The Code
Page 48
AWS might have looked like a happy accident, a step outside Jeff Bezos’s carefully cultivated long-range plan. Yet it made complete sense. Amazon had always been a big-data company, even when Bezos was the guy with the goofy laugh in the aging Honda, selling books on the Internet. He had stocked Amazon with crack engineering talent, people capable of creating a sophisticated and knowing platform that encouraged people to buy again and again, and a shipping-and-fulfillment system of unmatched logistical sophistication. Decades of dealing with seasonal fluctuations of American retail—pandemonium at the Christmas season, gentle undulations the rest of the year—gave Amazon smarter infrastructure, capable of accommodating sharp spikes in usage. The online giant’s brick-and-mortar retail competitors had seriously underestimated what a software-powered platform could do to upend their business model. AWS scaled this sensibility upward, becoming a service that helped a whole new crop of software-driven companies take on traditional market incumbents, from hotels to taxis to broadcast TV. It wasn’t quite flying cars (yet), but it certainly was more than 140 characters.
AWS provided the back-end support for some of the biggest new consumer companies of the century’s second decade, and it also brought Amazon into the realm of big-ticket defense contracting, providing storage and analytics services for a data-gobbling intelligence community. “We decided we needed to buy innovation,” explained one Pentagon official, and in 2014 they bought it from Amazon: a $600 million contract to build a cloud platform for seventeen national security agencies. The CIA’s chief information officer called it “one of the most important technology procurements in recent history.”19
Recognizing an enormously lucrative opportunity, other tech giants crowded in, further eroding the hardware-driven business that had ruled enterprise computing for decades. Large corporate customers no longer needed to buy stand-alone computers to fill their own data centers or server rooms, nor did they need to buy packages of enterprise software. As a signal of the changing times, IBM—the company once synonymous with mainframe computing, the fearsome hardware juggernaut that once considered software to be an add-on rather than a stand-alone product—launched its own cloud computing division.
While IBM was able to become a dominant provider for the Fortune 500, the very biggest players in the cloud market had been software companies from the start: Amazon, Google, and Microsoft. After years of struggling to find a new hit product, Microsoft hit the jackpot with its cloud service, Azure, whose growth spiked over 70 percent between 2016 and 2017 alone. In a remarkable turn of events for a company that had once been the cathedral of proprietary software, 40 percent of the virtual machines in Azure ran on open-source Linux. It felt like Day One in Redmond, too.20
THE SEEING STONE
Search and social, mobile and cloud: the tech story of the 2000s seemed to be more of a free-market success story than ever before. Yet the nine-figure defense contracts secured by AWS and other cloud computing giants were one sign of renewed collaboration between the tech industry and the Pentagon after two decades of relative chill. This time, it was all about software.
The Valley’s defense economy had never disappeared, of course. The Cold Warriors still held court at the Hoover Institution; the quiet hulk of Lockheed still brooded alongside Highway 101. But the investors and entrepreneurs of the Internet era had seen little upside or need for defense work, and the Pentagon’s investments in academic research had dwindled. Lockheed moved much of its missile and space division to Colorado. Cold War–era names—Raytheon, Boeing—continued to dominate the list of the Pentagon’s biggest contractors. While Internet giants pushed into new frontiers, military software lagged behind. “The cutting edge in information technology,” Harvard physicist and future Secretary of Defense Ash Carter observed in 2001, “has passed from defense to commercial companies.”21
The U.S. thus entered the post-9/11 era with a technological dilemma. The emergence of stateless, widely dispersed terrorist networks meant that the American way of war demanded higher-tech tools than ever. Added into that were large-scale data breaches and surveillance by foreign agents that had started to plague American corporations and government agencies. Hacking wasn’t about teenage cyberpunks in suburban bedrooms anymore; it was information warfare waged by the West’s most dangerous foes. With the outlays for conventional warfare surging, military and intelligence leaders needed a fast, relatively cheap way to ramp up the military’s technological capacity.
