The New New Thing: A Silicon Valley Story

by Michael Lewis

  Actually, the venture capitalists were just the first item on Clark’s list of what needed fixing in American capitalism. One day in late 1984, just as Silicon Graphics was preparing to ship its first machines, Glenn Mueller called a meeting in his office. It had taken two years from the founding of the company to its first product. The new computers were not easy to sell, at least not at first. SGI delivered far more computer than most people could handle. A cheap one cost seventy thousand dollars. The founding engineer Mark Grossman recalls “endless meetings and debates about how to squeeze what we had into a smaller box. Basically we couldn’t figure out how to make it cheap.” But even a cheap one was complicated to use, even for someone who knew his way around a computer. People with modest programming skills in front of a Silicon Graphics work station sometimes felt a bit like a man in a Lamborghini on a two-lane road.

  The result was a struggle between the engineers, who built ever niftier computers, and the managers, who had been brought in to market them. The managers wanted the engineers to make their products simpler to use. The engineers, like their leader, thought their managers were all idiots. “They were promising stuff we couldn’t deliver,” says the engineer Tom Davis, “and they didn’t understand what we had. Of course, fifteen more years of experience showed me that this is the normal situation.” At the time, however, the engineers were outraged. The managers would tell them to do one thing, the engineers would do the opposite, and, no matter how it turned out, the engineers thought they were right until someone proved them wrong. “It was like an especially contentious academic community,” recalls Tom Jermoluk. “You were encouraged to shred each other’s stuff. And everyone had an opinion about everything.”

  Glenn Mueller’s meeting was an attempt to sort out the problem and to transform Jim Clark’s company into a stable, well-adjusted, lasting institution. At the meeting were Clark, Vern Anderson, the CEO Mueller had told Clark he needed to have in order for the Mayfield Fund to supply him venture capital, a few other senior executives, and the venture capitalists bankrolling the enterprise. Dick Kramlich of New Enterprise Associates was the newest of these. It was Kramlich’s first encounter with Clark’s corporate culture, and it was the most contentious he’d ever seen. “People were yelling and screaming at each other, over the most petty things—not even business things,” he recalls. “I waited until it was five o’clock, and then I got out. I told them that I really didn’t want to go through any more of this and to call me when they’d worked out their emotional problems.”

  After the meeting Vern Anderson stepped down as CEO. This should have surprised no one: after Mueller had told him that, to obtain venture capital, he needed a CEO, Clark had literally walked outside, found Anderson on the street in Palo Alto, and asked him if he wanted the job. (“I bargained for a temperamental handful,” says Anderson.) Clark never had any serious ambition to manage the company himself; even then he knew he’d be poorly suited for the job. He couldn’t stand the small steps, the details, the tedium. He thought they should bring in some person to handle the details while he made sure that his beloved engineers kept on building the machine of the future. He didn’t fully appreciate just how much power he had to relinquish for the detail person to do his job well. Soon enough, he found out. Mueller found another man to run the company. His name was Ed McCracken, a highly regarded vice president at Hewlett-Packard. When Mueller told Clark about McCracken, and explained that Clark would have to give up the illusion that he still controlled Silicon Graphics, Clark broke down and cried. “Jim always had a sense that we weren’t doing things right,” said Anderson. “He knew in his heart that we needed someone like Ed.”

  With the understanding that he would enjoy complete control, Ed McCracken left his safe comfortable job at Hewlett-Packard and took a risky one at Silicon Graphics. He and Clark taken together were proof of the limits of physiognomy. To look at them, they could have been brothers—tall and blond with wire-rimmed spectacles perched on surprisingly delicate features. To watch them in action, you would think they came from different planets. Hewlett-Packard was the closet thing in the Valley to an enormous gray corporation, and Ed McCracken was the Valley’s version of the Organization Man. He had mastered the unnaturally sincere tone of voice of the Professional Man. When he wished to indicate seriousness, he dropped his chin down into his throat. When he took you aside to have a word, he looked and sounded as if he was giving a speech to an audience of a thousand people. To stress his points, which were rarely pointed or stressful, he’d press his thumb against his index finger as if he had just caught a fly by its wing. He wore suits. He hated strife. He loved consensus, or at any rate the idea of it.

