A History of Silicon Valley

Table of Contents | Timeline of Silicon Valley | A photographic tour
History pages | Editor | Correspondence
Purchase the book

These are excerpts from Piero Scaruffi's book
"A History of Silicon Valley"


(Copyright © 2010 Piero Scaruffi)

13. The iBoomers (1995-1998)

by Piero Scaruffi

Searching

Netscape's billion-dollar IPO started the "dot-com bubble" of the late 1990s. Visionaries all over the world envisioned the Internet as a vehicle that would revolutionize business, reducing the importance of the "brick and mortar company" (for example, of the physical store) and creating new ways for companies to market/sell their products online and for consumers to shop online from a vast worldwide catalog.

Sensing a lethal threat, in august 1996 Microsoft, which so far had contributed nothing to the Internet, introduced its own browser, Internet Explorer, bundled for free with the Windows operating system. It was widely interpreted as a strategy to kill Netscape (whose browser was not available for free) while it was still small. Microsoft's move not only fatally wounded Netscape (eventually bought out by America OnLine for $4.2 billion), but de facto preempted any other company from developing a browser for the Windows operating system. Internet Explorer had been quickly packaged by recycling code from the commercial version of Mosaic marketed by Spyglass (the business arm of the University of Illinois). Microsoft also acquired (january 1996) the tool developed by Boston's Vermeer Technologies for building websites, FrontPage.

Eventually (in 1998) the government would force Microsoft to unbundle Internet Explorer from the Windows operating system, but it would be too late to save Netscape.

Search engines were still at the forefront of the Internet revolution, and the Bay Area quickly generated start-ups in that field.

In february 1995, Steve Kirsch launched the search engine Infoseek, based in Sunnyvale. It pioneered "cost-per-impression" and "cost-per-click" advertising. Yet the company's main claim to fame might be that it raised Li Yanhong, the engineer who in 1999 moved to China and co-founded the Chinese search engine Baidu.

At Stanford six students started the project Architext in february 1993. It changed its name to Excite when it launched in december 1995. DEC had opened a research center in Palo Alto, which created an internal search engine called Alta Vista, made available to the Web in december 1995. Alta Vista boasted a more sophisticated technology, notably Louis Monier's crawler (the software that roams the Web looking for documents) and it allowed users to input natural-language queries. In 1997 it even introduced automatic language translation, Babel Fish, based on the old Systran system. David Patterson and others at U.C. Berkeley had obtained DARPA funding for a project called Network of Workstations (NOW) that envisioned a supercomputer built out of networked personal computers and workstations (a forefather of cluster computing). That architecture was used by Eric Brewer and his graduate student Paul Gauthier to create the fastest search engine yet, Inktomi, and spin off a namesake company. HotBot, launched in may 1996, was based on Inktomi's technology and quickly overtook AltaVista as the number-one search engine. Yahoo! went public in april 1996: in the first hour of trading its value reached $1 billion. Not bad for a company that so far had revenues of $1.4 million and had lost $643,000.

An alternative approach to indexing the Web was pursued by the Open Directory Project (originally Gnuhoo), a public index of websites launched in June 1998 by SUN's employees Rich Skrenta (the inventor of the first personal computer virus) and Bob Truel, and acquired by Netscape in October 1998. It was a collaborative effort by thousands of volunteer editors, initially modeled after the old Usenet. It surpassed Yahoo!'s directory sometime in 2000.

In 1998 two Stanford students, Larry Page and Russian-born Sergey Brin, launched a new search engine running on Linux, Google, an offshoot of a research project begun in 1996. The amount of information available on the Web was already causing information overload, and the issue was how to find the information that was really relevant. Google ranked webpages according to how "popular" they were on the Web (i.e. how many webpages linked to them). Google went against the trend of providing ever more sophisticated graphical user interfaces: the Google user interface was just text (and very little text). In 1999 Google had eight employees. Their first "angel" investor (before the company even existed) was Andy Bechtolsheim of SUN. Then (in june 1999) they obtained $25 million from Sequoia Capital and Kleiner-Perkins.

Outside the Bay Area, the most significant player was perhaps InfoSpace, founded in march 1996 by former Microsoft's employee Naveen Jain, that built an online "yellow pages" service, also offering "chat rooms" where users could exchange live text messages.

These companies did not invent search technology. Search technology had existed for a while. Verity was the main vendor of tools for searching text archives. However, companies like Verity, that had pioneered the field and were getting rich by selling text search to the corporate world, were too busy maintaining their legacy applications to think about making a Web version of their tools. There was still a huge divide between software companies working in the old off-line world and the software companies working in the new on-line world. The former did not quite understand the Web yet and, in fact, would never catch up. The wheel had to be reinvented by young inexperienced kids because the older, experienced, skilled, and competent experts were stuck in a time warp.

The concept of the "chat room" had been pioneered on the Web by at least Bianca's Smut Shack, started in february 1994 in Chicago by David Thau and Chris Miller.

Another application that would become popular on the Internet was born in Israel: in 1996 ICQ debuted. Invented by Arik Vardi, it would become the first instant messaging program to spread worldwide.

Creating a Webpage

Creating a website and finding a place that would host it was still a bit too complicated for the masses, but it got a lot easier and even free in 1995 when David Bohnett and John Rezner launched GeoCities in Los Angeles (the original name was BHI, Beverly Hills Internet). GeoCities basically created a web within the Web. The GeoCities web was structured in six virtual neighborhoods. Users (known as "homesteaders") could choose a neighborhood and an address within that neighborhood. GeoCities automatically created a webpage for each homesteader and provided an easy way for the homesteader to customize it. In october 1997 Geocities reached the milestone of one million users. In august 1998 it went public. By the end of the year it had become the third-most visited website on the Web after AOL and Yahoo. In january 1999 Yahoo purchased GeoCities for $3.57 billion.

Two Williams College classmates, Bo Peabody and Brett Hershey, and an economics professor, Dick Sabot, launched the website Tripod in 1994 out of their office in northwestern Massachusetts Tripod, like Geocities, became popular as a place where Internet users could easily create webpages, but its most lasting invention would be the "pop-up" ad, "invented" by Ethan Zuckerman in 1998.

