Andreessen, a brilliant young computer scientist, had just spearheaded a small software project at the National Center for Supercomputing Applications (NC SA), based at the University of Illinois, that developed the first really effective Web browser, also called Mosaic. Clark and Andreessen quickly understood the huge potential for Web-browsing software and decided to partner up to commercialize it. As Netscape began to grow, they reached out to Barksdale for guidance and insight into how best to go public.

Today we take this browser technology for granted, but it was actually one of the most important inventions in modern history. When Andreessen was back at the University of Illinois NCSA lab, he found that he had PCs, workstations, and the basic network connectivity to move files around the Internet, but it was still not very exciting-because there was nothing to browse with, no user interface to pull up and display the contents of other people's Web sites. So Andreessen and his team developed the Mosaic browser, making Web sites viewable for any idiot, scientist, student, or grandma. Marc Andreessen did not invent the Internet, but he did as much as any single person to bring it alive and popularize it.

“The Mosaic browser started out in 1993 with twelve users, and I knew all twelve,” said Andreessen. There were only about fifty Web sites at the time and they were mostly just single Web pages. “Mosaic,” he explained, “was funded by the National Science Foundation. The money wasn't actually allocated to build Mosaic. Our specific group was to build software that would enable scientists to use supercomputers that were in remote locations, and to connect to them by the NSF network. So we built [the first browsers as] software tools to enable researchers to 'browse' each other's research. I looked at it as a positive feedback loop: The more people had the browser, the more people would want to be interconnected, and the more incentive there would be to create content and applications and tools. Once that kind of thing gets started, it just takes off and virtually nothing can stop it. When you are developing it, you are not sure anyone is going to use it, but once it started we realized that if anyone is going to use it everyone is going to use it, and the only question then was how fast it would spread and what would be the barriers along the way.”

Indeed, everyone who tried the browser, including Barksdale, had the same initial reaction: Wow! “Every summer, Fortune magazine had an article about the twenty-five coolest companies around,” Barksdale recalled. “That year [1994] Mosaic was one of them. I not only had read about Clark and Andreessen but had turned to my wife and said, 'Honey, this a great idea.' And then just a few weeks later I get this call from the headhunter. So I went down and spoke to Doerr and Jim Clark, and I began using the beta version of the Mosaic browser. I became more and more intrigued the more I used it.” Since the late 1980s, people had been putting up databases with Internet access. Barksdale said that after speaking to Doerr and Clark, he went home, gathered his three children around his computer, and asked them each to suggest a topic he could browse the Internet for-and wowed them by coming up with something for each of them. “That convinced me,” said Barksdale. “So I called back the headhunter and said, Tm your man.'”

Netscape's first commercial browser-which could work on an IBM PC, an Apple Macintosh, or a Unix computer-was released in December 1994, and within a year it completely dominated the market. You could download Netscape for free if you were in education or a nonprofit. If you were an individual, you could evaluate the software for free to your heart's content and buy it on disk if you wanted it. If you were a company, you could evaluate the software for ninety days. “The underlying rationale,” said Andreessen, “was: If you can afford to pay for it, please do so. If not, use it anyway.” Why? Because all the free usage stimulated a massive growth in the network, which was valuable to all the paying customers. It worked.

We put up the Netscape browser, said barksdale, and people were downloading it for three-month trials. I've never seen volume like this. For big businesses and government it was allowing them to connect and unlock all their information, and the point-and-click system that Marc Andreessen invented allowed mere mortals to use it, not just scientists. And that made it a true revolution. And we said, 'This thing will just grow and grow and grow.'“

Nothing did stop it, and that is why Netscape played another hugely important flattening role: It helped make the Internet truly interoperable. You will recall that in the Berlin Wall-PC-Windows phase, individuals who had e-mail and companies that had internal e-mail could not connect very far. The first Cisco Internet router, in fact, was built by a husband and wife at Stanford who wanted to exchange e-mail; one was working off a mainframe and the other on a PC, and they couldn't connect. “The corporate networks at the time were proprietary and disconnected from each other,” said Andreessen. “Each one had its own formats, data protocols, and different ways of doing content. So there were all these islands of information out there that were disconnected. And as the Internet emerged as a public, commercial venture, there was a real danger that it would emerge in the same disconnected way.”

