How the Internet Happened

While they sold their project to the generals this way, the academics behind the ARPANET then spent the next twenty years evolving the network into a system that better suited their own needs: a distributed, nonhierarchical computer and communications network that facilitated discussion and exchange among the research and scientific community.

PC Windows had penetrated all the desktops, the Mac was a huge success, and point-and-click interfaces had become part of everyday life. But to use the Net you still had to understand Unix. . . . And the current users had little interest in making it easier. In fact, there was a definite element of not wanting to make it easier, of actually wanting to keep the riffraff out.

As the first browser designed for the common computer user, Mosaic had a symbiotic sort of chicken-and-egg relationship with the web. For millions of PC and Mac users, Mosaic was their first glimpse of the web. Once they saw what the web could do, they wanted to go off and code their own websites.

They didn’t know they were the vanguard of a newfangled gold rush. They were, literally, corn-fed midwesterners. They were used to making six bucks an hour for their coding and had little inkling that software development could pay much more than that. When Jim Clark dangled a high-five-figure salary in front of them, they almost thought he was joking.

Website after website on the still immature web started posting little buttons that read “Best viewed in Netscape Navigator” with a link that sent you to the download page. Just as had happened with Mosaic, webmasters and web creators wanted to show off the cool new things that Navigator allowed them to do, so they steered their users to the new browser organically.

The information superhighway was interactive, sure. It let you talk back to your TV. But it didn’t allow you to create your own television program. The web, by contrast, allowed users to consume content, and create it. Any user. Anywhere. Any kind of content. And anyone could do so outside the control of a major media corporation or gatekeepers like the cable companies or Microsoft.

lots of people inside and outside of Netscape were already seeing what Marc Andreessen had seen: the browser could be a software platform capable of supplanting traditional operating systems like Windows. If, in the future, people could live their lives and do their work entirely online, then what would be the need for a desktop OS?

do was interact with each other.

And first and foremost, AOL fixated on building a sense of community among its membership. AOL users were encouraged to email, argue, play and above all chat.

In a way, AOL embodied that most American of dichotomies: wholesome, friendly, mainstream on the outside, with all sorts of prurient stuff going on behind closed doors.

There were AOL discs left on seats at football games. At one point, Brandt even tested whether or not discs could survive flash freezing so that she could give away AOL discs with Omaha Steaks.

At one point, 50% of the CDs produced worldwide had AOL logos printed on them.18

“We didn’t really have an internalized grip on how important we were to people’s daily lives,” Jan Brandt says of the crisis. “What we didn’t calibrate was the ferociousness of the response. It was crazy and it was really enlightening. It was like, ‘Oh my God! People love us! They really love us!’ Or, at that point, they love-hated us.”

The first banner ads got click-through rates in astronomical percentages. “People just clicked on anything to see what might lead them somewhere,” Joe McCambley says. “It bordered between the high 70s and low 80s* [in terms of click-through percentages] for about 2–3 weeks.”22

The pair began working on the directory to the exclusion of almost everything else. They would toil away for dozens of hours at a stretch, trading off sleeping on the floor. For Yang and Filo, it wasn’t work; it was fun. “We wanted to avoid doing our dissertations,” Yang admitted.5 By September 1994, Yang and Filo had compiled a directory of more than 2,000 sites. What was more impressive was the fact that Jerry’s Guide to the World Wide Web was getting 50,000 hits (searches) a day.

The Yahoos held their breath to see if the ads chased searchers away. But the protests quieted down after only a few weeks. The directory was just as helpful as it always was. The users stayed loyal.

By the end of the long weekend, he had cobbled together a crude website that allowed users to do three simple things: list items for sale,

view items that were on sale, and place bids on those items.

Another way to help the system regulate itself was the Feedback Forum. This was a public online message board where users were encouraged to leave written feedback about other buyers or sellers, in addition to a numerical rating: plus one, minus one or neutral. Once a user’s rating on the feedback forum surpassed a negative four, they were banned from the site.

