The Idea Factory: Bell Labs and the Great Age of American Innovation

by Jon Gertner

  The first working laser—the name came from a man named Gordon Gould, a former associate of Townes, who also made a successful legal claim to the invention—was not built at Bell Labs. Nor was it built by Schawlow and Townes. Rather, it was developed at Hughes Aircraft, in Malibu, California, by an engineer named Ted Maiman. In Maiman’s design, which began functioning in mid-May 1960, a flash of bright light stimulated a small pink ruby that emitted a short and powerful pulse of focused light. The Maiman laser didn’t prove that the laser was guaranteed to have any practical value—that question was unresolved and still far off in the distance. But it did prove that the laser could actually exist beyond the theory outlined by Townes and Schawlow. Indeed, almost as soon as the optical researchers at Bell Labs heard of Maiman’s work, they built a near-exact copy of Maiman’s device to replicate and verify his results.10

  At around the same time, another group of researchers at Murray Hill—Ali Javan, Donald Herriott, and William Bennett—were trying to build a different kind of device. On December 13, 1960, the men succeeded in operating a laser that used a gas medium of helium and neon, rather than a ruby. Aided by a series of mirrors that could focus the energy emitted by the stimulated gases, the men succeeded in creating a narrow beam of light. Most important, it wasn’t a pulse; it was a steady and continuous beam.11 The day after Javan’s team got their laser working, a team of Labs engineers used the focused beam of light to carry a telephone call.12 That sort of thing made AT&T executives actually sit up and pay attention.

  THE LASER WAS NOT so much a single invention. Rather, it was the result of a storm of inventions during the 1960s. Noteworthy improvements (like a new design) or variations (like getting a new material to “lase”) followed one after another in rapid succession. Sometimes only a few days separated the announcements of new developments. “The whole business of [lasers] is one of these things you just can’t afford to let go,” John Pierce told an interviewer around that time. “You can’t clearly see that it will be of any use in communication. I mean, we certainly can’t guarantee that. But it has that potentiality.” Pierce added that “when something as closely related to signaling and communication as this comes along, and something is new and little understood, and you have the people who can do something about it, you’d just better do it, and worry later just about the details of why you went into it.”

  Rudi Kompfner, Pierce’s deputy and close friend, shared that sense of urgency. And so Kompfner, with Pierce’s and Bill Baker’s endorsement, started making a list. “He went around the world at that time, in 1960, trying to find good people—that’s all he wanted, good people,” recalls Herwig Kogelnik. “And then he would try to persuade them to switch their disciplines to take on what he called ‘laser and optical communications research.’ ” Even within the scientific community, the terms were new and strange. Kogelnik was finishing up his PhD at Oxford when Kompfner showed up one day during his worldwide recruiting tour. Kogelnik was a plasma physicist, which meant he studied the physics of certain gases. Rudi Kompfner didn’t care. His simple argument to Kogelnik on the day they met was that the young man should think of the high frequency of light and what that could mean in terms of its capacity for information. Colleagues all recall the charm as well as the passion that animated Kompfner. To Kogelnik, he just said, “Think of all this bandwidth!”—a line that inspired Kogelnik to switch from plasma to lasers. “I had invested many, many years in plasma physics. And he persuaded people like me to totally throw away their past and start in a new field.”13

  Simply put, if the Bell scientists could figure out how to use light’s vast capacity to transmit phone calls, data, and TV, they could avoid future worries about congestion.14 What was also attractive were the economics of optical transmission. For decades, the Bell System had realized that it was far more cost-efficient to mix together many hundreds of conversations on an intercity copper cable—by a complex technical means, the signals could be sent together at a higher frequency and then teased apart at the receiving ends. Sending more information and sending it more economically were often the same thing.

