The Participants
Bran Ferren, moderator, is president of R & D and the Creative Technology Group at Walt Disney Imagineering, which melds science, engineering, art, and design in virtual reality attractions and interactive television. Ferren's work has also been seen on stage (Evita, Cats) and in films (Little Shop of Horrors, Star Trek V: The Final Frontier).
Marvin Minsky, professor of electrical engineering and computer science at MIT, is often called the father of artificial intelligence. His provocative book The Society of Mind offers a constructive explanation of thought and action, leading one reader to laud Minsky as "the Bob Vila of consciousness."
Dean Kamen, president and owner of DEKA Research & Development Corporation, is a physicist and inventor. He holds more than 100 patents, many of them for critical medical devices such as the portable dialysis machine and arterial stents, and a radical new wheelchair that offers dramatically increased mobility.
Story Musgrave has logged more than 1,200 hours in orbit during his three decades as a NASA astronaut-scientist. His most notable assignment: making key repairs on the Hubble telescope. Musgrave is also an artist, mathematician, photographer, physician, physicist, and poet.
Joy Mountford has spent most of her career exploring the idea that there is an art, not just a science, to the relationship between people and machines. Formerly head of Apple Computer's Human Interface Group, she now leads consumer music development projects at Palo Alto's Interval Research Corporation.
Chuck House, executive vice president of Dialogic, is a leader in video and computer conferencing. A vigorous advocate of multimedia literacy, he led the recent creation of the Center for Information Technologies and Society at the University of California at Santa Barbara.
Bill Joy, cofounder and chief scientist at Sun Microsystems of Palo Alto, California, has been the driving force behind UNIX and such powerful computer languages as Java and, most recently, Jini. He is a champion of open-source and user-friendly systems.
Danny Hillis pioneered the concept of parallel computers, the basis for most supercomputers, and has designed toys for Milton Bradley. Now a fellow with Walt Disney Imagineering, Hillis is building a giant self-powered millennium clock that will tick once a year and chime every 1,000 years.
Don Norman's groundbreaking books, including The Psychology of Everyday Things, Things That Make Us Smart, and The Invisible Computer, examine poor technological design. Professor emeritus of cognitive science at the University of California, San Diego, Norman is president of UNext Learning Systems.
The Discussion
The computer age has been characterized as both the greatest wonder of technology and our worst nightmare. In only a few decades, these often mystifying machines have transformed our existence, easing daily tasks, taking over humdrum work, and making a lot of the other machinery around us run more reliably and efficiently. Yet there is something in most people that does not love a computer. We remain suspicious of its power--and potential. Recently, Discover, in conjunction with the Disney Institute, invited a group of outstanding scientists to Orlando, Florida, to discuss the issues of the computer age in a daylong debate. Their illuminating and provocative dialogue is excerpted below.
BRAN FERREN: How long will it be before the computer as we know it now--a box with a screen and a keyboard--will cease to be the predominant platform?
MARVIN MINSKY: It ought to go away pretty soon, because there's no reason you should have to be looking at a screen, and there is no reason you should have to be typing. Computers are almost good enough now to transcribe what you say. They make a lot of errors, but when people talk to other people, they make errors all the time, and no one can tell.
DEAN KAMEN: It's going to disappear in different aspects of life over different periods of time. For the average person who's using it now for computation and Internet access, it will become a different appliance in the next three years. For people who really do what are hard-core computational things for which it's an adequate piece of hardware, it will be a little longer than that.
STORY MUSGRAVE: Ten to 15 years.
JOY MOUNTFORD: I would say 20 to 30 years, maybe. Because after all, it is well suited to today's tasks and has been refined over a long time to do these things.
CHUCK HOUSE: Look at a far older technology for your answer: The telephone is so pervasive and so remarkable, yet 50 percent of the world has yet to talk on one, and it's been around for 125 years. My colleagues will use a computer, I suspect, until they die.
BILL JOY: Five-ish.
DANNY HILLIS: I say negative two years. I think it has already happened. Most of the computers you deal with don't look like that. You don't even notice them. It's the ones that are lousy that you still notice.
DEAN KAMEN: I agree with Danny that it's been about two years since most computers vanished from sight. The largest users of computers in the world are the phone companies, and you never see theirs. But in other ways, the computer is still like the telephone was in its first few years. It's still an adventure to use a PC. You feel pretty macho if you can actually run it and don't have to reboot it, if the application works, if the connection really goes somewhere.
BRAN FERREN: What changes in personal computing do each of you see coming, or would you like to see?
JOY MOUNTFORD: One of the things that's very sad about computers is that they don't actually see or react to me at all. Listen to the words we use, like "window." That's all we get, a little box to see the world through. I'd rather have something that allows me to see the person on the other side. Digital cameras are quite cheap; why not put cameras around the interface box? That would be a step forward.
STORY MUSGRAVE: We're still very primitive in terms of the interface between the person and machine and the richness that could be developed. To give one example, engineers have been working hard to incorporate voice commands, with a limited vocabulary of 100 or 200 words, into operating systems and word processors. Of course, none of you ever have to use the backspace key to correct anything. [Laughter] But I happen to use it fairly often. If I could simply say "back" and instantly get a backspace, that would be incredibly nice. Or "back, back," and I'd get two backspaces. Think of all the little functions that take a shift and a letter, or a shift and a control and maybe another key. What if there were a simple word that you could have in your application that would accomplish all that? We could also do more of an immersion experience between the person and the machine. As a little example, instead of communicating with someone over the Internet with words, you could have a device that your hand fits on and that you could squeeze at the bottom or top. At the other end, the person would have a hand device that would pick up your manipulations. You could have a form of physical contact that would add a lot to the computer experience.