Hotmail.com launched on the web on July 4, 1996. In little more than a year and a half, Hotmail would claim 25 million users.32

This struck Beninato as a bit nutty since When.com had not even launched to the public yet. “We didn’t really have anything,” Beninato recalled. “We were a couple of months old.”39 Yahoo didn’t end up buying When.com, but AOL eventually did. For $225 million.

But from another perspective, it’s worth wondering how much the flowering of the dot-com era was enabled by the fact that the most dominant, rapacious player in the industry was distracted while the new era was taking shape. The fact is, while Microsoft made plenty of moves during the dot-com era (MSN, Expedia, Hotmail, WebTV, just to name a few), it largely refrained from engaging in direct combat with the major dot-com players.

(AOL never got much respect from true techies in Silicon Valley), but by 1998, their jaundiced view was not shared by Wall Street. AOL was how around 40% of U.S. users got online.19

their weaknesses and intensified their strengths. I won’t deny I really believed that, as did many others—many of whom now pretend they never did.”42

But what Wolmar’s account also points out is to what degree the bubble, and the railroads constructed because of it, ultimately created the infrastructure that would enable the high Industrial Revolution in Victorian Britain. The mileage of rail schemes authorized during the bubble years came to represent 90% of the total route mileage on Britain’s rail system.

As the Internet was born in the midst of a great scientific effort to win the Cold War, the web was born in the midst of a great scientific effort to reveal the secrets of the Big Bang.

At the time, he was mainly concerned about bandwidth issues (this was the era of dial-up modems; images could take entire minutes to load onscreen) but he was also worried that he and Andreessen were opening the floodgates to frivolity and junk. “And I was right! People abused it horribly,” Bina said later. “But Marc was also right. As a result of the glitz and glitter, thousands of people wasted time to put in pretty pictures and valuable information on the Web, and millions of people use it.”19

Prodigy’s commercial efforts quickly proved to be a bust. It turned out that when people went online what they really wanted to

In January 1995, the Mercury News launched a website, with access originally $4.95 a month, though the paywall was later dropped in an effort to land more advertisers.

But the web was a different sort of animal, as Time Warner would learn to its great consternation. Time Warner wasn’t quite comfortable with people talking back. Early on in the site’s history, the O. J. Simpson case was raging and Pathfinder found huge success with an O. J. Central section where users could debate the case. But Time Warner executives feared that the free-for-all of comments and user-generated debates might expose the company liability-wise. Management slowly began to discourage community efforts. Community editors were tasked with policing community sections, discouraging users who saw their comments censored or deleted.

The web was decidedly not passive; the web was about links, about clicking. So, what should these new banner ads do? What would happen if users clicked on them? “Not only were there not a lot of really big corporate websites, at the time there was really a debate whether a corporate website would actually be the thing that people wanted,” says Kanarick. “Like, who would want to go to a website like Pampers, where they’re just going to talk about diapers all the time?”20

the time of the IPO, Amazon was recording a 900% growth in revenue.30 The promised efficiencies of the ecommerce model that Bezos had so much faith in were actually panning out. Amazon was turning over its inventory 150 times a year; traditional physical bookstores like Barnes & Noble turned inventory only 3 or 4 times a year.31

Every time a user sent an email using Hotmail’s free web mail accounts, a small link was appended at the bottom that read: “Hotmail: Free, trusted and rich email service. Get it now.” So, every time an email was sent, the sender was promoting Hotmail’s service. The very act of using Hotmail helped spread the word about Hotmail.