Social networking was born in april 1995 when Cornell University's students Stephan Paternot and Todd Krizelman launched theGlobe.com, an online community that expanded the old concept of the Usenet. In november 1998 theGlobe.com went public. On its first trading day the price of its stock closed at 606% the initial share price, setting an all-time record for IPOs.

Getting Online

The explosion of interest for the Internet meant, first and foremost, an explosion of ISPs (Internet Service Providers). The government helped the commercial boom of the Internet in yet another way. In 1994 the National Science Foundation commissioned four private companies to build four public Internet access points to replace the government-run Internet backbone: WorldCom in Washington, Pacific Bell in San Francisco, Sprint in New Jersey, and Ameritech in Chicago. Then other telecom giants entered the market with their own Internet services, which they often subcontracted to smaller companies. By 1995 there were more than 100 commercial ISPs in the USA. The user had to pay a monthly fee and was sometimes also charged by the hour. Some ISPs simply gave users remote access to a Unix "shell" running on the ISP's server. Services based on SLIP (Serial Line Internet Protocol), the first protocol (1988) devised for relaying Internet Protocol packets over "dial-up" lines, and later PPP (Point-to-Point Protocol) implied that the user was given a modem and had to dial a phone number (sometimes a long-distance number) to reach the ISP. In 1995 the leading ISPs were Uunet (yearly revenue of almost $100 million), that targeted companies (not individuals) willing to pay $1,000 a month for Internet service, Netcom (just over $50 million), which skyrocketed from zero to 400,000 subscribers in its first year thanks to a simpler pricing scheme (a flat fee for 400 hours) targeting the consumer market, and PSINet. Also in 1995 telecom giant AT&T unveiled its ISP service, called WorldNet, which copied Netcom's flat-rate model. In april 1996 Pacific Bell, a regional telephone company, began offering Internet access to most of the four metropolitan areas of California (Bay Area, Los Angeles, Sacramento, San Diego). The number of ISPs in the nation passed 3,000 at the beginning of 1997, and about 1,000 more were born in the next six months. By the year 2000, Uunet, the first and still the largest ISP, running 500,000 kilometers of fiber and cable, owned about 30% of the Internet's infrastructure. Dial-up pioneer CompuServe (acquired by AOL in 1997) was a distant second.

Metricom, founded in Los Gatos (south of Cupertino) in 1985, pioneered the business of wireless ISP in 1994 when it launched Ricochet Networks and offered service to Cupertino households. By 1996 it was covering the whole Bay Area.

At the other end of the dial-up connection, companies like Ascend Communications of Alameda (near Oakland) made fortunes by selling equipment to manage dial-up lines (acquired in 1999 by Lucent for $20 billion).

The one piece of equipment that almost everybody needed was the modem. U.S. Robotics was an example of a company that never bent to the standards (its modems used proprietary protocols) and managed to become the leader in its sector. In 1997 3Com purchased U.S. Robotics for $6.6 billion.

Browser-based Computing

Unless one worked for a university or research laboratory, an individual's e-mail was run by her or his ISP. When a user subscribed to the service, the ISP gave the user one or more e-mail addresses. The user had to install software on the home computer that periodically downloaded email messages. Starting with Soren Vejrum's WWWMail (launched in Denmark in february 1995), several attempts at been made at providing an email service that could be accessed from a browser. In july 1996 two former Apple hardware engineers, Indian-born Sabeer Bhatia and Jack Smith, funded by venture capitalist Tim Draper of Draper Fisher Jurvetson, launched Hotmail, which liberated the Internet user from the slavery of the ISP. Hotmail was a website. Hotmail users checked their e-mail on the Web, regardless of where they were. E-mail had just become location-independent. Even better: e-mail "traveled" with the user, because it could be accessed anywhere there was a web browser. Its model was so easy to understand that Hotmail signed up more than 8.5 million subscribers by december 1997, in just 18 months. It helped that Hotmail was free. Microsoft bought Hotmail in december 1997 (for about $400 million).

Another secondary business created by the emergence of the browser had to do with a browser's "plug-ins". A plug-in was a piece of software written by someone else that was incorporated into the browser to add a functionality that was not present in the original browser. One could add and remove plug-ins at will, thus customizing the browser. Macromedia released two of the most popular and innovative plug-ins for Netscape's browser Navigator: Shockwave in 1995, an evolution of Director to display videos, and Flash Player in 1996, an evolution of the animation tool FutureSplash acquired from Jonathan Gay's FutureWave. Flash would quickly become the world standard for embedding multimedia files into HTML pages (in other words, for displaying videos within webpages). Most plug-ins were made available for free.

In 1995 SUN launched Java, originally developed by James Gosling. It was both an elegant object-oriented programming language and a virtual machine. The advantage of a virtual machine is that any application written for it runs on any physical machine on which the virtual machine has been ported. Java specifically targeted the new world of the Internet, for no other reason than portability of applications. SUN gave away Java for free and Netscape soon introduced a feature to process webpages that contained "applets" written in Java. The promise of Java as the esperanto of the Internet was such that venture-capital firm Kleiner-Perkins hastily harnessed a $100 million fund to invest in Java start-ups. Supported by a growing number of companies, notably IBM and Oracle, Java indeed became the language of choice for Internet applications. The more the Internet grew, the more Java grew. Eventually it started threatening the decade-old supremacy of the C language.

WebLogic, formed in 1995 in San Francisco, launched an application server for the Java 2 Enterprise Edition (J2EE). It was acquired in 1998 by San Jose-based BEA, founded by Chinese-born former SUN manager Alfred Chuang. Another pioneering Java application server was delivered by Steve Jobs' NeXT in 1995: WebObjects, that was also an object-oriented rapid application development environment.

Building Bricks

Many of the tools required by Internet-based applications were available for free. In 1995 in Oregon a veteran of the Smalltalk environment, Ward Cunningham, who had his own software consulting business, created WikiWikiWeb, the first "wiki", an online manual (of reusable software programming methods) maintained in a collaborative manner on the Internet. Soon there would be "wikis" (user-maintained online knowledge bases) for everything. A wiki is an online knowledge-base that is created, edited, and maintained by a community of users. Any member of the community is allowed to change the content of any webpage of the knowledge-base from within an ordinary browser. The knowledge-base (in this case the manual) has no single author. It continuously evolves as more knowledge is provided by the users.

