Daniel Hillis
Imagine coming home after a long day in the year 2030. You plop down in your couch, exhausted, and the couch gathers around you, rocking and relaxing you. But this couch is not an elaborately designed amalgamation of leather and servo-motors. It has been grown, rather than manufactured, and it’s alive.
While this fanciful bio-lounger may never be part of our future, it’s the kind of thing many scientists and futurists can imagine. The 20th century was marked by advances in electronic technology, and many see the next century as the age of biological technology--as we learn to tinker with, reconstruct and reconstitute life itself. The building blocks are already in place; they include advances in the technology of human reproduction, the ability to clone organisms and the use of genetic engineering to produce drugs and pest-resistant crops.
Meanwhile, as scientists attempt to engineer biological organisms, engineers are trying to grow computer programs in simulated environments, using the principles of Darwinian evolution. It’s already becoming harder to distinguish between what is designed and manufactured, and what simply grows and evolves.
Straddling that blurring boundary is Dr. W. Daniel (Danny) Hillis, 41, a legend among computer hackers as the MIT Media Lab whiz who, among other things, built a computer made entirely of Tinkertoys that could beat humans at tick-tack-toe. More important, he designed the Connection Machine, a blazingly fast supercomputer with 64,000 parallel processors. These days, Hillis spends as much time thinking about biology as he does about computers, and he has created a series of experimental but highly efficient computer programs through a process he calls “artificial evolution.”
A native of Baltimore, Hillis was the son of an Air Force epidemiologist. He originally planned to study biology, but his love of building things led him into computer science. In the mid-1980s, he formed Thinking Machines Corp. to build his super-computer and, though the company eventually folded, its technology still forms the basis for much of today’s high-level computer design. Last year, Hillis joined the Walt Disney Co. as a Disney fellow, the first of a group of influential technologists the company has hired. He describes his job there as helping the company use technology to create magic. Married and the father of 5-year-old twin boys and a 2-year-old daughter, Hillis is also championing another strange idea to build what he calls a “millennium clock,” designed to last 10,000 years and force us to think more seriously about the future.
*
Question: Most people can understand the basic concepts of things like genetic engineering, and even cloning, but they’ll roll their eyes at futurist concepts you might embrace, like the idea of a living, biological computer made out of organic molecules.
Answer: Which is interesting, because people are biological computers. Electronic computers are just a cheap imitation of the original, which is the brain. And as we develop the technology of biology, I think we’ll have to start thinking of our own bodies as machines. That’s because we’ll be able to understand how they work, replace broken parts, and fix malfunctioning systems. At the same time, it will be possible to build machines out of the same components our bodies are built out of. Right now, we’re able to separate ourselves from machines because we’re not built from the same things. But as the technology of biology develops, we’ll have to learn that it’s not what we’re made of that separates us, it’s what we think.
Q: How will we change the way we think when the manufactured stuff around us is no longer steel or silicon or Formica?
A: Right now we have a kind of prejudice that says the good stuff is built out of squishy organic molecules, and the stupid stuff is built out of steel and Steve Proffitt, a contributing editor to Opinion, is director of the JSM+ New Media Lab. He spoke with Daniel Hillis at the technologist’s offices at Disney Imagineering in Glendale.
circuits. But two things are going to happen. We’re going to build far more complex stuff out of metal and circuits, but we’re also going to be able to build stupid stuff out of living matter--a chair, or a telephone, or a computer. Also, until now, electronic things were designed, and living things evolved. Now we’re applying evolution to electronics, and design to biology.
We’re now creating computer programs using the principles of natural selection and random mutation--the basics of Darwinian evolution. We start with a bunch of programs which are just random series of instructions, and place them in a simulated environment in the computer. Then we choose a task, for instance, organizing a list in alphabetical order. Because of the way we create the environment, the programs that complete the task are the most likely to survive. Then we let them grow, mutate, and even have sex--by that I mean they produce children which are combinations of the two parents. And eventually we end up with programs that all perform the task. What’s interesting is that I don’t always know how they work, I just know they do. Using this method we’ve been able to come up with code which is better than any human has been able to write.
This is a simple example. You might also create another set of random programs, and assign them the task of testing the first group of programs to see how well they perform their task of alphabetizing. The testing programs will only survive if they find bugs in the alphabetizing programs, which will only survive if they don’t have bugs. So we end up with very good alphabetizing code, because it grew up with a bunch of other code whose life depended on finding problems in it.
Q: How are we humans going to adapt to these radically changing ideas about what constitutes intelligence, how life is defined, and what we mean by consciousness?
A: I bet rather easily. We’re far more adaptable than we think we are. There was a time people couldn’t imagine society being able to accept that the Earth wasn’t at the center of the universe. And yet we swallowed that one. The same thing with Darwin and the idea of evolution--it only took a generation, and people accepted the idea. So when biological machines come along, people will be shocked, and then they’ll get used to it.
