A stunning ignorance of mass tastes was a common problem in high technology. When a brilliant engineer dreamed up a product, he tended to build the sort of things only a brilliant engineer would appreciate. Typically, he overestimated the average person’s willingness to learn how to use some new machine and underestimated the cost of making the machine. When you page through the history of computing, you find a lot of very weird examples of just this. In a warehouse behind the Computer Museum just north of San Jose, for instance, stands a gleaming red four-feet-high computer built by Honeywell in the mid-1960s. The Kitchen Computer, it was called. The Organization Man’s housewife was meant to program her recipes into it. “If only she could cook as well as Honeywell can compute,” read the ad in the Neiman Marcus catalog. The computer was premised on the dubious assumptions that every American housewife would (a) want a massive computer in her kitchen and (b) know how to program it. Neiman Marcus failed to sell a single unit.
The idea Clark became wedded to, albeit briefly, was that the computer would become the most important household appliance. Information appliance, was one term used to describe it back then. The company that built the first information appliance would sit in the middle of all human communication; it would be the McDonald’s of information. It could play the same role on the television that Microsoft played on the PC. “The telecomputer was a direct result of the frustration I felt watching Silicon Graphics continually fall behind the PC in market share,” Clark says. “I was trying to do an underbelly thing with Microsoft—come in under their monopoly and take it away.” The telecomputer would be the most fabulous toll booth ever built. But before it happened, a question needed answering: Why would people rush out to buy a telecomputer? What would a telecomputer do that people simply could not live without?
Clark had no great hope that Americans wanted their computers to educate themselves. He assumed they wanted their computers to play Nintendo and otherwise divert themselves from the poverty of their existences. The answer he finally came up with was that people wanted to watch any movies they pleased, whenever they pleased. The telecomputer would be many things, but at first blush it would be a virtual VCR.
Of course, Ed McCracken wanted nothing to do with a virtual VCR, at least initially. It was just another flaky idea from his flaky chairman, who refused to leave him to run the company alone. And so Clark found another outlet for his ambition, the media. Once he’d recovered from his motorcycle accident, he shared his thoughts freely with journalists. “I might not have any power at Silicon Graphics,” he says, “but they didn’t know that. I was still called the chairman. And most journalists think a chairman of a big company is an important person.” He knew perfectly well that McCracken disapproved of his talking to the press. But hype had its own wonderful generative power: the more he talked about his telecomputer, the more people wanted to hear about it. It was almost as if by talking about the telecomputer he made it happen. “I would go out, and I would just say all this shit to reporters,” he recalls. “And they’d print it! And people inside SGI started to talk about it. And I thought, ‘Fuck Ed McCracken. I can say whatever I want. And by God if I go out and talk about it enough, Ed won’t have any choice but to build it.’”
Clark’s opinions found their final published form in 1992 at the annual trade show for the computer graphics industry, called Siggraph. Clark delivered as a speech the paper he’d written while he was laid up in bed recovering from his motorcycle accident. The paper described “the consumer’s computer.” Clark guessed it could be built in “two to three years.” He explained that, although computer memory in 1992 made such a device too costly to be mass-marketed, computer memory in 1995 would be a different story. He outlined the basic technologies—digital audio, digital video, transmission-reception decoupling, resolution decoupling—all of which existed in one form or another. But he dealt with the technical side of things in only the sketchiest form. Mostly his paper was a bit of political propaganda, aimed at raising the heat on McCracken. In Silicon Valley political propaganda took the form of futurology. “Over the next four to five years,” Clark told his audience,
unprecedented change will occur in the computer, telecommunications and television industries as Multi-Media technology enters the home via the telecomputer…. The present “local loop” of the telephone system and the “cable franchise” for television will become one Multi-Media server loop. Each loop will represent tens of thousands of clients, each using a telecomputer. In the loop will be high-speed computer systems for serving audio and movies on demand, virtual reality games, digital forms of daily newspapers, weekly and monthly magazines, libraries, encyclopedias and interactive books. In time, all media will be available in dynamic form.