To do it, they turned once again to Silicon Valley, but flipped the Cold War supply chain on its head. Instead of government-funded academic labs and contracts producing military tech that later could be commercialized, now the defense establishment created VC firms to seed private software companies that could one day become contractors. Instead of the traditional research and procurement process, the Pentagon sponsored hackathons and design charrettes to get government bureaucracies to behave more like start-ups.
The swelling number of defense contracts were hard to see at first—just as in the Valley’s early years, the top secret nature of so much of this activity kept observers from fully understanding its size and scope—but it soon became impossible to ignore the growing amount of work Big Tech had started to do for the military. What also became clear was a continuing irony: that some of those most enriched by the new-style military-industrial complex were also some of the tech industry’s most outspoken critics of big government, and champions of the free market. In the space-age Valley, the person embodying this contradiction was Dave Packard. In the cyber age, it was Peter Thiel.
In contrast to his tech brethren who rallied around Barack Obama in 2008, Thiel remained unwavering in his belief that modern politics was a dead-end pursuit. “Politics is about interfering with other people’s lives without their consent,” Thiel wrote not too long after Obama entered office. “I advocate focusing energy elsewhere, onto peaceful projects that some consider utopian.”22
Thiel wasn’t alone. He was one of several tech titans making bets on private space travel, “a limitless possibility for escape from world politics.” He was one of a number who had bought compounds in New Zealand as extra insurance in case of social collapse. (A few years later Thiel went one step further and became a New Zealand citizen, just in case things really went sideways.) Thiel also spent millions on more-personal pursuits: fellowships to encourage smart young people to drop out of college and try entrepreneurship instead; a foundation dedicated to reversing human aging; a think tank devoted to preparing for the “singularity”—the moment when thinking machines would be able to self-replicate and, possibly, displace humans altogether. He became a major backer of a “seasteading” effort to build a floating city, free of government control, in international waters. The libertarian utopia was the brainchild of a former Googler who happened to also be a grandson of economist Milton Friedman, and perhaps the ultimate expression of the techno-libertarian project to escape the tentacles of bureaucratic control.23
That had been the original idea fueling PayPal, of course: an alternative system of Internet-based currency, unharnessed from government-controlled money. The fact that the company had devolved into a mere online payment-processing system had been extremely lucrative for Thiel and his colleagues, but he never had let go of the notion that a bigger and more disruptive system was possible. Along with the millions he pocketed from the PayPal sale, Thiel also took away an idea for fraud-detection software that he believed could be repurposed to root out potential terrorist attacks “while protecting civil liberties.” Mining his networks of old friends and Stanford connections to build a sharp, young leadership team, Thiel in 2003 bankrolled a new data-mining company he called Palantir, after the “seeing stones” in Tolkien’s Lord of the Rings. (And a hat tip to Silicon Valley’s legend, referring back to those Tolkien-themed offices at early Xerox PARC.) Established VCs were lukewarm on the start-up—Sequoia passed, as did Kleiner—despite the fact that the team believed
it was, in the words of CEO Alex Karp, “building the most important company in the world.”24
Then in swooped an unlikely angel: the U.S. Central Intelligence Agency. Keen for access to topflight software engineering, the CIA had gotten into the venture business at the tail end of the dot-com boom, creating an entity it called In-Q-Tel, and hiring the former CEO of Lockheed to run it. The CIA became Palantir’s first and sole customer from 2005 to 2008; after seeing the impressive results delivered by the company’s tracking software, other parts of the intelligence community joined as Palantir clients. Former FBI Director George Tenet became an advisor, lamenting that the intelligence community hadn’t had “a tool of its power” before 9/11. Condoleezza Rice signed on as an advisor as well. While the exact nature of Palantir’s intelligence work remained top secret, rumors swirled that its software was the literal “killer app” that helped track down Osama bin Laden. Palantir executives did little to quash the speculation.25