Meanwhile, a community of Unix developers (including Brian Behlendorf, a Berkeley student who volunteered for the Burning Man festival) launched the Apache server in 1996 as open-source software. The project extended and refined the web server implemented by Robert McCool at the government-funded National Center for Supercomputing Applications of the University of Illinois at Urbana-Champaign. A web server is a software application that transfers ("serves") pages of the World Wide Web to clients via the Hypertext Transfer Protocol (HTTP). Until then the only web server had de facto been the Netscape web server, that cost $5,000. Apache made it possible for anybody to start a website on their own home or office computer, thus increasing the amount of content and the number of applications that reached the Web.

In 1997 the World Wide Web Consortium (W3C), founded in october 1994 at the MIT by Tim Berners-Lee as the international standards organization for the World-wide Web, introduced the XML standard for exchanging documents on the Web, thereby recasting EDI (Electronic Data Interchange) for the age of the Web.

In 1998 a struggling Netscape that was being strangled by the free Microsoft browser and the free Apache server decided to license the code of its browser for free as open source so that anybody could use it to build a better browser. Thus began the project "Mozilla". It was a desperate survival strategy.

In 1998 Adam Stiles (based in the Los Angeles area) introduced a new web-browser called SimulBrowse (later renamed NetCaptor) that allowed users to open multiple webpages at the same time, i.e. the first browser with "tabs". Opera followed suit in 1999 and Mozilla would include tabs in 2002 in the very first release of its Phoenix browser (later renamed Firefox).

Hotmail's Lessons

Hotmail had grown its user base faster than any media company in history. The founders were both hardware engineers, not software engineers. This is a case in which the idea came from a user viewpoint, not from a technology viewpoint. They needed a way to bypass the company's firewall and figured out how to do it. Many software engineers could have done the same because they knew the Web technology well enough but simply did not come up with the idea because they personally didn't need it. Being hardware engineers, helped Hotmail's founder deliver a sturdy no-nonsense "product". Software engineers are artists who proceed by trial and error, and are not terribly concerned if their product has bugs (one can always fix the bugs later). Hardware engineers cannot afford bugs: every time a mistake is made, the cost is colossal. Therefore Hotmail's founders naturally delivered something that was ready to be used by millions of people. About 100,000 people signed up for Hotmail in the first three months. Most new software needs a lot of tweaking before it gets it right: theirs didn't need any tweaking. Hotmail was a case in which the cultures of hardware and software converged and leveraged each other. Hotmail, like Yahoo, was also another story that proved the unprecedented power of the Internet to spread news by word of mouth. Email, the very nature of Hotmail's business, was also the most powerful grass-roots tool ever to spread information. Hotmail did two things to further improve its efficiency: 1. They added a tagline to every email that invited the recipient to join Hotmail; 2. They gave each and every user an email address that contained "hotmail.com" so that each recipient would know (even without reading the tagline) that Hotmail existed. Both were useful for "branding". But neither would have worked if Hotmail had not been useful and usable: users spread the gospel that there was a simpler, better and cheaper way to do email.

Hotmail's concept (Web-based email, or "webmail") was a very easy concept to copy. In fact, it was copied. It was copied by many and by several companies that were much larger. Hotmail, however, proved how difficult it was to catch up. Precisely because the Internet medium was so unique in rapidly multiplying users, it created a de-facto standard even before the industry could do so. Once become a de-facto standard, a startup (such as Hotmail) enjoyed a huge advantage even over much larger companies (such as Microsoft). The advantage was such that the larger company would have to invest millions in R&D and marketing for several years to match the startup. Indirectly this also ended up discouraging large companies from trying to catch up with successful startups: much cheaper to just buy them for outrageous prices. Rocketmail became Hotmail's main competitor after being a partner (they build Hotmail's directory of registered users) and being funded by the same venture capitalist. Eventually, Yahoo bought Rocketmail (in 1997) the same way Microsoft had bought Hotmail.

Finally, Hotmail bet on advertising as a viable source of revenues. Hotmail failed to prove that concept (it was never profitable) but represented one more step towards proving it. Hotmail's founders realized that email was even more powerful than Yahoo to generate "click-throughs" (actual visits to the advertised website).

Indirectly, Internet start-ups realized that advertising (and marketing in general) had become such a key component of the capitalist society. One could almost claim that more creativity was going into marketing a product than in designing it. That industry had been consistently looking for innovative advertising vehicles: the newspaper, the nickelodeon, the radio and the television. The Internet boom took place at precisely the time when advertising was booming too. Cable television revenues had staged an 82% growth rate in 1994-95. The Web was also capable of delivering interactive advertising. and advertising that leads straight to the store (through a mouse click, i.e. a "click-through").

Ads were perceived by the masses as a negative feature. Many startups failed because they were perceived to be too "commercial" once their web-based services started displaying too many ads. It was a tricky balance: on one hand startups needed the revenues from ads, that were proportional to the number of people visiting their websites, and on the other hand ads were making their websites less attractive and therefore reducing the number of visitors. Websites without any ad were more likely to earn the trust of the Internet public than websites with ads. However, the Internet public was not willing to turn that "trust" into financial support: websites that tried to charge a fee for each visit were even less likely to survive than websites that displayed ads. The Internet public was an odd beast: it wanted information without advertising, but it was not willing to pay for that information in any other manner. The good news for Internet startups is that the costs of operation were very low (Hotmail had 15 engineers when it was bought by Microsoft for $400 million).

Internet startups that followed the "free service" model indirectly adopted the view that their real product was the user base. A car manufacturer makes a car and makes money by selling the car. Many websites offered a web-based service that was only an excuse to create a large user base, and then they made money by selling advertising to corporations interested in selling whatever products/services to that user base; or upselling a different kind of service to that user base.

