“When it’s tough, will you give up, or will you be relentless?”
In his talk in 1994 to wannabe Web entrepreneurs, Marc Andreessen imagined a system of e-commerce that resembled small-town life. “The corner pizza store in a box” is how Andreessen pitched Netscape’s product that included a server and software. “You put up your corner pizza store, your flower shop, your bookstore. I really believe that this is where things are heading. It is going to be interesting when it arrives.”1 This system wouldn’t be quite as decentralized as it looked, however. Netscape, as the provider of the browser, server, and related software, would become the beneficiary of the network effect that comes from holding the dominant position online. Netscape products would be what everyone used, and therefore what millions of new arrivals would choose to use when they joined the Internet party. This was the dream, at any rate, albeit a short-lived one. The next year, 1995, the Web took a sharp turn toward centralized online marketplaces as Craigslist, eBay, and most notably Amazon got their first taste of how being popular online only made you more popular.2
Craigslist and eBay both grew organically, almost by accident. Craig Newmark, a former IBM programmer living in the Bay Area, began maintaining and distributing his “list” of local events and job openings in early 1995. He sent the listings as a mass email to friends, at first just ten to twelve people. Word spread quickly, and to meet the growing demand Newmark soon professionalized his system for sending the emails. In early 1996, Craigslist—its nickname from the start—migrated to the Web. By the end of 1997, Craigslist was getting about a million page views a month and Microsoft approached him about running banner ads on the site. Newmark recalls turning down the offer, knowing in doing so he had “stepped away from a huge amount of money.”3 Though Craigslist was a for-profit enterprise (except for a short stint when it switched to nonprofit status), Newmark wasn’t in a particular hurry to grow Craigslist or to cash in. He was making a good living as a programming contractor and his priority was Craigslist’s users, not making a bunch of money.
To this day, the site has been slow to assess fees of any kind, limiting them to certain job and real estate listings as it has steadily added cities and countries to its service. In 2000, Craigslist expanded from San Francisco to nine other United States cities, including Boston, New York, and Chicago; the next year, Vancouver became the first city outside of the United States. In 2003, the London site launched, the first outside of the Americas.4 “Remember,” Newmark said, “in the conventional sense, we were never a startup. In the conventional sense, a startup is a company, maybe with great ideas, that becomes a serious corporation. It usually takes serious investment, has a strategy, and they want to make a lot of money. We’ve done something very different.”5 As Craigslist’s chief executive, Jim Buckmaster, once explained to a room full of equity analysts and fund managers, Craigslist has no business development team and politely declines any offer that shows up in its inbox.6 Newmark describes himself as the company’s founder and customer service representative.
EBay was born a bit later in 1995 as AuctionWeb, the project of another programmer in the Bay Area, Pierre Omidyar, who at first hosted the online marketplace on his personal Web site.7 His plan for publicizing AuctionWeb included posting notices on Usenet groups and getting listed on the What’s New page at the supercomputing center at the University of Illinois, which was created by Marc Andreessen when he was still an undergraduate. Through such informal channels, Omidyar reached enough Web users to have hosted thousands of free auctions with tens of thousands of bids by the end of 1995. Word of mouth did the rest. His personal Internet service provider noticed the spike in activity and raised his monthly fee from $30 a month to $250—the business rate—even though the site wasn’t yet operating as a business. “That’s when I said, ‘You know, this is kind of a fun hobby, but $250 a month is a lot of money,’” Omidyar recalled.8
