by Rudy Rucker
Although Michael could be stern when arguing a logical point, he had a very warm and humane side as well. His face would light up with pleasure when we met, and he’d hug me when we parted. He greatly enjoyed the jokes I’d tell him, like when I told him, “Our dog Arf’s so smart he can say his own name,” and then, several weeks later added, “Arf’s so famous that the other dogs talk about him.”
Both Michael and Jon were married, each with young children. Jon was California mellowness incarnate. He scattered his speech with words like “groovy,” “awesome” and “trippy,” and was a master at inserting the word “like” into a sentence. Somehow he could get away with this.
Jon said he required very little sleep. He was constantly mastering great new swaths of computer science and writing up lecture notes on them. At times his notes seemed to bear the imprint of his experiences as a Berkeley logician—for instance he might diverge into a discussion of an imaginary “metalanguage” when an outline of some actual programming language would have been more to the point. But if I mentioned this, Jon would argue that it was a mistake to teach about actual programming languages, since they changed every couple of years anyway, going in and out of fashions like hairstyles. The platonic ideal of the underlying metalanguage could serve as a programmers’ beacon throughout his or her working life.
It was such fun to listen to Jon’s slow, oozing voice that he could say almost anything. Like one time, he was talking to me and the kids, and I asked him about the secret of his alleged appeal to our female students, and Jon made a joking remark that the kids and I would repeat to each other for years.
“I tend to dazzle.”
Cellular Automata
So how did I get that job in California, and what did I end up doing there?
A year before we left Lynchburg, in 1985, I had become interested in some computer programs called cellular automata, or CAs for short. My first contact with the modern CA mind-virus was through an article by Stephen Wolfram in the Scientific American. He was displaying his cellular automata as changing patterns of pixels on the computer screen. The colorful images had an organic, natural look, neither too orderly nor too random. Wolfram seemed to think that CAs were rich enough to emulate anything at all, be it a society, a brain, a pond, or a galaxy.
The patterns spoke to me at a deep level. You might say they were a trigger that sent me into a metamorphosis—like a full moon that changes a werewolf from a man into monster. These fascinating graphics set me on a path to becoming a computer hacker.
Here I need to pause and explain that I use the word “hacker” in its original, positive sense of “clever person who delights in doing complex things with computers.” Unfortunately, by the end of the 1980s, our ignorant and hysterical mass media had begun using “hacker” exclusively in the negative sense of “computer intruder, electronic embezzler, or cybercriminal.” It’s as if the media had arbitrarily decided to begin using “realtor” to mean exactly the same thing as “sleazeball, scam artist, swindler.” Well, maybe that’s not such a good analogy!
But please indulge me, and understand that in the times I’m reminiscing about, computer hackers were good people—and far from dull. Hollywood often depicts computer types as nerdy, inhibited types, but that’s not generally accurate. It’s more common that hackers are like acid freaks or mad scientists or car mechanics or surrealist artists.
From Wolfram’s article, I learned that his CAs, or cellular automata, were based on the idea of dividing a region of space into a grid cells, and then letting a tiny program run inside each of the cells. A CA is a parallel computation, in that each of the thousands of cells acts like an independent computer. With each tick of the system clock, the cells look at their nearest neighbors and use their tiny programs to decide what to do next. You can put random starting values into the cells or input some particular pattern. And you can change the little program that goes into each cell. Incredibly rich patterns arise: tapestries, spacetime diagrams, bubble chamber photos, mandalas…
As I say, I felt a sense of recognition when seeing these pictures, as if I’d been waiting to see them for my whole life. Come to think of it, I’d actually been writing about these kinds of patterns for several years. In my novels Software and Wetware, I’d wrapped my bopper robots in colorful plastic that acted as a computational tissue, generating unpredictable patterns. I’d called the stuff flickercladding. In hindsight, I could now see that I’d covered my boppers with CAs.
Early in 1985, I convinced the magazine Science 85 to pay my expenses for a journalistic trip up the East coast to visit the cellular automata researchers. Wolfram was first on my list. He’d gotten his Ph. D. in physics at age twenty, and now, at the ripe old age of twenty-six, he was a visiting fellow at my grad-school haunt, the fabled Institute for Advanced Study in Princeton.
Wolfram was stocky and tousled, with the directness of a man who knows what he’s doing and doesn’t much care what others think. I felt a kinship to him right away. We were even wearing the same kind of clothes: oxford cloth button-down shirts and chino pants. I started out by having lunch with him in the good old Institute cafeteria—where I used to eat with Gaisi Takeuti when I was on the path to becoming a set theorist.
One of Wolfram’s hopes for CAs was that they could be used to analyze certain kinds of phenomena that resist being reduced to any simple formulas. To get the conversation going, I asked him what engineers thought about modeling air turbulence with CAs.
“Some say it’s wrong, and some say it’s trivial,” said Wolfram in his thoughtful Oxford accent. “If you can get people to say both those things, you’re in quite good shape.”