The Net Economy: 1995

Dotcom companies appeared in every sector of society, and electronic commerce became widespread. In 1995 Jeff Bezos, a former Wall Street investment banker who had relocated to Seattle, launched Amazon.com as the "world's largest bookstore", except that it was not a bookstore but a website: Amazon was the quintessential declaration of war to the "brick and mortar" store. Craigslist.com, launched in 1995 by former IBM salesperson Craig Newmark from his San Francisco residence, provided a regional advertising platform (initially only for the Bay Area) that quickly made newspaper classified obsolete (since it was free and reached more people). It became a cult phenomenon, spreading by pure word of mouth. Newmark refused investments and offers to sell. Xing Technology, in southern California, developed the first live audio and video delivery system over the Internet, StreamWorks, i.e. a system to play an audio or video file while it is downloaded from the Internet (Xing was acquired by RealNetworks in 1999). Progressive Networks (later renamed RealNetworks), formed in Seattle by ex-Microsoft executive Rob Glaser in 1995, introduced the RealAudio "streaming" audio software (that could only stream audio before the acquisition of Xing). These two companies enabled live broadcasts on the Internet. In 1993 a concert by rock band Severe Tire Damage at the legendary high-tech laboratory of Xerox PARC was streamed live around the world, the first Internet live broadcast of music. In 1995 the online magazine The Word routinely streamed live content, and also in 1995 the television channel ESPN SportsZone used RealNetworks to stream live a baseball game. Live streaming went truly mainstream in 1996 when Marc Scarpa streamed the Tibetan Freedom Concert from San Francisco's Golden Gate Park, and the term "webcast" became commonplace. In 1998 RealNetworks, Netscape and Columbia University finalized the Real Time Streaming Protocol (RTSP), a protocol that, unlike the "stateless" HTTP, had "state" and was therefore better suited to live streaming. Viaweb (Yahoo Stores): In 1995 New York-based computer scientists Paul Graham and Robert Morris (famous for having created the "Morris worm" in 1988 when he was still a student at Cornell University) started Viaweb (sold to Yahoo in 1998). It was a website to create online stores which made it easier for many people to get into e-commerce. It was also one of the first Web-based applications: it ran on a server and was controlled by the user by clicking on the links of a webpage displayed on a browser. Viaweb was similar to what had been done with X-terminals to run X-Windows (the browser being the equivalent of the X terminal). In 1995 French-born Armenian-Iranian and former General Magic engineer Pierre Omidyar founded AuctionWeb (renamed eBay in 1997), a website to auction items. By the end of 1998 eBay's auctions totaled $740 million.

The online dating system Match.com was launched in 1995 in Los Altos by Gary Kremen, one of the pioneers of Internet-based classified advertising with his company Electric Classifieds (1993)

Other companies ventured into Internet telephone, security, and website development tools. In 1995 the Israeli company VocalTec released the first commercial Internet phone software, i.e. a system capable of dispatching telephone calls over the Internet. In 1995 Kevin O'Connor, with the money he made at Internet Security Systems, started Internet Advertising Network (IAN), which later acquired and retained the name of the DoubleClick system of New York's Internet ad agency Poppe Tyson, largely the creation of Tom Wharton. Indian-born Samir Arora, who had managed Hypercard at Apple, founded NetObjects in 1995 in Redwood City to develop tools for people to build their own websites (at the time not an easy matter). In 1995 only 15% of Internet users in the USA were women, but two New York media executives saw that the number was growing and that women were not served adequately by the new male-dominated medium: Candice Carpenter (Time Warner) and (Doubleday) Nancy Evans founded the female-oriented portal iVillage in june 1995. In 1995 corporate lawyer Stacy Stern, Stanford law and economic student Tim Stanley and German-born Stanford student Martin Roscheisen created FindLaw, a portal for law-related information, the first time that case law was made easily accessible to the USA public.

Some also started thinking about the possibility of combining Internet and television. In july 1995 former Apple's employee Steve Perlman founded Artemis Research, later renamed WebTV, to build a set-top box based on custom hardware and software which would allow a television set attached to a telephone line to plug into an Internet service using a dial-up modem. The goal was to turn the World-wide Web into a home appliance. The WebTV set-top box (introduced in september 1996 by Sony and Philips) was not very successful but it pioneered the idea of accessing the Web via a consumer electronics device instead of a workstation or personal computer. In april 1997 WebTV was acquired by Microsoft.

Internet IPOs escalated rapidly during the decade. AOL had gone public in 1992. UUNet and Netscape went public in 1995; Lycos, Excite, Yahoo!, CompuServe, Infoseek and E*Trade in 1996; Amazon in 1997. There had been ten Internet IPOs in 1995 but there were already 40 in 1998, and there would be 272 in 1999. In 1999 "dotcom" startups would receive more than 40% of venture capital investment.

The dotcom boom also boosted the stock of e-learning, a field pioneered by Stanford in the 1960s. However, very few of the startups of the era survived, notably Saba, founded in 1997 in Redwood Shores by Oracle's executive Bobby Yazdani.

At this point the "users" of a dotcom service were typically technology-savvy professionals, especially those raised on Unix. At the time not everybody was connected to the Internet, not everybody had access to high-speed lines, and not everybody understood what the World-Wide Web was. In 1996 there were 14 million households in the US with Internet access, almost all of them on dial-up lines. Even in 1998, when the number of users in the US had skyrocketed to 75 million, the vast majority still used dial-up services. Ordinary people were still reluctant to use email, let alone sophisticated ecommerce websites. Many households were just beginning to get familiarized with computers.

The Net Economy: 1996-97

Significant innovations for the Web were push technology and subscription models. In february 1996 Pointcast, founded by Christopher Hassett, started a fad for "push" technology with its software that gathered information from the Web and then displayed it on a personal computer, basically the opposite of surfing (a software agent sending information to the user instead of the user searching the Web). Marimba, a spin-off of SUN's Java group founded in 1996 in Mountain View by Arthur van Hoff and Jonathan Payne and headed by Kim Polese (at the time a rare female CEO of a high-tech startup), offered subscription-based software distribution so that one could automatically get updates to its applications. The movie industry was helped by an early Web subscription-model business named Netflix. In august 1997 Reed Hastings, who had been selling tools for programmers with his first start-up Pure Software, founded Netflix in Scotts Valley (between San Jose and Santa Cruz) to rent videos (initially on DVD) via the Internet. DVD rentals were a direct strike against movie theaters, one of the pillars of social life in the 20th century. Netflix upped the ante by removing even the DVD store and allowing movie buffs to simply order movies online. The movie still required a physical device, the DVD, but was delivered at home.

