__ For WebTV's hyperactive founder, never growing up is the best revenge. __
Armed with a fistful of colored markers, Steve Perlman launches himself at the whiteboard - again. Assembled behind him in the Mountain View, California, offices of WebTV, a dozen employees and consultants issue a low, collective groan as he furiously starts drawing little green houses, satellite dishes, box-shaped broadcast TV stations, and programming companies, all interconnected to a cartoon depicting the Internet as a cloud. Some people let their pens drop. Others, who only seconds before were perched attentively, fall back on a soft black leather sofa and squish into place for the long haul. This could take a while.
Perlman, facing the board, is oblivious. In a matter of seconds he's gone ... away. He scribbles and chats with himself, erasing things, adding buildings, moving the cloud, using a red marker to make dotted lines that seem to connect everything. He's sketching out the broadband future of interactive TV, and about four minutes into it he gets so excited that one listener jokingly accuses him of getting high on fumes from the markers. Hearing the wisecrack, Perlman turns around and jams a green marker up his nose. Then he makes a more sober attempt to explain the meaning of his illustration, but to no avail. He alone sees a vision in the scrawl, though everybody else sincerely wants to see it too.
The moment is typical of Perlman. He can be effusive, playful, self-absorbed, manic, and silly all at once - but whatever's going on, you'd better listen, because he just might say something priceless amid the noise. After all, it was Perlman's 1995 startup, WebTV Networks - and not Time Warner, TCI, or US West - that first made good on the promise of interactive television. So it stands to reason that he might be the first to divine how the field, which at this point is crowded with a couple dozen companies worldwide, will play out in the years ahead.
For Perlman, part of the answer is obvious: Convergence and increased interactivity will prevail. "The broadcast-TV experience as we know it today will be viewed as an anachronism of the 20th century," he declares. "Within a decade, almost all TV programming - essentially everything except for live events - will be available on demand and combined with interactive content. All channels will be virtual, and, just like with the Web, you'll have access to what you want, when you want it. Prime time as we know it will cease to exist."
By now that rant is somewhat familiar, and Perlman's success is a big reason why. WebTV has undeniably delivered what other companies with huge R&D budgets did not: subscription-based interactive TV that people are willing to pay for. The company manufactures a $99 set-top box that is easier to use than a VCR, and for $19.95 a month it has pushed the Internet - and with it email and ecommerce - down the tubes of 800,000 subscribers, 70 percent of whom do not own a PC. That fact is especially notable given that much of the company's growth was made possible by the marketing power of Microsoft, a company whose fortunes are tied to the PC.
In April 1997, Microsoft decided that if you can't beat 'em, buy 'em, and spent nearly half a billion for WebTV, a move that made Perlman a wealthy man. Despite differences in management style and company culture, the acquisition has been good for both sides. Microsoft, which had almost no presence in America's living rooms, bought a dominant position in the emerging interactive-TV market, as well as a chance to counter the conventional wisdom that it is too rigid to channel the energies of upstarts like Perlman and his cofounders. WebTV, in turn, received lots of cash - it couldn't have survived without it - along with some priceless credibility.
Perlman will enthusiastically tell you that several innovations have sprung from the alliance, including the WebTV/EchoStar DishPlayer announced in January. This more advanced WebTV system makes it possible to search for and record programming from multiple networks and to create the personally tailored "virtual channels" that Perlman likes to gush about. Perlman will also talk endlessly about Redmond South, the Microsoft campus under construction in Mountain View, which will be home to WebTV, Hotmail, and Microsoft's other Silicon Valley properties. He'll tell you there's no better company in the world than Microsoft to deliver his technology to the masses, or to have behind him when powerhouses like AOL and @Home Networks enter his market, as they have announced they will.
For all that, Perlman has appeared to be restless for quite a while now: In fact, by the time you read this, he may have left Microsoft. According to sources there and at WebTV, he quietly tendered his resignation to the company back in March, when Microsoft president Steve Ballmer announced a reorganization that put two Redmond-trained executives between Perlman and Ballmer - that is, in charge of Perlman and WebTV.
Not that anyone in Redmond seemed eager to lose Perlman. Rumor had it that Microsoft executives offered him millions more to stay and even create his own job - a visionary-in-residence slot. For his part, Perlman was keeping mum, but he wasn't denying the rumors either. Asked about his possible departure, he simply offered quips.