Omidyar started charging a fee that was a percent of the final sale price, a system well designed for ensuring that you had the resources to grow as fast as your audience would take you. The project’s first employee and first president came the following year; in June 1997, a Silicon Valley firm, Benchmark Capital, invested $6.7 million for what became approximately a quarter of the company. Omidyar deposited the check, but didn’t touch the money, he said. More than its capital, Benchmark was providing prestige, connections in hiring, branding, and marketing advice9; in September, AuctionWeb was officially renamed eBay.10 After eBay had its successful IPO in 1998,11 many called the Benchmark deal “the best-performing Silicon Valley investment ever.”12
Amazon.com offers a stark contrast to these hacker-founded marketplaces, which bloomed naturally, one as a popular, profit-agnostic Web site, the other as a Silicon Valley–endorsed tech titan. Amazon, too, started small in July 1995 as an eager-beaver online bookseller, but the intention was always to become the one-stop shop for all online commerce. Nothing accidental about its path to greatness. First books, then everything else. The plan was hatched by a young Wall Street analyst, Jeff Bezos, while he worked for D. E. Shaw & Co., an elite “quantitative” hedge fund founded by David E. Shaw, who was a veteran of McCarthy’s artificial intelligence lab at Stanford in the 1970s. “The idea,” which Shaw says occurred to him and Bezos, “was always that someone would be allowed to make a profit as an intermediary. The key question is, ‘Who will get to be that middleman?’”13
At Stanford, Shaw had researched how to improve the underlying architecture of thinking computers, periodically leaving the lab to start high-tech ventures. He earned his computer science PhD in 1980 after he “was eventually coerced into going out of business by his thesis supervisor.”14 Shaw’s years at the lab were spent “living in the future,” he recalled, with robots wandering the grounds and music streaming from the lab’s PDP-10 computer. The future appeared well confined within those walls, however. In a message to his former colleagues, Shaw wrote: “Although I’m sure that some of you foresaw more than I did, I can’t remember sharing a collective vision, for example, of how the Arpanet we took for granted then might someday turn into anything remotely resembling the Internet we take for granted now.”15 For all its obvious innovation, the Stanford lab of that era was inward looking, skeptical of entrepreneurism, and proudly peculiar, like its leader, John McCarthy. The message was clear: to make a mark in society and put some cash in one’s pocket, best to leave the lab and its quixotic pursuit of artificial intelligence and look for a market to disrupt.
After earning his degree, Shaw initially continued his computer science research at Columbia, constructing the supercomputers that he had designed while a Stanford graduate student. In time, he drew up a business plan based on that research, which is how Shaw first crossed paths with Morgan Stanley investment bankers. Shaw’s startup idea fell through, but the bankers’ pitch to him to bring his computer skills to investing found a receptive audience. “I couldn’t help wondering whether state of the art methods that were being explored in academia could be used to discover the other investment opportunities that weren’t visible to the human eye,” he recalled.16 In 1986, Shaw joined Morgan Stanley, where he applied his high-speed computers and sophisticated algorithms to financial markets, earning six times his assistant professor’s salary.17 When things were going well, Shaw’s computers and algorithms could find assets that hadn’t reached a stable, global price, allowing Morgan Stanley to buy that asset where it was undervalued and sell it where it was overvalued. In other words, artificial intelligence could discover a sure thing, the dream of everyone who plays the horses or the stock market.
Though Shaw’s computers and accompanying algorithms produced a handsome return on investment, they didn’t represent an important breakthrough about the nature of intelligence. Like the machines of that era that ran chess programs that could beat even elite players, Shaw’s investing algorithms may have been much smarter than their predecessors, and more capable than people, but not because they were better at independent thought. They benefited from faster computers that could process more information, combined with the fact that there was so much more financial information to be processed from an increasingly global economy. However, there was one obvious difference between chess and investing—the extent of the prize money. After a couple of years’ success at Morgan Stanley, Shaw decided to strike out on his own and was able to attract $28 million in capital.
