RICHARD POWERS

by Unknown

  No one could say why, after thirty years of research in obscure labs across the Northern Hemisphere, VR overnight became 1990's cover girl. A couple of research outfits let the ghost out of the machine before it was time. Here and there, universities began to demo projects that suddenly had the whole world talking as if full-body dives into wraparound LSD, robotic prostitution, and long-distance teledildonics would hit the toy store shelves by Christmas. Two or three start-up firms, eager to appease their serial venture capitalists, began to sell cheap telegloves, stripped-down head-mounted displays, and even body suits whose performance amounted to little more than faint holograms of their hinted potential.

  In the Santa Barbara Sheraton, at March's research conference for virtual environment and telerobotics interfaces, Freese stood at the back of a packed grand ballroom. Just looking out on the sea of charged participants rearranged his viscera. His cobbled-up cottage craft had graduated beyond an esoteric discipline. Ready or not, reality engineering was about to become a full-fledged industry.

  A world begging for deliverance cared nothing for a porcelain jug sitting on a rickety wooden bed stand. But the inexorable market machine that had, just the previous year, swallowed up the globe's last holdout nations already knew what it wanted from virtuality. It wanted holophonic videoconferencing. It wanted the Ferris wheel-cum— feature film. All-talking, all-singing, incarnate sex fantasies. Interplanetary mining from the comfort and safety of our own back yard. But the market craved something more significant as well. Something more fundamental.

  Industry saw, in the Cavern and comparable virtual vistas, the race's next launchpad. The first commercial use of virtual space would be as a three-dimensional workbench for designing every physical trinket from saucepan to space station. In that lucid crucible, any conceivable device could be probed from all angles before incurring the expense of manufacture. Even the components for the next generation of Cavern itself could be taken for a 3-D test drive, revealing, in conceptual space, their optimal form before coming into the world. The amount of cash waiting to be thrown at the magic workbench, the sums waiting to be made could swallow the Rl, budget many times over, for generations upon generations. For the human project had many more goods to make, before its final triumph over goods.

  It was here already: the Pong of Things to Come. Downtown, a dozen blocks from Pioneer Square, where Spiegel and Adie had strolled only months before, a technopalace opened up where, for ten bucks, University of Washington kids and frustrated Boeing execs could sit in networked cubicles and blast deep-animated representations of one another out of the infinite vacuum of space. And the month after that arcade opened, Hollywood released the first of several feature-length spawn—a heavily chromed rendition of the new Aladdin and his wonderful data glove. The grand future vision that the RL pioneered was rapidly being left in collective imagination's dust. Must you Americans oversell everything? Rajasundaran asked. Freese liked the aggressive ones. Oversell? You can't oversell this. We're engineering the end of human existence as we know it. Not as I know it, White Man.

  Still, Freese insisted: it was the end. The end of something. An end to the limits of symbolic knowledge. Beyond the hype, past the immediate feeding frenzy, the press had gotten at least that much right. But even in the thick of the current mania, no one had yet guessed how big this thing was going to get. No one.

  The Cavern threatened the final disappearance of interface. Future operators would engage simulation in the same way that humanity's current version engaged material existence: using all the degrees of freedom built into their sovereign bodies. The right way to grasp the planet's mounting sandcastles of data was to step inside and poke around.

  As the scramble for funds broke out everywhere, Freese took his ideas on the lecture circuit. The computer would go transparent, more invisible than all its crude, qualified precursors in representation. Talking to data would be like talking to a friend over the phone. Explorers would move through a literal forest of numbers, strolling through their woody representations and singling out by sight or sound or smell the significant trees, the hidden arbors.

  Freese's techno-evangelizing carried a strong dose of private salesmanship. No other start-up in the fledgling reality industry had yet shown anything remotely in the league of the prototype Cavern. The Cavern, Freese teased, would make head-mounted displays and cumbersome gloves seem like Smell-O-Matic, SensaVision, or any other doomed evolutionary backwater. He always ended his speeches with the coy suggestion that everyone stay alive long enough to see the thing that their imaginations couldn't quite visualize yet.

