But the BIAC did much more than simply enable data and instructions to be fed into the machine more quickly; it enabled the machine to accept input material of a completely new type. Whereas classical computers had required every item of input information to be explicitly specified in numerical or encoded form, the BIAC, incorporating the most up-to-date advances in adaptive learning techniques, could respond to generalized concepts—concepts visualized in the operator's mind—and convert them into forms suitable for internal manipulation.
It thus functioned more as a supercomputing extension of the operator's own natural abilities, its feedback facilities evoking in him a direct perceptual insight to complex phenomena in a way that could never have been rivaled by mere symbols written on pieces of paper. The dynamics of riding a bicycle can be represented as a complicated string of differential equations, the solutions of which will infallibly tell the rider what he should do to avoid falling off when confronted by a given set of conditions—speed, curve of road, weight of rider, etc. The young child, however, does not concern himself with any of this; he simply feels the right thing to do—given some practice—and does it. In an analogous fashion, the BIAC operator could feel and steer his way through his problem. It was the perfect tool for handling Clifford's k-function solutions.
Only a handful of BIACs had been built, and all of them were undergoing government evaluation trials under conditions of strictest security. The offer to make available to Sudbury one of the next three scheduled to be built provided, therefore, as convincing a measure as anyone could ask for of the significance attached to the Institute's work. Even so, it would take three months or so for the machine to become available.
Security of the BIAC posed a problem that had to be solved during that period. Dismantling the GRASER and the detector and shipping them elsewhere would have been possible as a last resort, but the magnitude of the task promised to be horrendous. Eventually Peter Hughes suggested an arrangement that, although falling below the requirements usually stipulated for that type of situation, was granted a special dispensation. Structural alterations were made to the GRASER building to seal off all entry points apart from the main door and a fire exit at the rear, which was operable from the inside only. Everything and everybody not directly involved with the project were moved into other accommodations elsewhere at the Institute. Then, finally, access to the building was limited to a few specially designated people, and two officers of the State Police were to be stationed at the door around the clock to insure that the rules were observed.
Clifford saw these developments as portents of things to come, and his misgivings intensified. Life took an unexpected turn, however, and soon he was too preoccupied with other things to brood about such matters. He was sent away for six weeks to undergo an intensive course in BIAC operation on a machine already installed at the Navy's equipment evaluation laboratories in Baltimore. Aub remained at Sudbury, being too immersed in the design details and preparations for Mark II to afford any time away. He would follow later.
* * *
For the first couple of days after his arrival in Baltimore, Clifford sat through a series of lectures and tutorials aimed at imparting some essential concepts of BIAC operation and at giving the class some preliminary benefits from the techniques that others had developed.
"The BIAC becomes an efficient tool when you've learned to forget that it's there," one of the instructors told them. "Treat it as if you were learning to play the piano—concentrate on accuracy and let speed come in its own time. Once you can play a piano well, you let your hands do all the work and just sit back and enjoy the music. The same thing happens with a BIAC."
Eventually Clifford found himself sitting before the operator's console in one of the cubicles adjacent to the machine room while an instructor adjusted the lightweight skull-harness around his head for the first time. For about a half-hour they went through the routine of calibrating the machine to Clifford's brain patterns, and then the instructor keyed in a command string and sat back in his chair.
"Okay," the instructor pronounced. "It's live now. All yours, Brad."
An eerie sensation seemed to take possession of Clifford's mind, as if a bottomless chasm had suddenly opened up beside it to leave it perched precariously on the brink. He had once stood in the center of the parabolic dish of a large radio telescope and had never forgotten the experience of being able to shout at the top of his voice and hear only a whisper as the sound was reflected away. Now he was experiencing the same kind of feeling, but this time it was his thoughts that were being snatched away.
And then chaos came tumbling back in the opposite direction—numbers, shapes, patterns, colors . . . twisting, bending, whirling, merging . . . growing, shrinking . . . lines, curves. . . . His mind plunged into the whirlpool of thought kaleidoscoping inside his head. And suddenly it was gone.
He looked around and blinked. Bob, the Navy instructor, was watching him and grinning.
"It's okay; I just switched it off," he said. "That blow your mind?"
"You knew that would happen," Clifford said after he had collected himself again. "What was it all about?"
"Everybody gets that the first time," Bob told him. "It was only a couple of seconds . . . gives you an idea of the way it works, though. See, the BIAC acts like a gigantic feedback system for mental processes, only it amplifies them round the loop. It will pick up vague ideas that are flickering around in your head, extrapolate them into precisely defined and quantitive interpretations, and throw them straight back at you. If you're not ready for it and you give it some junk, you get back superjunk; before you know it, the BIAC's picked that up out of your head too, processed it the same way, and come back with super-superjunk. You get a huge positive feedback effect that builds up in no time at all. BIAC people call it a 'garbage loop.' "
"That's all very well," Clifford said. "But what the hell do I do about it?"