The world was being mapped and made available. In 1997 Jim Gray of Microsoft, working with Aerial Images of North Carolina, created the web-based mapping service TerraServer that also offered satellite images from the United States Geological Survey (USGS) and Russia's space agency Sovinformsputnik. In 1996 Chicago-based GeoSystems Global launched the web-based mapping service MapQuest that also provided address matching. In 1996 Jim Clark of Silicon Graphics fame founded Healtheon (originally Healthscape) to create software for the health-care system (to "map" treatment data).

In October 1996 Microsoft launched Expedia, an online travel agency (that would become the largest travel agency in the world in the 2010s).

Stanford students Anand Rajaraman and Venky Harinarayan started in 1996 in Sunnyvale a service that helped consumers find products on the Internet: Junglee.

A few companies even ventured into early social networking. In 1997 Stanford University's engineering students Al Lieb and Selina Tobaccowala founded Evite, a free website to manage invitations. A milestone for web-based social networking software was SixDegrees.com, launched in 1997 by New York's corporate lawyer Andrew Weinreich and named after the hypothesis that all human beings are linked by at most six connections. A user could link to friends and family and reach people beyond the first level of connection. At the peak it had one million registered users. Technically speaking, it had been preceded by PlanetAll, a social-networking site launched in 1996 in Boston by Warren Adams and Brian Robertson.

The world of finance was affected too. Online stock brokerage moved to the Web with E*Trade, launched in 1992 in Palo Alto by William Porter and Bernard Newcomb. Internet banking (a bank with no physical branch) was pioneered by ING Direct, founded in 1996 in Canada by Arkadi Kuhlmann.

The Net Economy: 1998-99

As the technology became more reliable, the applications became more sophisticated too, spreading to broadcasting, books, and groceries.

In 1998 Berkeley's professor Abhay Parekh founded FastForward Networks to provide radio and television broadcasting over the Web (FastForward's first customer was RealNetworks when it set up its Real Broadcast Network). In 1998 NuvoMedia, founded in Palo Alto by Martin Eberhard (one of the founders of Network Computing Devices) and Marc Tarpenning, introduced the Rocket eBook, a paperback-sized handheld device to read digital books downloaded from online bookstores ("ebooks"). Another e-book reader was introduced at the same time by rival SoftBook Press, founded by Amiga's videogame guru James Sachs and by publishing industry executive Tom Pomeroy in 1996 in Menlo Park. The two start-ups were eventually purchased by Rupert Murdoch's media conglomerate.

Webvan, born in 1999 to deliver groceries purchased online (initially only in San Francisco), raised $375 million in 18 months and at one point it was worth $1.2 billion. Kozmo in New York, started in 1998, was even more ambitious, planning to deliver all sorts of goods within one hour: it raised $280 million.

In 1998 Amazon started expanding beyond books. It acquired the Internet Movie Database, which had originated in Britain as Col Needham's newsgroup on the Usenet, the Boston-based PlanetAll, which already boasted 1.5 million members (which at the time was a lot, especially since Amazon only had 3.1 million customers), and Silicon Valley-based Junglee, and within one year it became the biggest online sales platform in the world. In 1999 it also patented 1-Click, the technology that created hassle-free online shopping: Amazon's customers could now make purchases with a click (or little more than a click). 1-Click was soon licensed by Apple (in 2000). In 2000 Amazon launched the Amazon Marketplace, which opened its online platform to small businesses all over the country. A bold strategy of expansion was creating the juggernaut of e-commerce.

Internet banking was pioneered by ING Direct (1996, Canada), Everbank (1998, Florida) and Ally Bank (2001, Utah).

The moment that companies started doing business on the Web, it became important to have reliable services to display webpages in real time. Akamai, founded in 1998 by MIT mathematician Thomson Leighton and his student, Israeli-born Technion alumnus Daniel Lewin, replicated content to servers located around the world in order to minimize the time it took to deliver it to the end users.

The New Nature of Innovation

The semiconductor boom was largely the making of one legendary company. Fairchild Semiconductor birthed more than 50 semiconductor startups. The vast majority of Silicon Valley engineers working in semiconductors during the 1960s had worked at Fairchild at one time or another. The history of the semiconductor industry in Silicon Valley is the Fairchild genealogical tree. The semiconductor boom largely consisted in the refinement of one technology. It was a very vertical kind of technological development.

The age of Apple, Cisco, SUN and Oracle was already different. This boom was more diversified: the inventions of Apple, Cisco, SUN and Oracle had little in common. Furthermore, neither of them gave rise to a (significant) genealogical tree. Relatively few startups were created by former Apple, Cisco, SUN or Oracle engineers; and no major company of the size of Intel emerged from any of these. What each of them created was a chain of suppliers. There was a vertical economy that relied on them but the boom was becoming more horizontal.

The dotcom age was completely horizontal. The dotcom boom was wildly diversified. It wasn't just the refinement of a technology, but the application of a technology to many different applications in wildly different fields. The dozens of dotcoms were exploring a vast landscape. Some dotcoms refined previous ideas (for example, search or social networking), but most of them did not refine a previous application at all: they were the first to implement that application on the Web. If someone else had preceded them, there was no relationship between the two sets of engineers (it was not a Fairchild-Intel kind of relationship). The Internet resettled Silicon Valley in a new landscape, and the dotcom boom was mostly about exploring this new landscape and finding gold mines in it.

One reason that dotcoms did not bother to refine previous dotcoms' inventions (the way semiconductor startups had done in the 1970s) was that there were so many opportunities: why compete with an existing dotcom? Another reason that dotcoms did not bother to refine previous dotcoms' inventions was outsourcing: the job of refinement was best left to India.