  But privately, back at the mountain, he fretted. He sent off an anxious e-mail to the brass at TeraSys. The whole fad may quite simply fade before we get the real thing to market- In the current climatei potential clients for genuine immersion environments could well feel burned by their own expectations and sour on all subsequent demos-i once the bubble bursts...

  As project administrator, Freese managed a delicate balancing act between come-on and kiss-off. He could say nothing of the project and risk being lost in a sea of false claims. Or he could promise the world and risk failing to satisfy. Already, knots of prospective cyber-nauts were queuing up in the RL's parking lot, cash in hand. But the Realization Lab was worlds away from showing anything that resembled a finished product. All they had was proof of concept.

  Freese called a general meeting. Programmers, hardware jockeys, scientists, and designers assembled in the central atrium, the only nonvirtual auditorium large enough to contain them all.

  This may be the first time I've seen some of you in the daytime, Freese said. I'm surprised at how healthy you all look in natural light.

  Those are called monitor tans, Sue Loque called out from the gallery.

  I figured it wasn't the diet. First off, I want to applaud every member of this group for the distance we've already come. When I think of our technical and aesthetic advances in the two years since we put together the Crayon World, it feels ... He breathed in slowly and rolled his eyes. It feels as if I'm watching a film about evolution on fast-forward. Those of you who spend night after night chained to the workstation may have started to take monthly or even weekly breakthroughs for granted. I don't, I assure you. If this project were to move any faster, I'd be unable to keep up.

  He's about to tell us that it's time to pick up the pace, Rajan stage-whispered. The room exploded in laughter.

  Freese screwed up his mouth. It's time to pick up the pace. The room erupted again. Well, not so much the pace. I doubt any one of you could work any harder or more ... happily than you already do. What we need to accelerate, I suppose, is the release schedule.

  Vulgamott raised a hand. Run that one by us again, Chief? More, quieter laughter.

  Don't sweat the details, Michael. Here's the problem. We're all over ourselves, shattering yesterday's landmarks. We've gotten the polygon budget up from—what?—a couple of thousand per second? He looked at Spider Lim, who gave an infinitesimal nod. To ... what are we running now? I can't even keep track anymore.

  Spider cleared his throat. Over a hundred thousand per wall.

  From ten to the third to ten to the fifth. In two dozen months. I'd call that impressive.

  Lim, sensing the blow, stood up. Actually, we'll have to step up just as many more orders of magnitude before we can start to deliver believability without a lag.

  I agree. Jackdaw addressed his calculator watch. Reality demands something on the order of a hundred million. Reality ... is ten to the eighth surface-filled polygons a second.

  Minimum, Spider agreed, and sat down. Freese nodded. You see? This is the problem. Reality is always a problem, Spiegel said.

  The question is: when does the show stabilize and our act hit the road? At our current rate of change, the answer is never. The product would be forever obsolete before we got it out the door.

  O'Reilly raised his hand. It sounds like you're saying that Deep Pockets is wanting to see some
more near-term return?

  I'm afraid they want a public press conference for the spring. By next year's SIGGRAPH convention, we're to do a grand rollout. The popular press between now and then will be whipping the public imagination into a frenzy. We'll need to show something, just to compete with the rumors.

  Hardware, Software, and Design all took the floor to lodge their official reservations. But by the time the party broke up, the rules of the game had changed.

  Freese saluted them as they left. March of'91, then. Delivery Day.

  He wants us to be salesmen? Adie asked Ebesen, back in the cubicles. This is all just about selling iron?

  The old guy hunched his flannel shoulders. You knew it had to happen someday.

  No, I didn't, she said. It never occurred to me.

  Know what you should make? Lim told her. He was gutting Rembrandt again, tossing the machine's outdated entrails into cardboard boxes full of priceless scrap. A RAM room. You know: a huge blow-up representation of everything happening inside the real computer down at silicon level, right as it's running the simulation.