Joe in the accounting department would get on his office PC and try to get the latest sales numbers for 1995, but he couldn't do that because the sales department was on a different system from the one accounting was using. It was as if one was speaking German and the other French. And then Joe would say, “Get me the latest shipment information from Goodyear on what tires they have sent us,” and he would find that Goodyear was using a different system altogether, and the dealer in Topeka was running yet another system. Then Joe would go home and find his seventh-grader on the World Wide Web researching a term paper, using open protocols, and looking at the holdings of some art museum in France. And Joe would say, “This is crazy. There has to be one totally interconnected network.”

In the years before the Internet became commercial, explained Andreessen, scientists developed a series of “open protocols” meant to make everyone's e-mail system or university computer network connect seamlessly with everyone else's-to ensure that no one had some special advantage. These mathematical-based protocols, which enable digital devices to talk to each other, were like magical pipes that, once you adopted them for your network, made you compatible with everyone else, no matter what kind of computer they were running. These protocols were (and still are) known by their alphabet soup names: mainly FTP, HTTP, SSL, SMTP, POP, and TCP/IP. Together, they form a system for transporting data around the Internet in a relatively secure manner, no matter what network your company or household has or what computer or cell phone or handheld device you are using. Each protocol had a different function: TCP/IP was the basic plumbing of the Internet, or the basic railroad tracks, on which everything else above it was built and moved around. FTP moved files; SMTP and POP moved e-mail messages, which became standardized, so that they could be written and read on different e-mail systems. HTML was a language that allowed even ordinary people to author Web pages that anyone with a Web browser could display. But it was the introduction of HTTP to move HTML documents around that gave birth to the World Wide Web as we know it. Finally, as people began to use these Web pages for electronic commerce, SSL was created to provide security for Web-based transactions.

As browsing and the Internet in general grew, Netscape wanted to make sure that Microsoft, with its huge market dominance, would not be able to shift these Web protocols from open to proprietary standards that only Microsoft's servers would be able to handle. “Netscape helped to guarantee that these open protocols would not be proprietary by commercializing them for the public,” said Andreessen. “Netscape came along not only with the browser but with a family of software products that implemented all these open standards so that the scientists could communicate with each other no matter what system they were on-a Cray supercomputer, a Macintosh, or a PC. Netscape was able to provide a real reason for everyone to say, 'I want to be on open standards for everything I do and for all the systems I work on.' Once we created a way to browse the Internet, people wanted a universal way to access what was out there. So anyone who wanted to work on open standards went to Netscape, where we supported them, or they went to the open-source world and got the same standards for free but unsupported, or they went to their private vendors and said, 'I am not going to buy your proprietary stuff anymore... I am not going to sign up to your walled garden anymore. I am only going to stay with you if you interconnect to the Internet with these open protocols.'”

Netscape began pushing these open standards through the sale of its browsers, and the public responded enthusiastically. Sun started to do the same with its servers, and Microsoft started to do the same with Windows 95, considering browsing so critical that it famously built its own browser directly into Windows with the addition of Internet Explorer. Each realized that the public, which suddenly could not get enough of e-mail and browsing, wanted the Internet companies to work together and create one interoperable network. They wanted companies to compete with each other over different applications, that is, over what consumers could do once they were on the Internet-not over how they got on the Internet in the first place. As a result, after quite a few “format wars” among the big companies, by the late 1990s the Internet computing platform became seamlessly integrated. Soon anyone was able to connect with anyone else anywhere on any machine. It turned out that the value of compatibility was much higher for everyone than the value of trying to maintain your own little walled network. This integration was a huge flattener, because it enabled so many more people to get connected with so many more other people.

