Seek!: Selected Nonfiction Page 5

  Q35: In your Ware tetralogy, artificial life and biological life increasingly coincide. With Software, you were way ahead of your time, but writers like Hans Moravec and Kevin Kelly have done much to make your ideas more acceptable in America. Do you think that people distinguish too much between human and machine?

  A35: I remember when I was writing Software, I was wrestling with the notion of whether a machine can ever be alive like a person. How can chips have soul? But then I hit on the idea that the "soul" is a universal mystical jelly that imbues everything. A rock is already alive like a person. This said, of course there is a big difference between a machine and a person. But if machines became soft and wet, that would be a step toward being more like us. That's why in Freeware I liked having the moldies.

  Q36: One of the sites in the Ware tetralogy is a colony on the moon, built by robots. The Dutch astronaut Wubbo Ockels works on a similar idea in the project Euromoon (http://www.estec.esa.nl/euromoon/), but the ultimate goal of Euromoon is human settlement. Is human presence on the moon necessary?

  A36: It would certainly be interesting to have a human colony on the moon. I went and looked at that the Euromoon page of Wubbo Ockels - what a wonderful name he has! The page refers to the discovery of ice at the lunar South Pole; this is indeed something which is very encouraging. As a practical matter it would be easier in the near future to have a human colony on the moon than to have a colony of self-reproducing robots. But a middle path might be the

  Page 30

  best: to have robots with fairly low level of intelligence that are instructed by the (slow) remote link to people on the Earth. Given that there's a several-second-lag in the communication with Earth, the robots have to be smart enough not to fall off a cliff, and so on. I think this could be a very popular form of entertainment, to rent time running an actual lunar robot, especially if a good virtual reality interface were in place.

  Q37: Studly in The Hacker and the Ants is a speaking household robot you can relate to as a friend. Do you think there's a real chance that such a tool will be developed in the next ten or twenty years?

  A37: Oh, yes, I think so for sure. Descendants of the Furby. Your robot friend would not really have to be so very intelligent. We humans anthropomorphize relentlessly and can already easily image ourselves to be having a conversation with, say a cat or a dog. Why not a machine?

  Page 31

  PART I:

  SCIENCE

  Page 32

  The Mandelbrot set. (Generated by Chaos.)

  Page 33

  Welcome to Silicon Valley

  In 1986, it became unfeasible for me to continue living as an unemployed cyberpunk writer in Lynchburg, Virginia. I was broke and getting deeper into debt, while our children were needing braces and college. Even if it was peaceful and cozy in Lynchburg, the bandwidth always seemed way too low - where the ''bandwidth" of some information source means the number of bits per second that it delivers.

  What was really chafing me the most was my strong sense that I was missing out on a great intellectual revolution: the dawn of computer-aided experimental mathematics. Fractals, chaotic iterations, cellular automata - it was everywhere. I clicked over the final switchpoint when I went as a journalist to Princeton and to Cambridge, Massachusetts, to interview computer scientists for an article about cellular automata. Those guys were having so much fun, looking at such neat things, and making up such great theories about what they saw! I decided to become one of them.

  If you're a mathematician, becoming a computer scientist is not so much a matter of new knowledge as a matter of new attitude. Born again. Willing to commit to the machine. By way of preparation, I wrote Mind Tools, a book which surveys mathematics from the standpoint that everything is information. So when I got the chance to interview for a job in the Department of Mathematics and Computer Science at San Jose State University, I had thought enough about computers to give a good talk on information theory. They hired me and I started teaching there in the fall of 1986.

  Most people in the East don't know where San Jose is. Put your right hand so the palm faces down. Think of the left edge of your arm as the coast of California. San Francisco is the tip of your thumb. The space between thumb and forefinger is San Francisco Bay. The thumb's first knuckle is Palo Alto. San Jose is at the bottom of your thumb, near the bay. Silicon Valley is the thumb's sec-

  Page 34

  ond joint, between San Jose and Palo Alto. There're a lot of roads and a lot of traffic. And for the first seven years I lived there, it never rained.

  One of the courses I had to teach in my first semester at SJSU was assembly language. Assembly language is a very stark and simple language - a bit like Basic - with about a hundred elementary commands. What makes assembly language tricky is that in order to use it properly, you need to have a very clear image of what is going on inside the specific family of machines you are writing for (our course is for PC clones). You have to interact with the machine a little before you can get an assembly language program to run. I got the textbook: Dan Rollins, 8088 Macro Assembler Programming, and I couldn't understand what it was about at all. The only computer I'd used at this point was an Epson machine I bought for word-processing. I didn't know that 8088 was the name of a processor made by Intel. I didn't know that you say it "eighty-eighty-eight" and not "eight-thousand-and-eighty-eight" or "eight-oh-eight-eight.'' If I were the type to panic, I would have done so.