  For many of the founding engineers it was the first experience with a Serious American Executive, and with the vaguely phony emotional postures that seem to be, for whatever reason, necessary for the success of the leader of a large organization. “He came off as being awkward and manipulative,” says Mark Grossman. “There was no idle chitchat.” “He had a weird way of using silence in conversation,” says Kurt Akeley. “Ed was one of those people who likes to design questionnaires for others to fill out,” a third engineer says. McCracken liked to deal with people indirectly, through intermediaries. As yet another of the engineers who left Stanford with Clark to found Silicon Graphics puts it, “Ed had this phrase for all our problems. He said we were ‘highly oppositional.’” It was that phrase—“highly oppositional”—that stuck in people’s minds.

  In late 1984 McCracken took over the company from Clark, who stayed on as chairman, whatever that meant. He looked at the books and discovered that the company had only seven million dollars left, and was running through money at the rate of two million a month, which meant it had exactly three and a half more months to run. He cast one long look at the contentious engineers and their warring opinions. He decided that Jim Clark’s corporate culture needed more than a chief executive officer. It needed a therapist.

  McCracken hired a corporate psychologist. The psychologist together with forty Silicon Graphics employees, including the senior engineers and their leader, Jim Clark, retreated for three days to a resort not far outside of the Valley. There they submitted to a battery of psychological tests designed to make them better Organization Men.

  Before the retreat Clark and his engineers were required to find two people to fill out psychological evaluations on their behalf, and mail them to the corporate psychologist, confidentially. Once they’d gathered together at the resort, the engineers filled out two more forms. The first was designed to classify them into one of four psychological types. These types probably said more about psychology as filtered through the American business mind than about the engineers. At any rate, the tests broke the world down into introverts and extroverts, then into right-brained versions and left-brained versions. The four types were given the following names:

  introvert + right brained = “supportive”

  introvert + left brained = “analytical”

  extrovert + right brained = “promoter”

  extrovert + left brained = “controller”

  Depending on the degree to which you possessed these qualities, you were classified as a “strong” or “weak” version of your type. Ed McCracken was a “weak analytical.” Jim Clark was a “strong controller.” (As Clark recalls it, “The psychologist determined that everyone else on the executive committee was passive aggressive and I was just aggressive.”) The engineers Clark had hired all scored pretty much the same as he did. Of the thirty engineers who took the test, only two registered as “supportives” and two more were “analyticals.” The rest were “controllers.”

  The second test required Clark’s engineers to answer 250 questions designed to further parse their psyches. The results were laid out on a pie chart with twelve slices. Again, each slice of pie represented a different personality type. The psychologist had odd little names for these, too: “humanistic,” “optimistic,” “vague,” and so on. When t
he chart was completely filled in, eleven of the categories remained empty. The names of all the engineers were crammed into a single sliver of pie. “Highly oppositional,” it was labeled. (That is when it occurred to the engineers where their new CEO got his lingo.)

  Afterward each engineer was required to explain his psychological profile. That is where tempers finally flared: the charge that they were “highly oppositional” actually pleased a lot of Clark’s engineers. To this day it pleases them. For instance, Tom Davis, one of the seven founding engineers who had followed Clark out of Stanford, says, “Their definition of oppositional was basically someone who stuck to his guns. That’s clearly a bad trait if you stick to your guns when you’re clearly wrong, but at the time it was probably the best group of engineers I’d ever seen assembled, and everyone knew that they were almost always right. From my point of view, they had exactly the right qualities to produce great products—they were almost always right, and if you disagreed, you had a big argument to change their minds. But if you were right, after some struggle they almost always would. If somebody insists that 2 + 2 = 5, I’ll never back down, no matter how bad it makes them feel. I think that the shrink thought that you should. In nonengineering areas where things are a lot fuzzier, giving in more often is clearly a reasonable thing to do.”