  But as the 1960s wore on, this possibility seemed increasingly remote. Several years after Townes and Schawlow outlined the laser in their paper, a variety of ticklish problems, as one Bell Labs research director described them, remained unsolved.15 For starters, there were a host of technical challenges in finding the best laser materials with the most useful frequencies for communications. At the same time, Bell engineers were still working on a technology that could modulate voice and data signals and then “impress” those signals upon the laser beam. Above all, there were nagging questions about transmission. Without question, light could carry voice and data—but how would it be sent around the country? Through the air? In a hollow pipe? In September 1960, just a few months after the ruby laser was invented, several scientists at Bell Labs succeeded in sending a pulse of laser light twenty-five miles through the air from Crawford Hill in Holmdel to the Murray Hill labs.16 The laser group then began to realize that the earth’s atmosphere creates all types of interference. “Rain, snow and fog can cause heavy power losses,” a Bell Labs director, Stewart Miller, explained in Scientific American. That left another possibility—the waveguide, those hollow pipes, beset with problems and still in development, that were ultimately supposed to relieve the future traffic on the Bell System by carrying millimeter radio waves long distances. They could conceivably carry light waves, too. But sending light through a waveguide would be immensely difficult: The beam would have to go around turns and up hills and down hills and remain perfectly focused within the walls of the hollow pipe.17 An intercity waveguide might have to do this for hundreds of miles. Nobody at Bell Labs—not Fisk, not Baker, not Pierce, not Kompfner—seemed to think this was imminent. Light waveguides were, at best, a next-generation technology.18

  There was, perhaps, another option. It was the idea of a scientist well outside the elite ranks of Bell Labs. In 1966, an engineer named Charles Kao, employed in England by International Telephone and Telegraph, visited the Labs to talk about a technology he was researching in Europe. Kao had recently delivered a paper at an engineering conference in London suggesting that transparent glass fibers, carrying waves of light, might solve the transmission problem. Some of the scientists at Bell Labs had already toyed with this possibility. It had long been clear to engineers that thin glass strands could transport light for short distances; such fibers, in fact, were already being used in medicine, where they were proving useful for gastrointestinal examinations. But those working in the communications field doubted that glass fibers could transport signals the much greater distances the phone system required. The glass just wasn’t clear enough. A signal would be lost—technically speaking, it would scatter and attenuate—after only a few dozen meters. Fibers were, as the engineers put it, too “lossy.”

  Kao was more optimistic. He had spent the past few years in Europe doing deep exploratory research, and he had determined there was no fundamental reason why strands of exceedingly pure glass couldn’t carry signals substantial distances. His paper was partly theoretical. Glass of the sort Kao was talking about didn’t really exist; even the clearest glass on earth would effectively absorb or scatter light and thus kill a signal in a few dozen meters. But Kao hadn’t said that pure glass would replace wires or waveguides immediately. He had only concluded it was possible.

  He was also liberated to some extent from the pressures that shaped the views of the Bell Labs scientists. Labs upper management had bet the future on waveguides, but Kao had not. The fiber optic historian Jeff Hecht would later point out that Kao (unlike the accountants at AT&T) had no incentive to make years of investment, in both time and effort, pay off. What’s more, there was a pull on Kao in a different direction. Innovations are to a great extent a response to need. Phone engineers in Europe—Kao included—weren’t looking for a complex new technology, such as the waveguide, for intercity communication. They needed int
racity communication. Generally speaking, Europe’s metropolitan areas were both denser than America’s and closer to other metro areas. The British telecom planners, Jeff Hecht notes, “wanted better technology to send signals between local switching centers that typically were a few miles apart. They wanted something easy and inexpensive to install in heavily developed areas, not high priced, huge capacity systems to span vast distances.”19

  When Kao visited Bell Labs and urged the people there to follow up on this line of research, the optical team under Kompfner heeded him. “I remember very well that day,” says Tingye Li, who would eventually help lead Bell Labs’ optical research efforts and would become a close friend of Kao’s. “We were having a picnic on top of Crawford Hill and he joined us. I had not met him before.”20 Kao would later say that he received a skeptical welcome at Bell Labs; but Li and several researchers who were there at the time recall it differently. “I don’t think anyone pooh-poohed it,” recalls Ira Jacobs, who would later become deeply involved in the Labs’ fiber optics development. “But I think there were a lot of people who thought it would move slowly.”21 As the 1960s wore on, Kompfner’s group became increasingly interested in the possibilities of fiber, and by 1969 there was a fiber research program up and running.22

