We unlock the door under the stairs and walk into the Shop. The stairs themselves constitute part of the ceiling, and there are gaps in the walls sufficient for viewing the setting sun. The air is damp and acrid with the smell of solder. The workbenches and shelves lining the walls are piled high with electronic gear, including two oscilloscopes, the KIM computer, breadboarded circuits, and row upon row of plastic ice cream containers filled with batteries and components. Doyne tunes a radio to KFAT, the local country station, and Willie Nelson comes over the air singing a love song.
Most of the Shop floor is taken up with workbenches, a Heath power supply, a drill press, and a band saw. There is also a bookshelf holding a small library of reference works, lab books, microswitch catalogues, casino brochures, and novels. In between The Radio Amateur’s Handbook and a volume on transistor-to-transistor logic is Joan Didion’s White Album. While most of the room is given over to Science, the western wall is reserved for Art. Hanging on it is a life-sized buffalo head fired out of clay and mounted on a trophy board. Surveying the room through red glass eyes, the buffalo wears a gold plaque under his chin whiskers that says “Manifest Destiny.” Below the buffalo is a circle of chairs surrounding a table made out of cinder blocks and a wooden shipping crate. Stationed on top of the crate, and balanced on three stainless steel jacks, is the Project’s B. C. Wills roulette wheel.
Doyne leans over the wheel and points to where the varnish has been burned by a stroboscope that exploded during the feasibility tests. He touches the chrome capstan and spins the rotor. Even under a bare light bulb, the red and black numbers meld into the hypnotic whirl of a roulette wheel in play.
But every few seconds the wheel goes thunk, thunk, until Doyne takes a piece of sandpaper and slips it into the tight spot where rotor and stator are rubbing against each other. “It’s sad to see the wheel looking like this,” he says. “We stored it for the winter under the stairs, and when I uncovered it in the spring I found a puddle of water sitting in the middle of it. The brass band around the edge of the rotor is badly warped. I should take the wheel back to Reno and get it refurbished, but for now we’re making do with sandpaper.”
When the rotor spins without rubbing against the housing, Doyne puts away the sandpaper and wipes the dust off the wheel and the mahogany rail that rises around it like the wall of a stadium. This venerable piece of wood, burnished through years of use, supports along its surface the parts to another game—a counter-force, an antiroulette—which will be played against the wheel in a contest of wits.
Not yet mounted onto a PC board, the Eudaemonic computer is still pinned to two beds of Styrofoam. Like the tendrils of a Day-Glo creeper, multicolored wires trail from the computer over the railing to several battery packs, two microswitches, and three solenoids. These make the system complete with a brain, power, input, and output. The switches, fastened onto pieces of wood, are worked with the index finger. Later, when built into shoes, they’ll be operated by toe. The solenoids are mounted into a metal plate designed to be worn against the stomach, but they too will eventually fit into a shoe. Doyne plugs the batteries into the computer, and I hear the tinking of relays as it powers up. He sets the rotor spinning at an easy lope. As it revolves on its steel spindle without a sound, we’re ready to play roulette.
With a practiced flick of the wrist Doyne launches a roulette ball onto the grooved track under the rail. It circles fast, passing in front of us a dozen times with the rattling sound of a coin rolling over a hardwood floor. It slows from friction and air resistance and then hangs for an instant on the lip of the track before surrendering finally to the pull of gravity. The descent toward one of the thirty-eight numbered pockets spinning below is charged with significance. While loose in space, the free-falling ball is open to chance, and as long as it struggles against entrapment on the central disk the game is still infused with hope.
The ball picks its way through the eight chromed diamonds dotting the stator. It whirls over the varnished bowl, strikes a fret separating two pockets, bounces and hangs for another instant before finally coming to rest. What was once an open field of possibility is now fixed into the determinacy of a number between 00 and 36. Over and over again, Doyne shoots the ball along the track. He taps the microswitches next to the computer with his left and right index fingers and watches the ball spin and fall.