D. E. Shaw & Co., which began trading in 1989, moved into offices in Manhattan’s Flatiron district, above a famous Marxist bookstore.18 The dress code was straight out of the computer lab, and even in those pre-Web days the firm’s employees were equipped like the members of a lab—Sun SPARCstations with Internet access, which could be used for email and analysis.19 At this computer lab/investment firm, however, there was a single, overarching research goal: “To look at the intersection of computers and capital and find as many interesting and profitable things to do in that intersection as we can.”20
Hiring at the firm was treated with academic rigor, focused exclusively on finding the biggest brains with the best degrees. Shaw didn’t wait to hear from interested candidates and instead sent unsolicited letters to top students that explained, “We approach our recruiting in unapologetically elitist fashion.”21 The interviews at Shaw were grueling and meant to explore how a candidate thinks by asking gnarly questions like, How many gas stations are there in the United States? This method of detecting genius by posing riddles has been popular in Silicon Valley at least since the 1950s when the transistor pioneer William Shockley began hiring for his influential start-up, Shockley Semiconductors. Shockley grew up in Palo Alto, where he knew the Termans. To his lifelong shame, Shockley was tested for Lewis Terman’s study of gifted children and came up short. Shockley nonetheless relied on tests of all kind—personality, intelligence, lie detector—to assess job candidates as well as to keep tabs on current employees. The more about a person that could be converted into a number, from Shockley’s perspective, the better.
For Shockley, however, part of the point was that no candidate should score higher than he did. One time, a young physicist, Jim Gibbons, was asked during his interview to figure out how many tennis matches were required to settle a singles elimination tournament with 127 players. Shockley started his stopwatch, expecting to see a bunch of calculations, but in a barely a moment, Gibbons gave his answer: 126. The candidate explained his logic: there is only one winner, with 126 others eliminated; a match is needed to eliminate a player; therefore, 126 matches to eliminate 126 players. Shockley replied in fury, “That’s how I’d do it!” He wanted to know if Gibbons had been tipped off. Only when the next question stumped Gibbons was peace restored to the hiring process, though in truth, Shockley never found peace from tests.22 He spent his later years in the 1970s and 1980s as a reviled figure, an emeritus professor of engineering at Stanford who would travel the country to promote his claims that blacks were intellectually inferior and advocating voluntary sterilization programs for mothers with low IQs.23 Without any personal sense of irony, the Nobel prize–winning Shockley would cite results from a test that found him not-quite-genius material to make his case for group differences in intelligence.24
When Jeff Bezos arrived for his interview at the Shaw offices, he was already prepared to leave Wall Street to somehow start his own business, but he discovered a soul mate in David Shaw.25 Shaw soon promoted Bezos to cover the Internet for the firm. In no time, Bezos and Shaw became convinced of the Web’s nearly boundless commercial potential, which was reflected in the unprecedented burst in traffic in 1993 as the Mosaic browser was quickly being adopted. The question for Bezos became whether to start a Web commerce site as part of Shaw’s company or to begin his own. In the spring of 1994, Bezos scheduled a heart-to-heart with his mentor, who “took me on a long walk in Central Park, listened carefully to me, and finally said, ‘That sounds like a really good idea, but it would be an even better idea for someone who didn’t already have a good job.’” Bezos considered the point for forty-eight hours, and decided to leave Shaw and his 1994 bonus. Looking back on that crucial meeting as part of a speech to Princeton graduates, Bezos concluded that life was best considered as a series of choices: “Will you follow dogma, or will you be original? . . . Will you play it safe, or will you be a little bit swashbuckling? When it’s tough, will you give up, or will you be relentless? Will you be a cynic, or will you be a builder?”26
Bezos was a builder, clearly, and was prepared to play the long game with his Web site, which wouldn’t appear online for nearly a year. At thirty years old and after nearly a decade on Wall Street, Bezos wasn’t some young hacker who stumbled on a great Web idea and then would shyly go about inquiring how to produce a business plan. He was a young investor who enlisted hackers to carry out the business plan he had mapped out with spreadsheet projections and a general’s sweep. Books were the perfect beachhead for an aspiring e-commerce giant trying to build a loyal customer base, he concluded. A book, unlike a razor, a pizza, or a blender, could inspire. People loved books. People gave books as presents. “We’ve all had books that have changed our lives,” Bezos said early on, “and part of what we’re doing at Amazon.com is trying to help you find and discover that next one.”27