  Fortunately, there was another mathematician-turned-computer scientist at SJSU who was teaching assembly language, and his class met the period before me mine. I went to his classes and wrote down everything he said, and then I would teach that to my class. I enjoyed sitting in his class like a student again, soaking up info for free. The only thing about his class I didn't like was this jerk who sat in front of me, a guy named Farley.

  Farley was fat and petulant. His upper lip stuck out like on the man in that crummy Sunday funnies cartoon, "The Lockhorns," if you've ever seen it. Farley would get into big arguments with the teacher about arcane features of assembly language. He would interrupt without even raising his hand. And after class he was always trying to cozy up to the girls. Remember Farley; I'll come back to him at the end.

  I could never get enough time on the machines at school to do the assembly language homework, so after the first semester I went and bought the then-maximum personal computer - it had a twelve megahertz processor, a forty megabyte hard disk, and a sixteen-color graphics card. Some of my friends on the faculty were real computer jocks, and they helped me get psyched up for it. One

  Page 35

  professor in particular liked to say, "Computers are to the '80s what LSD was to the '60s."

  The first program I ran was a Mandelbrot set program that a fan had sent me. The Mandelbrot set is a fantastically complex pattern that arises from applying a lot of computing power to a very simple rule having to do with repeatedly taking the square of a complex number. It looks a bit like a black beetle with a long stinger on one end. You can use a computer to endlessly zoom in on its details, and the remarkable thing is that there are endless levels of detail to examine. Just like the irrational decimal number pi, the Mandelbrot set goes on forever, to as many levels of magnification as your computer can examine.

  I was so happy watching the colored little dots of my Mandelbrot zooms accumulate. I didn't know any other programs yet, but I could make this one look different by screwing with the monitor controls. If you messed up the vertical hold and set the monitor to analog instead of digital mode, for instance, the picture looked sort of like Antarctica, with more and more new little pixels moving in, men in boats, penguins, real deep info being born.

  The next program I played with a lot was SF-writer-and-computer-hacker Charles Platt's "Cell Systems" program for showing cellular automata. Charles and I went to a CA (cellular automata) conference together at MIT right before I came to SJSU. I liked to look at Charles's program all the time; in the morning or at night, e
specially at night.

  Cellular automata came to seem rich enough to symbolize everything: society, the brain, physics, whatever. The whole thing with a cellular automaton is that you have a tiny tiny program that is obeyed by each pixel or screen cell. With each tick of the system clock, the cells all look at their nearest neighbors and use the tiny program to decide what to do next. Incredibly rich patterns arise: tapestries, spacetime diagrams, bubble chamber photos, mandalas, you name it. Each pattern is a screenful of info, about 100,000 bits, but the pattern is specified by a very short rule, sometimes as short as eight bits. The "extra" information comes from time flow, from the runtime invested, from the logical depth of the computation actually done. The same thing is true for the Mandelbrot set, by the way.

  That next semester - this would be the spring of '87 - I taught

  Page 36

  A cellular automaton rule called "Tree."

  (Generated by CAPOW.)

  assembly language again, plus an advanced course in Pascal. With Pascal I couldn't find a teacher to copy, so it was pretty grim. I spent a lot a lot a lot of time trying to get my programs to work, or at least trying to figure out what I could lecture on the next day. Assembly language was starting to be fun, though. Making it up as I went along, I showed my class how to write a program to show simple cellular automata, and it worked, and we were all really happy. One of my programs made a pattern that looked like elephants and giraffes. Shirley Temple used to sing "Animal Crackers in My Soup," and in Gravity's Rainbow, Pynchon has someone call that song "Super Animals in My Crack." That was a joke that my new pattern made me think of. I bought a 24-pin dot matrix printer so I could start saving the pictures I made.

  In the summer of '87, I persuaded SJSU to buy me a CAM-6 "cellular automaton machine." This was a chip-laden card you could plug into a slot in any DOS-based personal computer. It had the effect of making my computer screen become a window into incredible new worlds. The CAM-6 made patterns that looked alive. And fast? Imagine globs of oil oozing around on your screen like a light show. Sixty updates a second!

  So in the fall of '87, I was ready to go to some computer conferences. I went to the first workshop on artificial life, in Los Alamos, not quite sure what it was. Artificial life turned out to be such a great

  Page 37

  A cellular automaton rule called "Ranch."

  (Generated by Cellab.)

  concept. I mean, forget artificial intelligence, let's do artificial life. Simple programs that grow and get more interesting as time goes on. Programs that eat computational energy! It was great at the Los Alamos conference. It was the first time I'd ever felt comfortable at an academic conference. We were all interested in the same thing: evolving artificially alive systems. And it was exactly what I'd been writing about in my SF novels Software and Wetware. Really happening at a government lab!