As early as December 1998, internal AOL emails show that Steve Case and his lieutenants began kicking around the idea of purchasing a safe lily pad to land their company on. AOL came very close to acquiring eBay, but Case was wary of doubling down on the Internet space. A merger with AT&T was floated as a way for AOL to claim direct ownership of distribution pipes, but Ma Bell declined. After approaching Disney and getting rebuffed by its CEO, Michael Eisner, AOL turned its focus to arguably the biggest media company in the world: Time Warner. A deal with Time Warner would allow AOL to marry its new media savvy to the toniest of old-media content. Aside from its numerous, tangible and lucrative assets (magazines, TV channels, movie studios and more) Time Warner had one key piece of the puzzle that AOL craved: the second-largest cable network in the country.

The AOL/Time Warner merger was announced on January 10, 2000. On April 3, 2000, Judge Jackson’s final ruling suggesting the breakup of Microsoft was announced. At the time, these two events felt epochal—clarion signals ushering in a new era in the technology and even media industries. Instead, from the perspective of hindsight, they look more like historical footnotes, bracketing the weeks when the dot-com bubble finally burst. ■

That was the coolest thing about doing Netscape after Mosaic. We literally started from scratch and were able to avoid many of the same mistakes (while of course making new ones).”36 This time they would do it better and get it right. All hands were tasked with speedily producing what the team of young coders had dubbed “Mozilla,” suggesting that the new browser was a monster set to devour their previous brainchild, Mosaic.

This open attitude toward the technology is what allowed the first ecommerce activities to begin flowering across the web. Netscape benefited as the underlying platform that was the most trusted and valued by users. Netscape also eagerly supported and incorporated the advances of others—for instance, the Java programming language when it emerged. Netscape would even encourage others to build add-ons and plugins that would interact with Netscape’s own software, adding features and functions that Netscape itself couldn’t dream up.

As AOL executive Ted Leonsis put it, AOL could become “the Carnival Cruise Lines” of the Internet, the trusted guide to places unknown.28

But Wired was a magazine flush with success primarily as an advertiser-supported operation, so, the HotWired brain trust logically turned to the notion of merely replicating the model they already knew. It was decided that HotWired’s launch on the web would be sponsored by advertisers Wired had existing relationships with from the print side of its operation. Like the launch of a print magazine, advertisers would be asked to sponsor specific content sections of the new website for a flat fee. “$10,000

The HotWired team was attempting something more ambitious, both technically and aesthetically. Two advertising and digital design firms, Modem Media and Organic, were brought on board and tasked with designing and selling something that felt closer to a magazine-style ad. Big. Colorful. Eye-catching.

Most important, a cadre of new hires was fashioned into a team of professional web surfers who would help build out the Yahoo directory and stay on top of the exploding web. The surfers, who would eventually number more than fifty, were each expected to add as many as a thousand new sites a day to the directory.16

Amazon offered the bestsellers too, of course, and heavily discounted them as loss leaders. But it would be more obscure titles like Fluid Concepts and Creative Analogies that would allow Amazon to create a rabid following among early adopters.

“Their systems were programmed in such a way that you didn’t have to receive ten books, you only had to order ten books.”12 So the Amazon team found an obscure book about lichens that was listed in the system but was regularly out of stock. They began ordering the one book they wanted and nine copies of the lichens book.

The Bulletin Board accomplished this as well. It was the place where users could ask questions: “How do I upload pictures?” or “What do you think is the proper minimum bid I should set for this item?” Fellow eBay users could chime in with their input. Very quickly, as often happens in online communities, a select group of users prominently stepped forward to become regular advice gurus and trusted “experts.” Omidyar had accidentally stumbled upon one of the longer-term factors in AuctionWeb’s eventual success. A focus on community, on empowering the users and allowing them to function autonomously would prove to be absolutely vital.

For most of the twentieth century, the “product cycle” was something that happened in a comparatively leisurely, plodding, measurable pace. But as has become common to the point of cliché over the last twenty years, in the Internet Era, change—whether to products, industries or entire economies—would come literally overnight.

Marc Andreessen described it this way: “You keep kicking versions out the door, making them better. Any individual product is less important than the basic idea. If a beta turns people off, you put out a beta that turns them back on.”