"The only things I will do," he said, "is stuff that I am interested in doing."
Opinion split on whether Perlman would stay or go. Those who considered him good-as-gone pointed out that he has walked out on (or struck out with) three other companies during his storied and turbulent career: Apple, General Magic, and Catapult. Others said he would never leave his "baby."
"I don't believe he's going to leave," said Bruce Leak, WebTV's cofounder. "I just don't see that happening. For a guy who might leave, I got so many emails from him last night, there's not enough time in the day to answer them all."
One compelling alternative viewpoint went like so: It doesn't really matter. WebTV is at the stage where its leadership needs to shift focus away from innovation (Perlman's all-consuming passion) and toward market performance (Microsoft's specialty). As such, it might be time Perlman moved on, and he has lately seemed to be bound for something new. He's started building what may be his next office - a sort of Fortress of Solitude, a state-of-the-art digital film production studio in San Francisco's South of Market district.
Ultimately, whether Perlman stays or goes, the current moment is mostly about him doing what he does best: being Perlman. He is busily living up, once again, to his self-cultivated image as the Silicon Valley savant - the brilliant, mercurial programmer/hacker who does things his own way, the whimsical, unyielding pain in the ass who is always lurching toward his next breakthrough.
"Do you like Happy Donuts?"
It's a bright afternoon in downtown San Francisco, and Perlman, dressed like a hiker in fleece and boots, is feeling demonstrably chipper. We've spent the morning roaming the construction site where his film studio is being built. Now we're just tooling around; right before spotting the Happy Donuts, he was giving a disquisition on his Land Rover's built-in GPS. I know there is only one right answer to his question, but two thoughts form in my mind: First, we just ate lunch. Second, wealth apparently does not alter the cravings of a programmer.
Happy Donuts lacks any Starbucks pretense; it's like a low-rent Dunkin'. Flies swirl in the doorways of its tacky orange-and-brown storefront; greasy windows cast the place in a weird, dirty aquarium light. The videogames inside aren't new or classic - just over.
Perlman doesn't seem to notice. He strides briskly to the counter; it's clear he knows the doughnuts in the case like he knows the connectors on a circuit board. He opts for a maple glaze, recommends a jelly for me, and rips into his with a child's delight. As he snarfs, it's suddenly quite easy to imagine him as a bright kid on a sugar high.
Brought up in an upper-middle-class household, Perlman excelled in academics but says he felt like an outsider among his peer group. He found sanctuary at the Talcott Mountain Science Center, a school for gifted kids just a few miles from his home in West Hartford, Connecticut. Perlman's parents enrolled him in weekend and summer courses at the center when he was 10. "It's the only place that I really fit into," says Perlman, who donated $1 million to the center after he hit the mother lode with WebTV. "It's the only place that could keep up with me and I could keep up with it."
The school is founded on the principle of "fitting programs to kids, not kids to programs," and there Perlman was introduced to the scientific method and tons of technology, including a black-and-white reel-to-reel videotape system, radio electronics, and a time-share computer system running Basic and Fortran.
"He practically lived here," says Talcott president Don La Salle, who has known Perlman since the boy wonder first came to the center almost 30 years ago. Even as a child, La Salle says, Perlman exhibited unusual focus, "what we call perseverance and stick-to-it-ness in gifted children. He has always been passionate with a capital P."
While at Talcott in the early '70s, Perlman invented his first commercially viable product, a device that converted analog information to digital data for a company studying the effectiveness of solar-panel technology. According to Perlman, it took him about a year to build the system.
High school was another matter. Until he was 15, Perlman earned straight As, enrolled in every Advanced Placement class he could, and played alto sax in the jazz band. But then he built his first home computer, a Southwest Technical Products 6800, a text-only terminal. In 1976, he taught himself how the system worked, staying up late reading manuals and programming. He even founded a company, Innovative Softronics, to get access to prototype chips and software. Eventually, he bought himself a book on TV innards and designed his first major digital circuit, a color graphics display, so that he could use a TV set to play the videogames he'd programmed for the system. In a contest between computing and high school, high school lost.