Soon other government clients signed on. Large police departments sought out Palantir’s visualizations and graphs and data-mining techniques to track criminals. The U.S. Immigration and Customs Enforcement agency, or ICE, bought software to profile its targets. To help break into the tight circle of favored government contractors, Palantir spent heftily on Washington lobbyists and cultivated House and Senate lawmakers responsible for defense appropriations, ultimately landing over $1 billion in federal contracts. Data-hungry and privacy-obsessed corporations swelled Palantir’s billings even further; by 2013, the company was drawing in 60 percent of its revenue from the private sector.26
By 2016, Palantir had a $20 billion valuation, the third largest of any private company in the Valley, and thousands of employees. Its corporate culture was as distinctive and quirky as early Google. Lord of the Rings references abounded; Palantir’s downtown Palo Alto offices were “the Shire” and its two thousand employees “Palantirians.” The hiring process was notoriously demanding. “I interviewed at Facebook, Google, D. E. Shaw, and a bunch of other places,” reported one engineer. “Without a doubt, Palantir’s questions were the hardest and they asked more of them than anyone else.” If you make it through our gauntlet, Palantir’s leaders seemed to be saying, you are the best of the best.27
Underneath the geeky normalcy, the firm unnerved the industry’s privacy watchdogs. “They’re in a scary business,” remarked one EFF attorney. And it was a business made scarier by how connected its founders and funders were to the Valley elite. Chamath Palihapitiya was an investor; Thiel remained on the board of Facebook. One year, to some members’ horror, Palantir sponsored the EFF’s annual awards ceremony. Questions swirled about whether the company’s surveillance software used too broad a brush; were the innocent also getting ensnared in a net designed to catch terrorists and thieves? But the contracts continued to roll in.28
Peter Thiel always had been a figure of contradictions: a gay man who rejected special treatment for minorities; a defender of free speech who funded a lawsuit that drove a prominent online publication out of existence; a steadfast libertarian who was close to some of the Valley’s biggest liberals. His halting manner and sparing public pronouncements added to the mystery. “He really is like a chess master,” said one young admirer, “planning his moves several steps ahead.” Now, as Palantir soared, Thiel became a latter-day H. Ross Perot: a champion of free enterprise who was simultaneously reaping a great fortune from the government he disdained.29
CHAPTER 25
Masters of the Universe
I wanted to see with my own eyes the origin of success,” Russian President Dmitry Medvedev declared from the Stanford stage one bright summer’s day in 2010. American and Russian flags stood at attention behind the jeans-and-blazer-clad leader, looking as sleekly casual as any Valley venture capitalist. A member of the glasnost and perestroika generation who became president at forty-two, Medvedev was on a quest to end the brain drain that had plagued Russia since the end of the Cold War. His country was one of the fastest-growing Internet markets in the world, with sixty million citizens online and counting. Its financiers and oligarchs were investing millions in American tech giants. Now it was time for this onetime technical superpower to build some world-changing companies of its own.
Only weeks before, the Russian leader had rolled out splashy plans for a high-tech “Innograd” on Moscow’s suburban outskirts, little more than a twenty-minute drive away from the auditorium where Ronald Reagan had praised the high-tech revolution two decades earlier. He then hopped on a plane to San Francisco, following the trail of so many world leaders before him. He visited Twitter and sent his first presidential tweet (handle: @KremlinRussia). He met Steve Jobs, then back at Apple after his liver transplant, although visibly ill. He sat down with Stanford’s leadership. “Unfortunately for us,” he confessed to Provost John Etchemendy, “venture capitalism is not going well so far.” There simply wasn’t enough appetite for risk. “It’s a problem of culture, as Steve Jobs told me today. We need to change the mentality.”1
The president’s road show generated plenty of skepticism. No sooner had he sent out his first tweets than a parody account called @KermlinRussia began mercilessly poking fun at the earnest high-tech experiment. “One needs to understand that money given to modernization and innovation will be spent on corruption and swindling,” read one post. “We are aware that you are aware that we are thieves,” read another. In the raucous world of social media, nothing—and no one—was sacred.2
Medvedev was discovering that the quest to build another Silicon Valley rarely went as planned. Despite the billions spent around the world by national governments on high-tech ventures—research parks, venture funds, broadband networks, even entire cities—the United States continued to out-innovate them all, pumping out one market-altering tech company after another. China was the only place that had managed to produce companies of comparable size and reach to Google or Facebook, and that was largely because the government placed stringent and often censorious barriers that kept American tech giants from entering.