Throughout the Internet Era, company after company would become obsessed with the idea of creating or owning a platform. If you are a platform, you can create an ecosystem of developers and software and apps all dependent on the underlying platform. To own a platform is to own the ball field, the rule book, the turnstiles, and the broadcast rights. Netscape did not originate the obsession with platforms, but it would provide the template.

Netscape was valued at $150 million. Barksdale also put the legal issues with the University of Illinois to bed by settling out of court. Netscape agreed to pay the university $2.2 million in damages, with an additional payment of $1.4 million depending on future business deals. The university split the money with Spyglass, the Mosaic licensee. Netscape offered to pay with shares of the company in lieu of cash, but this was rebuffed. That refusal would cost the University of Illinois tens of millions of dollars when Netscape had its initial public offering.

“You had to show that you were the company not of the present, but of the future. The most appealing companies became those in a state of pure possibility.”

So, Microsoft did what it had to do: it cut corners. Having lost BookLink to AOL and having been rebuffed so arrogantly by Netscape, Microsoft was forced to turn to the most logical remaining choice: Spyglass, Inc., the company approved by the University of Illinois to commercialize the original Mosaic web browser. Microsoft signed a $2 million licensing agreement with Spyglass to use Mosaic code for Windows 95. Irony of ironies, the code that would be the basis for Microsoft’s web browser (and the weapon Microsoft would soon wield against Netscape) was a descendant of the same code written by Marc Andreessen and Eric Bina a few years before at the NCSA.

The company that would truly take online services mainstream was another early online pioneer that would concentrate almost religiously on allowing users to interact with each other in whatever way they wished.

Selling consumers on the virtue of one online service over another was not as important as educating consumers on just what an online service was. It was during market research that she realized she needed to go back to basics. According to Brandt, during a focus-group study, “Someone took a computer mouse and started pointing it at the computer like a remote control. And one person put it on the floor and tried to use it like a sewing machine pedal.”15

“One disk to install. One price to pay. One customer service number to call. Building web sites and hoping people will find them is a significant leap of faith.”

“The fundamental bet we are making is that we are a media company, not a tools company,” Yang told Fortune magazine. “If we are a tools company, we are not going to survive. Microsoft will just take over our space. If we are a publication, like a Fortune or a Time, and we create brand loyalty, then we have a sustainable business.”21 Making Yahoo the first great brand of the Internet Era would serve the company well throughout the entire dot-com period.

The business plan Bezos shopped around to these local investors was projecting $74 million to $114 million in sales for Amazon by the year 2000.17 On the strength of these projections, Bezos was able to raise $981,000 by the end of 1995, giving away around 20% of the company.18 Of course, those investors would do quite well, because that best-case scenario they bought into was not even close to what Amazon would eventually achieve. By the year 2000, Amazon would record $1.64 billion in net sales, more than fourteen times Bezos’s rosiest estimate.

eBay was nothing more than a virtual marketplace, and by being virtual, it didn’t actually do anything other than facilitate the interactions between buyers and sellers. It didn’t store goods. It didn’t ship goods. It didn’t even guarantee the exchange of goods between buyers and sellers! The one truly tangible thing that eBay had was the goodwill of those buyers and sellers and the community they were creating—on their own—to make the buying and selling happen.

eBay’s only asset, in fact, was its users, and therefore the only important thing for the company to do was to make sure the buyers and sellers were happy so that they would keep coming back.

Within a month, that single Beanie Baby category was responsible for 6.6% of the entire site’s sales volume.10 eBay was not exactly the company that Beanie Babies built, but Beanie Babies certainly brought eBay to the world’s attention.

And eBay embraced its image as the hobbyists’ mecca. Many people are familiar with eBay’s founding myth: how Pierre Omidyar created the site so his fiancée could expand her Pez dispenser collection. But like many company creation stories, the Pez story is a fiction. The Pez story was created by Mary Lou Song to get reporters interested in covering eBay’s role in the collectibles phenomenon. As she put it later, “Nobody wants to hear about a thirty-year-old genius who wanted to create a perfect market. They want to hear he did it for his fiancée.”