  Yet the Bell Labs brain trust had also concluded that their present strategy of hollow pipes remained more feasible. Apparently, Jim Fisk and Bill Baker agreed. “A telecommunications network is emerging in our nation which will match the historic upsurge of our society’s demand for the transmission and distribution of human knowledge and personal experience,” Baker said in a speech about networks for the future at a Chicago symposium in 1968. The “new science which will provide the massive telecommunications capabilities of the future,” he explained, still depended on waveguide pipes, which would carry all the traffic and easily handle the demand for Picturephone circuits. The future, in other words, still looked the same to Baker as it had to Fisk at the beginning of the decade: the Picturephone, waveguides, electronic switching. These were the Bell System’s bets, and they were sticking with them.

  ANY SCIENTIST WHO WORKED at Bell Labs—especially anyone in Bill Baker’s research department, whose job was probing the unknown—understood that failure was a large part of the job. Experiments sometimes literally exploded; results often disappointed; gut feelings frequently turned out to be indigestion. Moreover, new innovations that portended a grand future—the germanium point-contact transistor, for instance—could quickly be rendered irrelevant by a new iteration of a similar idea, such as the silicon transistor or (later still) the integrated circuit. In retrospect, of course, the evolution of technology looks like an ever-ascending staircase, with one novel development set atop another, leading incrementally and inevitably to all the benefits of modern life. Bill Baker was expert at crafting precisely this public narrative. “Within my lifetime,” he testified to the Federal Communications Commission in 1966, “the United States has progressed from a nation held together by post and telegraph, in which the ability to span the country by telephone had barely been demonstrated, to one in which the complex telephone network is an indispensable component, intimately linked with the growth and operation of communities, private organizations, government, public safety and national security.”23 In truth, Baker, having spent years at the bench working on chemistry experiments, knew that science and technology weren’t a matter of assured upward progress. The waveguide, for example—those hollow pipes that had entailed years of research and millions of dollars—effectively died as a possible future in the fall of 1970. That September, the Corning glass company announced that it had succeeded in creating thin glass fibers so pure that they could transmit light with very low losses for thousands of feet. Corning did not yet have a manufacturable product, but they had demonstrated that Charles Kao’s research and intuition were correct: There was a future for lightwave communications, and it was almost certainly in fiber optics rather than in waveguide pipes. By the end of November 1970, Rudi Kompfner agreed: “Optical fibers will be at the center of the stage in the future,” he wrote to a colleague.24

  No one was keeping an actual ledger during the 1960s on how well Bell Labs was doing in planning for the next century, but had they been, they might have put electronic switching and the laser work in the column of spectacular and prescient successes. During those same years, there were other achievements at Bell Labs that would, in time, alter the world. One occurred when several computer scientists at Murray Hill got together to write a revolutionary computer operating system they called Unix, which would serve as the basis for a host of other computer languages.25 Another breakthrough—the charge-coupled device, or CCD—was invented during work on new forms of computer memory. The CCD was a light-sensitive electronic sensor that used the varying responses of electrons to different amounts of light to create photographs and images of extraordinary detail.26 The CCD would become the foundation of digital photography as we now know it, but Bill Baker perceived its value to national security long before its commercial potential was understood. “I took it out to the NRO immediately,” Baker would later recall, knowing that the National Reconnaissance Office, the spy satellite agency he had helped found, could use it for espionage.27