Network Neutrality

The dotcom boom owed a lot to a fundamental principle of the Internet, never encoded in a law but tacitly accepted by the whole community. The network had to be company-neutral and application-agnostic. The network was simply a highway that anybody could use. There was no VIP lane. No matter how large and powerful a corporation was, it had to use for its ecommerce activities the very same network that the poorest student was using to set up her personal website. The Internet had been conceived to be as general as possible and with no particular application in mind. The dotcom boom was a boom of applications, not of platforms. The innovation went into the applications. The platform was roughly still the same as in the 1980s if not the 1970s. The application boom was driven by the neutrality of the Internet. First of all, that neutrality protected the small Internet startup from the "brick and mortar" corporation. Everybody could get a fair chance in the marketplace, a fact that was not true in the "brick and mortar" economy where large businesses could use a plethora of tactics to hurt smaller newcomers. Secondly, the network allowed the producer (the creator of a Web-based service) to offer its "product" directly to the consumer (the user). There was no need for an intermediary anymore. In the traditional economy users shopped for the products that an intermediary had decided to deliver to the store. It was not the user who decided which products would show up at the store: it was a chain of intermediaries starting from the product planning division all the way down to the store owner. The Internet removed the whole chain. People were creating applications on the Web, and users were deciding which ones became successful. In the "brick and mortar" economy a corporation decided which product to sell and then proceeded to publicize it. In the "net" economy, instead, it was the user who decided which "product" (website) to use. The marketing for that product, its website, was the "buzz" created within the community of users. There was no advertisment publicizing a website, no salesman selling it door to door, and no store displaying it on its window.

Getting Online Part II

Internet dial-up services were multiplying, but dial-up access was slow. Two technologies appeared at the end of the 1990s to remedy the problem. The USA government made another important decision for the future of the Internet in 1996: the "Telecommunications Act" allowed cable television providers to offer Internet services. Milo Medin, a former NASA scientist, and tycoon William Randolph Hearst III started At Home Network (@Home), technically a joint venture between TCI (the largest cable operator in the USA) and the venture capital firm Kleiner Perkins Caufield & Byers. Their mission was to deliver Internet broadband to a million homes by the first quarter of 1997: the high-speed cable ISP was born. In 1997 US West launched the first commercial DSL (Digital Subscriber Line) service in Phoenix. Cable Internet used the cables laid down for cable television. DSL used the ordinary telephone lines. Neither was widely available at the time.

The obvious beneficiaries of the age of networking were the companies that specialized in networking hardware: by the mid-1990s the three Silicon Valley giants Cisco, 3Com and Bay Networks (formed by the merger of SynOptics and Billerica) had achieved $1 billion in revenues. Indian-born Pradeep Sindhu, a former semiconductor scientist at Xerox's PARC, founded Juniper Networks in 1996 in Sunnyvale to manufacture high-end routers in direct competition with Cisco. When it went public in 1999, its IPO was one of the most successful in history, turning it overnight into a $4.9 billion company.

The astronomical growth of Internet communications also created demand for high-speed fiber-optic cables to connect not only research centers but also millions of ordinary homes. Fiber optics are made of thin glass wires. Optical switches convert the digital signals into light pulses. The light pulses are transmitted over these optical cables instead of the copper wires of the traditional telephone lines. A fiber-optic cable has a much broader bandwidth than a copper cable. Therefore optical fibers can be used to transmit large amounts of data over large distances. The story of fiber optics goes back many decades, but applications were unfeasible until more progress in both optical fibers and laser technology (the only light source capable of high-speed modulation) took place in the 1960s and 1970s. Seven months after Theodore Maiman's demonstration of the first laser (the ruby laser), a team at Bell Labs (Ali Javan, William Bennett and Donald Herriott) demonstrated the first continuous-wave laser. Robert Hall at General Electric created the first semiconductor diode laser already in 1962, but it was not a practical one. Only in 1970 did someone invent a semiconductor diode laser capable of emitting continuously and at room temperature (Zhores Alferov in Russia, followed by Bell Labs), and only in 1975 did a company begin commercialization of the technology (Laser Diode Labs, a spinoff of RCA Labs). By the following year Bell Labs was demonstrating laser diodes with a long life. Throughout the 1960s fiber optics communication was promoted mainly by Charles Kao at Stanford, but its feasibility had to wait until 1970 when the first practical optical fiber was developed by glass maker Corning Glass Works. In 1975 Manfred Boerner at Telefunken Labs in Germany demonstrated the first fiber-optical data transmission. Commercial applications had to wait for optical fiber amplifiers, which came of age with pump laser diodes. The boom of optical cables began in 1996, when the US government enacted the Telecommunications Act that deregulated the telecom market, thus creating ferocious competition among local and global telecom companies at the same time that the pundits predicted an exponential rise in broadband need due to the advent of the Web. The age of fiber optics truly began in 1988 with the first transatlantic cable to use optical fibers, the TAT-8 (a collaboration among AT&T, France Telecom and British Telecom), the eighth transatlantic communications cable, capable of carrying 40,000 telephone simultaneous conversations (as opposed to the 138 conversations of the 1966 transatlantic cable).
It is not a coincidence that the capacity of optical cables almost doubled every six months starting in 1992 until in 2001 it achieved a bit rate of 10 terabit per second. Lucent, a 1995 spin-off of AT&T (including the Bell Labs), that between 1997 and 1999 had spent a fortune to acquire some 30 startups, and Canadian giant Nortel, which had deployed the world's first commercial fiber-optic link in 1980, dominated the market. In 1999 Lucent introduced the all-optical LambdaRouter, the first optical network switch, advertised as 16 times faster than electronic switches of the time. In 1999 Nortel and Lucent together owned more than 50% of the $12.3 billion market, a market that had just grown by 56% from 1998. In 1999 Cisco acquired Cerent of Petaluma (north of San Francisco, formed in 1996 with funding from Vinod Khosla) for $7.4 billion, and became the third power, while Nortel bought Xros of Sunnyvale in 2000 (for $3.25 billion). In 1988 the Southern Pacific Railroad began installing fiber-optic cable along its lines and in 1995 it spawned Qwest, headquartered in Denver (Colorado), which by 1999 had revenues of $3.92 billion. Aerie Networks of Denver (Colorado) raised $100 million of venture capital in its first two years (1999-2000). Sycamore Networks, located in Massachusetts, was worth $29.1 billion in september 2000. Silicon Valley became one of the hotbeds of optical technology. A spin-off of Optivision founded in 1997 in San Jose by Indian-born Rohit Sharma, ONI Systems was one of the first to go public. Then came the deluge: Kestrel Solutions of Mountain View, Lightera Networks of Cupertino, Calient of San Jose, Mayan Networks of Mountain View, Amber Networks of Fremont, Zaffire of San Jose, Luminous Networks of Cupertino, Luxn of Sunnyvale, etc.
The world was suddenly awash in fiber-optic cables. This overcapacity, not mandated by governments but simply the result of business miscalculations, dramatically lowered the cost of broadcasting information, thereby increasing the motivation to broadcast information. The fiber-optic rush created on the Internet the equivalent of the freeway system created by the US government in the 1950s. It did so rapidly and at no cost to the taxpayer. The cost was political: the vast fiber-optic infrastructure did more than connect the nation electronically: it connected the nation to India too, thus accelerating the process of outsourcing IT jobs to India.

