card number wouldn’t get stolen. There wasn’t an obvious way to pay for shipping, or even to choose how quickly you’d receive your purchase. The solution arrived in the Valley via an under-the-radar federal grant, clocking in at a relatively modest $2.5 million, that at last answered Marty Tenenbaum’s dreams of making the Internet into a bustling marketplace.

  As often happens in politics, the money from the U.S. Commerce Department hadn’t been designed with the Internet in mind at all. Instead, it was part of the blandly named “Technology Reinvestment Program,” or TRP, created at the tail end of the Bush Administration to alleviate the economic pain felt by defense-dependent regions—like California—as the Cold War wound down. In the Valley, however, the money that the Clinton Administration awarded in early 1994 went toward creating CommerceNet, a new, Tenenbaum-run industry association dedicated to e-commerce. “Capitalism is coming to the Internet,” declared The Wall Street Journal. Under Tenenbaum’s management, CommerceNet developed software to protect credit card data and ensure safe shipping. It helped companies and engineers develop new e-commerce security tools. Most important, the enterprise got lumbering old-economy companies comfortable with conducting transactions online. At one point the association had 800 corporate members. In those early days, “everyone who did e-commerce was part of it,” said Tenenbaum.14

  By the end of the year, a host of other ventures were competing to develop their own electronic-payment software, while new companies were coming into the world that gave users their first taste of buying and selling on the Internet. “Right behind sex, commerce on the Internet seems to excite people the most,” remarked Jim Bidzos, whose data-encryption technology was at the core of CommerceNet’s software. In 1995, Bidzos commercialized his technology via the Web authentication company VeriSign—whose seal of approval became a familia
r sight on many a payment page, assuring online consumers that they could enter their credit card numbers without hackers snooping in.

  That same year, an Iranian-American engineer named Pierre Omidyar launched online auction service eBay out of his one-bedroom Silicon Valley apartment. eBay wasn’t fully electronic in its commerce at the very first—to buy auction items, most customers opted to put a paper check in the mail—but that changed as more technologies spun out into the world and new users flocked to these portals. Omidyar’s innovation was to blend the community-building power of the BBS or the Usenet group with the platform of e-commerce. He wasn’t selling things; he was creating a community of people who sold things to one another. And with the auction model, he created an addictive product, with buyers returning to the site again and again to see if they could make the winning bid. eBay grew by 40 percent per month for fourteen straight months. It and other e-commerce portals developed and bought their own proprietary technology, but many of the basics of doing credit-card transactions and shipping online came from CommerceNet. Given the trillions of dollars eventually generated by online retailers, those TRP dollars invested at the start of 1994 may have had the greatest return on investment in Valley history.15