Overdrive, page 10
Nathan Myhrvold was raised by his mother, a schoolteacher in Santa Monica, California, who would recall years later that he was 2 years old when he proudly announced that he was going to be a scientist.
A child prodigy who taught himself programming, the gifted Myhrvold graduated from high school at age 14, and from UCLA five years later with a bachelor of science in mathematics and a master’s in geophysics and space physics. From UCLA he went on to Princeton, where the great Albert Einstein had taught. By age 23, Myhrvold had a master’s in mathematical economics to go with his doctorate in theoretical physics. His thesis, titled “Vistas in Curved Space Quantum Field Theory,” dealt with some of the problems confronting scientists in their search for the origins of the universe. After Princeton, the young Myhrvold crossed the ocean to England, where he had accepted a postdoctoral fellowship at Cambridge to work with the great theoretical physicist Stephen Hawking, who held the prestigious Isaac Newton chair and was generally regarded as having the best mind since Einstein when it came to understanding the concepts of time, space, quantum physics, and the origins of the universe.
Inevitably, Microsoft’s immense gravitational field finally attracted Myhrvold. In 1986, he was head of his own small software company in Berkeley that he had started with a hand- fill of his Princeton physicist pals. The company, called Dynamical Systems Research, had recently launched a graphical user interface, and Microsoft became interested in licensing the technology for Windows. Instead, it bought the company and hired its six programmers, including Myhrvold, then 27, and his brother Cameron. For the next several years, Myhrvold worked for Steve Ballmer on Windows, and later on the joint effort with IBM to develop what would become Big Blue’s operating system known as OS/2.
In Microspeak, Myhrvold had super-high bandwidth, just like the boss, who, before Comdex, named Myhrvold the company’s chief technologist. Other than Gates, no one at Microsoft would have a greater impact on its long-term strategy than Myhrvold. A prolific writer, Myhrvold was soon knocking out long memos about future strategy and technology. It was not unusual for him to write a dozen or more single-spaced, 30-page memos in a single month. Some went on for 100 pages. Myhrvold was passionate about many subjects, and his mind raced along like the Formula I car he occasionally took for a spin around the track at a driving school in Blakeslee, Pennsylvania.
One lengthy memo in 1990 explained to Gates why Microsoft needed to be developing new technology for the coming information highway. Myhrvold saw the highway as much more than just movies on demand. He foresaw a highway rich in computer networks over which hundreds of services would be offered, and through which people, businesses, and government would freely interact. Discount warehouses such as Wal- Mart and Costco had succeeded by eliminating many of the distribution points in the retailing food chain. Myhrvold believed the information highway would do the same thing. “After being at Costco a few times, I’ve concluded many of the things you’d be willing to do there, you’d be willing to do remotely,” he would later say in an interview.
That 1990 memo had led to the formation the following year of Microsoft’s research think tank, known as the Advanced Technology Group. Gates put Myhrvold in charge. For the first time in its history, Microsoft had a true R&D effort. For years, Gates had been criticized by his peers within the industry for taking innovations that had been developed elsewhere and simply improving on them. It was a sore point with Gates, who would go to great lengths to argue that Microsoft created innovative software. But the fact was that Microsoft bought its cash cow, DOS, for $50,000 in 1981, when IBM needed an operating system. And Microsoft’s graphical user interface that became Windows was so similar to what Apple had developed for its Macintosh computer that Apple sued Microsoft. Of course, the idea had not originated with Apple. The graphical user interface was the work of a team of brilliant scientists at Xerox’s Palo Alto Research Center, or PARC. Apple’s Steve Jobs had taken one look at the interface technology during a visit to PARC and realized its commercial potential. PARC was known for its innovative technology, but it didn’t know how to bring that technology to the marketplace. Gates and Myhrvold were determined that Microsoft’s Advanced Technology Group would not make the same mistakes as PARC, and leave its best work in the lab.
They divided the Advanced Technology Group into three subgroups. The basic research fell to the Advanced Research group. It was dedicated to developing “smart” software that would make computers easier to use, and was headed by Rick Rashid, who had been recruited by Microsoft from Carnegie Mellon University, where he had been a computer science professor for a dozen years. Rashid had developed Mach, a version of the UNIX operating system.
Another subdivision was an unnamed special projects group within the Advanced Technology Group, but by far the largest of the three subgroups was the Advanced Consumer Technology group.
Gates was well aware that no high-tech company had ever made the transition from one technological era to the next. IBM had once ruled the industry with its “big iron” mainframes, only to stumble when the personal computer came along. If Microsoft was going to make that transition, it would do so because of the ideas and work coming out of the various divisions within the Advanced Research Group. As Myhrvold was the heavyweight thinker of the bunch, it was his job to peek into the future, identify trends, and help position Microsoft so it could capitalize on those trends. His philosophy was simple. The future, he said, belonged to those willing to make fundamental investments in technology.