He was wrong about all of this, at least in his timing. He was off and running down a dark tunnel that ran directly into a brick wall. But at least he was running.
Clark’s paper attracted a lot of attention from everywhere except Silicon Graphics’ boardroom. The cardboard box in Clark’s guest room contains a tall stack of fan mail and his responses. For instance, on August 4, 1992, he replied to a letter from Lance Glasser at the U.S. Department of Defense, who agreed with Clark and wanted to know why he didn’t just build his new machine. “Silicon Graphics Management considers pursuit of Digital TV a distraction,” Clark wrote. “I am the driving force but…left to its own, SGI will not pursue this for at least four or five years…. I believe the solution is to form a new company.”
By far the most important letter he had came from Jim Chiddix. Chiddix was the chief technology officer at Time Warner Cable. Chiddix told Clark that he had heard his talk and that Time Warner shared his interest in the telecomputer. He agreed with Clark that now, at last, the world was ready for the new machine. Time Warner had just decided to yank out all of its copper cables and replace them with fiber-optic cables, which transmitted data much more efficiently. In other words, the infrastructure would be laid for a telecomputer to traffic in moving images. More to the point, Time Warner was willing to pay someone to build a telecomputer. Chiddix suspected Clark was just the man for the job.
The deeper Chiddix dug into the problem, the more certain he became that Clark was the only man for the job. Or, at least, the company Clark had founded was the only company for the job. In late 1992 Chiddix toured the companies that might conceivably build the black box he needed: Microsoft, Oracle, Sun Microsystems, Silicon Graphics. They contained the biggest egos and brightest minds in technology. They sold themselves as hard as they could to Chiddix, and yet Chiddix came away with a very clear opinion: the engineers Jim Clark had brought together were in a class by themselves. “It was clear that Clark himself had no power in the company,” says Chiddix. “But it was also clear that if anyone could build an interactive television it was Silicon Graphics. Everyone at all these companies was smart. But the engineers at SGI, they were the real cowboys.”
In October 1992 Clark finally made his pitch for his telecomputer to the Silicon Graphics board of directors. He opened with a slide that read “How We Can Be to Entertainment Computing What Microsoft Is to Productivity Computing.” He laid out his view of the market, which he thought could be as big as 10 billion dollars a year. He discussed potential competitors—Sun Microsystems, IBM, DEC. And then he pulled a kind of bait and switch: rather than create the telecomputer inside SGI, he suggested, SGI should finance a new company to do it. In exchange for money and technical support, SGI would be granted a large equity stake in the new enterprise, which Clark would control. No more Fucking Ed McCracken. Clark called his new company an Entertainment Computer Company. As a kicker he mentioned that Time Warner was willing to pay for the telecomputer to be built.
Not long after Clark’s presentation to the SGI board, Pavan Nigam was called in to see Tom Jermoluk—T. J. as he was called. Improbably, T. J. had been able to win Clark’s friendship without alienating Ed McCracken. While McCracken was off playing corporate statesman, T.J. ran the compan
y; and it was through T. J. that Clark exercised what influence he had on SGI. “When I got the call that T. J. wanted to meet with me, I thought he wanted to complain about the bugs in our software,” says Pavan. “He was always complaining about our bugs. So I brought a list of the bugs and what we were going to do about them.” Instead T. J. had an offer for Pavan. “How would you like to build the world’s first interactive television set?” he asked. He explained that Ed McCracken had come around to Jim Clark’s way of thinking, but did not like the idea of Clark’s having his own company. Time Warner was willing to pay SGI $30 million to build what it was calling an interactive television (ITV). Whatever SGI built would be installed in four thousand homes that Time Warner was wiring for the occasion in Orlando, Florida. And on the recommendation of Jim Clark it wanted the thirty-three-year-old Pavan Nigam to run the project.
