LEC, of course…or at least you should, if you pay attention to TV commercials, browse the web, or visit shopping malls. Lang Electronics Corporation is one of the three major U.S. manufacturers of consumer robots; it started out as a maker of IBM-clones in the early ’80s, then diversified into robotics shortly after the turn of the century, introducing its first-generation robot vacuum cleaners and home sentries about the same time that its closest competitors, CybeServe and Cranberry, entered the market with their own household ’bots. CybeServe was the leading company, and solidified that position after it was bought out by Mitsubishi; Cranberry, on other hand, was hurt by poor sales and a reputation for making second-rate ’bots that tended to forget instructions, burn actuators, and taser the mailman. By the time CybeServe and Mitsubishi merged, Cranberry had laid off one-third of its employees and was on the verge of declaring bankruptcy.
This left LEC in somewhere in the middle. It remained strong enough to fight off hostile takeover attempts by larger electronics companies in both America and Japan, and its Valet and Guardian series of home ’bots held their own in the marketplace, not only selling as many units as CybeServe but even surpassing their sales in Europe. The success of its first-generation robots prompted LEC to invest considerable capital in developing a second-generation series of universal robots. Biocybe Resources in Worcester, Massachusetts, had recently introduced its Oz 100 biochips, pseudo-organic microprocessors capable of handling 100,000 MIPS—Millions of Instructions Per Second, the robotic equivalent of megabytes—and LEC had built them into its Gourmand, Guardian III, and Companion ’bots, successfully bringing them to market nearly two months before CybeServe brought out their rival systems. It also helped that CybeServe’s ’bots were more expensive and that their CybeServe Butler had an embarrassing tendency to misunderstand questions or commands given in less than perfect English (e.g., “Is the dishwasher running?” No, it’s still in the kitchen. “Answer the door, please.” But it hasn’t asked me anything. And so forth.).
(If all this is beginning to make your eyes glaze over, please be patient. Home ’bots may be rather commonplace these days—if you don’t already own one, chances are one of your neighbors does, and your kids may be dropping hints about how nice it would be to find a CybeServe Silver Retriever or a LEC Prince barking and wagging its tail beneath the Christmas tree—but I’m relating events which occurred about ten years ago. It may seem like business talk, but it has quite a bit to do with the story at hand, so bear with me, okay?)
CybeServe wasn’t about to let itself get stampeded the way Cranberry was several years earlier, so after it spent a small fortune working out the bugs in its second-generation ’bots and an even larger fortune in consumer advertising, it took the next logical step: the development of a third-generation, all-purpose universal robot, one which could serve as butler, housekeeper, sentry, cook, chess-player, dog-walker, babysitter…you name it. And just to put the icing on the cake, CybeServe intended its new ’bot to be humanlike: bipedal, about six feet in height, with multijointed arms and legs and five fingers on each hand.
This was probably the most significant factor, for with the exception of a few experimental prototypes like Honda’s P2 of the late ’90s, virtually every robot on the market looked like a fire hydrant, an oversized turtle, or a vacuum cleaner with arms. A humanlike robot, however, would not only be aesthetically familiar, but it would also be able to adapt more readily to a household environment, since it would be able to climb stairs or place objects on tables.
Although CybeServe tried to keep their R3G program secret, the cybernetics industry is small enough—and the Robot Belt along Route 9 in Massachusetts short enough—that it was only a matter of time before word leaked out of its Framingham headquarters. The fact that their R3G project was codenamed Metropolis, an ironic allusion to the robot in the 1927 silent film directed by Fritz Lang, was a clear signal that CybeServe meant to pull an end-run around its rival in Westboro.
When Jim Lang, LEC’s founder and CEO, learned that CybeServe was actively engaged in the development of a third-generation ’bot, the lights stayed on all night in the fourth-floor boardroom. The following morning, Slim Jim summoned his department heads to the executive suite, where he read them the riot act: LEC was now in a race with CybeServe to be the first company to produce a third-generation universal robot.
As luck would have it, though, the company wasn’t caught flat-footed: during their spare time, two of its senior engineers had already been working on third-generation robots.
Where Phil Burton or Kathy Veder managed to find any spare time at a company where everyone in the R&D divisions typically puts in a 7-by-14 work week is beyond me, yet nonetheless these two had been using their downtime to tinker in their labs. On their own initiative, both Phil and Kathy had drafted plans for universal ’bots which would utilize the new Oz chips being produced by Biocybe. Since the Oz 3Megs were capable of processing three million MIPS, this meant that a third-generation robot could have the approximate learning ability of a Rhesus monkey, as opposed to a second-generation ’bot with the IQ of a well-trained mouse.
The fact that they had designed their robots independently of each other, without one being aware of what the other was doing, was no great surprise to anyone. Phil Burton was in charge of the division which developed the Companion robot, while Kathy Veder was the senior engineer behind the Guardian III. Their departments were located at opposite ends of the LEC quad, and their staffs shared little more in common than the company cafeteria. Not only that, but the two couldn’t be more unalike: Phil Burton, tall and rather skinny, with thinning blond hair, and a lifelong stutter which betrayed his shyness, and Kathy Veder, short, plump, with unruly black hair which was seldom combed and an aggressive manner which bordered on outright hostility (hence the nickname). A pair of über-geeks who couldn’t have agreed on the proper pronunciation of banana if someone threatened to take away their Usenet accounts.
Nonetheless, Lang was delighted that they already had a head-start, and asked them to show him their work. However, Kathy was a little more reluctant than Phil to comply; in fact, rumor had it that Jim had to memo Darth three times before she finally coughed up her notes and blueprints, while Phil delivered his material almost immediately. The rest of us chalked up her reticence to peer rivalry, never realizing that there was something else going on just under the surface.
Lang carefully studied their plans, talked to some of his other geeks—myself included—and eventually reached the conclusion that, although each robot was designed differently, they were so fundamentally similar that either could serve as LEC’s entry in the R3G race. However, since the company didn’t have the time, money or resources to manufacture two third-generation ’bots, it was one or the other. To make matters worse, there was no accord among the brain trust upon which robot should be chosen; Kathy’s people were solidly behind her Guardian IV design, while Phil’s group was equally convinced that Companion II was the superior system.
Jim Lang loved strategy games. He collected antique chess sets and backgammon boards, and was renowned among Go enthusiasts as something of a master. Indeed, when LEC was a small start-up company in the late ’70s, its first major product had been a modular pocket game system, the now-forgotten Lang Buddy. So it came as no great surprise that his solution took the form of a competition.
Dr. Burton’s group and Dr. Veder’s group were divided into two teams, respectively code-named Samson and Delilah, with Dr. Burton and Dr. Veder as their leaders. Each team was given a substantial R&D budget and access to the same material resources, not the least of which were copies of the Oz 3Meg chips. However, the members of each team would not be allowed to talk to one another or share notes; only the team leaders were given that privilege, if they saw fit to do so.
The objective of Slim Jim’s game was the fast-track development of a fully-operational, self-learning universal robot within six months. At the end of this period, each team would let their robots be tested—first by themselves, then interacting with each other—in a series of environments approximating real-world conditions. The team which produced the best robot