In the year between deciding to start his own e-commerce Web site and when that site would go live, Bezos raised capital from friends and family, hired employees, and prepped and tested the Web site. Bezos was proficient enough technically to know that he needed to hire professionals to design a site that could win an audience and accommodate rapid growth. His favorite to do that initial build would have been Jeff Holden, who had recently graduated with undergraduate and master’s degrees in computer science from the University of Illinois, Urbana-Champaign. An alumni newsletter traced Holden’s path out of Illinois: “As a good, hard-working student, Holden caught the attention of a recruiter for D. E. Shaw, the investment powerhouse. Holden had never given investment banking a thought, but not one to say no to something he didn’t understand, he decided to check out the company. Mesmerized by New York City, the amazing people at Shaw, and the physical presence of Shaw’s corporate headquarters, which he described as ‘a super sexy, totally artistic view of the office space,’ he took the job.” At D. E. Shaw, Holden and Bezos worked together, and when Holden heard his colleague would be starting an e-commerce site he was eager to join. Unfortunately, Bezos couldn’t act until two years had passed as part of a nonpoaching agreement with Shaw.28
Holden would join Bezos “two years and four seconds” later, but in the interim Bezos signed up a West Coast hacker named Shel Kaphan, who built the first version of the Amazon Web site. Kaphan came with less pedigree and was older—he was self-taught and had been hacking since the 1970s.29 Kaphan felt he had missed out on the earlier computer revolution, in part because he lacked a strong business sense. Seemingly, nothing had changed. Even before speaking with Bezos, Kaphan was thinking about how the Web could help people find the books they were interested in. But his idea was to hack the clunky notecard system used in libraries by creating a clickable digital version of Books in Print. “I wasn’t thinking about it in the context of selling books,” he explained, “but I was thinking, ‘Man, I hate going to the library and ruffling through those card catalogues and trying to find that thing that I’m looking for.’”30
Bezos, however, was only thinking of selling books. Creating an online index along the lines Kaphan described was crucial to Amazon.com’s impressive claim to give its customers access to “a million titles.” As a Web start-up, the company wouldn’t actually stock copies of those million titles, but neither did the biggest bookstore, for that matter. Book buyers were accustomed to learning that a book they wanted wasn’t in stock and would be delivered later. Bezos was introduced to Kaphan through a friend of a friend who worked with Bezos at D. E. Shaw. After a few meetings in Santa Cruz, California, Kaphan agreed to build the Web site for a company then still known as Cadabra, as in the magical abracadabra. The meetings were in late spring 1994, and Kaphan’s participation was contingent on where Bezos decided to put down roots. At the time, the choice was between Seattle and somewhere in Nevada, with California out of the picture, Kaphan recalled, because of its sales tax. When Bezos chose Seattle over Nevada, Kaphan was in—employee No. 1. He arrived in October 1994 and went about acquiring the computers, software, and expertise to build the site.31
In a month, Kaphan was joined by another programmer and in spring 1995, Amazon had a “friends and family” soft launch. In July, the site had its hard launch as Amazon.com. Among the many attributes Bezos liked about the new name—besides conveying the size and breadth of the world’s largest river—was that it began with an A, and would be at the top of the What’s Cool, What’s New pages on the Netscape home page, the heir to that first list on the Mosaic home page. Barely a week after launching, Yahoo, an up-and-coming Web index created on the Stanford campus by two engineering graduate students, offered to feature Amazon.com on the site. Kaphan worried if Amazon.com could handle the surge in traffic, but Bezos said yes, setting the stage for a turbo-charged start.
Within its first month, Amazon.com had sold books to residents of all fifty states and of forty-five countries.32 Even so, the book distribution system posed an early challenge because distributors had a rule that they had to ship a minimum of ten copies at a time. At the start, Amazon may only have needed a single book from a certain distributor. Bezos discovered a hack: the distributors required that ten books be ordered at a time, not that ten books be delivered to a client. “So we found an obscure book about lichens that they had in their system but was out of stock,” he explained. “We began ordering the one book we wanted and nine copies of the lichen book. They would ship out the book we needed and a note that said, ‘Sorry, but we’re out of the lichen book.’”33 That would not be a lingering problem. The site’s traffic was doubling every quarter as Amazon expanded its purview, first by adding music and videos, which necessitated changing its motto from “Earth’s Biggest Bookstore” to “Books, Music and More.”
