Many successes have risen from high-tech failures

Send us your failures. I don't mean the wimpy startups that sputter and die before they get off the ground. I mean the big successes that turn into big craters.

That's one lesson of Fairchild Semiconductor's 50th anniversary last week that we shouldn't forget.

Silicon Valley is built on the rubble of failures, not the spires of its successes, says Paul Saffo, a technology forecaster in Palo Alto, Calif.

Bad managers drive out talented workers. In other places, that can be disastrous. And I don't dismiss all of the pain and suffering that come with events such as the dot-com crash. But the key to our high-tech ecosystem is mobility, which means displaced workers can often find work down the street.

Fairchild's bad management, particularly its failure to reward its chip pioneers, led directly to the founding of Intel in 1968, according to Intel co-founder Gordon Moore.

Dozens more companies spun out of Fairchild, resulting in a diaspora that created tens of thousands of jobs and wealth beyond imagination.

In the 1980s, the folks at the Semiconductor Materials and Equipment International trade group assembled a flow chart that tracked "Fairchildren" who created spinoffs such as Advanced Micro Devices, LSI Logic and Electroglas. Perhaps the most important was the venture firm Kleiner Perkins Caufield & Byers.

A more recent parallel to the Fairchild diaspora is Netscape Communications, which was funded in part by Kleiner Perkins. While Fairchild helped inaugurate the era of the integrated circuit, Netscape ushered in the era of the World Wide Web.

Netscape had more than 5,000 employees when it was purchased by AOL in 1998 for $4.2 billion, but it went into a tailspin and was shut down in July 2003 by its eventual owner, Time Warner.

As it unraveled, Netscape created a big diaspora of its own. Marc Andreessen, who co-founded Netscape (as Mosaic Communications) with Jim Clark in 1994, provides an example of the economic benefits that result from big craters. He founded LoudCloud, which became OpsWare.

Last month, Hewlett-Packard paid $1.6 billion for Andreessen's company. He also has another company, Ning, and invested in social-networking startup Digg.

Other companies founded by "Netscapees" are the Barksdale Group, Epinions, Backflip, Friendster, Tellme, Kontiki, Hopelink, OurPictures, Viralon, Meernet, RxCentric, Medicalogic, Cloudmark, Xoopit, Responsys, the nonprofit Mozilla Foundation (which gave us the open source Firefox browser) and Ram Shriram's Sherpalo Ventures, which invested in Google.

Saffo argues that you get better economic results out of the diasporas of failures.

Big companies such as Intel, Microsoft and Google also shed employees who start new companies, but overall these big companies do a decent job of retaining their employees. It's the explosions where everybody gets laid off that often result in more new companies.

Andreessen, however, says that the "failure fetish" in Silicon Valley is silliness. He noted that Fairchild was successful in its day and Netscape was profitable when it was sold. Only in the aftermath were they branded failures.

"Other companies then spun off and were inspired by it, riding many of the ideas that Netscape originated," Andreessen said in an e-mail.

"Same is now true of Google, of course. The number of new companies coming out of Google is growing and is going to be awe-inspiring. To counter Paul [Saffo], who to my knowledge has never put himself in a position where he was exposed to failure by trying to build something, I'd quote Jim Clark: 'They say that in Silicon Valley failure is OK, but they're wrong — failure is never OK.' "

Which of them is right? Probably both. Many people, including employees and shareholders, hope companies such as Google and Yahoo hold together rather than fall apart for the sake of Silicon Valley's future. But we all might benefit in the long run even if they eventually collapse.

Dean Takahashi is a technology columnist at the San Jose Mercury News.