The minute Time Warner announced that it had chosen Silicon Graphics, just about every big company that had anything to do with information or entertainment leaped into action and hired its own engineers to build them a telecomputer. “There was a mad scramble,” recalls Chiddix. AT&T and Viacom announced that they were building their interactive television pilot in Castro Valley, California. TCI and Microsoft announced their pilot in Seattle. The phone company U.S. West, the computer company DEC, and the computer animation company 3DO announced their own project in Omaha, Nebraska. Oracle announced its vague intentions to get involved. When you added it all up, it implicated thousands of people and hundreds of millions of dollars. The telecomputer was the first big step in a new direction. Very large companies, and a lot of important investors, became swept up in the idea of an intelligent home appliance. They bought into Clark’s notion that people would shop, communicate, and amuse themselves through their televisions.
It took about three days for Ed McCracken to make the telecomputer his own and for the Orlando project to become the place to work at Silicon Graphics. Before long, Michael Douglas and Arnold Schwarzenegger had dropped into SGI’s offices to hear about the technology that would change their business. The CEO of Time Warner, Gerald Levin, rolled up to the front door of Silicon Graphics in the longest limousines anyone had ever seen. The scramble to get assigned to the project became so vicious that McCracken sent out an e-mail warning engineers about sabotaging each other’s careers. At the same time he let Pavan Nigam know that the project was critical to the future of the company. “‘Spend what you need to spend; the thing must work,’ was what Ed said,” recalls Pavan. “All of a sudden everyone thought ITV was the future of the company. All of us thought it was going to be the next big thing.” Pavan believed that “a thousand people don’t build anything; if you need to build something really complicated really fast, you hire fifty of the smartest people you can find.” That is exactly what he did. He started by hiring Kittu Kolluri. The telecomputer had been entrusted to Clark’s private software tutors.
To the engineers the main appeal of the telecomputer was its complexity. This was in itself an ominous sign. It’s a good rule of the technology business that the more intellectually appealing a machine, the less likely anyone will pay for it. It was not trivial to rig a system so that the fat guy on the sofa with the beer in his hand no longer needed to drive to Blockbuster to get the movie he wanted. Every fat guy required his own video stream, and a single video stream contained a huge quantity of data. The computer needed to process information more quickly than information had ever been processed. “The Orlando project turned out to be the most aggressive supercomputer project ever put together in the history of the world,” says Jim Barton, one of the senior engineers on the job. “As a pure engineering challenge,” says Marco Framba, another engineer on the project, “it was more difficult than any I’ve ever heard of.” Jim Chiddix at Time Warner says, “It required more lines of computer code than was required to put a man on the moon.”
Pavan and Kittu were told that they had eighteen months to build the new machine. They said it would take them twenty-four. It took them twenty-three. On December 12, 1994, the press piled into Orlando for the demonstration. Other companies had spent more money and employed more engineers, but none of them got a system that worked. Time Warner’s Gerald Levin sat on the stage with Ed McCracken and a couple of other suits. Their presence was a kind of assurance that technology moved in a stable and predictable manner, the way the men in suits directed it to move. “Never in the history of telecommunications,” Levin said, “has a medium of such complexity been designed, developed, and given its debut in such a concentrated space of time.” Then he picked up the remote control and pressed a button and…it worked! (“Like a charm,” says Pavan.) The engineers received standing ovations. Levin cut the tape and announced that Time Warner intended to spend five billion dollars to deploy this fantastic new technology. All Americans would one day live even more of their lives than they currently did through their television sets. The new technology would change the way people shopped, voted, worked, and loved.
Ed McCracken followed Gerald Levin onstage. Without once mentioning Jim Clark, he explained to the audience how his team at Silicon Graphics was now poised to control the future.