  Meanwhile, the other side of the ledger was filling up, too. The integrated circuit of Kilby and Noyce, which built on the engineering and materials-science achievements of Bell Labs, seemed more a missed opportunity than a misstep.28 But nobody could offer such a mitigating rationale to explain the waveguide and Picturephone, two interrelated and fabulously expensive follies.29 It seems worth considering not only how those endeavors failed, but what those failures represented. Innovators make different kinds of mistakes. The waveguide, for instance, might be considered a mistake of perception. It was an instance where a technology of legitimate promise is eclipsed by a breakthrough elsewhere—in another corporate department, at another company, at a university, wherever—that solves a particular problem better. It was perhaps understandable, moreover, that a breakthrough in the creation of pure glass fibers wouldn’t come from an organization such as Bell Labs, where materials scientists were experts on the behaviors of metals, polymers, and semiconductor crystals. Rather, it would come from a company like Corning, with over a century of expertise in glass and ceramics.

  Mistakes of perception are not the same as mistakes of judgment, though. In the latter, an idea that developers think will satisfy a need or want does not. It may prove useless because of its functional shortcomings, or because it’s too expensive in relation to its modest appeal, or because it arrives in the marketplace too early or too late. Or because of all those reasons combined. The Picturephone was a mistake in judgment.

  The Picturephone began on a high note of optimism. “We have now received a clear go-ahead from AT&T on the Picturephone program we proposed,” Julius Molnar, Bell Labs’ executive vice president, informed the staff in late summer of 1966.30 The actual Picturephone technology was being upgraded and redesigned; instead of the egg-shaped futuristic device that had made a splash at the World’s Fair, Molnar told the staff, the set would be a “Model 2,” or Mod 2, as it was called at Bell Labs, a squarish device, designed by a renowned industrial designer, that was both more elegant and more functional. Molnar’s goal was to field test the device in 1968 and begin a rollout of “commercial face-to-face picturephone service” soon after.

  A small exploratory marketing study of the Picturephone, comprising ninety-nine employees from major corporations and nonprofit institutions, was compiled at the end of 1967. Meant to investigate the views of “a cross-section of business customers,” the study’s conclusions sounded no alarms. The market potential for Picturephone service appeared to be strong among the survey participants, the study concluded: “Many would be willing to pay more than $50 monthly for a Picturephone service designed to meet their needs.”31 By the spring of 1968, there was no turning back. “The trials of Picturephone have now progresse
d to a point,” Bell Labs president Jim Fisk said in a speech at the time, “where any skepticism as to its interest and utility is only a replay of the skeptical response which greeted [Alexander Graham] Bell when he tried to promote the telephone over 90 years ago.”32

  OFFICIALLY, the Picturephone rollout began with a trial at the Westinghouse Corporation in Pittsburgh, starting in February 1969; during the summer of that year Bell Labs devoted an entire issue of its magazine, the Bell Laboratories Record, to explain the science and engineering of the new launch. The possible impacts of the Picturephone, Julius Molnar suggested in an introductory note, could well be seismic: By lessening the need for shopping trips or for conducting in-person business, “there will be less need for dense population centers,” as well as reduced traffic. “Picturephone is therefore much more than just another means of communication,” Molnar wrote. “It may in fact help solve many social problems.” AT&T began offering Picturephone service in Pittsburgh and Chicago at the end of 1970. Other electronics companies—RCA and GTE, for instance—readied similar technologies. If video telephones were the future, they saw no reason to let the Bell System reap all the rewards.33

  But within about twelve months, Bell executives saw that the anticipated demand for the Picturephone service was not materializing. In a speech to Bell Labs’ department heads in March 1972, Julius Molnar went through the results: “Most of you probably know that attempts to introduce it in Pittsburgh and Chicago have hardly been howling successes,” he began. “In Pittsburgh after a year and a half there are only eight paying customers with 30 sets in service.” The monthly price in that city for Picturephone service was $160. But in Chicago, where the price was set at a cut-rate figure of $75 per month, the results were nearly as worrisome. “After a vigorous sales campaign,” Molnar acknowledged about the Chicago business, “they have 46 customers with 166 sets in service, with another 128 sets on order. Much better than in Pittsburgh, but still not quite encouraging.”34