There was no shortage of skeptics at the time, who said that none of this would work because it was all too complicated, recalled Andreessen. 'Tou had to go out and get a PC and a dial-up modem. The skeptics all said, 'It takes people a long time to change their habits and learn a new technology.' [But] people did it very quickly, and ten years later there were eight hundred million people on the Internet.“ The reason? ”People will change their habits quickly when they have a strong reason to do so, and people have an innate urge to connect with other people,“ said Andreessen. ”And when you give people a new way to connect with other people, they will punch through any technical barrier, they will learn new languages-people are wired to want to connect with other people and they find it objectionable not to be able to. That is what Netscape unlocked.“ As Joel Cawley, IBM's vice president of corporate strategy, put it, ”Netscape created a standard around how data would be transported and rendered on the screen that was so simple and compelling that anyone and everyone could innovate on top of it. It quickly scaled around the world and to everyone from kids to corporations.“

In the summer of 1995, Barksdale and his Netscape colleagues went on an old-fashioned road show with their investment bankers from Morgan Stanley to try to entice investors around the country to buy Netscape stock once it went public. “When we went out on the road,” said Barksdale, “Morgan Stanley said the stock could sell for as high as $14. But after the road show got going, they were getting such demand for the stock, they decided to double the opening price to $28. The last afternoon before the offering, we were all in Maryland. It was our last stop. We had this caravan of black limousines. We looked like some kind of Mafia group. We needed to be in touch with Morgan Stanley [headquarters], but we were somewhere where our cell phones didn't work. So we pulled into these two filling stations across from each other, all these black limos, to use the phones. We called up Morgan Stanley, and they said, 'We're thinking of bringing it out at $31.' I said, 'No, let's keep it at $28,' because I wanted people to remember it as a $20 stock, not a $30 stock, just in case it didn't go so well. So then the next morning I get on the conference call and the thing opened at $71. It closed the day at $56, exactly twice the price I set.”

Netscape eventually fell victim to overwhelming (and, the courts decided, monopolistic) competitive pressure from Microsoft. Microsoft's decision to give away its browser, Internet Explorer, as part of its dominant Windows operating system, combined with its ability to throw more programmers at Web browsing than Netscape, led to the increasing slippage of Netscape's market share. In the end, Netscape was sold for $10 billion to AOL, which never did much with it. But though Netscape may have been only a shooting star in commercial terms, what a star it was, and what a trail it left.

“We were profitable almost from the start,” said Barksdale. “Netscape was not a dot-com. We did not participate in the dot-com bubble. We started the dot-com bubble.”

And what a bubble it was. “Netscape going public stimulated a lot of things,” said Barksdale. “The technologists loved the new technology things it could do, and the businesspeople and regular folks got excited about how much money they could make. People saw all those young kids making money out of this and said, 'If those young kids can do this and make all that money, I can too.' Greed can be a bad thing-folks thought they could make a lot of money without a lot of work. It certainly led to a degree of overinvestment, putting it mildly. Every sillier and sillier idea got funded.”

What was it that stimulated investors to believe that demand for Internet usage and Internet-related products would be infinite? The short answer is digitization. Once the PC-Windows revolution demonstrated to everyone the value of being able to digitize information and manipulate it on computers and word processors, and once the browser brought the Internet alive and made Web pages sing and dance and display, everyone wanted everything digitized as much as possible so they could send it to someone else down the Internet pipes. Thus began the digitization revolution. Digitization is that magic process by which words, music, data, films, files, and pictures are turned into bits and bytes-combinations of Is and Os-that can be manipulated on a computer screen, stored on a microprocessor, or transmitted over satellites and fiber-optic lines. It used to be the post office was where I went to send my mail, but once the Internet came alive, I wanted my mail digitized so I could e-mail it. Photography used to be a cumbersome process involving film coated with silver dug up from mines halfway across the world. I used to take some pictures with my camera, then bring the film to the drugstore to be sent off to a big plant somewhere for processing. But once the Internet made it possible to send pictures around the world, attached to or in e-mails, I didn't want to use silver film anymore. I wanted to take pictures in the digital format, which could be uploaded, not developed. (And by the way, I didn't want to be confined to using a camera to take them. I wanted to be able to use my cell phone to do it.) I used to have to go to Barnes & Noble to buy and browse books, but once the Internet came alive, I wanted to browse for books digitally on Amazon.com as well. I used to go to the library to do research, but now I wanted to do it digitally through Google or Yahoo!, not just by roaming the stacks. I used to buy a CD to listen to Simon and Garfunkel-CDs had already replaced albums as a form of digitized music-but once the Internet came alive, I wanted those music bits to be even more malleable and mobile. I wanted to be able to download them into an iPod. In recent years the digitization technology evolved so I could do just that.

