by Alec Foege
Griffith said feelings of innovation insecurity exist virtually everywhere that he visits around the world for speaking engagements. “In Australia, people say they wish they had the British education system and the American innovation system,” he said. “You’re in Britain and they say, we wish we had the American education system and the German innovation system. You’re in Germany, and they say, we wish we had the Australian innovation system and the Japanese education system. Every country in the world is suddenly paranoid they have lost their advantage to someone else.”
Despite Griffith’s fascination with the pregnant possibility of the new, he is quick to acknowledge that the paranoia nations exhibit when it comes to the future of innovation is just that: paranoia. “The number of things that are genuinely new are amazingly close to zero,” he said. A healthy tinkering culture, he argues, is transparent about that reality, and uses it as way to connect the potential of tomorrow with the accomplishments of yesteryear. “We were building inverted pendulums at the turn of the last century. Henry Bessemer was building gyroscopically stabilized monorails a century ago.”
Frantically tinkering to develop the new, new thing should constantly create friction with the innovators of the past. That’s an advantage that American tinkerers have over all others; when we create something new and practical, we acknowledge it with an odd combination of hubris and humility that many have done something quite similar before us.
And with that acknowledgement, we fit the fruits of our labor into the long history of American exceptionalism, for better or for worse.
CHAPTER 8
PARC AND THE POWER OF THE GROUP
TRUE TINKERING IS ALL ABOUT RISK and unusual behavior. The far-flung fanaticism that world-class tinkering requires rarely thrives in an institutional frame work. As noncorporate and freewheeling as the world Saul Griffith describes is, it has its roots in what is perhaps the prime example of institutionalized tinkering. Of all the corporate research facilities established over the last fifty years, the PARC Corporation, an innovation laboratory born out of Xerox, is often mentioned as the one that held the most promise. And yet PARC, an acronym fo the Palo Alto Research Center, also has come to represent the best evidence that corporate tinkering in all its shapes and forms is ultimately doomed to stagnancy and failure. The very nature of corporations, and American corporations in particular, is to minimize risk and behavior that stands out. Research and development at those corporations tends to be narrow in focus and product oriented. Yet in its heyday, PARC exuded a unique frontier spirit that rarely shows its head in today’s metrics-minded boardrooms. Understanding what happened and didn’t happen at PARC over the past forty or so years is important in assessing the viability of tinkering in a corporate framework going forward.
Established by Xerox in 1969, PARC is best known for developing the first personal computer, which Xerox then promptly ignored, leaving the field wide open for Apple and IBM to rip off its prototypes and make billions from the results. PARC made its name by attracting the best engineers and scientists in their respective fields and allowing them to express their creativity in as wild and radical ways as they could without concern for their corporate parent.
In the wake of the Xerox 914 copier, at that point the most successful industrial product in history, the company, then based in Rochester, New York, decided to acquire Scientific Data Systems (SDS), a scientific computer concern in Southern California. Newly anointed chief executive C. Peter McColough, unfamiliar with the nascent computer industry, was looking for an acquisition to allow Xerox to compete with IBM and others in the business data processing sector.
He did not consult Xerox’s engineers before offering $918 million in stock for SDS. In May 1969, Xerox shareholders approved the purchase and ushered in the modern computer era. Operating out of the old Encyclopaedia Britannica building at 3180 Porter Avenue in Palo Alto, PARC began its history with the unusual organizational move of establishing three separate divisions, despite employing only a handful of staff. PARC’s first director, George E. Pake, a former physics professor and provost of Washington University in St. Louis, felt strongly that the research center should have a computer science laboratory, a systems science laboratory, which was developing the world’s first laser-equipped computer printer, and a general science laboratory. By putting computer science on the same level as the traditional hard sciences, Pake and Xerox’s chief scientist, Jacob “Jack” Goldman, hoped to encourage the scientists at SDS to rise to the level of innovation and productivity that had helped Xerox to become a leader in the photocopier market. The idea was to populate the relatively new science of computers with the methods and rigor of classic scientific inquiry.
Initially set up more like a university than a corporate research department, PARC was populated mostly with former academics who had never before experienced a genuine corporate culture. Thanks to a recent cutback in military spending due to the political backlash from the Vietnam War and a brutal recession, Xerox had its pick of the top research and engineering talent of the day. What distinguished PARC from other industrial development departments was its lack of clear purpose, by design. That and premium salaries that easily surpassed those offered by even the most profligate universities.
In the early 1970s, computer science did not have the pedigree it does today, so the idea of offering $30,000 to $35,000 in salaries to computer geeks with PhDs was pretty much unprecedented. But by 1970, there was not even a consensus among the nation’s top computer scientists that devising a computer with a high-powered display for personal use was a worthwhile goal. The prevailing model at the time involved building the largest computer technically feasible and allowing only for professional computer operators to time-share on the unit.