From their seats in the crowd Pavan Nigam and Kittu Kolluri looked up, and they knew it wasn’t true. They had spent the last eighteen months pulling off one of the great engineering feats of the century, and they had nothing to show for it but black boxes that cost five grand apiece. Few would pay that much for the technology, and the value of the technology depended in part on a lot of people’s owning it. A VCR was as valuable if one person owned it as it was if a hundred million did. But a television that interacted with other televisions demanded an audience. “ITV was one large academic exercise,” says Kittu. “We solved a problem that was not that important to many people. I’m not terribly proud of what we accomplished. We started out to change the world. Did we change the world? No fucking way.”
There was another bad sign—though just how bad they didn’t fully appreciate. Jim Clark was gone. He’d up and left the company he’d created and said he was going to start another. “It was good that Jim left,” says Ed McCracken, in what must rank as one of the troughs in a career of understatement. “No big company could accommodate such financial ambition.” Clark’s lawyer had advised him to quit bitching and moaning about McCracken for six months, and then walk into a board meeting and quit. And that’s what he did. Publicly, he said he was leaving on the best of terms. Privately, he told friends that he was going to show Fucking Ed McCracken that it was he, not McCracken, who had been responsible for the success of Silicon Graphics. He told the engineers who had helped him create SGI that he was going to get rich. Rocky Rhodes recalls him stopping by to talk about his future plans: “Jim said right out, ‘I’m going to make $100 million.’ And I said, ‘That’s great, Jim,’ and also something about how pleased I was with the way SGI had worked out. And then I said, ‘But even if it hadn’t worked out, I’ve had a great time, met great people, learned a lot.’ And Jim got mad. He said, ‘No! If I go and do this next thing and don’t make $100 million, it’ll be a failure. I’ll be a failure.’ And that’s kind of the way I remember leaving it with him.”
Pavan and Kittu knew most of this, and it troubled them—but not nearly as much as it should have. They still believed that economic history was moved along slowly, by big companies like Silicon Graphics.
By the time Clark walked out of SGI in early 1994, the list of grievances he was nursing had grown too long for anyone but him to keep track of. First Glenn Mueller screws him. Then Ed McCracken takes full credit for things he was at best partially responsible for. Then Mueller and the other board members conspire to shut him out of the company while McCracken fucks it up. Then Mueller and McCracken cut him out of the stock options pool. Then he puts the company right out in front of the future—arranges for Time Warner to pay for them to be there!—and is then promptly shut out of all further discussion. The minute Time Warner’s limousines turned up at
Silicon Graphics, Ed McCracken had taken over the project. There was so much ill will between him and McCracken that when he found his image missing from a picture in the SGI annual report of the founding engineers, he assumed McCracken had ordered it removed.
Success was his chosen form of revenge. Clark was intent on inventing a new role for himself that would not allow the Muellers and McCrackens to take advantage of him. In retrospect, there are many ways to describe this new role. But maybe the simplest is the way Clark described it to himself: the guy who finds the new new thing and makes it happen wins. The engineers who help him to do it finish second. The financiers and the corporate statesmen, the sucker fish of economic growth, finish a distant third. Clark intended to sit on top of Silicon Valley as surely as Microsoft sat on top of the personal computer. Of course, he was not the first engineer with a taste for money and power. He was just the one who finally said, “Now’s the time to take it.”
But that version of events is misleadingly neat. Clark didn’t conceptualize his new role: he groped for it. He had an animal desire to have what he wanted and not to have what he did not want. He wanted Silicon Valley to be even more suited than it already was to his talent for anarchy. He wanted to harness the forces of creation and destruction. He did not want to manage a large company. He did not even want to be a venture capitalist who vetted thousands of business plans, backed dozens of companies, and then sat back with Olympian detachment and hoped that a few became big. He wanted to create the company that invented the future. Once he’d done that, he wanted to do it again and again and again and again. For his services he wanted to be treated better, and paid more, than anyone else.
The New New Thing: A Silicon Valley Story Page 8