The Innovators

by Walter Isaacson

  Gates ignored the exam cramming he was supposed to be doing and even stopped playing poker. For eight weeks he and Allen and Davidoff holed up day and night at Harvard’s Aiken Lab making history on the PDP-10 that the Defense Department was funding. Occasionally they would break for dinner at Harvard House of Pizza or Aku Aku, an ersatz Polynesian restaurant. In the wee hours of the morning, Gates would sometimes fall asleep at the terminal. “He’d be in the middle of a line of code when he’d gradually tilt forward until his nose touched the keyboard,” Allen said. “After dozing an hour or two, he’d open his eyes, squint at the screen, blink twice, and resume precisely where he’d left off—a prodigious feat of concentration.”

  They would scribble away at their notepads, competing at times to see who could execute a subroutine in the fewest lines. “I can do it in nine,” one would shout. Another would shoot back, “Well, I can do it in five!” Allen noted, “We knew that each byte saved would leave that much more room for users to add to their applications.” The goal was to get the program into less than the 4K of memory that an enhanced Altair would have, so there would still be room left for the consumer to use. (A 16GB smartphone has four million times that amount of memory.) At night they would fan out the printouts on the floor and search for ways to make it more elegant, compact, and efficient.59

  By late February 1975, after eight weeks of intense coding, they got it down, brilliantly, into 3.2K. “It wasn’t a question of whether I could write the program, but rather a question of whether I could squeeze it into under 4k and make it super fast,” said Gates. “It was the coolest program I ever wrote.”60 Gates checked it for errors one last time, then commanded the Aiken Lab’s PDP-10 to spew out a punch tape of it so Allen could take it to Albuquerque.

  On the flight down, Allen remembered he hadn’t written a loader, the sequence of commands that would instruct the Altair how to put the BASIC interpreter into its memory. As the plane was preparing to land, he grabbed a pad and wrote twenty-one lines in the machine language used by the Intel microprocessor, each line a three-digit number in base-8. He was sweating by the time he left the terminal, wearing a tan Ultrasuede polyester suit and looking for Ed Roberts. Eventually he spotted a jowly three-hundred-pound man in jeans and a string tie in a pickup truck. “I’d expected a high-powered executive from some cutting-edge entrepreneurial firm, like the ones clustered along Route 128, the high-tech beltway around Boston,” Allen recalled.

  The MITS world headquarters was likewise not quite what Allen expected. It was in a low-rent strip mall, and the only Altair with enough memory to run BASIC was still being tested. So they put off until the next morning trying out the program and headed off “to a three-dollar buffet at a Mexican place called Pancho’s, where you got what you paid for,” Allen said. Roberts drove him to the local Sheraton, where the desk clerk told him that his room would be $50. That was $10 more than Allen had brought with him, so after an awkward stare Roberts had to pay for the room. “I guess I wasn’t what he’d been expecting, either,” said Allen.61

  The next morning, Allen returned to MITS for the big test. It took almost ten minutes to load in the code for the BASIC interpreter that he and Gates had written. Roberts and his colleagues exchanged amused glances, already suspecting that the show would be a fiasco. But then the Teletype clacked to life. “MEMORY SIZE?” it asked. “Hey, it typed something!” shouted one of the MITS team. Allen was happily flabbergasted. He typed in the answer: 7168. The Altair responded: “OK.” Allen typed, “PRINT 2+2.” It was the simplest of all commands, but it would test not only Gates’s coding but also Davidoff’s floating-point math routines. The Altair responded: “4.”