"I didn't know what was going on," Perlman recalls. "I was working so goddamn hard doing all these different things, and the only outcome was to get a grade. With that computer, when I created something, I had something real - something that was working."
He eventually completed the minimum coursework to receive a diploma and was accepted to Columbia University. While there, he hacked into the school's mainframe so he could run its word-processing programs on his trusty 6800. According to Perlman, other students who used his PC for papers formed his first focus group.
"I would watch the students work and I'd note the things they got hung up on," he says. "I kept modifying the software until it was easier to use." Hearing that, it comes as no surprise that shortly after graduating in 1983, he ended up in Cupertino.
At Apple, Perlman joined the Advanced Technology Group, an elite R&D team established in 1986 and led by Larry Tessler, a Xerox PARC alum. Staffed with as many as 500 engineers, the ATG's charter was to think big thoughts about future technologies, but few of its big ideas, says Perlman, were ever developed into products. (Apple declined to comment.)
By 1990 Perlman had quit - disillusioned, he says, because none of his toys ever made it out to play. Apple chair and CEO John Sculley called him and asked him to return. According to Perlman, he said he would - on one condition: "I wouldn't develop any more technology for Apple unless I retained the nonexclusive rights to it if they canceled the projects." Sculley agreed and Perlman began developing a low-cost set-top box with high-performance graphics and a fast software-based modem.
He code-named the project Tucker - after the innovative, ill-fated car - because he says he fully expected Apple to kill it. He was right, and later in 1990, Perlman left again, this time carrying away the rights to his TV technology - the start of a controversial trend in Perlman's career. Behind closed doors, some former coworkers from Apple and his next two employers, General Magic and Catapult Entertainment, have accused Perlman of walking off with intellectual property that belonged to the respective companies. Perlman denies this, though he has prepared extensive archives should anyone ever sue. So far, no one has.
"What I did," he says, "was fashion an agreement based on intellectual-property laws in other parts of the world, where basically they take the approach of, 'If you use it, you can have it. If you don't, you lose it.'"
That may work for Perlman, but it doesn't fly with those who paid his salary, and who lay claim not to what he brought to their companies, but what he did while he was in their employ. In any event, his former associates are diplomatic. For instance, Catapult cofounder Steve Roskowski is respectful yet candid about Perlman. "He's absolutely brilliant," Roskowski says. "And I never want to work with him again."
As Perlman sees it, his detractors prove the existence of what he calls the Salieri syndrome. A favorite riff of his, the syndrome is named for the antagonist of Amadeus and refers to people of mediocre talent who, out of jealousy, dedicate their lives to thwarting genius. Perlman first wrote about the Salieri syndrome in a white paper he crafted during his last days at Apple titled "Keeping Innovation at Apple, or Growing Up Without Growing Old." In it, he passionately outlined the dangers of Salieris who take in unsuspecting technologists in order to undermine their work.
"Innovators don't subject themselves to the [creative process] for money or power," he wrote. "They do it because they are driven, committed, and obsessed with reaching that solution." Uninspired innovators, he adds, "are worse than useless."
As Valley legend has it, it was Campbell's Soup that served as the immediate inspiration for WebTV. Shown the door in 1995 by his Catapult cofounders, Perlman was licking his wounds and hanging out, surfing the Web, when he came across the Campbell's Soup site. It struck him as ridiculous that this icon of Middle America was on the Web, a place few Middle Americans knew existed. He began toying with ideas of how to bring the Web to a mass audience - to bring all those Campbell's users online.
For three days straight Perlman toiled nonstop; he finally created a working prototype of a WebTV box. He called his friend Bruce Leak, inviting him over for a first look at his creation. In a "Hey kids, let's put on a show!" moment that day, the two decided to start a company, which they did, in a guest bedroom of Perlman's modest Silicon Valley home. Code-named Artemis Research, after the Greek goddess of the hunt, the startup operated out of Perlman's house for six months and then moved to an empty BMW dealership in Palo Alto. Top engineering talent came fairly easily. (The third WebTV cofounder, Phil Goldman, created Multifinder, one of the first software programs to enable multitasking on the Mac.) But money did not, and what investments WebTV did land - from oil and media mogul Marvin Davis and, later, Paul Allen's Vulcan Ventures, Citibank, the Washington Post Group, Microsoft, and Bill Gates himself - it quickly burned through.