A handful of fortuitous developments made PARC a catalyst for change in this environment. The first had its roots in the US Defense Department’s Advanced Research Projects Agency, or ARPA (and renamed DARPA in 1972), originally formed to create new missile technology in the national panic that followed the launch of Sputnik in 1957 by the Soviet Union. By the early 1960s, the space program had been pulled out from under the military umbrella and given its own agency, NASA, leaving ARPA to concentrate on civilian scientific research. Despite the fact that ARPA’s mission was supposed to be related to national defense, with ample funding to suggest it was something of a governmental priority, it lacked a clear mandate.
J. C. R. Licklider, the former behavioral psychologist who first headed ARPA, suggested that the world’s largest user of computers, the Defense Department, should fund a world-class computer science research program. The result was the Information Processing Techniques Office, with a then astronomical $14 million budget (worth roughly $100 million in today’s dollars) and none of the bureaucratic red tape that other federal agencies had to deal with. While Licklider funded mostly large-scale time-sharing computer projects during the mid-1960s, his successor, Bob Taylor, was keen on directing ARPA’s capital into smaller projects, the most prominent being Project Genie, conducted at the University of California at Berkeley. Project Genie’s goal was to construct a computer system for use by ten to twenty operators simultaneously, rather than the hundreds required for larger-scale systems. The theory was that a smaller, more affordable computer could be distributed more widely and thus empower more users overall.
The core of the Genie system was the SDS 930, made by Scientific Data Systems, which retailed for $73,000. It was later upgraded with about $5,000 of additional hardware and sold as the SDS 940, for $173,000 per unit. It became one of SDS’s best-selling products.
But again, SDS’s strength was scientific computing, not business computing, which had somehow eluded Xerox’s Connecticut-based management team. When PARC’s engineers recommended purchasing a PDP-10, made by the Digital Equipment Corporation and rapidly becoming the computer of choice for research laboratories nationwide, the purchase order was declined by management on the East Coast, based on the stub
born but incorrect belief that the SDS 940 could be modified to match the abilities of the PDP-10.
In a desperate move, the computer geeks in Palo Alto decided to make their own version of the PDP-10. Armed with PARC’s unique creative philosophy, crafted by Bob Taylor—that everything they designed should be designed for everyday use—a core group of scientists and engineers swiftly entered a heretofore unimaginable environment in which they were granted the authority to erect their own computer system to their own specifications.
The second key factor in PARC’s ascent was the hiring of Alan Kay, a rumpled, eccentric computer scientist, erstwhile jazz guitarist, and acolyte of Seymour Papert, inventor of the LOGO educational programming language. LOGO was designed to teach children about computers by allowing them to see the immediate effect of typing simple programming commands that would appear on a screen and move a toy robot around the floor.
Kay, often cited as the archetypal computer nerd, arrived at PARC in 1970. Kay was a new kind of computer scientist who did not fit the stereotype of the previous era, a timid, clean-cut Poindexter in a lab coat. Kay, to the contrary, had a wild shock of curly hair and a moustache, accompanied by a swagger that somehow injected the computer world with a shot of coolness that it never really recovered from. Prior to being hired at PARC, while still a graduate student at the University of Utah, he had envisioned a device he called the Dynabook, even going so far as to build a nonworking model of it; it looked remarkably like a cross between an Amazon Kindle and a laptop computer. Kay was attracted to the job at PARC because some of the computer whizzes he admired had recently taken positions in the rapidly expanding Palo Alto lab.
The third development, in 1971, was the emergence of silicon semiconductors, introduced by Intel, which quickly replaced the bulky ferrite core memory that had been the industry standard since the early 1950s.
The combined force of these factors resulted in the MAXC, Xerox’s answer to the PDP-10. The actuality of the MAXC, which at the time had the largest semiconductor memory of any computer on earth, was ultimately less important than the team and methods that created it. The MAXC took eighteen months to deploy, in contrast to the decades it had taken for most computers of the time to be assembled.
For all the potential success that the MAXC represented, it became a source of conflict between its designers at PARC and the Xerox executives back east. Xerox’s corporate officers typically viewed technological change as something to be monitored in order to protect the company’s business plan. They were interested in predicting future trends in a general sense, in order to understand what the world was going to be like. That way, Xerox could defend its existing product line against impending competitors. On the other hand, Alan Kay and his colleagues saw only innovation and opportunity ahead. Kay famously stated that “the best way to predict the future is to invent it.”
It is that faith in serendipity that made PARC’s accomplishments so impressive over the next decade. When headquarters demanded a concrete plan for what was ahead, PARC delivered a folder with seven reports, each solely written by a PARC scientist, outlining what he hoped to accomplish.
The vision outlined in that folder showed remarkable prescience. From portable flat-screens to CD-ROM-like photo optical media, the reports described innovations that are now humdrummedly mainstream but at the time seemed nothing short of futuristic. Over the next decade, as it failed to capitalize on nearly all of the innovations it clearly anticipated, PARC came to embody all that was wrong with the corporatization of American tinkering. Certainly, it is clear from chronicles of the era, that was not the fault of Alan Kay and his merry band of programmers.
In a fascinating article by Stewart Brand, the founder of the Whole Earth Catalog, that ran in Rolling Stone in December 1972, Kay, Taylor, and others at PARC were portrayed as staggeringly brash, knowledge-fueled hippies determined to wrest control of computers from the hands of stern corporate factotums who were only interested in the technology’s value as a high-tech abacus. Their esprit de corps was based less on some political notion or ideological cant than on the idea that computers could only achieve their full potential in the hands of individuals, not corporations.