In order to make it easier for small merchants to set up shop in its mall, Yahoo purchased a company called Viaweb from a young British programmer named Paul Graham. By the holiday season of 1998, more than 3,000 different storefronts had opened shop, with Yahoo raking in monthly fees and a percentage of every sale.

Ironically enough, far from engendering customer loyalty, consumers tended to treat the dot-coms as a fly-by-night bonanza, taking the deals when they presented themselves, but often not repeating the experience. “They all e-mail me specials,” one New York pet owner told Businessweek of her experience with the pet dot-coms. “I order from whoever has the special. Sometimes, it’s even free.”31

Pixelon was a company that raised $35 million in venture financing, promising to develop “full-screen, TV-quality video and audio streaming technology” in an era of dial-up modems.57 It promptly turned around and blew $16 million of that on a company launch party at the MGM Grand in Las Vegas that featured performances by KISS, the Dixie Chicks, Sugar Ray, and a reunion concert by the Who. It was later revealed that Pixelon founder Michael Fenne was really a man named Paul Stanley (no relation to the guitarist from KISS) who was wanted by the State of Virginia on stock fraud charges. Pixelon never released a product before it was eventually forced into bankruptcy.

And so, one must consider to what degree Microsoft was distracted by the trial, allowing it to miss, say, the development of paid search as a dynamic new market, or the rise of social networks as an entirely new paradigm.

“For weeks it was, ‘You’re great, you’re great, you’re great,’ and then one day [we had to] give them every last dollar we had in the bank and 20 percent of our company.” Another dot-commer said AOL demanded 30% of her company, “and then for good measure they tell us, ‘These are our terms. You have 24 hours to respond, and if you don’t, screw you, we’ll go to your competitor.’

Priceline had cratered 94%. Yahoo was down 97%, from an all-time high of $432 per share to $11.86 on August 31, 2001, its market cap down to $6.7 billion from $93 billion. That $1,000 put into Amazon’s IPO, which had climbed in value to more than $61,000 at the bubble’s height, was worth about $3,400 at the end of September 2001, when Amazon was trading under $6.

Another key enabler of the Netscape IPO was the fact that Wall Street was buying into Marc Andreessen’s platform strategy. The investing community believed Navigator was building a platform on the web, and therefore, Netscape could become the next Microsoft. “A lot of people had missed out on the Microsoft IPO because they didn’t believe in PCs” said Frank Quattrone, a Morgan Stanley banker who would help take Netscape public. Buying Netscape stock as soon as it IPOed was, in a lot of people’s estimation, a once-in-a-lifetime chance to jump on board the Microsoft of the next technology era.

Omidyar had no idea if charging a fee would bring an end to his little experiment or not. Furthermore, he had no way of actually enforcing payment. He didn’t have a credit card merchant account or even a method for validating auction results. In keeping with his libertarian ethos, however, he refused to impose any governance or policing of his system. He simply relied on sellers to be honest. It turned out that his faith in humanity was justified, because envelopes started showing up in his mailbox with checks inside them. By the end of that first month of February, when Omidyar tallied up the envelopes, he found that he had made more than the $250 he needed to cover his web hosting. And just like that, eBay became that rarest of things: the first-ever meaningfully profitable ecommerce company.

eBay was one of the first websites to show that a largely anonymous community, carefully constrained by a few guidelines and regulations, but invested in a system of online reputation, could actually work.

But all lost out to Microsoft, who, on New Year’s Eve 1997, purchased Hotmail for $400 million in stock. Not bad for two years of work, and an idea that even its founders thought was so obvious that anyone could have done it.