"Steve's vision can test any limit," says Randy Komisar, a longtime friend of Perlman's who served as the company's "virtual CEO" prior to the Microsoft acquisition. "He knows no limits. I felt like a Depression baby. He'd say, 'I need $25 million more,' and I'd say, 'Steve, what if we can't get it?' He'd just look at me and say, 'What do you mean?'"
In January 1997, Perlman was once again trawling for cash. He'd already raised and spent $46 million. This time he wanted to raise $100 million, with an IPO to follow. But later that month, Microsoft executive Craig Mundie, then head of its consumer group, invited Perlman and WebTV cofounders Leak and Goldman to Microsoft.
"These guys came up to Redmond, and we started talking and it was an epiphany," Mundie says. "Perlman and I were finishing each other's sentences; we had an identical vision of how this evolution of TV was going to take place."
According to Mundie, the original WebTV device, now called the Classic, did a couple of things right out of the box that made it great for consumers and thus great for Microsoft: It provided cheap Internet and email access for non-PC users, and it delivered programming that more closely mapped to a TV experience than to a Web-surfing experience. More important, the WebTV service was built on top of what Mundie calls a "managed appliance" platform. That is, each box can automatically dial up a local POP server during off hours to update its software as needed. For Microsoft, learning how to take that technology across to the PC platform has tremendous value. It would mean that owners of Windows PCs could get software upgrades, bug fixes, and system tune-ups while they sleep.
Mundie told Microsoft executives that WebTV Networks was where Microsoft wanted to go yesterday. Six weeks later, Bill Gates paid WebTV a visit. Perlman says he was blown away by Gates' technical acumen. As Komisar puts it, "Perlman and Gates seduced one another."
Not surprisingly, one sticking point for Microsoft was Perlman's previous deals regarding his intellectual-property rights. The attorneys from Redmond removed any ambiguity surrounding ownership of WebTV's patents and intellectual property. They belong to Microsoft. Period. "Everything Steve does or has done at WebTV is the property exclusively of Microsoft," explains Mundie, who makes it clear that proprietary status was never a negotiable point.
The WebTV team had the usual emotional and logistical concerns over being bought. How would WebTV operate once assimilated? Would WebTV be forced to move? Would there be layoffs?
The deal that the team negotiated kept the company more or less autonomous and stable. As to who got what, the purchase made Perlman, Leak, and Goldman quite a bundle. Based on share percentages listed in the S-4 filings, the three netted around $70 million apiece. Still, each insists the windfall did not close the deal.
"The reality is, we didn't do it for the money," says Komisar, who helped settle the arrangement. "We did it for the financing. Nobody was excited about taking the money, but it was the only opportunity to fund the vision."
Despite some starts, fits, and backfires - "we spent the first year imposing a level of discipline on Steve and Bruce," says Mundie - the two companies made evident progress together.
Microsoft, Perlman says, enabled the startup to negotiate deals with strategic partners like EchoStar Communications and Scientific-Atlanta, as well as content providers and advertisers, including E! Entertainment, the Weather Channel, Ford, Ricola, and Volvo Cars of North America. It didn't hurt distribution, either. WebTV devices are now on the shelves of 21,000 stores, including Wal-Mart.
While WebTV continues to support its telco-based analog systems, it is really the company's broadband strategy that defines its future - and quite possibly the future of digital television. According to Perlman, the WebTV/EchoStar DishPlayer satellite receiver is essentially the precursor to what he calls "smart TV." The DishPlayer delivers Internet and email access, a satellite receiver, and a digital VCR with an 8.6-Gbyte drive for downloading and storing movies, music, and old episodes of Star Trek. Similar to newcomer TiVo's and Replay Networks' digital VCRs, the DishPlayer can pause live programs (by storing the signal to disk). Unlike the other two devices, however, the DishPlayer has built-in Web access.
The specs for a device like the EchoStar system, Perlman says, have been a part of the WebTV business model from the start. In part, Perlman says, his vision has been hampered by limitations of disk-drive capacity, a problem that is quickly disappearing - the storage size of drives increases 61 percent a year, while the cost declines. By 2006, Perlman predicts, the average mass-market disk-drive size will be about a terabyte, capable of storing roughly 1,000 hours of video.