Well, as investors watched this mad rush to digitize everything, they said to themselves, “Holy cow. If everyone wants all this stuff digitized and turned into bits and transmitted over the Internet, the demand for Web service companies and the demand for fiber-optic cables to handle all this digitized stuff around the world is going to be limitless! You cannot lose if you invest in this!”

And thus was the bubble born.

Overinvestment is not necessarily a bad thing-provided that it is eventually corrected. I'll always remember a news conference that Microsoft chairman Bill Gates held at the 1999 World Economic Forum in Davos, at the height of the tech bubble. Over and over again, Gates was bombarded by reporters with versions of the question, “Mr. Gates, these Internet stocks, they're a bubble, right? Surely they're a bubble. They must be a bubble?” Finally an exasperated Gates said to the reporters something to the effect of, “Look, you bozos, of course they're a bubble, but you're all missing the point. This bubble is attracting so much new capital to this Internet industry, it is going to drive innovation faster and faster.” Gates compared the Internet to the gold rush, the idea being that more money was made selling Levi's, picks, shovels, and hotel rooms to the gold diggers than from digging up gold from the earth. Gates was right: Booms and bubbles may be economically dangerous; they may end up with many people losing money and a lot of companies going bankrupt. But they also often do drive innovation faster and faster, and the sheer overcapacity that they spur-whether it is in railroad lines or automobiles-can create its own unintended positive consequences.

That is what happened with the Internet stock boom. It sparked a huge overinvestment in fiber-optic cable companies, which then laid massive amounts of fiber-optic cable on land and under the oceans, which dramatically drove down the cost of making a phone call or transmitting data anywhere in the world.

The first commercial installation of a fiber-optic system was in 1977, after which fiber slowly began to replace copper telephone wires, because it could carry data and digitized voices much farther and faster in larger quantities. According to Howstuffworks.com, fiber optics are made up of strands of optically pure glass each “as thin as a human hair,” which are arranged in bundles, called “optical cables,” to carry digitized packets of information over long distances. Because these optical fibers are so much thinner than copper wires, more fibers can be bundled into a given diameter of cable than can copper wires, which means that much more data or many more voices can be sent over the same cable at a lower cost. The most important benefit of fiber, though, derives from the dramatically higher bandwidth of the signals it can transport over long distances. Copper wires can carry very high frequencies too, but only for a few feet before the signal starts to degrade in strength due to certain parasitic effects. Optical fibers, by contrast, can carry very high-frequency optical pulses on the same individual fiber without substantial signal degradation for many, many miles.

The way fiber-optic cables work, explains one of the manufacturers, ARC Electronics, on its Web site, is by converting data or voices into light pulses and then transmitting them down fiber lines, instead of using electronic pulses to transmit information down copper lines. At one end of the fiber-optic system is a transmitter. The transmitter accepts coded electronic pulse information-words or data-coming from copper wire out of your home telephone or office computer. The transmitter then processes and translates those digitized, electronically coded words or data into equivalently coded light pulses. A light-emitting diode (LED) or an injection-laser diode (ILD) can be used to generate the light pulses, which are then funneled down the fiber-optic cable. The cable functions as a kind of light guide, guiding the light pulses introduced at one end of the cable through to the other end, where a light-sensitive receiver converts the pulses back into the electronic digital Is and Os of the original signal, so they can then show up on your computer screen as e-mail or in your cell phone as a voice. Fiber-optic cable is also ideal for secure communications, because it is very difficult to tap.