Something else was brewing amid all the startup frenzy, but mostly elsewhere. In 1988 Mark Weiser had proposed the vision of a future in which computers will be integrated into everyday objects ("ubiquitous computing") and these objects are connected with each other. This became known as the "Internet of Things" after 1998 when two experts in RFID at the Massachusetts Institute of Technology, David Brock and Sanjay Sarma, figured out a way to track products through the supply chain with an RFID tag linking to an online database. In 1996 Japan's Sony had introduced the chip FeliCa for RFID technology. The foundations had been laid down. Silicon Valley had pioneered RFID applicatoins with Identronix, but little had come out of it.

A Brief History of Open-Source Software

The World-wide Web was a major boost to the open-source movement. Open-source software was born as a reaction by the "counterculture" to the big corporations and investment firms that had started changing the software industry in the 1980s. Until 1970 there had been basically no software industry. In the 1980s software emerged as both a "mission-critical product", for which corporations were willing to pay a lot of money and a "consumer good" that millions of computer users were willing to purchase. The software engineers who embarked in open-source projects were idealists who disliked the commercial success of software. In the 1980s it was not easy to share and distribute software. It was mainly done on Unix machines connected via the Internet. The "social media" of the Unix world were the "newsgroups" of the Usenet invented in 1980. For example, the most influential "scripting" language of the Unix world, Perl, was released by its creator Larry Wall in 1987 on the comp.sources.misc newsgroup. In 1991 Guido van Rossum, based in Holland, released Python, soon to become a very famous programming language, on the alt.sources newsgroup. Another example of free software that became very popular before the age of the World-wide Web was the X Window System, developed in 1984 at the MIT by Jim Gettys and Bob Scheifler. It was the most popular windowing system on Unix machines at the time when Apple and Microsoft where beginning to make the "windows-mouse" paradigm popular. At that time Richard Stallman launched the GNU Project and founded the Free Software Foundation. The first GNU was released in 1989 and contained a complete set of software development tools that were completely free for anybody. The first famous GNU success was Linus Torvalds' Linux, a variant of the Unix operating system, that was released in 1991 and became a GNU item in 1992. Much of the impulse towards free software came from UC Berkeley, where in 1977 Bill Joy had created the version of Unix called BSD (Berkeley Software Distribution). That version of Unix became very popular on the workstations of the 1980s. As part of BSD, in 1979 Eric Allman created the program "delivermail" that became "sendmail" and that managed most of the email traffic on the Internet throughout the 1980s. Much of the free software was developed by students. In 1992 William and Lynne Jolitz created a version of BSD called 386BSD that was completely free. They named it that way because they wanted a Unix for the personal computers that ran the Intel microprocessor 80386, small cheap computers that were faster than the traditional (and much more expensive) DEC machines that ran BSD Unix. Later their 386BSD would evolve into FreeBSD, NetBSD, and OpenBSD, all of them available for free. What had become as a leftist insurrection against capitalism was mutating into an academic practice. A former Berkeley student, Brian Behlendorf, was one of the software engineers who wrote the Apache HTTP Server in 1995. They were known as the "Apache Group" that later became the Apache Software Foundation. The Apache software was extremely important for the success of the World-wide Web. Incidentally, when he was not working on software, Behlendorf was helping the Burning Man festival. In 2006 the same Behlendorf would be invited to speak at the World Economic Forum in Davos. His story was typical of how people in the Bay Area could mix counterculture, art, business and technology. In 1998 Netscape created Mozilla, the open source version of its Internet software. Jamie Zawinski was one of the leaders of that project. The term "open source" was probably invented in 1998 by nanotech guru Christine Peterson, co-founder of the Foresight Institute in Palo Alto. The first conference on free software, the "Freeware Summit" organized in 1998 in Palo Alto by the publisher Tim O'Reilly. was quickly nicknamed the "Open Source Summit". Participants included Linus Torvalds, Guido van Rossum, Brian Behlendorf, Jamie Zawinski, Guido van Rossum, and Eric Raymond (who had written an influential article titled "The Cathedral and the Bazaar" in 1997). By now the beneficial effects of open-source software were becoming obvious: collaboration was increasing the ability of the community as a whole to innovate. It was the spirit of collaboration of the San Francisco counterculture transported into the world of computers. The only thing that was missing, in order to cement the open-source community, was a central catalog of all the available free software. SourceForge was founded for that purpose in 1999 by VA Linux (later VA Research and VA Software), a Sunnyvale-based maker of Linux personal computers founded by Larry Augustin and James Vera.