  Up until then, Roberts had been watching quietly. He had taken his failing company further into debt on the wild surmise that he could create a computer that a home hobbyist could use and afford. Now he was watching as history was being made. For the first time, a software program had run on a home computer. “Oh my God,” he shouted. “It printed ‘4’!”62

  Roberts invited Allen into his office and agreed to license the BASIC interpreter for inclusion on all Altair machines. “I couldn’t stop grinning,” Allen confessed. When he arrived back in Cambridge, bringing with him a working Altair to install in Gates’s dorm room, they went out to celebrate. Gates had his usual: a Shirley Temple, ginger ale with maraschino cherry juice.63

  A month later, Roberts offered Allen a job at MITS as director of software. His colleagues at Honeywell thought he was crazy to consider it. “Your job’s safe at Honeywell,” they told him. “You can work here for years.” But career safety was not an ideal embraced by those eager to lead the computer revolution. So in the spring of 1975, Allen moved to Albuquerque, a city he had only recently learned was not in Arizona.

  Gates decided to stay at Harvard, at least for the time being. There he endured what has become a rite of passage, amusing only in retrospect, for many of its most successful students: being hauled before the university’s secretive Administrative Board for a disciplinary process, known as being “Ad Boarded.” Gates’s case arose when auditors from the Defense Department decided to check the use of the PDP-10 that it was funding in Harvard’s Aiken Lab. They discovered that one sophomore, W. H. Gates, was using most of the time. After much fretting, Gates prepared a paper defending himself and describing how he had created a version of BASIC using the PDP-10 as an emulator. He ended up being exonerated for his use of the machine, but he was “admonished” for allowing a nonstudent, Paul Allen, to log on with his password. He accepted that minor reprimand and agreed to put his early version of the BASIC interpreter (but not the refined one he and Allen were then working on) into the public domain.64

  By that time Gates was focusing more on his software partnership with Allen than his course work at Harvard. He finished his sophomore year that spring of 1975, then flew down to Albuquerque for the summer and decided to stay there rather than return for the first semester of his junior year that fall. He went back to Harvard for two more semesters, in the spring and fall of 1976, but then left Harvard for good, two semesters shy of graduating. In June 2007, when he returned to Harvard to get an honorary degree, he began his speech by directing a comment to his father in the audience. “I’ve been waiting more than 30 years to say this: Dad, I always told you I’d come back and get my degree.”65

  MICRO-SOFT

  When Gates arrived in Albuquerque in the summer of 1975, he and Allen were still supplying BASIC for the Altair on a handshake deal with Ed Roberts. Gates insisted on a formal agreement and, after much haggling, agreed to license the software to MITS for ten years, to be bundled with each Altair, for $30 in royalty per copy. Gates was able to win two provisions that would be historically significant. He insisted that he and Allen would retain ownership of the software; MITS would merely have rights to license it. He also required that MITS use its “best efforts” to sublicense the software to other computer makers, splitting the revenues with Gates and Allen. It set a precedent for the deal Gates would make six years later with IBM. “We were able to make sure our software worked on many types of machines,” he said. “That allowed us and not the hardware makers to define the market.”66

  Now they needed a name. They kicked around a few ideas, including Allen & Gates, which they decided sounded too much like a law firm. Eventually they picked one that was not particularly exciting or inspiring but did convey that they were writing software for microcomputers. In the final documents for the MITS deal, they referred to themselves as “Paul Allen and Bill Gates doing business as MicroSoft.” A credit line appeared in the source code of what was then their only product: “Micro-Soft BASIC: Paul Allen wrote the non-runtime stuff. Bill Gates wrote the runtime stuff. Monte Davidoff wrote the math package.” Within a couple of years, the name was simplified to Microsoft.

  After bunking for a while at the Sundowner Motel on a strip of Route 66 known more for prostitutes than programmers, Gates and Allen moved to a cheap furnished apartmen
t. Monte Davidoff, of floating-point math fame, and Chris Larson, a younger student from Lakeside High, moved in, turning the apartment into a frat house doing business as a geek bunker. In the evenings Allen would crank up his Stratocaster guitar and play along with Aerosmith or Jimi Hendrix, and Gates would retaliate by loudly singing Frank Sinatra’s “My Way.”67

  Of them all, Gates was the prime example of the innovator’s personality. “An innovator is probably a fanatic, somebody who loves what they do, works day and night, may ignore normal things to some degree and therefore be viewed as a bit imbalanced,” he said. “Certainly in my teens and 20s, I fit that model.”68 He would work, as he had at Harvard, in bursts that could last up to thirty-six hours, and then curl up on the floor of his office and fall asleep. Said Allen, “He lived in binary states: either bursting with nervous energy on his dozen Cokes a day, or dead to the world.”