In a memo to Microsoft’s board when the Advanced Technology Group was formed, Myhrvold proposed that the company spend about $10 million a year on research. In fact, it would quickly spend more than 40 times that amount. But for a company with some $2.5 billion in cash reserves, spending $400 million a year on research hardly seemed exorbitant.
As Microsoft’s elite technology shock troops, the Advanced Technology Group had a twofold mission: to coordinate various technology strategies across business lines, and to conduct basic research for software likely to be needed in the future. The group’s goal was to develop products that would become important to Microsoft as future business and technology markets opened up. One of those products was the software to make interactive television a reality. In 1993, U.S. consumers were spending about $12 billion a year on movie rentals and videocassettes. With a full-scale interactive system, they could become true couch potatoes and never leave their homes. Gates was banking on the next DOS being the software that would be used to run those millions of television set-top boxes.
But Myhrvold did not believe that video on demand was the be-all and end-all application. Instead, he thought Microsoft should develop a new class of applications that mixed communications and data storage and computing. “If all you want to do is watch movies, then 500 channels may be good enough,” he said in a 1993 interview. “But I don’t think that’s what it’s all about. It’s not the thing that really restructures our economy. That’s not what creates an information highway where goods and services are marketed. That’s not what allows us to reach out with video telephony and e-mail and other rich forms of communication. It’s not what brings distance learning to people, or remote medical diagnosis, or any of those other wild and exciting applications. Our belief at Microsoft is that all of the stuff I mentioned, as well as many other things we can’t even guess at today, [comprise] the exciting things.”
Myhrvold had become an oft-quoted and high-profile figure as the media hyped the information highway. After Clinton took office, Myhrvold was invited to join Secretary of Commerce Ron Brown’s industry advisory group on the formation of infrastructure for the information highway. The position gave Myhrvold an opportunity to lobby for an open and competitive marketplace. As Myhrvold liked to point out, the information highway was illegal, strictly speaking, because the various players, like cable and phone companies, were prevented by the Communications Act of 1934 from offering the same services.
While Myhrvold was shuttling back and forth between Redmond and Washington, D.C., the job of developing Microsoft’s interactive television software fell to Craig Mundie of the Advanced Consumer Technology group. Despite the name, the group was totally separate from the company’s consumer products division, which was developing multimedia and PC software titles.
Mundie was the former chief executive officer of Alliant, a supercomputer company that he had co-founded in 1982 and that had gone belly-up. He was hired by Microsoft in 1992. That he had headed a failed company was not an issue. In fact, as far as Gates was concerned, it counted in Mundie’s favor, because he had learned from failure—a rare experience around Microsoft.
The Advanced Consumer Technology group had been started with about two dozen people, but after Mundie arrived it quickly expanded into the biggest division within Myhrvold’s Advanced Technology Group. Mundie said it was the largest start-up in Microsoft’s history. He would need all the smart troops he could get. Like Gates, Mundie was convinced that interactive television would drive the digital revolution and foster the creation of the information highway. Microsoft would have to spend a lot of money to develop not just the software to run in the box that would be on top of the television, but the much bigger and more complex software at the other end of the fiber-optic cable, the program that would run on the computers that stored the movie database, the directory, and everything else.
In April 1993, Microsoft, Intel, and General Instrument Corporation announced a deal to develop a set-top box for managing all the games, shopping, movies, and other options that would be available to homes through interactive television. General Instrument was a major supplier of hardware for the cable industry.
At the heart of Microsoft’s interactive software strategy was a hush-hush project code-named Tiger,, which would permit the sending of digitized video streams over advanced cable and telephone networks. Tiger would be able to store massive libraries of digital information such as movies on laser disks or on hard drives; powerful PCs would deliver the information to a set-top box. Tiger was being designed by Mundie’s group to work with Microsoft’s powerful new operating system, Windows NT. Rashid’s research group was building and testing models of the complex tools Tiger would need.
Microsoft’s chief competition to develop interactive software was an old adversary, Oracle Systems Corporation, the industry’s number-one producer of software for corporate databases. The battle promised to be fierce, with the industry’s two biggest and most aggressive companies going head-to- head in the race to set the industry standard for interactive software. Oracle’s flamboyant founder, Lawrence Ellison, was the software industry’s “other” billionaire, after the Microsoft triumvirate of Gates, Paul Allen, and Steve Ballmer. Ellison had committed Oracle to an interactive strategy that was the opposite of that being pushed by Microsoft. It was, as Microsoft’s Craig Mundie, the onetime creator of supercomputers, liked to say, the “big iron approach.” Oracle would use lots of supercomputers working together—massively parallel supercomputers in industry terms—to handle the overwhelming amounts of video data that would be sent to set-top boxes. The hardware would come from a 10-year-old supercomputer company in Foster City, California, called nCube Corporation, in which Ellison had a controlling interest. nCube was perfectly suited for the interactive job. And with a little rewriting, Oracle’s database software, originally designed to run on high-end computers, seemed ready-made to work as the operating system for the processors from nCube.
At this point in 1993, Oracle had the early lead over Microsoft. It already had a project under way to test video on demand through Bell Atlantic, British Tele-Communications, and nCube. And Bell Atlantic and US West were eyeing Oracle’s system for an interactive test scheduled in late 1994.