After two years, Bezos was free to hire D. E. Shaw employees, and Kaphan found himself being pushed aside. “The company started growing and we started attracting zillions of MBAs,” he recalled. “It just stopped being fun.”34 He remained at the company for five years, so that all his stock could vest, but he was well out of the loop. Kaphan bears a grudge toward Bezos about how he was treated, even if that intense period of work left him very wealthy. He wonders if his work benefited society. “At this point, I don’t know,” he told an interviewer. “When I look at technology these days, I see that it’s either doing something to connect people or it’s doing something that isolates people.” Amazon, in particular, he says, “is more on the isolating people side. Everything caters to convenience so much that you don’t even have to get out of bed to take care of your day-to-day business. To me, that’s a step too far.”35
In early 1999, Amazon bought a controlling share in Drugstore.com and began a service called “Shop the Web,” which added still more products for online purchase. Bezos was frequently raising money, which allowed Amazon to propose “megadeals” with the largest Web portals like AOL, Yahoo, MSN, and Excite; Amazon would pay tens of millions of dollars to be the exclusive bookseller and to have its results show up on searches. This was inorganic growth, to be sure, but Bezos considered growth too important to leave to chance. Around the time of the Drugstore.com acquisition, a journalist tried to explain Bezos’s drive to expand before even returning a profit. “While his beleaguered physical rivals are mired in the present, where they have to attend to sticky details like making money, the specifics of Amazon. com’s ultimate form remain forever elusive, a lovely shimmering at the edge of the horizon,” he wrote. “In this way Amazon.com truly is a virtual company, existing only in the imagination.”36
Amazon has continued to expand to encompass all aspects of commerce, including an increasingly lucrative business selling cloud-based computing. The company now has a dominant share of online retail sales and has become for all intents and purposes the “middleman” for online commerce that Bezos and Shaw imagined, shaping our economy and our lives. In 2015, Amazon surpassed $100 billion in sales and registered 300 million customers; analysts estimate that more than half of any growth in e-commerce will go directly to Amazon.37 Upon achieving that $100 billion milestone, Bezos was prompted to look back twenty years to when he was “driving the packages to the post office myself and hoping we might one day afford a forklift.” Even as he celebrated how far Amazon had come, he noted that, “measured by the dynamism we see everywhere in the marketplace and by the ever-expanding opportunities we see to invent on behalf of customers, it feels every bit like Day 1.”38
There is something rejuvenating about approaching every day as if it were your first. But something unnatural, too. The workplace pressure never ceases—each day you must return to the marketplace to have the public render its verdict. Again. Again. And again. The Web-based marketplace keeps everybody productive and on their toes by instantaneously shuttling information between customers and businesses. If Amazon somehow isn’t attentive to its customers, the dollars quickly go elsewhere. The tricky issue for Bezos has been to look beyond how this system serves customers to consider how the system serves Amazon’s workers.
In 2015, the New York Times published a detailed report that portrayed Amazon as a ruthless employer on behalf of its customers the company’s white-collar workers were crying at their desk from the stress and pressure.39 And this account didn’t begin to address the treatment of blue-collar workers in Amazon’s warehouses around the world, which, depending on where on the globe they are and depending on the season, can be unbearably cold or unbearably hot.40 Bezos expressed disbelief at the Times account—he was confident, he wrote in a note to employees, that he would never, could never, lead the kind of “soulless, dystopian workplace where no fun is had and no laughter heard,” as described in the Times.
The market simply wouldn’t allow it, he explained to his employees. “I don’t think any company adopting the approach portrayed could survive, much less thrive, in today’s highly competitive tech hiring market,” he wrote. “The people we hire here are the best of the best. You are recruited every day by other world-class companies, and you can work anywhere you want.”41 This response appeared to be willfully ignorant of the pressures on employees to fight through unpleasant work experiences, to resist switching jobs. But more so, Bezos’s response revealed his unalloyed libertarian faith in markets as the path to a moral workplace and a moral society. He had convinced himself that the market would somehow protect workers.