“The switch on the left is for changing modes,” he says. “I’m using it to drive around the mode map. The other switch is for entering data. It allows me to increment and decrement parameters.”
He points above the wheel to Manifest Destiny. Taped to the buffalo’s nose is a copy of the mode map, which is nothing more than a series of interconnecting loops, with a clicking code that allows for switching from one loop to another. The loops represent domains in the computer program, and each domain specializes in handling a different part of the roulette algorithm. This separation of the program into modes allows it to be fine-tuned to the unique characteristics, or parameters of motion, that vary from wheel to wheel.
“I’ve just driven into mode five,” Doyne tells me. “That’s the one for setting the rotor parameter. Like all the other modes, it has a preset standard value that has to be incremented or decremented according to what we find in the casinos. Tonight, for instance, the wheel is slowing down faster than usual. So I want to jack the parameter around to take account of an increase in friction.
“Conditions on a wheel can shift dramatically from one day to the next. The casino can change balls on you, or oil the wheel, or move it to vacuum the carpets. That’s why it’s important to hone the parameters just before playing. Otherwise your predictions are going to be skewed. You might find the computer telling you to bet in the shadow—that part of the rotor where the ball almost never lands, and in that case you’re going to get creamed. Instead of a forty percent advantage, you’re losing at that rate.”
Doyne goes back to tapping microswitches and motoring the computer from mode to mode. I follow his path on the diagram taped to Manifest Destiny. As he steps through subroutines and adjusts parameters, the computer gets progressively smarter at tracking the game.
“It takes me about fifteen minutes to motor around the map, but it sometimes gets confusing. I get lost and can’t figure out where I am, or the computer crashes. You can’t pull out your mode map in the middle of a casino; so you have to have the instructions in your head. I’ve got to the point where I can cruise around the program without too many problems, but if I really get in trouble, I can do a total reset, which wipes out the parameters and dumps the computer back at the preset values.”
With the practiced fingers of a pro, Doyne steers the computer through the eight subroutines in its program. He taps an ambidextrous code for entering, adjusting, and leaving modes. He calculates the deceleration of the rotor by clocking two of its revolutions past a fixed reference point. He does the same for the ball. At still another location in the program he calibrates the time at which the ball tends to leave the track. Five other modes establish variables for playing on flat, tilted, or slightly tilted wheels.
“It looks tonight like mode four is the ticket,” Doyne says, on finishing his drive around the map. “Let’s see how it does against the wheel.”
He nudges the capstan and sets the rotor turning at a gentle clip. He launches the ball up on the track. Once the game is in play, every second counts toward beating its outcome with a prediction. Doyne clocks the green 00 on the rotor as it passes the reference point. He does the same for the ball. Four clicks…and the computer in microseconds can play through to its conclusion a game that in real life takes ten to twenty seconds to enact. The computer calculates the coefficients of friction, wind resistance, and drag at work on the ball. It figures the rate of deceleration, position, and time of fall from orbit. It knows in advance the speed, distance, and path of the ball’s trajectory down the sloping sides of the stator. It tracks the speed and relative position of the disk spinning below. It pinpoi
nts one of eight octants of numbers spaced around the perimeter of the wheel and announces, several seconds in advance of its happening, where the ball will finally land on the revolving disk.
I hear the faint buzz of a solenoid coming from under Doyne’s shirt. “Nine,” he says, translating a high-frequency buzz on solenoid number three into a “no bet” signal. “I must have clicked too late.”
Doyne again spins the ball on the track and clocks its progress over the varnished surface. Looking like a yogi meditating on a lower chakra, or a colitis victim suffering gas pains, he concentrates on the thumpers held to his stomach. “Six,” he sings out, referring to a high-frequency pulse on solenoid number two. Octant six on the wheel holds the numbers 30, 26, 9, 28, and 0. Doyne and I wait as the ball slows on the track, hangs for a moment, drops from the lip, arcs down the bowl, skids among the frets, and come to rest in pocket number 9.