  She gave the notion three seconds. Bad idea. Am I to take it, from this mound of scrap metal, that we are obsolete again?

  They say that the Great Wall was obsolete before it was halfway finished. Your average printed circuit hoard is obsolete before it's even begun.

  Someone should go through all these junk piles, she said. It's getting hard to walk in here.

  Lim looked up, horrified. We can't throw any of that away. We might have to ... refer to it.

  Why? Why? She picked up a shoebox-sized assembly, once a miracle of miniaturization, a whole interplanetary system. Now incompatible with everything. She dropped the chunk of parts. Worthless.

  We can use some of those old motherboards as souped-up serial ports.

  Spider, these things are all dead. Killed by bigger, faster, better. Cartons of milk past the stamped expiration date. Tickets for last night's concert.

  TeraSys might be able to sell them to places that are still back at earlier machine levels.

  You mean that Bulgaria might be interested in running its own experimental virtual reality program, now that it's joined the Free World?

  I was thinking more like, you know, Arkansas?

  She mentioned the elephants' graveyard to Jackdaw and Rajan. This world digitization thing is the single most wasteful expenditure of effort in history.

  The kid bared his palms. You think the hardware side is wasteful? At least you can make those things into doorstops and paperweights. At least you can pirate last year's million-dollar state-of-the-art research tool for its edge connectors.

  Rajan chuckled. Right. Smelt them down to reclaim the two dollars and fifty cents' worth of gold plating on the pins.

  But software... ? Jackdaw said. Nothing is more pitiful than Version I. The biggest sinkhole of human genius in existence. The average lifetime of a given release is now shorter than the time it takes to leam its features. And as soon as Version 2 comes out, Version 1 turns into a time bomb in the operating system, just waiting to foul up any improvements in other software that postdates it.

  Let us put this in your terms, Rajan said. Suppose all of world art came down to the last three months of images. Every time an artist painted a painting, it invalidated all previous paintings of the same subject.

  Sure. That's called commercial design. I did that for half a dozen years.

  It's worse than that, Jackdaw cut in. Most development has no coordination to speak of. The wheel gets reinvented a million times a day, even in the same company. Even at the same workstation. It took half a century of coding to come up with reusable objects. And even now, they're not all that reusable, because, you know, the APIs and the hardware standards are changing the ground underneath them by the nanosecond.

  Rajan's cranium went into sympathetic oscillations. Truly demoralizing. Every basket of subroutines has to be invented dozens of times, each one doing the same thing in slightly different ways. Then all of these maddening, incompatible variants are thrown into public battle to determine which one will become the de facto standard, and everyone who puts money on the wrong flavor has to throw it out and start over again.

  Kaladjian came and stood nearby, cleaning off his glasses. Fortunately for all of us, waste is this culture's greatest engine.

  The others looked up, snagged by some expansive departure from his usual tone.

  Is that supposed to be ironic? Raj asked.

  Kaladjian gave a victimized shrug. Progress is destruction with a compass.

  Raj's nods accelerated a couple of hertz. It does make one wonder what the finish line looks like.

  Adie dear, Spiegel said. You've come to a world where truth is stamped with its own expiration date.

  Jackdaw grimaced. Not to mention the obsolete media. We still have these ancient tapes from before we ported to the Cavern? They can't be read anymore. The machines that used them have all been upgraded beyond compatibility. And even when we rebuild an antique drive from scratch? The tape has decayed; it spits out check-sum errors every three records.

  The world is losing its memory. Raj toyed with a stack of printouts headed for the shredder. Whole areas of the collective brain are being wiped out as its storage degrades. We've contracted a slow virus. Global Alzheimer's.

  Kaladjian lifted one shoulder. His tilted ear met it halfway. Perhaps. But look how far we managed to get, from flint to silicon, before the enterprise shut down.

  The Cavern caved in for several days, while Lim and company finished debugging a new generation of graphics accelerators. Deprived of their magic testing chamber, imagination's prototypers hit a wall.

  Maybe we should do a retreat or something, Vulgamott suggested to his fellow designers.