By 1999, nearly 40% of retail security trades were being done online.9 If Joe Smith saw a stock like Lycos profiled on CNBC, he could jump online and place an order for Lycos stock within minutes. There was no longer any middleman to talk him out of it. And if Mr. Smith wanted to spend his days discussing the relative merits and future prospects of Lycos, he could do so on message boards at sites like Yahoo Finance that had many thousands of forums devoted to discussing individual stocks.

Of course, even this unfolding high-profile disaster didn’t stop other entrepreneurs from chasing the same dream. On the East Coast, two companies, Kozmo.com and UrbanFetch, took instant gratification a step further: both promised same-day delivery. But the question was, could anyone make money doing that? That pint of Ben & Jerry’s a customer ordered on a rainy afternoon? Kozmo would send it to them for less than it would cost to buy at the local bodega across the street. And Kozmo still had to pay the army of bike couriers who made the delivery. It was retail without the overhead of real estate, sure, but what about the costs of warehousing, of labor, of the website and logistical back-end systems? Neither Kozmo nor UrbanFetch were much worried about this. Ubiquity came first. Profits later.

Excite@Home’s acquisition of Blue Mountain Arts for $740 million dollars in cash and stock. Excite@Home was a company formed when the broadband ISP @Home merged with the search portal Excite.com. Blue Mountain Arts operated the website Bluemountain.com, where users could send each other electronic greeting cards by email. That’s right. Bluemountain did nothing but send Grandma electronic “get-well-soon” greetings. But Bluemountain.com was getting 9 million unique users a month to do this, and at the time, traffic was the sine qua non for a Yahoo-chasing portal player like the Excite half of Excite@Home.50 As the New York Times noted in its article announcing the deal, Excite@Home “predicted that the acquisition would increase its audience by 40%, to encompass approximately 34% of Internet traffic.”51 So, Excite@Home was willing to pay $82 per user to attract additional eyeballs to its network of properties and try to keep pace in the portal race.

Nonetheless, Drkoop.com enjoyed a nearly 100% first-day IPO pop and raised $85 million from investors despite reporting lifetime revenue totaling only $43,000.62

No company took greater advantage of the bubble madness than AOL, by straight-up cannibalizing other dot-coms. You might remember that the Drkoop.com IPO raised $85 million for the health-information website. A month after its stock market debut, Drkoop turned around and basically spent all that money by agreeing to pay AOL $89 million over four years to provide health content to AOL users.26 And that wasn’t even the biggest deal AOL struck in those days. A long-distance phone provider named Tel-Save ponied up $100 million.27 AOL

skillfully played one competitor off another: Barnes & Noble paid $40 million to be the bookselling partner on AOL’s online service; Amazon paid $19 million just to be on the AOL.com portal; in the midst of fending off auction competition from Amazon, eBay ponied up $75 million for a four-year auctions exclusive. And Wall Street rewarded such tie-ups. Tel-Save’s shares leapt from $13 to $19 after announcing its deal; Drkoop.com’s deal announcement caused its stock to surge 56%.28

Locking down guaranteed traffic from AOL became a box that dot-coms had to check in order to begin the Get Big Fast sweepstakes. And as AOL realized the position of power it had over the dot-coms, the deal-making only got more aggressive.

The definitive book on the merger, There Must Be a Pony in Here Somewhere, was written by Kara Swisher some years after the deal was consummated. In it, she claims that, at the time, the merger seemed like a home run to her and nearly everyone else: “In one major move, the two companies had seemingly addressed

For one thing, the Fed had finally begun to raise interest rates: three times in 1999 and then twice more in early 2000, the most sustained round of fiscal tightening over the whole of the late 1990s. And just as suddenly, the language from the Fed had shifted to an open attempt to rein in equity prices.