“It looks good,” he says. “But you never know for sure how you’re doing until you compile a success histogram with hundreds of trials. The only thing we’re interested in is the law of large numbers.”
We spin the wheel and ball through dozens of mock games. Doyne calls out the predictions pulsed onto his stomach, while I plot a histogram of data points recording whether the ball lands in front of, behind, or dead center in the range of predicted numbers. Doyne establishes a rhythm for spinning the wheel, launching the ball, clicking the switches, and reading the buzzes off his stomach. Spin, launch, click, buzz. His body is sprung tight with concentration. His reflexes are perfectly sensitized for translating the motion of wheel and ball into the digitized clicks of a computer program. While simultaneously playing the roles of croupier and data taker, he looks like an athlete, maybe a basketball player, practicing lay-ups and free throws, alone on the court late at night, stretching and jumping for the boards over and over again. After years of practice he is loose-jointed and easy, but always in control and faultless in timing each move. Spin, launch, click, buzz. We record trial after trial while accumulating data points that rise in columns high over the x axis on our histogram. If one tower in the center of the graph rises higher than the others, then we have a clear-cut advantage over the casino; we should lace the computer into our shoes and head straight for Las Vegas.
“Give it a try,” Doyne says, backing away from the wheel. “You play, and I’ll keep the histogram.”
Stepping up to the computer, it takes me a while to get the hang of it. The data switch—a sliver of steel sprung over a contact point—is light to the touch. Trying to depress it exactly as the ball passes in front of the reference point, I sometimes click the switch too early. Other times I click too late.
“Don’t worry about clicking early or late,” Doyne tells me. “Just be consistent. At some point we’ll hook you up to the biofeedback machine so you can work on your hand-eye coordination.”
Reading the solenoids also requires some skill. The little mechanical thumpers are held in place under a layer of rip-stop nylon, and it takes me a while to gauge the difference between slow, moderate, and fast buzzes. Rather than slipping them under my belt, I find it easier to read the buzzers with my palm. Built into magic shoes, these little ticklers are eventually going to be popping off on the soles of my feet.
After a few false starts, I begin to get the rhythm. I learn not to hurry the clicks. I breathe easily and pace myself. I get better at deciphering solenoid buzzes and translating them into numbers. Standing under a bare light bulb, Doyne and I lean over the wheel for another hour. We listen to the ball rattle around the track and clatter down into the rotor’s metal pockets. We’re mesmerized, locked into the game so deeply that someone walking in on us might have thought he’d stumbled on the local numbers racket. This is my maiden voyage at the controls of the computer. I’m transported by the precision of the machine, the rhythm of the game, the whirl of numbers at the bottom of the varnished bowl, and the shine of the ball as it sails time and again for its rendezvous with destiny.
Rather than merely predicting the ball’s trajectory, it feels as if I’m actually steering it. I can nudge the controls of the computer and bring the ball in for a perfect landing time and again on the surface of the spinning disk. I work the switches, and the ball drops from orbit to touch down dead-center in the predicted octant of numbers coming around to meet it. It’s as if I’m controlling the outcome of the game, nursing the little white ball by remote control through the Newtonian cosmos of a roulette wheel whose laws I’ve mastered.
“Take a look at this,” Doyne says, showing me the histogram for the night’s play. A single tower of data points rises high over the x axis. The two columns flanking it to either side are significantly smaller. This is exactly what we’re looking for—a frequency distribution in the shape of an inverted V. Most of the data points are right on target: at the congruence of prediction and outcome, the jackpot, the winner-take-all-break-the-bank beauty spot.
“According to the data,” Doyne announces, “the computer is killing them. We’re mopping up the house.”