  Adie snorted. Maybe we should do a full-scale rout.

  Don't bail on me, please. I'm skidding out, here. Real deadlines. Real demos. No real place to test them. What's the imaginary world coming to?

  Ebesen said nothing. He was ready to accommodate—always the path of least resistance.

  Vulgamott got hold of a small cabin that TeraSys maintained up on the south fork of the Stillaguamish, near Mount Pilchuck. Art and Design booked the place for a forty-eight-hour stay. Ebesen's dirty flannel and corduroys, so squalid under fluorescent light, seemed almost indigenous, outdoors. Vulgamott, after two hours of the upland air, ceased twitching and began to breathe deeper. Adie went through a small sketch pad on the first afternoon. Thereafter she simply looked, with no more point than looking.

  In wildness, description fell away into its parent density. The three of them walked out in the woods, into the network of living agents, rooted, burrowing, and airborne. They drifted their feet in the bone-mashing cold of the river current, the rushing fluid still imprinted with its past life as mountain snow. At night, the curl of their campfire smoke rose and obscured itself in the Milky Way's fainter smear. The haunt of owls on the hollow night turned the listening heart against all hope of representation.

  They talked about what they had done, what they were doing, and what they would need to do before being anywhere near ready to release their work to the public. Months of mock-up had not yet even blocked out the floor plan of that furnished rec room of the cerebrum

  they pictured.

  The vines of Rousseau's Dream had spread, lovely and profuse. Its creatures had scampered in modest For-Next loops through the coded undergrowth. But the forest had remained a thought without a deed, a look without a behavior, flat and planar, less a living thing than a cadaver's cross section. A visitor could walk into the jungle moonlight, but only along fixed paths, strolling past the successive cardboard props of a tableau vivant.

  Out of this dream, they'd awakened to perspective. Their tools had all scaled up: frame rate, color depth, resolution, vertices per second. And the Aries bedroom exceeded the sum of these leaps. It zeroed in on that longed-for locale that no one had yet seen but everyone knew by sight. Its b
ed lay thick with invitation. The sun streamed through its casements, swelling and decaying in the length of a single visit. The wood floor bent to the weight of the current tenant. And yet even that humming space was no more than a single stereo slide. The bedroom filled out its frame, but no farther, refusing to venture beyond the grotto that housed it.

  Now Design had to plan its next escape. Under the sap-heavy trees, the chilled antics of a Cascade stream between their toes, the digital artists turned over the problem, less through talk than with shared scribbles. The task was obvious. They needed a way to wed inimical worlds, to combine the dream of these two chambers.

  Half a dozen months, Vulgamott repeated, past the point when either of his colleagues heard him any longer. We're in a situation here, people. It's demo or die.

  Or both, Ebesen said. "Both" remains a distinct possibility. In her mind, Adie wanded off down hinted-at ravines, lost in the extensions of sight, looking for the room they had to reach. The trick was how to find it without clues. How to resolve the place, without knowing what it looked like. In rapid succession, they torched each proposal put forward. All possible rooms either cloyed or curdled, too banal or too vaporous, too mundane or too incorporeal. Nothing both satisfied desire and yielded to available technique.

  No more paintings, Adie said. We tried that twice. We want something that will break out of the frame.

  All three knew the medium they would have to inhabit, already laid out for them. Vulgamott and Ebesen's architectural tool chest—now numbering in the hundreds of modular components, from the simple I beam to the ornate ogee molding—all but forced their hand. Their resizable image library had grown into an encyclopedia of smart architectural elements, one that made it possible for any reasonably patient person who could manage a pipe-cleaner sculpture or a box of Lincoln Logs to build her own pan-and-zoom Versailles.

  With the suite of Palladian tools, prototyping a simple architectural fly-through shrank from months to weeks. The kit had never been meant as anything but its own demo, a proof of concept rather than a mission-critical development tool. Now it represented their only chance at hewing out a substantial show by press date. Even here, on the verge of the virtual, they were condemned by those absurd constraints, time and practicality.