  The town of Los Alamos is very weird, like a Twilight Zone movie set. They have a little museum with full-scale white-painted models of Little Boy and Fat Man. It made me just a little anxious why the government would be interested in artificial life. But I'll trust those artificially alive robots of the future to get free - just like the boppers in Software.

  Even more fun than the A-life workshop was a meeting I went to a month later, something called Hackers 3.0, the third of a presumably annual meeting of Silicon Valley hackers.1 I was a little nervous going - I mean, was I a poser? But it was the most welcoming

  1. I should note here that "hackers" was being used in the older sense of "someone who loves to do things with computers, and not in the newer sense of "computer criminal."

  Page 38

  atmosphere I've felt since I went to my very first science fiction convention, Seacon at Brighton in '80. In straight academia there's not enough money and they usually don't welcome newcomers. But in science fiction, and again in the hackers world, I got a feeling of "Come on in! The more the merrier! There's enough for all of us! We're having fun, yeeeee-haw!"

  Some of the guys at Hackers had read some of my books, which made me happy, and we stayed up all night playing with my CAM-6. Like many others, I'd brought my machine with me. One guy explained to me why he wanted to have his head frozen. He had a zit on his nose, and I had to wonder about freezing the zit, too. At the end of the conference we posed for a big group picture. To get the right expression on our faces, we chanted, "Hack, hack, hack, hack . . . " They all seemed like such contented guys - happy because they actually knew how to do something.

  As I write this essay, it's spring '88, and I'm teaching courses in computer graphics, assembly language, and cellular automata. Teaching CAs has been the greatest, and I've just finished writing my first disk of programs, nice fast color cellular automata programs that run on DOS computers.

  Yesterday I was at another computer meeting, this one mostly chip designers, in Asilomar near Monterey. One of the guys was giving a talk about a great new chip he's building and someone asks, "How much will it cost?" and he comes back real fast, "Hey, I'd like to give them away." Another guy had a bottle of liquid nitrogen to show off a superconductor he'd gotten from Edmund Scientific. When we got tired of that he poured a lot of liquid nitrogen into a reflecting pool. The liquid nitrogen froze itself little boats of water that it sat on, boiling, finally leaving one small crystal of dry ice. Another guy took me out to the garage and showed me an electronic lock that he'd designed for his Corvette. There's a three-position toggle switch by the door, and to unlock the car, you jiggle the switch sixteen times up or down from center. The whole glove compartment was full of chips to make the system work. It was all he could do to keep from telling me the combination. Someone else had robot cars that seek light. Another one had programmed flashing electronic jewelry . . . and of course I brought my CAM-6.

  Page 39

  A lot of play, but beyond that, there's a real sense here of being engaged in something I'm starting to think of as "the Great Work," some kind of noble overarching all-encompassing quest. But it's all high stress, too, in a California kind of way. If you're not plugged in and working at staying that way, you can slip down real fast. Take Farley.

  A couple of months ago I saw Farley's picture in the paper. It took me a minute to understand what he'd done. He'd gone to a company that had fired him, and had killed seven people because some girl there wouldn't go out with him. I thought of all the times I'd wanted to tell Farley what an asshole he was. I was glad I hadn't. And then I was scared - what if he'd fallen in love with someone in the math department and had gone on his rampage there?

  Something that really got me was the newspaper descriptions of the seven people who'd died. For four of them, there were no facts available. They were simply additional human computer fodder who'd drifted out here to make some money. No friends, no connections, just a tiny expensive room in a garden apartment complex.

  One of my students in the CA course works at the place where Farley shot the people. "We heard the shooting," he told me, "and we went and hid behind the big computer." Somehow that's very heartbreaking to me - the people here can be so fucked and unreliable - and the only place to hide is behind the mainframe.

  Appeared in Science Fiction Eye, #4, August, 1988.

  Page 40

  A Brief History of Computers

  Calculating Devices

  One of the simplest kinds of computations is adding numbers. There are two ancient technologies used for this: the abacus and the counting board.

  In the familiar abacus, you have columns of beads corresponding to separate powers of ten - though often abacuses are designed with each column of beads broken into two parts, a lower part of five "unit" beads and an upper part with one or two "fives" beads that stand for five of the "unit" beads.

  A counting board is a more primitive idea; here instead of having beads on wires, you have loose tokens that can be placed into columns standing for successive powers of ten. Whenever you build up more than ten counters in one column, you can remove the counters fro
m that column and perform a "carry" operation by adding a counter to the next higher column. Often counting boards would have successive rows to stand for different quantities, in the style of a ledger-book, and sometimes the alternating rows and columns would be marked in different colors like a checkerboard. This is the origin of the British word "Exchequer" for their (rough) equivalent of the U.S. Treasury.