At that point, he says, "it is feasible to download just about every show you could possibly be interested in watching during the week on, say, Sunday night. Then, when you sit down to watch TV, instead of seeing what is being broadcast at that moment, you would navigate a Web-like series of pages to look through what shows you are interested in among the hundreds that have been recorded." Since all of these shows are on disk, says Perlman, you'll be able to zip through them, rewind, and pause.
A number of WebTV's competitors are making similar promises, and Perlman knows it. At the National Association of Broadcasters conference in Las Vegas this spring, Perlman outlined five categories under the umbrella of interactive TV. They include Internet TV (basic Web and email access), response TV (simple TV-audience surveys), enhanced TV (Web access, plus TV listings and some VCR control), digital VCRs (able to store and record digital video and to recognize user preferences), and personal TV (virtual channels, audio and video on demand, and more). There's an interactive TV for every media diet.
Though the motivations behind Microsoft's March reorg remain obscure, the move is seen, in part, as a response to this kind of market segmentation - as a way to establish a platform that can be tailored to compete in each category. Indeed, the longtime Microsoft execs inserted into the org chart between Steve P. and Steve B. are Jon DeVaan - former VP of desktop apps, including Microsoft Office - and Brad Chase, the marketing muscle behind Windows. Ballmer named the two as the top executives of the new Consumer and Commerce division, which will focus on bringing consumers and businesses together online. And yet another executive is expected to join the team. As one source close to Microsoft says, they are "in search of the Great White Hope," a veteran from the consumer-electronics biz.
In keeping with the company's unstated strategy to leverage the hell out of all the software to promote the platform, this move seems intended to help integrate WebTV and Microsoft groups, including MSN and Hotmail. Integration of various software groups within Microsoft is DeVaan's métier. And this, sources say, is why Perlman wanted an exit strategy.
As you'd expect, some people at Microsoft view Perlman's threats to leave as juvenile at best. Says one, "Steve is brilliant. No one has an argument with that. The real question is, how do you support a brilliant guy who doesn't want to grow up? He wants power without responsibility, and at Microsoft that just doesn't fly."
"There are horses, and then there are racehorses," says Microsoft's Mundie. "There are cars, and then there are race cars. At Microsoft, we tend to value good, reliable Lincoln Continentals that get down the road every day. Perlman is definitely a sports car." Mundie says Microsoft entered the agreement with WebTV with its eyes wide open. Hopping metaphors, he says that "like a racehorse, Steve tended to bridle" when either leadership or supervision was imposed on him. But, he adds, at this point in WebTV's life it needs a leader who's a manager. "America Online is going to launch AOLTV," says Mundie. "Our asset is really the ability to bring together MSN, WebTV, Hotmail, and other related portals."
Whatever happens, Perlman doesn't need to be at WebTV to do what he loves. He's loaded, and he has big plans for his digital production studio. When complete, it will be outfitted with a motion-capture system, large-scale plasma viewing screens, and an HDTV projector. A history buff and amateur videographer, he intends to make a Civil War movie based on a screenplay he's writing. He says it's a romantic adventure, à la Elizabeth, and that much of it will be created using 3-D effects and bluescreen technology - the next phase, he says, in the evolution of television. In short, having developed a mass-market device to deliver smart TV, he wants to pioneer programming for it.
"The line between what we today think of as a videogame and what we today think of as a TV show will blur," he says. "Some shows will have people from around the country, and to some degree around the world, participating in stories and experiences together. Today, we are just starting to have the capability of rendering realistic hair, fabric, and faces. Within the decade, we'll be rendering images of people and animals that look organic and move expressively." Call it Role-Play TV - it's just the latest vision he will gladly share over a maple-glazed doughnut.
"I believe that people like Steve have pure talent; they are the artists of my field," says Komisar. "In turn, it does take some talent to deal with them. You have to understand their thinking process: They are not dismissive; they are focused. They are not mean-spirited; they are intent on the success of their ideas."
Or, as Steve Perlman will tell you, "I don't see myself as a bad boy who is rocking Microsoft's world. I see myself as an innovator who is operating in the way that innovators operate ... It's always a question for me of opportunities for innovation. I am not the best person for continuing engineering. It's not the best use of my time."