We went up to his office and he introduced me to another CA researcher, a cool guy called Norman Packard. They began showing me cellular automata on the computer screen. Some of the patterns were predictable as wallpaper, some were confusingly random, but every now and then we’d hit that pleasing balance between order and chaos that characterizes gnarly CAs. We found one that was shaped like a pyramid, with red and blue lace down the sides, and a symmetrical yellow pattern in the middle—a pattern vaguely like an Indian goddess.
“What’s the number on that one?” asked Wolfram.
“398312,” answered Packard.
“This is the way to do scientific research,” I remarked. “Sit and watch patterns, and write down the numbers of the ones you like.”
“Oh, this isn’t for science,” said Wolfram. “This is for art. Usually I just talk to scientists and businessmen, and now I’m trying to meet some artists. Wouldn’t that last one make a good poster?”
At this point a small baggie of what appeared to be pot dropped out of Packard’s jeans pocket. Wolfram picked it up, impishly sniffed it and raised his eyebrows. “What’s this?” he demanded, as if pretending to be an irate schoolmaster. “What’s this?”
Packard didn’t say a word. After all, he was in the presence of a journalist! Wolfram handed Norman his tiny baggie and he stashed it away.
Over the coming years, I’d learn that computer hackers are very tolerant people. All they really care about is whether you can make machines do interesting things.
I’d also learn that it’s common for hackers to begin seeing the entire world in terms of the computer concepts that they’re working on. You don’t hear carpenters saying that everything is made of lumber—but there’s something about computers that gets deep into a hacker’s head.
I’m reminded of a guy I’d meet at an IBM research lab later on. We’d been in his office talking about his work, and then we walked outside together and were looking at a range of low, wooded mountains. He began telling me that the ridgeline of the hills—with trees included—was essentially a version of the same exact noise graph that he was studying in his lab. He was saying this with a complete lack of irony. I felt sorry for him then. But later, as I became more of a hacker, I’d often be just as bad off.
On that first journalistic cellular automata tour in 1985,
I continued up to Boston and met some more CA hackers. The guys at MIT had built what they called a “cellular automata machine,” which was a special board filled with chips. You could insert it into an ordinary PC computer and then watch graphical images of cellular automata rules running really fast—like light-shows.
A Hungarian computer scientist showed me a screen full of boiling red cottage cheese. Despite the boiling, the cheese was staying mostly red. To him, this represented the persistence of memories in the seething human brain.
“With the cellular automaton machine, we can see many very alien scenes,” he remarked. “We have a new world to look at, and it may tell us a lot about our world. It is like looking first into a microscope.”
Seeing the CAs running in real time fully converted me. These hackers were having so much fun, looking at such neat things, and making up such great theories about what they saw! I started wondering if I might become one of them.
By the way, my editor at Science 85 didn’t get it about cellular automata. He shelved my article—and a few years later it later it appeared in Isaac Asimov’s Science Fiction Magazine. The science fiction world takes care if its own.
As I mentioned earlier, in Virginia I had been running low on money and getting worn down by the freelance life. In the fall of 1995, I happened to be complaining on the phone to Craig Smorynski, a mathematician friend in California, about how broke I was, and he told me that they had an opening where he worked, and that several of the faculty admired my book Infinity and the Mind.
The place was the Department of Mathematics and Computer Science at San Jose State University in San Jose, California—which lies in the heart of Silicon Valley, at the southern end of the San Francisco Bay, some seventy miles south of San Francisco. It had never before occurred to me before that I might possibly move to California. It seemed as unlikely as moving to China, or to Mars.
I flew out to San Jose State for an interview early in 1986, and gave a talk, for which I was very well-prepared, as the ideas were drawn from the book Mind Tools that I’d been working on for the past year. My ideas had to do with complexity, computation, and cellular automata.
Mathematicians and philosophers had begun wondering if there might be some precise way to quantify how complex a given pattern is. A blank wall or a checkerboard seems to have very little complexity. A completely random mess of colored dots has complexity of a sort—but in an uninteresting way. What people were after was a measure of complexity that would give the highest ratings to the kinds of things that humans find interesting and beautiful, things like living plants and animals, works of art, or scientific theories.
A cellular automata hacker named Charles H. Bennett was spreading the notion that we’d do well to measure an object’s complexity in terms of how much computation goes into generating the object from its shortest possible description. I’d met Bennett by now, and he’d impressed me as a very brilliant man. So I’d worked a lot of his ideas into Mind Tools, and into that talk I gave at San Jose.
Bennett used the term “logical depth” for his new measure of complexity. A checkerboard has no real logical depth: you use a short description to make a simple pattern, with no heavy computation involved. Random fuzz doesn’t have a big logical depth either—and this is the interesting part. For the shortest description of completely random pattern is simply going to be a listing of the pattern itself: you use a long description to make the complicated pattern, and again no great amount of computation is needed to pass from the description to the object itself.
But a living organism is logically deep. Why? On the one hand, it has a rather simple description: the DNA. But the organism itself is intricate, and it results from a prolonged process of biological growth which, in some sense, uses the DNA as its starting program.