  The psychologist ran the subsequent three-day drill like an AA meeting. He asked each of Clark’s engineers to stand up, explain how he scored on his test, and tell everyone else how he planned to change. The second test, which the shrink kept calling “the instrument,” had reduced each engineer’s personality to a three-dimensional “shape,” which turned out to be a piece of paper with warts and divots and ridges. One of the company’s founders grabbed his shape and led off. He rose and said that he had spent his whole life getting himself to the point where he could tell someone he was an idiot and that he had no intention of changing that now. Did anyone want to fight about it? The other engineers cheered. Another engineer said that he had figured out how the questions related to the profile, and he wanted to take the test again so he could get a perfect score. The shrink insisted that there was no such thing as a perfect score. “He went on and on and on how there is no right answer,” says Rocky Rhodes. “How this was just for us to get to know each other better. That there were no bad people and no bad shapes. That there was nothing judgmental about it. It was just a shape. You sat in front of the group on a little chair and held up your shape. One by one. Then Greg Chesson walks up and the shrink almost gasps. He says, ‘Wow! that’s perfect!’”

  One way of viewing Silicon Graphics in the mid-1980s is as an answer to a pair of questions. The first question was: If an extraordinarily willful human being with great technical aptitude is permitted to create a large business organization, how will that organization behave? By 1984 everyone understood that it would behave like Jim Clark, which is to say that it would behave as no big, successful American company had ever behaved. It would be a loose collection of argumentative, brilliant, bullheaded engineers who might or might not make money but almost certainly would build something wonderful.

  The second question was: How would such a place ever grow old? The answer was: painfully.

  After the retreat Ed McCracken quickly set about making his company less like Jim Clark. This is just how it always went with one of these new Silicon Valley hardware companies: once it showed promise, it ditched its visionary founder, who everyone deep down thought was a psycho anyway, and became a sane, ordinary place. With the support of Glenn Mueller and the other venture capitalists on the board of directors, McCracken brought in layer upon layer of people more like him: indirect, managerial, diplomatic, politically minded. These people could never build the machines of the future, but they could sell the machines of the present. And they did this very well. For the next six years Silicon Graphics was perhaps the most successful company in Silicon Valley. The stock rose from three dollars a share to more than thirty dollars a share. The company grew from two hundred employees to more than six thousand. The annual revenues swelled from a few million to billions. Therapy aside, the company remained the most desirable place to work if you were a certain kind of computer cowboy who wanted to live on the edge of the technology. The technology Clark and his brilliant engineers had invented turned out to be just what the world hungered for.

  At the same time McCracken dealt Clark out of his own business. From McCracken’s point of view Clark was just wildly disruptive. He’d wander around the company stirring up all manner of trouble and cause all the engineers to become even more highly oppositional than they were on their own. Senior engineers turned up in his office and told him that the chairman (Clark) had persuaded them they’d be better off quitting Silicon Graphics and starting their own businesses, and McCracken would have to spend hours talking them into staying. But he found outlets for his own frustration. When McCracken joined the company in 1985, Glenn Mueller had assured both him and Clark that they would be paid the same amounts each year. In late 1989 McCracken called Mueller, who now chaired the executive compensation committee, to complain about this rule. Mueller phoned the other members of the committee and changed the policy. The next year McCracken was paid hundreds of thousands of dollars more than Clark. Not long after that Clark ceased to be granted stock options in his company. He was the only member of the executive committee so treated.

  Given what happened later, it is easy to feel sorry for Ed McCracken. At the time it was harder. Having established himself as the captain of the ship, he was doing what captains of industry have done since they were invented. He was transforming himself into an Important Person. He chaired conferences of the future of American industry. He hobnobbed with U.S. senators and testified before Congress. He encouraged the 1992 Clinton campaign to use Silicon Graphics as a backdrop for an important speech by Clinton on economic policy—only to find, too late, that Clark had decorated the halls that Clinton strode along with the Fortune magazine cover that featured Clark. Once Clinton was elected president, McCracken became a regular at White House dinners. Such highfalutin behavior might not sit well with Clark’s engineers, who accused him of not paying attention to the business, but it had a purpose: it created an air of permanence not merely about Ed McCracken but also about Silicon Graphics. McCracken was trying to build something enduring.