But PARC’s style of group tinkering was perhaps its most valuable asset and contribution to the annals of innovation. One of the best examples of the form was known as “Beat the Dealer,” or just “Dealer,” after the book called Beat the Dealer by Edward O. Thorp, a professor at MIT who devised a card-counting system to win at blackjack. PARC’s version involved twenty or so of its researchers assembled on mustard-colored beanbag chairs as one of their ranks pitched a new project and the rest tried to find its flaws. The presenter, or dealer, though left to his own defenses to support his idea, had the advantage of being allowed to set the rules of the debate as well as the topic. One dealer used his time to show how to disassemble a bicycle and apply the proper lubricant. Another discussed at length how similar computer algorithms were to cooking recipes.
Along a similar line of reasoning grew a collaborative process known as “Tom Sawyering,” after the enterprising protagonist from the Twain novel. The concept here was that when a researcher came up with an idea for a new project or device, he would try to make it a reality by rallying those who were interested to help put it together. If the project began to show promise, the informal team would work on it for the next six months or so; but if it failed to gel, the participants could slowly migrate back to their own work and the project would simply dissolve.
In late 1972, PARC’s most infamous product would emerge from this ragtag tinkering process and forever change the course of computing for good. Alan Kay had spent the earlier part of the year sketching out a simple programming language for his modified Dynabook project, now known as the miniCOM, and instructed his engineers to make it a reality, which he named Smalltalk.
Kay’s follow-up proposal, to build a personal computer that ran Smalltalk, was met with chilly disdain by the suits at Xerox headquarters. So Kay waited until the executive responsible for the computer lab’s budget was out of the office for a couple of months on a special task force and then told his team to build the personal computer as fast as they could.
The computer’s design began in November 1972 and was completed by February 1973. All of those dealer meetings and Tom Sawyering sessions finally paid off, allowing the engineers to pull bits of knowledge from problems they had already explored and deploy them in building a new prototype. In classic tinkering fashion, the engineers assembled the new device from stray parts they had lying around. They repurposed memory boards that had been built for MAXC; the display monitors were from another project, a large networked system called POLOS that had more to do with the old world of massive shared computers than the compact individual units they were now creating.
The result was the Alto, the world’s first personal computer. The Alto included a screen about the same dimensions as a letter-size piece of paper held vertically. The monitor was large and bulky by today’s standards, but otherwise the Alto was way ahead of its time—it had a mouse that moved a cursor on a screen that showed exactly what the user would be printing. And while hardly the most powerful or fastest computer around, the Alto had other virtues, the most significant being the freedom it allowed operators, who could now sit at desks wherever they liked rather than in isolated rooms filled with cumbersome mainframes. Xerox manufactured around two thousand Altos in the coming months.
While the Alto was never introduced as a commercial product, it earned Xerox a place in the annals of computer history, though not exactly the one its creators at PARC had envisioned. The reason was the result of one soon to be famous visitor to the Xerox offices in West Hollywood, California.
His name was Steve Jobs.
By the summer of 1979, Jobs’s company, Apple Computer, was already a presence on the West Coast computer industry landscape. The Apple II had already been released and the fledgling company was readying itself for an initial p
ublic offering. At this point, Xerox and the folks at PARC had little awareness of Jobs and his distinctly countercultural operation. As for Jobs, he was unlikely to consider a partnership with Xerox, due to his suspicions of large, faceless corporations.
But then an Apple engineer told him about the project some of his buddies at PARC had been working on, and Jobs was intrigued. So when Xerox requested participation in the last round of financing at Apple before the public offering, he made them an offer. He agreed to sell the company 100,000 private shares of Apple at $10.50 per share in exchange for a simple visit to PARC’s research lab and an explanatory tour. Confident in their company’s supremacy, Xerox’s executives believed they had made a ridiculously advantageous deal. As developers of one of the largest and most powerful computers on the market, they regarded Apple as a manufacturer of technology for hobbyists. Besides, the PARC researchers had shown the Alto and Smalltalk to representatives from the Central Intelligence Agency with little consequence.
Little did Xerox suspect what Jobs’s true intentions were. Apple had been developing Lisa, its follow up to the Apple II, but Jobs had been dissatisfied with some of the more user-friendly aspects of it. After a number of skirmishes over how much of what they were working on they were required to show him, the PARC engineers demonstrated their graphical user interface, a new kind of computer interface that used graphic images instead of words, including a series of overlapping “windows,” which could be dynamically moved with a small, rounded pointing device known as a “mouse.” They also revealed that they could scroll text on the screen as if it were a piece of paper.
Jobs was blown away. In 1980, he requested a license to use Smalltalk in the Lisa. Xerox refused to grant it, having already cashed out its investment in Apple. So Jobs hired away one of Smalltalk’s creators, Larry Tesler, who would become a key developer of the Lisa and Macintosh computers, eventually rising to the position of Apple’s chief technology officer.