"Learn to concentrate and to continue concentrating," Bob told him. "It's the stray, undisciplined thoughts that trigger it . . . the kinds of thing that run around in your head when you've got nothing in particular to focus on. Those are the things you have to learn to suppress."
"That's easy to say," Clifford muttered, then shrugged helplessly. "But how do I start?" Bob grinned.
"Okay," he said. "Let's start by giving you some easy exercises for practice. Try ordinary simple arithmetic. Visualize the numbers you want to operate on, concentrate hard on them and also on the operation you want to perform, and exclude everything else. Get it fixed in your mind before I switch you in again. Okay?"
"Just anything?" Clifford shrugged. "Okay." He mentally selected the digits 4 and 5 and elected to multiply them together, just to see what happened. The torrent of chaos hit him again before he realized Bob had hit the key.
"That was a bit sneaky of me," Bob confessed. "The best time to slot in is often when the problem is clear in your mind. Try again?"
"Sure."
After three more excursions round the garbage loop, Clifford sensed something different. Just for a split-second it was there; the concept of the number 20 seemed to explode in his brain, impressing itself with a clarity and a forcefulness that excluded everything else from his perceptions. Never before in his life had he experienced anything so vividly as that one simple number for that one brief moment. Then the garbage came at him again and swallowed it up. For a while he just sat there dumbstruck.
"Got it that time, huh?" Bob's voice brought him back to reality.
"I think so, at least for a second."
"That's good," Bob stated, encouraging his pupil. "You'll find for a while that the shock of realizing it's working distracts you enough to blow it. You'll get over that though. Don't try and fight it—just ride it easy. Try again?"
An hour later Bob posed the problem, "Two hundred seventy-three point five six multiplied by one hundred ninety-eight point seven one?"
Clifford gazed hard at the console, visualized the numb
ers, and almost immediately recited, "Fifty-four thousand, three hundred fifty-nine point one zero seven six."
"Great stuff, Brad. I reckon that'll do for a first session. Let's break off for lunch and go have a beer."
* * *
A week later Clifford was learning to cope with problems in elementary mechanics—situations involving concepts of shape, space, and motion as well as numerical relationships. He found, as his skills improved, that he could create a dynamic conceptual model of a multibody collision and instantly evaluate any of the variables involved. Not only that, he could, by simply willing it, replay the abstract experiment as many times as he liked from any perspective and in any variation that he pleased. He could "feel" the changing stress pattern in a mechanical structure subjected to moving loads, "see" the flow of currents in an electrical circuit as plainly as that of liquid in a network of glass tubes. By the end of the fourth week he could guide himself through to the solution of a tensor analysis as unerringly as he could guide his finger out of a maze in a child's coloring book.
The BIAC's adaptive learning system grew steadily more attuned to his particular methods of working and automatically remembered the routines that it had flagged as yielding desired results. As time went on it proceeded to string these routines together into complete procedures that could be invoked instantly without their having to be assembled all over again. In this way the machine automated progressively more of the mundane mechanics of solving a whole variety of problems, leaving him ever more free to concentrate on the more creative activity of evolving the problem-solving strategy. It therefore built up its own programs as it went along; and it was all the time expanding and refining its collection. Programming in the classical sense, even with respect to the parallel programming used in the distributed computing systems of the 1980s and '90s, no longer meant very much.
Clifford imagined a single cube. He imagined that he was looking at it from the direction of one of the corners and down on to it. Having fixed the picture in his mind, he opened his eyes and found a fair representation of it staring back at him from the BIAC graphic screen. It was not bad—a bit ragged at one of the corners and the lines were a little wavy here and there, but . . . not bad. Even as he thought about it, the subconscious part of his mind took its cue from his visual perceptions and the imperfections in the displayed image dissolved away.
"Try adding some color," Aggie suggested. She was the graphics instructor taking Clifford through the final part of the course. He mentally selected opposite faces red, blue, and green, consolidated the thought, then used the knack that he had developed and projected it at the view in front of him. The hollow cube promptly became solid—and colored.
"Good," Aggie pronounced. "Now try rotating it."
Clifford hesitated for a second, felt the first surge that forewarned the bio-link was beginning to become unstable, and caught it deftly before it could run away into positive feedback. The reaction was by now reflex. He settled down again and tried lifting one corner of the cube, but instead of pivoting about its opposite corner as if it were a rigid body, the shape deformed and flowed like a piece of plasticene. He emitted a short involuntary laugh, reformed the smear of colors back into a cube, fired a command at the BIAC to lock the display, relaxed and sat back in his seat.
"Went off the rails there somewhere," he remarked. "What should I do?"
"You let the idea that it was rigid slip," Aggie told him. "But even if you hadn't, trying to rotate it by stimulating external forces is a pretty difficult thing to get right at first. That's what you were trying to do, isn't it?"
"Yes." Clifford was impressed. "How could you tell?"
"Oh . . ." She smiled and gestured as if throwing something away. "You learn to spot such things. Now, when you try it again, don't think of actually moving the cube. Imagine it's fixed and you're walking around it . . . as if it were a building and you're in a hoverjet, okay? You'll find that if you do it that way, rigidity and all the other implied concepts take care of themselves subconsciously. Right. So, unlock it and give it another whirl."