It was actually the coincidence of the dot-com boom and the Telecommunications Act of 1996 that launched the fiber-optic bubble. The act allowed local and long-distance companies to get into each other's businesses, and enabled all sorts of new local exchange carriers to compete head-to-head with the Baby Bells and AT&T in providing both phone services and infrastructure. As these new phone companies came online, offering their own local, long-distance, international, data, and Internet services, each sought to have its own infrastructure. And why not? The Internet boom led everyone to assume that the demand for bandwidth to carry all that Internet traffic would double every three months-indefinitely. For about two years that was true. But then the law of large numbers started to kick in, and the pace of doubling slowed. Unfortunately, the telecom companies weren't paying close attention to the developing mismatch between demand and reality. The market was in the grip of an Internet fever, and companies just kept building more and more capacity. And the stock market boom meant money was free! It was a party! So every one of these incredibly optimistic scenarios from every one of these new telecom companies got funded. In a period of about five or six years, these telecom companies invested about $ 1 trillion in wiring the world. And virtually no one questioned the demand projections.

Few companies got crazier than Global Crossing, one of the companies hired by all these new telecoms to lay fiber-optic cable for them around the world. Global Crossing was founded in 1997 by Gary Winnick and went public the next year. Robert Annunziata, who lasted only a year as CEO, had a contract that the Corporate Library's Nell Minow once picked as the worst (from the point of view of shareholders) in the United States. Among other things, it included Annunziata's mother's first-class airfare to visit him once a month. It also included a signing bonus of 2 million shares of stock at $10 a share below market.

Henry Schacht, a veteran industrialist now with E. M. Warburg, Pincus & Co., was brought in by Lucent, the successor of Western Electric, to help manage it through this crazy period. He recalled the atmosphere: “The telecom deregulation of 1996 was hugely important. It allowed competitive local exchange carriers to build their own capacities and sell in competition with each other and with the Baby Bells. These new telecoms went to companies like Global Crossing and had them install fiber networks for them so they could compete at the transport level with AT&T and MCI, particularly on overseas traffic... Everyone thought this was a new world, and it would never stop. [You had] competitive firms using free capital, and everyone thought the pie would expand infinitely. So [each company said,] 'I will put my fiber down before you do, and I will get a bigger share than you.' It was supposed to be just a vertical growth line, straight up, and we each thought we would get our share, so everybody built to the max projections and assumed that they would get their share.”

It turned out that while business-to-business and e-commerce developed as projected, and a lot of Web sites that no one anticipated exploded-like eBay, Amazon, and Google-they still devoured only a fraction of the capacity that was being made available. So when the dotcom bust came along, there was just way too much fiber-optic cable out there. Long-distance phone rates went from $2 a minute to 100. And the transmission of data was virtually free. “The telecom industry has invested itself right out of a business,” Mike McCue, chief operations officer of Tellme Networks, a voice-activated Internet service, told CNET News.com in June 2001. “They've laid so much fiber in the ground that they've basically commoditized themselves. They are going to get into massive price wars with everyone and it's going to be a disaster.”

It was a disaster for many of the companies and their investors (Global Crossing filed for bankruptcy in January 2002, with $12.4 billion in debt), but it turned out to be a great boon for consumers. Just as the national highway system that was built in the 1950s flattened the United States, broke down regional differences, and made it so much easier for companies to relocate in lower-wage regions, like the South, because it had become so much easier to move people and goods long distances, so the laying of global fiber highways flattened the developed world. It helped to break down global regionalism, create a more seamless global commercial network, and made it simple and almost free to move digitized labor-service jobs and knowledge work-to lower-cost countries.