The most famous example from this era is Mark Cuban, perhaps the quintessential dot-com billionaire. Cuban had already cashed out early by selling his company, Broadcast.com, to Yahoo. But he didn’t trust the insane valuation of the Yahoo stock he had been paid in, so he set up a hedge against his Yahoo holdings, called an “options collar.” When Yahoo’s stock subsequently collapsed, his entire fortune was protected.

there is a lot of talk about how the American public, especially the middle and working classes, have come to believe the economic structure of America is rigged against them, and everything is tilted in favor of the insider, the moneyed, the elite. An argument can be made that this was a belief that first took hold when the dot-com bubble burst, especially to a generation of investors who came to the stock market for the first time in those years. Baby boomers did what society told them: they invested in stocks; they bought and held. And for a time, they did well, seeing their nest eggs go up by five, even six, figures (or more if they were lucky). And then they watched it all evaporate. They watched the insiders and the bankers, the lucky and the elite, walk away scot-free while they, the hardworking Americans who did what they were told, lost everything. And all of that would happen to them again less than a decade later, only this time, in the housing market.

A browser like Navigator would be the Internet’s operating system, replacing the old PC operating systems like Windows.

The information superhighway was the fever dream of the telephone industry and the cable industry and the computer industry and even of Hollywood. The idea was that we’d all be linked together via a Frankenstein-like combination of the television and the PC. We’d be able to shop from home, and exchange video chats with each other, and rent movies on demand and receive personalized news and media based on our interests. I know. Sounds exactly like the Internet we know today. But all of this was supposed to happen on your television.

challenge anybody to say that video-on-demand isn’t what the consumer wants,”7 Jerry Levin, CEO of Time Warner, declared. He could have just asked the consumer. One by one, all of the interactive TV experiments failed spectacularly.

It all came down to this: no one in tech, no one in media, no one from Bill Gates to Jerry Levin to Hollywood ubermogul Barry Diller had realized what Marc Andreessen and Jim Clark had realized: the information superhighway was already here. The Internet and the World Wide Web were the information superhighway. The revolution was now, and it was being delivered not by the television, but by the computer.

To Gates’s mind, the biggest question of all was how Microsoft could make money on the Internet. Seemingly everything on the Internet was free. This was not a small point to overlook. Gates could see how Microsoft could make money on the information superhighway, by serving as the gatekeeper and toll collector. But the freeware, ungoverned, unsettled Internet didn’t seem to offer a similar opportunity.

But Internet Explorer was right there automatically on every Windows machine. To get a copy of Navigator, conversely, you had to search it out and download and install it yourself—not an easy feat for Internet newbies.

He made the calculation that the browser wars were more important to win than the scramble to connect people to the Internet.

Classifieds began migrating to the web almost from the beginning, because for years there had been digital antecedents on the message boards and newsgroups of the early Internet and online services.

In September of 1997, Fortune magazine had a story with the title “Why Barnes & Noble May Crush Amazon.” In the article, the author posited, “Anything Amazon.com can do on the Internet, so, too, can Barnes & Noble.”26 Famously, Forrester Research released a report in early 1997 entitled “Amazon.toast.”27

So, Silicon Valley has always been equal parts egghead libertarianism and acid-tinged hippie romanticism. Both of these worldviews mesh quite well actually when it comes to believing that technology can be used to better mankind and free it from all manner of oppression, repression and just everyday drudgery.

More important, Microsoft never had the chance to absorb any of the cream of the new crop, as it had shown it was wont to do in earlier technology eras. Microsoft never attempted to acquire Amazon, say, though it certainly had the money to do so early on. And, crucially, once the dot-com bubble burst, Microsoft was in no position to swoop in and gobble up the wounded survivors because it feared angering the government again.

Of course, AOL—for one brief shining moment, the embodiment of the Internet in American life—went away. But the Internet itself didn’t go away. And that’s why the railway example is so pertinent.

To this day, most of what we do online, with the exception of ecommerce and the rare subscription service, is all advertising-supported. It’s conceptually jarring to realize that a medium and an industry that we think of as being so futuristic and technological is sustained by a business model that is centuries old.