Cyberculture and Cybersociety

The Internet was also beginning to be used for cultural purposes. The online magazine Salon was founded in 1995 in San Francisco by David Talbot, a former editor for the San Francisco Examiner. In april 1999 Salon would purchase the glorious WELL. Slate was founded in 1996 in Seattle by Michael Kinsley, initially under the ownership of Microsoft. Slate and Salon basically transferred the print magazine to the Web, while The Word was trying a completely new format, specifically adapted for the Web. The term "blog" came to be associated with websites run by an individual who published content on a more or less regular basis, the Web equivalent of a television talk show or of a newspaper column. In 1998 Bob Somerby, an op-ed writer for the Baltimore Sun, started "The Daily Howler", the first major political blog. In 1998 Piero Scaruffi, formerly the manager of Olivetti's Artificial Intelligence Center in Cupertino, launched his own website Scaruffi.com to publish articles on music, cinema, politics, etc. In september 1997 Michigan's computer-science student Rob Malda launched Slashdot, a website catering to the audience of the open-source hobbyists ("news for nerds") that provided an index of stories published by other websites or magazines, the first news aggregator.

Anthropologists were excited about the convergence of the virtual communities enabled by cyberspace and the technology of virtual reality that had matured since the pioneering years. In 1994 Ron Britvich in southern California created WebWorld, later renamed AlphaWorld and then again renamed Active Worlds, in which people could communicate, travel and build. Bruce Damer, a former member of the Elixir team in L.A. who, inspired by the hippie communes of the 1960s, had purchased a ranch in the Santa Cruz mountains south of Silicon Valley, established the Contact Consortium in 1994 that in 1996 yielded a three-dimensional virtual-reality environments such as a virtual town (Sherwood Forest) and a virtual university (The U). A virtual-reality project that would have an impact throughout the country was the "Cave". The Electronic Visualization Lab at the University of Illinois Chicago created a "Cave Automatic Virtual Environment", a surround-screen and surround-sound virtual-reality environment in which graphics was projected from behind the walls of a room. The walls themselves are basically just very large screens, and ditto the floor. Many universities would build a Cave system in the next decade.

In 1994 Ron Britvich in southern California created WebWorld, later renamed AlphaWorld, in which people could communicate, travel and build; and in Silicon Valley a former hippy named Bruce Damer formed the Contact Consortium that in 1996 launched 3D virtual-reality environments such as a virtual town and a virtual university. At that time (about 1996) a new kind of "massively multiplayer game" appeared, the MMORPG. The genre was invented in 1996 by the Korean game "Baramue Nara" ("Baram" in the USA), followed by "Meridian 59" (1996), developed by brothers Andrew and Chris Kirmse in Virginia, and "Ultima Online" in 1997, developed by Electronic Arts' game designer Richard Garriott who also coined the term MMORPG. The emphasis here was on having many players at the same time, which indirectly laid the foundations for the software of virtual-reality.

Anthropology of Transience

The real society was characterized by transience at all levels. People were coming and going. Companies were being started and closed. This society of transience created a landscape with no monuments. Not even rich people cared for building monuments. In a place that had no great buildings a company's sign might become a landmark. Buildings were to be destroyed so nobody felt the need to start a building meant to last forever like all civilizations had done before.

This "flatness" of the world created a strange contradiction: there was no visible sign of Silicon Valley's grandeur. The greatness of Silicon Valley was defined by products designed in corporate rooms and built in laboratories that were hidden from the public. People could read about what made Silicon Valley great but there was no exterior manifestation to prove it. The office buildings of Intel, Hewlett-Packard, Oracle and Apple were almost incognito. Dubai was erecting one skyscraper after the other, and Shanghai would soon begin to do the same. The Roman and the British empire had left countless public buildings to celebrate their triumphs. Silicon Valley, by contrast, didn't even have a landmark of the kind that any midsize city in the USA could boast (like Seattle's Space Needle or St Louis' Arch). A tour of Silicon Valley's historical buildings was a tour of garages and offices. Silicon Valley was "inside", not "outside".

It is not that these rich individuals and companies were reluctant to spend their money: they did. In the 1990s Silicon Valley contributed more than one billion dollars to charitable causes. However, the typical contribution was in the form of philanthropy. A company or an individual was more likely to spend a fortune in a "project" (whether helping the poor in Africa or helping a local organization) than in a building. It was all part of the mindset of transience: a project does not depend on a physical location.

Civic life was, in fact, often shaped by philanthropy rather than planned by government. Silicon Valley's very beginning had been shaped by an act of philanthropy and not by an act planned by the government: Stanford University.

Bubbling

The business model that became ubiquitous among dotcom start-ups was about becoming popular and cool, not necessarily making money. The dotcom start-up was after market shares, not profits. In other words, the competition was about getting as many users as possible, typically by providing a service for free, hoping that eventually the large number of users would also bring in revenues. Therefore the typical dotcoms were operating at a net loss. Very few actually had a plan on how to make money. Most of them envisioned a day when they could charge a fee for their services, but the trend was obviously in the opposite direction, with more and more companies waiving fees from their online services. Netscape, for example, made Navigator available for free in january 1998. A different avenue to profitability was taken by the dotcoms that began to sell advertising space on their websites, notably Geocities in may 1997.

In the second half of the 1990s there were more than 100 venture capital firms at work in the Bay Area, the five most prominent being Accel (founded in Palo Alto in 1983 by Jim Swartz and Arthur Patterson, both former partners of Jay Adler's firm in New York), Kleiner Perkins Caufield & Byers, Crosspoint Venture, Sequoia Capital and Hambrecht & Quist. In 1998 Taiwanese computer manufacturer Acer opened Acer Technology Ventures to invest in Silicon Valley start-ups. Venture capitalists were becoming popular heroes too, notably John Doerr, a former Intel salesman, who at Kleiner-Perkins had invested in SUN, Compaq, Lotus, Genentech, Netscape, Amazon and Google.

The new wave of software companies had further changed the demographics of Silicon Valley with a significant injection of brains from Asia. By 1998 Chinese and Indian engineers ran about 25% of Silicon Valley's high-tech businesses, accounting for $16.8 billion in sales and 58,000 jobs.

In november 1998 Netscape surrendered and was purchased by America OnLine (AOL), a bad omen for the dotcoms.


(Copyright © 2010 Piero Scaruffi)

Table of Contents | Timeline of Silicon Valley | A photographic tour
History pages | Editor | Correspondence