(It should be noted, though, that those fiber highways in America tended to stop at the last mile-before connecting to households. While a huge amount of long-distance fiber cable was laid to connect India and America, virtually none of these new U.S. telecom companies laid any substantial new local loop infrastructure, due to a failure of the 1996 telecom deregulation act to permit real competition in the local loop between the cable companies and the telephone companies. Where the local broadband did get installed was in office buildings, which were already pretty well served by the old companies. So this pushed prices down for businesses-and for Indians who wanted to get online from Bangalore to do business with those businesses-but it didn't create the sort of competition that could bring cheap broadband capability to the American masses in their homes. That has started happening only more recently.)

The broad overinvestment in fiber cable is a gift that keeps on giving, thanks to the unique nature of fiber optics. Unlike other forms of Internet overinvestment, it was permanent: Once the fiber cables were laid, no one was going to dig them up and thereby eliminate the overcapacity. So when the telecom companies went bankrupt, the banks took them over and then sold their fiber cables for ten cents on the dollar to new companies, which continued to operate them, which they could do profitably, having bought them in a fire sale. But the way fiber cable works is that each cable has multiple strands of fiber in it with a potential capacity to transmit many terabits of data per second on each strand. When these fiber cables were originally laid, the optical switches-the transmitters and receivers-at each end of them could not take full advantage of the fiber's full capacity. But every year since then, the optical switches at each end of that fiber cable have gotten better and better, meaning that more and more voices and data can be transmitted down each fiber. So as the switches keep improving, the capacity of all the already installed fiber cables just keeps growing, making it cheaper and easier to transmit voices and data every year to any part of the world. It is as though we laid down a national highway system where people were first allowed to drive 50 mph, then 60 mph, then 70 mph, then 80 mph, then eventually 150 mph on the same highways without any fear of accidents. Only this highway wasn't just national. It was international.

“Every layer of innovation gets built on the next,” said Andreessen, who went on from Netscape to start another high-tech firm, Opsware Inc. “And today the most profound thing to me is the fact that a fourteen-year-old in Romania or Bangalore or the Soviet Union or Vietnam has all the information, all the tools, all the software easily available to apply knowledge however they want. That is why I am sure the next Napster is going to come out of left field. As bioscience becomes more computational and less about wet labs, and as all the genomic data becomes easily available on the Internet, at some point you will be able to design vaccines on your laptop.”

I think Andreessen touches on what is unique about the flat world and the era of Globalization 3.0. It is going to be driven by groups and individuals, but of a much more diverse background than those twelve scientists who made up Andreessen's world when he created Mosaic. Now we are going to see the real human mosaic emerge-from all over the world, from left field and right field, from West and East and North and South-to drive the next generation of innovation. Indeed, a few days after Andreessen and I talked, the following headline appeared on the front page of The New York Times (July 15, 2004): “U.S. Permits 3 Cancer Drugs from Cuba.” The story went on to say, “The federal government is permitting a California biotechnology company to license three experimental cancer drugs from Cuba-making an exception to the policy of tightly restricting trade with that country.” Executives of the company, CancerVex, said that “it was the first time an American biotechnology company had obtained permission to license a drug from Cuba, a country that some industry executives and scientists say is surprisingly strong in biotechnology for a developing nation... More than $1 billion was spent over the years to build and operate research institutes on the west side of Havana staffed by Cuban scientists, many of them educated in Europe.”

Just to summarize again: The PC-Windows flattening phase was about me interacting with my computer and me interacting with my own limited network inside my own company. Then came along this Internet-e-mail-browser phase, and it flattened the earth a little bit more. It was about me and my computer interacting with anyone anywhere on any machine, which is what e-mail is all about, and me and my computer interacting with anybody's Web site on the Internet, which is what browsing is all about. In short, the PC-Windows phase begat the Netscape browsing-e-mail phase and the two together enabled more people to communicate and interact with more other people anywhere on the planet than ever before.