Ellison was slamming Gates and Microsoft in the media, insisting that Microsoft would be the Johnny-come-lately of interactive technology. In the fall of 1993, he said publicly that there was no need to wait for Microsoft; that Oracle was ready for a serious rollout of its technology by the next year. Gates, meanwhile, was saying that video on demand and other services would not hit the market for at least two more years, and that companies that came out early with inadequate technology would just end up being embarrassed. Gates would prove to be right. The challenges were daunting; video on demand had to overcome major technological hurdles.
At Microsoft, figuring out the best way to store the huge amount of video and sound and serve it up to hundreds of thousands of users at peak demand times would test the high- bandwidth types being hired to work with the Advanced Technology Group. Gates and Myhrvold were hiring the best and the brightest. Mundie was just one. Others included Bob Frankston, who years earlier had developed the industry’s first PC spreadsheet, known as VisiCalc. When he was hired by Microsoft in 1993, Frankston’s mission was to research the infrastructure of the information highway. There were many such “big think” projects under way within the Advanced Technology Group, which would double in size to about 400 people in 1993. But only a handful were working on a project that would soon become far more important to Microsoft than the work being done to develop interactive television software. That project was Marvel.
Like many important talks at Microsoft, this one took place outside, on one of those postcard-perfect spring days in Seattle when the rain stops and the low-hanging cover of gunmetal gray clouds gives way to sunshine and deep blue sky. It was on such a day that a songwriter must have come up with the line: “The bluest skies you’ve ever seen are in Seattle.”
All over Microsoft’s sprawling campus, Microserfs in jeans, sport shirts, and tennis shoes were enjoying the weather. Some, like the two men walking slowly between Buildings 8 and 10 near the fountain in the huge courtyard at the core of the campus, were engaged in serious talk. Others played Frisbee or soccer on the acres of grass that gave the campus a parklike setting, with its explosion of color from the springtime flowers in bloom. Everywhere they looked, the lush green landscape was rich in hues of deep purple and electric pink.
“I’m going to recommend to Bill that we build our own,” said Russ Siegelman, one of the two men having the serious business discussion. “Boy, that’s great! I’d sure like to be involved if you get the okay,” said the other man, Jeff Lill. “Well,” said Siegelman, “I’m not sure I’ll be running things, but if I am, you will be the first guy I come looking for to help manage the project.”
A few weeks later, on May 11,1993, after Siegelman made his recommendations, Bill Gates gave the go-ahead for work to begin on the yet-unnamed project to build an on-line service to compete with the existing big three—America Online (AOL), Prodigy, and CompuServe. Even though he had no previous experience managing such a large, critically important development effort, Siegelman, 33, was put in charge.
As he had promised, Siegelman tapped Lill to be his number-two guy, the overall technical manager for the new project. Lill was finally getting the opportunity to do what he had wanted for some time—help take Microsoft into the wired world.
“That was real cool,” said Lill, who would remain with the project until shortly before he left Microsoft in January 1996 to form his own Internet company in Seattle. “We didn’t really know where we were going with the project, what it was going to look like, but at least something had finally been approved that would have direction from the top.”
Lill had joined Microsoft four years earlier. He and a friend had been running a very small software company when Microsoft came calling and wanted to hire them. They said no. “We kind of liked being on our own,” said Lill. But they did begin making a Windows product, which they sold to Microsoft on a royalty basis. Eventually, though, Microsoft bought them out, and Lill and his friend joined the growing ranks of Microserfs in Redmond.
“I really wanted to get into the on-line world,” said Lill, who in 1992 transferred to Myhrvold’s Advanced Technology Group, where there were several projects under way related to on-line services.
“These were pretty much ill-defined projects,” said Lill. “Typically, Nathan would say, ‘Here are some guys. Go do something really cool.’ ” Lill went to work on a project that would be called RIP, for remote information protocol. Their marching orders from Myhrvold: come up with the ultimate, patentable protocol. It was one of Myhrvold’s many nonsanc- tioned projects within the Advanced Technology Group, and that was just fine with Lill. It meant they would have less direct supervision and more freedom to figure out what, exactly, they were supposed to come up with.
“Once we got into this thing,” recalled Lill, “we quickly realized that without a customer base and not really knowing what the heck we were trying to do, we couldn’t just write it. We weren’t being watched very closely, so we decided why not just do an on-line service? So we went to work on lower-level protocol things.” Soon Lill took over as leader of the project, reporting to Craig Mundie in the Advanced Consumer Technology group.
Lill visited bank executives and other potential customers to tell them: “You should do on-line banking through our protocol.” Of course, he didn’t tell them that the protocol didn’t exist. At about the same time, Siegelman, under orders from Gates to find out all he could about the on-line world and report back on what Microsoft should do, was making the rounds within Microsoft to get a handle on what internal efforts were under way. He found several, all going in different directions with no central oversight. In Microspeak, they were “random.”