The mostly fruitless global chase showed exactly how much politics still mattered, even for an industry so long understood as a free-enterprise success story. A light government touch when it came to the online world had enabled Google and Facebook and Amazon to grow immense and ubiquitous without much worry about regulation or antitrust action. (America’s laissez-faire status quo was a striking contrast to the European Union, where tech giants faced continual pushback from the courts for privacy violations and anticompetitive practices.) Open doors to the world allowed Silicon Valley to draw from a global talent pool, even as politicians had begun to hotly debate the nation’s immigration policy. Over half of the companies founded in the Valley between 1995 and 2005 had a foreign-born founder. About 40 percent of the engineering degrees at Stanford were earned by international students. Drawing in this talent helped keep U.S. universities the best in the world.3
This was a distinctly American tale of state-building by stealth, over many generations—through defense contracts to private industry, grants to academic labs, tax breaks to venture capitalists, sustained boosterism by politicians of both parties, and more. No high-tech city built by presidential decree could match the exuberantly capitalist, slightly anarchic tech ecosystem that had evolved over several generations.
The thing that made Silicon Valley so alluring to foreign leaders like Medvedev was the same thing that allowed Twitter trolls to gleefully skewer his technotopian ambitions. Valley culture was American culture, allowing free flows of people, capital, and information like no other country in the world. And in the years immediately before and after Medvedev’s visit, it made a small group of people in Silicon Valley and Seattle very, very rich.
THE POWER OF PLACE
By the time Marc Andreessen took to the opinion pages of The Wall Street Journal in the summer of 2011 to pronounce that “software is eating the world,” the new tech platfo
rms were not only altering entire industries. They were transforming the geography of tech as well.4
Across North America and beyond, tech had inhabited a sprawling, suburbanized landscape of research parks and corporate campuses since the age of Eisenhower. It had continued to do so even as other white-collar industries and middle-class residents began returning to denser urban environments at the century’s end. Part of this suburban persistence had to do with the outsized influence of Silicon Valley, leading its imitators to assume that high-tech magic required extensive landscaping and low-rise buildings sheathed in mirrored glass. It also, however, reflected tech’s need for plenty of square footage to accommodate scores of coders in cubicles, roomfuls of server blades, and closets full of coaxial cable and routers. The up-front costs for real estate, personnel, and equipment were considerable, demanding early venture rounds of $10 million, $20 million, or more.
This changed in the new era. After two decades of concerted government-directed effort, broadband penetration was extensive not only in the U.S. but across the world. After a decade-plus of software and hardware innovation by tech’s largest operators, supercomputing power became ever smaller: the long-dominant desktop PC market shrank as lightweight laptops, tablets, and mobile devices delivered comparable amounts of computing power. The iPhone and other mobile platforms allowed a popular app to build a massive user base very quickly; the dreaded “valley of death” that tech start-ups had to travel between idea and go-to-market became a dip in the road. On top of it all, cloud services freed new companies from having to fork over precious venture capital for computing power and server rooms.
The cost of starting up plummeted and the variety of entrepreneurial enterprises exploded. Want to build an app for the iPhone? All you needed were some sharp coding skills, a good laptop, and a little cash each month for some AWS server space. Support services of all kinds could be outsourced to hourly contractors who went from gig to gig (and a whole new wave of software-powered start-ups had emerged to facilitate this matchmaking). Founders didn’t have to search for a suite in a fully-wired tech park anymore. They could rent a desk or two in a tech incubator or co-working space. Unleashed from real estate constraints, tech start-ups left the suburbs for the cities, popping up in high-density, high-rent districts favored by the young and hip from Brooklyn to Boulder, Munich to Melbourne to Mumbai. A small slice of entrepreneurs chose not to settle down anywhere, becoming high-tech nomads who could do their work from anywhere in the world with a decent Internet connection.