  Order and hierarchy were essential to this process. McCracken was uncomfortable with even the most trivial challenges to his authority—which is to say that he lived in a state of perpetual discomfort. The engineers, inspired by Clark, were constantly challenging his authority.

  The practical jokes were a case in point. An engineer’s idea of a joke is a practical joke, perhaps because a practical joke, unlike the less practical kind, needs to be designed. It requires the jokester to build the contraption to ensnare his victim. Silicon Graphics engineers loved their practical jokes. And they loved their chairman for the practical jokes he played, especially the ones he played on the boss. For instance, Clark had been struck by the inability of his teenage daughter to recall McCracken’s name. She kept calling him Ed McMuffin. So one day Clark bought one of the thin name plaques that were affixed to the doors in the executive suite, stenciled with the name ED MCMUFFIN. He replaced McCracken’s name plaque and waited for the response. For three days the new CEO walked in and out of an office with a door marked ED MCMUFFIN. The engineers would sneak up from their labs to watch him do it, then run back giggling to their work.

  Everyone waited for the glorious moment when the victim of an engineer’s practical joke realized what was happening and blushed and smiled and stammered and told everyone what a good joke it had been. It never came. On the fourth morning of the joke, the ED MCMUFFIN plaque was gone. McCracken never said a word about it. Along with a lot of other pranks Clark played on McCracken, it festered in company lore. One day, years later, at a meeting filled with Clark’s engineers and McCracken’s managers, Clark told the story of how for three whole days McCracken
was McMuffin. McCracken reddened, the managers swallowed their laughter. “It was like someone had played a joke on the dictator and you weren’t allowed to laugh,” says one of the engineers who was there.

  Ed McCracken was Jim Clark’s first intimate encounter with the American professional management class, and its politics. From it was born his conviction that there was a whole layer of people in American business who called themselves managers who were in fact designed to screw up his plans. Life was unfair: Jim Clark wasn’t the first person ever to feel that way. What is more surprising is that a man who grew up, as he put it, “sitting in a large pile of shit” would be as convinced as Clark was that life should be fair. He was so convinced of this that he set out to correct the problem, and to take what rightfully belonged to him and to his engineers. Of course, it took people a while to realize that the new rule in Silicon Valley was that Jim Clark always got his way. It took ten years, to be exact.

  Right through the golden years of Silicon Graphics, as McCracken took over the company and made it his own, Clark fought a civil war. He persuaded his fellow engineers that they should feel as mistreated as he did. “Tom [Davis, another founding engineer] and I would go in and talk to Jim,” recalls Rocky Rhodes, who had left Stanford with Clark to create SGI. “And we’d learn how shitty life was. When we’d leave his office, we’d say, ‘Yeah, I guess we really have been mistreated. I guess we should have been paid millions of more dollars.’ Before that, I had no idea. I walked into Silicon Valley with $500 a month from the government. I was a twenty-seven-year-old with essentially no computer experience. Now I had stock options that Jim said would one day be worth a million dollars. Plus I now had a big salary. Before I learned from Jim that I’d been mistreated, I was quite pleased.”

  A whispering campaign wasn’t Clark’s style. Anything he said to his engineers he also said directly to McCracken. Dick Kramlich recalls a meeting between Clark and McCracken, attended by himself and Glenn Mueller. “Jim just ripped Ed apart. He explained to Ed everything that was wrong with his character. Jim can be truly brutal—unfairly so. And that day he just took Ed apart into pieces. By the time he was finished, Ed was crying. No one knew what to say.” Clark’s friends who did not know McCracken came to believe the man’s name was Fucking Ed McCracken. “Fucking Ed McCracken,” Clark would say, “he may have helped to stabilize the company, but now he’s destroying it. He can’t see what’s happening.”