* * *
Three days later, early in the evening and after their serious business for the day was over, Aggie showed Clifford some games based on animated cartoons that she had produced to amuse herself during her spare time. The difference with these cartoons was that the sequence of events unfolding on the screen could be modified interactively from second to second by the players.
Clifford's mouse scurried along the floor by the baseboard with Aggie's black-and-white cat pursuing close behind. He instinctively read the speeds and distances and sensed via the BIAC's responses that his mouse would just make it with two point three seven seconds to spare. He slowed the mouse slightly to take the corner at the bottom of the stairs and then raced it flat out along the last straight to where its hole, and safety, lay.
Suddenly he screeched the mouse to a halt. The entrance to the mouse hole was barred by a tiny door bristling with solid-looking padlocks.
"Hey, that's cheating!" Clifford roared indignantly. "You can't do that!"
"Who says?" Aggie laughed. "There's no rules that say I can't."
"Christ!" Clifford accelerated the mouse away as the cat pounced on the spot it had just vacated. He ran it round behind the cat, who immediately began turning after it. For an agonizing second he stared helplessly searching for a way out, and then, seized by sudden inspiration, he created a second mouse hole in the baseboard and promptly shot the mouse through it.
"That's not fair!" Aggie shrieked. "You can't change the house!"
"There's no rule that says I can't," Clifford threw back. "I win."
"Like hell. That was a tie."
They were still laughing as they removed the skull-harnesses and shut off the operator station to finish the day.
"You know, Aggie," he said, shaking his head. "This really is an incredible machine. I'd never have dreamed this kind of thing could work."
"It's primitive yet," she replied. "I think all kinds of applications that even we can't imagine will grow out of this some day. . . ." She gestured vaguely in the direction of the screen. "For example, I wouldn't be surprised if a whole new art form developed from little things like that. Why hire actors to try and interpret what's in the scriptwriter's mind if you can get straight into his mind?" She shrugged and looked sideways at Clifford. "See the kind of thing I mean?"
"Make movies out of peoples' heads?" He gaped at her.
"Why not?" she said simply.
Why not? Somewhere, he remembered, he had heard that said before.
* * *
The final thing they showed him in Baltimore was the way in which the BIAC could function as a communications intermediary between man and man. Two or more human operators interacting simultaneously with the machine were able to exchange thought patterns among themselves in a way that was uncanny, using the computer as a common translator and message exchange. Even more remarkable was the fact that there was no particular reason why these operators had to be in close proximity to one another, and a number of experiments of this kind had been conducted in which the machine in Baltimore was linked to another BIAC, owned by the Air Force and located in California, thus coupling operators three thousand miles apart. Clifford found this the most astounding thing he had seen since coming to Baltimore. He thought about it all the way back to Boston.
* * *
Clifford returned to Sudbury to find that installation of the Institute's own BIAC was well under way and that construction of the Mark II had commenced. The latter operation would require more time to complete, however, and as an intermediate measure to gain some preliminary experience in using BIAC techniques to interpret k-functions, the new computer was connected on-line to the Mark I prototype.
He slowly learned to steer his way through the masses of data to ferret out and manipulate the space-like solutions of the equations and to project them as visual displays. To his astonishment he found that he could "move" his va
ntage point at will throughout the body of Earth and about its surface. The resolving power of the Mark I was still poor, preventing him from distinguishing much in the way of meaningful detail, but he did succeed in producing recognizable images of some prominent geographic features such as mountain ranges, continental margins, and ocean trenches. He managed to obtain some surface views of the Moon too, in which the ghostly outlines of the larger craters and ring-walled plains could just about be discerned. It was somewhat like viewing the transmission from a remote-TV space probe that could be moved instantly from place to place—a tantalizing foretaste of what might be possible with Mark II.
* * *
One evening, while they were out for a few drinks at their favorite bar in Marlboro, Clifford was describing his experiences in Baltimore to Aub and Morelli. Aub had at last reached the point of being able to leave the immediate work on Mark II in the hands of the rest of the team and had made arrangements to go on a BIAC training course himself, starting the following week. Naturally, he was interested to learn about what the Navy had in store for him.
"You mean there's this guy in Baltimore and there's this other guy out in California someplace, both plugged into BIACs that are hooked together, and they can exchange thoughts?" Aub stared over his beer in astonishment. "Man, that's crazy."
"You've gotta be joking, Brad," Morelli said.
"Really." Clifford nodded emphatically. "I've seen them doing it. One of them can read a list of numbers off a piece of paper and the other one will tell you what they are. . . . They can send pictures—one guy imagines a face that they both know and the other guy identifies it . . . all kinds of things."
"Sorta like telepathy by the sound of it," Morelli remarked. "I never had much time for that kinda stuff."
"It's not really, though, is it," Clifford pointed out. "Not in the way that people usually mean the word."
The Genesis Machine Page 19