Human-generated bodies of knowledge are also logically deep. Consider, for instance, the collection of all currently known mathematical theorems. There’s a simple underlying program; in this case it’s the axioms and definitions found in the first chapters of math books. But the computational process that produced our theorems is huge—it’s the work carried out by generations of mathematicians, doggedly figuring out proofs.
I think of music as another example of a message with high depth. The musical score and the words of the lyrics might be rudimentary, but performers can make the music deep by practicing a lot and adding mental spin. Along the same lines, you can argue that a novel is logically deep. It has a simple description—in the form of the author’s outline. But the work of writing the finished book is very long computation.
In all of these examples of logically deep patterns, we have phenomenon with a rather short description that generates a very intricate pattern. I would of course admit that pushing this kind of reasoning into the worlds of artistic creation is a bit risky. But staring at images of cellular automata makes this kind of mental leap seem increasingly reasonable. You can start up a CA rule with a seed no more complicated than a few dots—and a few minutes later, you’re looking at a braided macramé like a valley of rivers, or at nested scrolls resembling a pattern made by a dedicated ivory-carver.
My talk at San Jose State was well received. It was a milestone for me—my re-emergence into academia. The interview committee took me to an inexpensive outdoor cafe and we drank a couple of pitchers of beer, with dappled patches of shade and sun bobbing over us. Everyone was hip and mellow. The San Jose State faculty thought it was great that I wrote science fiction as well as popular science books. Welcome to California.
“If you want, you can teach computer science as well as math,” one of the guys told me. “And if you do that, we’ll pay you ten percent more.”
“I’d like that a lot.”
Starting in at San Jose State in the fall of 1986, I was assigned two computer science courses, plus a course on the good old history of mathematics. The department was using IBM-clone machines with Intel microprocessor chips, and I bought myself one of these beige boxes.
The harder of my CS courses was about assembly language. Assembly language is very stark and simple, 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’s going on inside the specific kind of computer that you’re writing your program for. Learning Intel assembly language was a little like learning, say, the complete set of the part numbers for a 1986 Ford Motors Company truck engine.
Fortunately for me, there was another mathematician-turned-computer scientist at San Jose State who was also teaching Assembly Language, a man named William Giles. His class met the period before mine, and he was kind enough to let me attend his lectures. Giles was a good guy who wrote a lot on the blackboard with chalk. By the end of the period the front of his body would be covered with white dust—and his back too, because he liked to lean against the board while he was talking. In a way, it was fun sitting in someone else’s class like a student again, soaking up info for free.
I wrote down everything Giles said, and then I would teach that to my students. Not that I understood everything. When my own assembly language students would ask me how to do the homework, I’d tell them the truth.
“Hell, I can’t do it either. I only figured out how to find the on/off switch on my computer last week.”
For some reason the students liked this. “You’re a good guy, Dr. Rucker,” they’d say. I got great student evaluation ratings in that course. I was really bringing the material down to their level. And by the second time I taught Assembly Language, I was ready to set the class to writing cellular automata programs.
There was a full spectrum of races among the computer science students, with many Indians and Chinese, a few Filipinos and Latinos, some Iranians and, more populous than any other group, the Vietnamese. San Jose has one of the largest Vietnamese communities of any town in America—the signs in buses and voting booths are in English, Spanish, and Vietnamese.
I really enjoyed h
anging around with my students and hearing their offbeat accents. Who needs intergalactic aliens, when you have computer science majors in San Jose? And, by the way, after meeting the Vietnamese, I was gladder than ever that I hadn’t gone to fight against them in that crazy war. First of all, I liked them—and second of all, they seemed very tough and tenacious.
In the summer after my first year at San Jose State, in 1987, I persuaded the department to buy me one of the cellular automaton machines that the gonzo MIT hackers had told me about on my journalism run back East. The “machine” turned out to be a memory-chip-laden card you could plug into a slot in an ordinary PC-clone computer. It had the effect of making my new home computer feel as powerful as the high-end Silicon Graphics and Sun machines that Wolfram had been using to look at his CAs.
With my new card, I could watch globs of red and blue oozing around, almost like oil-drops in a light show. And to make it the more geekily titillating, the language for programming the cellular automaton machine was a willfully obscure “Reverse Polish” dialect known as Forth. I soaked it right up. Programming was close enough to mathematical logic to be congenial for me.
In the past, I’d always envied laboratory scientists for having machines to play with. And now I had one too. I relished the hands-on, experimental nature of computer work.
From grad school I knew all too well that if you find a hole in a mathematical proof, you can very well be left with nothing at all. But if a computer program has something only slightly wrong with it, you still might see something interesting on the screen. Hackers have a saying for this: “It’s not a bug, it’s a feature.” In any case, the interactive, iterative process of programming meant that usually I could, in time, get my programs to do what I actually wanted them to.
I took to lugging my heavy PC-clone computer to parties to show cellular automata to whoever was there. Sometimes, if I was sure of finding the right kind of host machine, I’d just bring the bare cellular automata card and a connector cable. Musician-style, I took to calling the CA card my “axe.” It was wonderfully science fictional to bring my axe to parties and play symphonies of living colors.