Take the match between [Deep] Blue and Gary Kasparov. People speculated for years about what it would mean if a computer could ever beat a master chess player. People wrote dissertations about it, philosophers declared it to be impossible, and then when it happened, it was news for a couple of days and life went on. And in the same way, when there are living machines, and designed organic things, people will adapt.
Q: And yet a lot of what you foresee has the power to radically change what it means to be human, particularly if we can extend our lives, or perhaps even download the contents of our brains for safekeeping.
A: Absolutely. Things like the telephone radically changed the experience of being human, too. We can talk to each other across the globe. This has changed our lives, but it hasn’t changed us--we’re still humans, pretty much like those of pre-telephone times. There may be, in the future, people who live longer, or who think faster, but they’ll still be people.
Q: Having been born in the 1950s, you grew up in a world focused on the future, and on predicting what life would be like in years to come. A lot of those predictions were way off the mark. Does that experience temper you when you think about the future from today’s vantage point?
A: They did get some stuff right. I think much of what they got wrong grew out of obvious extrapolation--flying cars, for instance. It was easy to explain, and easy for people to imagine, but it didn’t happen. They pretty much missed personal computers, and thought robots would be big. It’s hard to explain what you might do with a computer, but a robot is easy to understand, I could explain it to a child in a minute. So the lesson there is just because something is easy to imagine doesn’t mean it’s going to happen.
Q: Given that lesson, can you venture a guess about the future--what kind of biomachine might be ubiquitous in 50 years?
A: This is dangerous business. There are all kinds of things I can imagine, but they’re probably all wrong. If you look back on the telephone, for instance, people were dead wrong. They thought there would be only one in every city, or that they would only be used to broadcast concerts from different cities. But people who understood the broad impact the telephone would have on electronics and communications were the ones who guessed right. And in a broad sense, in the future some things which we used to manufacture we will grow, and some things which we used to grow we will manufacture. We may grow telephones, but manufacture cabbage. The line between what’s built and what’s grown will be blurred. Maybe you’ll plant a house, let it grow, and then move into it. Perhaps not that exactly, but there will be things just that strange happening.
Q: What about in the near term? How will these principles of artificial evolution begin to be seen?
A: We’re already doing it to produce drugs. Rather than doing it laboriously by hand, trying by trial and error to find a tropical bark that has just the right properties we want, we’ll build a situation in which the drug we want will evolve. A little farther out, may be a tree which has gasoline or kerosene as its sap. So you could imagine tree farms which supply us with fuel. The same trees are already quite good at taking the carbon dioxide out of the atmosphere and fixing it as carbon, so you have a solution to two problems. I’m not saying that’s going to happen, but these are the kinds of things that are possible when you have engineered biology.
I think what excites me most right now is the interaction between biological thinking and computational thinking. Right now it’s two different sets of people. Biologists and computer hackers usually stop talking to each other sometime in high school. But because the two fields are merging, it’s going to be very interesting to see these two groups mix.
Q: What is it that so fascinates you about computers and machines?
A: The fascinating thing about working with computers is that they are such a reflection of ourselves that it’s really just thinking about our own minds. And it’s interesting--my religious friends think I’m crazy because I believe the mind is a machine, but my scientific friends accuse me of being a mystic, because I do believe that there are limits to scientific explanation and understanding. Just because something is the result of physical laws doesn’t mean we can understand it.
Q: I suppose the holy, or perhaps unholy Grail in the technology of biology would be the ability to download and store one’s consciousness--to be able to preserve the total of one’s experience and intelligence.
A: I’d love to do it, if it could be done, but there’s lots of people who wouldn’t be at all interested. If I had the option of continuing to think after my body was gone, I would take it. I’m as fond of my body as anyone else, but if I can be 200 with a body of silicon, I’ll take it.
Q: Moving back to the planet Earth, one of your tenets is that we’ve been too concerned with creating new technology and haven’t spent enough time figuring out what to do with it.
A: Yes, and I think if we halted all technological development, we could spend a couple of generations just finding new applications for existing technologies. One of the downsides of this emphasis on developing new technology is that it causes us to focus too much on the short term. We’re often so overwhelmed by the possibilities of our current technologies that we don’t think enough about the future. When we were kids, we thought about what would happen in the year 2000. Now it’s 1997 and people are still thinking about the year 2000. Yet nobody’s thinking about the year 2030.
Q: But I thought you were.
A: Well, that’s true--not nobody. I’m interested in getting people to understand and believe that there will be a year 3000. And once you start believing that, you start taking a little greater responsibility and being somewhat less opportunistic. The millennium clock is my way of trying to impress that on people. It’s my idea to build a clock that counts out the next 10,000 years. The cuckoo comes out in the year 3000, cuckoos three times, and then waits 1,000 years before coming out again. The clock must be wound every year, and people will have to take care of it.
Right away, it forces you to face the future. Where will we put the clock, for example. If we put it in Los Angeles, is Los Angeles still going to be there in 10,000 years? What’s the level of the sea going to be then? The point is that it makes you think of the future as something real, and it’s time we began doing that more seriously.