  Gates was also a rebel with little respect for authority, another trait of innovators. To folks like Roberts, a former Air Force officer with five sons who called him “Sir,” Gates came across as a brat. “He literally was a spoiled kid, that’s what the problem was,” said Roberts later. But it was more complex than that. Gates worked hard and lived frugally off his then-meager earnings, but he did not believe in being deferential. The scrawny Gates would go toe to toe with the brawny six-foot-four Roberts and engage in arguments so heated that, as Allen remembered, “you could hear them yelling throughout the plant, and it was a spectacle.”

  Allen assumed that his partnership with Gates would be fifty-fifty. They had always been a team, and it seemed unnecessary to fight over who had done more. But since their spat over the payroll program in high school, Gates had insisted on being in charge. “It’s not right for you to get half,” he told Allen. “You had your salary at MITS while I did almost everything on BASIC without one back in Boston. I should get more. I think it should be sixty-forty.” Whether or not Gates was right, it was in his nature to insist on such things, and it was in Allen’s nature not to. Allen was taken aback but agreed. Worse yet, Gates insisted on revising the split two years later. “I’ve done most of the work on BASIC, and I gave up a lot to leave Harvard,” he told Allen on a walk. “I deserve more than 60 percent.” His new demand was that the split be 64-36. Allen was furious. “It exposed the differences between the son of a librarian and the son of a lawyer,” he said. “I’d been taught that a deal was a deal and your word was your bond. Bill was more flexible.” But again Allen went along.69

  In fairness to Gates, he was the person who, by then, was actually running the fledgling company. Not only was he writing much of the code, but he also was in charge of sales, making most of the calls himself. He would kick around ideas about product strategy with Allen for hours, but he was the one who made the final decisions on which versions of Fortran or BASIC or COBOL would be built. He was also in charge of business deals with the hardware makers, and he was an even tougher negotiator with them than he had been with Allen. Plus he was in charge of personnel, which meant hiring, firing, and telling people in words of one syllable when their work sucked, which is something Allen would never do. He had the credibility to do so; when there were contests in the office to see who could write a program using the fewest lines of code, Gates usually won.

  Allen would sometimes come in late and might even think it was permissible to leave work in time for dinner. But not Gates and his close coterie. “It was hard-core,” he recalled. “A small group and me would work late into the night. And I’d sometimes stay all night and then sleep in the office and my secretary would come wake me up if we had a meeting.”70

  Born with a risk-taking gene, Gates would cut loose late at night by driving at terrifying speeds up the mountain roads to an abandoned cement plant. “Sometimes I wondered why Bill drove so fast,” Allen said. “I decided it was his way of letting off steam. He’d get so wound up in our work that he needed a way to stop thinking about the business and the code for a while. His breakneck driving wasn’t so different from table stakes poker or edge-of-the-envelope waterskiing.” Once they had made a little money, Gates splurged on a green Porsche 911, which he would race along the freeway after midnight. At one point he complained to his local dealer that the car’s top speed was supposed to be 126 miles per hour, but he could get it up only to 121. Late one night he was caught speeding and got into an argument with the cop about why he wasn’t carrying a driver’s license. He was thrown into jail. “Got arrested,” he said when Allen picked up the phone. He was released in a few hours, but his mug shot from that night became a memorable icon of geek history.71

  Gates’s intensity paid off. It allowed Microsoft to meet software deadlines that seemed insane, beat other competitors to the market for each new product, and charge such a low price that computer manufacturers rarely thought of writing or controlling their own software.