I imagine chips piling up in front of us. Stacks and mounds of them tossed around like Necco wafers. In this scenario the computer feels like an antigravity machine, a vacuum tube for sucking money from one side of the roulette table to the other. I grab the histogram for a closer look, while Doyne stuffs the solenoid plate under his shirt and launches the ball up on the track. There is a dangerous grin on his face as he steers the computer into another prediction.
14
Rebel Science
After Goliath’s defeat, giants ceased to command respect.
Freeman Dyson
Several months after our roulette session in the Shop, I get another phone call from Doyne. “We finished the first pair of magic shoes today. I’m wearing them right now.”
“How do they fit?”
“I was afraid to put them on. We still don’t know whether walking on the sandwich is going to work. But I’m wiggling the switches in my toes, and the shoes feel fine.”
By the time the Eudaemonic computer was ready to be compressed into a sandwich, I was no longer living in Santa Cruz. “How’s the weather out there?” I ask as a newly installed New Yorker inquiring after his old hometown.
“We’ve been working around the clock, really jamming,” Doyne says. “It looks like we’ll be leaving soon for a trip into the desert. What do you think about coming out for a visit, as soon as possible?” Telephone conversations about the Project are always oblique, and this one, beamed from coast to coast, is particularly understated.
It takes me a couple of days to disappear from the city and translate myself—under the guise of a sun worshipper going on holiday—from New York to California. Walking into 707 Riverside at dinnertime, I expect to find everyone sitting down to one of their copious meals. But to my surprise the kitchen is a free-for-all of people snacking on smoothies and tortillas, and the house is in tumult. Mobilized in a community alert, everyone is preparing for the final assault on the roulette tables of Nevada. Eudaemons walk around the house with computers on their feet. A roulette layout covers the dining room table. The click of roulette balls filters up from the basement. Everyday discourse has disappeared into technical argot about battery boats, mode maps, and histograms.
From April through the summer and into the fall of 1980 the Project had fought its way through a thicket of technical problems. There was dissension about who, if anyone, was to blame for the delays. Doyne had got sidetracked finishing his dissertation and then taken his first vacation in years—a six-week trip with Letty through Indonesia and Bali. Struggling in his spare time with the difficulties of intershoe communication, Norman had been preoccupied with the Chaos Cabal and strange attractors. Mark once again had gone out on strike, demanding more money and a larger slice of Pie.
Mark was also upset by the fact that Doyne had accepted a postdoctoral fellowship at the Los Alamos National Laboratory in New Mexico. The birthplace of Little Boy and Fat Man—the atomic b
ombs dropped on Hiroshima and Nagasaki—LANL has either produced or engineered most of their lethal offspring. Even though Doyne’s postdoc was at the Center for Nonlinear Studies, way off in the rarefied and supposedly harmless realms of the Theoretical Division, Mark still didn’t trust the arrangement. It merely confirmed his worst fears about physicists always ending up, one way or another, building bombs. He was even more upset when Norman won a NATO fellowship to spend a year studying at the Institut des Hautes Études outside Paris.
Mark demanded that Doyne phone Las Alamos and postpone his arrival for three months. Doyne agreed, realizing as everyone else did that it was now or never for finishing the Project. After their years spent perfecting the recipe, it was time to get the Eudaemonic Pie in the oven. The impending breakup of the household gave their work a special urgency. As long as the community held together, their dream of self-sufficiency remained alive. They still at the last minute might beat the wheel in a big way. This would give them the money they needed to break free of universities and governments, buy land together in the Coast Range, convert their scientific knowledge into tools for convivial living, and brainstorm a dozen other Eudaemonic projects.
The only thing standing between them and the dissolution of their enterprise was one small object: a microcomputer built into a shoe. On this slender piece of fiber glass and silicon rested their hopes for a last-minute reprieve. “I no longer have any romantic interest in playing roulette,” Norman said. “I’ve spent enough time in casinos. But if we could make twenty thousand dollars a month, then it would be worth it. I could kiss the NATO postdoc good-bye.”
The Eudaemonic Pie Page 26