  SOFTWARE WANTS TO BE FREE

  In June 1975, the month that Gates moved to Albuquerque, Roberts decided to send the Altair on the road as if it were a carnival show exhibit. His goal was to spread the word about the Altair’s wonders and create fan clubs in towns across America. He tricked out a Dodge camper van, dubbed it the MITS Mobile, and sent it on a sixty-town tour up the coast of California then down to the Southeast, hitting such hot spots as Little Rock, Baton Rouge, Macon, Huntsville, and Knoxville.

  Gates, who went along for part of the ride, thought it was a neat marketing ploy. “They bought this big blue van and they went around the country and created computer clubs everyplace they went,” he marveled.72 He was at the shows in Texas, and Allen joined them when they got to Alabama. At the Huntsville Holiday Inn, sixty people, a mix of hippyish hobbyists and crew-cut engineers, paid $10 to attend, then about four times the cost of a movie. The presentation lasted three hours. At the end of a display of a lunar landing game, doubters peered under the table suspecting that there were cables to some bigger minicomputer hidden underneath. “But once they saw it was real,” Allen recalled, “the engineers became almost giddy with enthusiasm.”73

  * * *

  One of the stops was at Rickeys Hyatt House hotel in Palo Alto on June 5. There a fateful encounter occurred after Microsoft BASIC was demonstrated to a group of hobbyists, including many from the recently formed Homebrew Computer Club. “The room was packed with amateurs and experimenters eager to find out about this new electronic toy,” the Homebrew’s newsletter reported.74 Some of them were also eager to act on the hacker credo that software should be free. This was not surprising given the social and cultural attitudes, so different from the entrepreneurial zeal in Albuquerque, that had flowed together in the early 1970s leading up to the formation of the Homebrew Club.

  Many of the Homebrew members who met the MITS Mobile had built an Altair and been waiting impatiently to get hold of the BASIC program that Gates and Allen had produced. Some had already sent checks to MITS for it. So they were thrilled to see that the Altairs on display were running a version of it. Indulging in the imperative of hackers, one of the members, Dan Sokol, “borrowed” the punched paper tape that had the program and used a DEC PDP-11 to make copies.75 At the next Homebrew meeting, there was a cardboard box filled with dozens of BASIC tapes for members to take.III There was one stipulation: you had to make a few copies to replenish the communal box. “Remember to bring back more copies than you take,” Lee Felsenstein joked. It was his signature line for any software sharing.76 Thus did Microsoft BASIC spread freely.

  This, not surprisingly, infuriated Gates. He wrote a passionate open letter, displaying all the tact of a nineteen-year-old, which served as the opening shot in the war over the protection of intellectual property in the age of personal computers:

  An Open Letter to Hobbyists . . .

  Almost a year ago, Paul Allen and myself, expecting the hobby market to expand, hired Monte Davidoff and developed Altair BASIC. Though the initial work took only two months, the three of us have spent most of the last year documenting, improving
and adding features to BASIC. Now we have 4K, 8K, EXTENDED, ROM and DISK BASIC. The value of the computer time we have used exceeds $40,000.

  The feedback we have gotten from the hundreds of people who say they are using BASIC has all been positive. Two surprising things are apparent, however, 1) Most of these “users” never bought BASIC (less than 10% of all Altair owners have bought BASIC), and 2) The amount of royalties we have received from sales to hobbyists makes the time spent on Altair BASIC worth less than $2 an hour.

  Why is this? As the majority of hobbyists must be aware, most of you steal your software. Hardware must be paid for, but software is something to share. Who cares if the people who worked on it get paid?

  Is this fair? One thing you don’t do by stealing software is get back at MITS for some problem you may have had. . . . One thing you do do is prevent good software from being written. Who can afford to do professional work for nothing? What hobbyist can put 3-man years into programming, finding all bugs, documenting his product and distribute for free? The fact is, no one besides us has invested a lot of money in hobby software. We have written 6800 BASIC, and are writing 8080 APL and 6800 APL, but there is very little incentive to make this software available to hobbyists. Most directly, the thing you do is theft. . . .