DON NORMAN: That's simply taking today's world and making it a little bit better. It's like making the steering wheel on a car better. The real impact of the computer, as with the car and the telephone, is that it is dramatically changing the way we interact. Automobiles and phones changed social life. They changed families, dispersing them throughout the nation. The real impact of computers today is on the communications network. The computer is the computational brain behind it. Now, suddenly, we can always be in touch with each other. That's what the real revolution is about. It's not about a better keyboard.
JOY MOUNTFORD: Don is right that these are steps transitioning us to something else, but we won't know what that other thing is until we start having a more active, conversational form of interaction. What is important is to put the tool in the hands of some new users, people who have more of an ability to express things, like artists. Think of what happened after Edison said, "Why would anyone want to watch more than eight minutes of a film?" If you put Webcams, for example, in the hands of some creative people, something else might happen.
BRAN FERREN: Marvin, what's your sense of when the first computer or network is going to become self-aware? And when are we going to become aware that it has become self-aware?
MARVIN MINSKY: And when are we going to become self-aware. People have this strange idea that we are self-aware, but we don't actually know how we remember what a word means or how we get ideas, and we don't have the slightest awareness of how we can tell [he lifts a can] this is a Classic Coke. There is something very peculiar about the idea that we're aware. As for computers, there are some programs that keep track of everything they do if you ask them to, and those programs can answer more questions about how they work than a person can. Computers really don't do very much yet except computations. I'm not sure that the networking is very powerful yet, because it depends so much on wasting the time of other people. Now, what I'm trying to do something about is getting the computer to be able to understand what you want and solve problems that you can't solve easily. Let me give you one example. In about 1970 there was a project trying to get a computer to understand a story about somebody's daughter being kidnapped by the Mafia and a ransom being demanded. The computer program, which had about four people working on it for three years, never got past the second sentence. Because what did it mean that he wanted his daughter back? Why would he pay this money? What is money? No computer knows what that is. Computers today don't understand the simplest thing that even a 5-year-old understands pretty well. At some point, people will figure out how to get computers to understand what words mean and how they fit together and represent ideas. And then, suddenly, there will be a new entity that's maybe as smart as you. Then, as many science fiction writers have noticed, if it can be as smart as a person or smarter, we'll have a new set of problems.
BRAN FERREN: And you see nothing that is a technical or conceptual obstacle to this happening?
MARVIN MINSKY: It's a social obstacle. There's a field called artificial intelligence in which people try to figure out how to get machines to do the kinds of things that if a person did them you would say, "That's pretty smart." But this field stopped moving in about 1980. The problem is: How do you get a bright person to go into the question of how to make computers smart instead of making money?
BRAN FERREN: This idea of our relationship with machines has always been complex, whether it's automobiles, aircraft, or telephones. Somehow, when we cross over into the realm of computers, we seem to be thinking about them differently. Maybe it's because we're kind of afraid that they are thinking about us differently. You don't hear a lot of discussion about us bonding with telephones, and yet we talk about bonding with computers. Story, is it your sense that there is something different about our relationship with computers? Are we on the cusp of a different moment in our evolution as a species, where our intersection and interaction with a tool become something that is deeper and more profound than it has been before?
STORY MUSGRAVE: I think you really strike on something there when you talk about emotional relationships with inanimate objects. We do have emotional relationships, at least I do, with different kinds of machinery.
MARVIN MINSKY: An hour or so ago I listened as you and Dean talked about flying your airplanes. I found it very boring, but I could see that you didn't.
STORY MUSGRAVE: I don't kiss my computer, but I have kissed airplanes. I know they are not sentient beings, but emotions are emotions, and I think it is very important to listen to one's body as well as one's mind. It's important that communication with machines not just be in abstract terms.
JOY MOUNTFORD: Well, the things that I'd like to kiss are things that I can sort of cuddle up with in bed. No computer comes close to anything that's warm and fuzzy enough that I would want to curl up with it. And the relationship? I have a relationship with my computer that is based almost exclusively on negative feelings: "Damn it, it's not working!" Very rarely do I have a positive feeling toward it except relief, like: "It found it, it got where I wanted to go." I feel I'm engaged in a battle. What's interesting to me is that gardening has become such a popular pastime. Silicon Valley is full of people getting into gardening. Similarly, cooking. Is that a desire to have your hands on something that's not as sterile and hard as today's computer? I think there are many questions about what a computer is anymore. Is the microwave a computer? Is my watch a computer? Is an autopilot a computer? We should spend more time trying to make computers feel a little bit more cuddly, a little bit softer. Something squooshier would make me feel a lot more positive about the interaction.
DANNY HILLIS: I don't think the problem is the interface, although we blame everything on the interface. We feel like there must be something wrong with the way we're talking with this thing, because it ought to be smarter somehow. The basic problem is that it's very unintelligent. It has enough intelligence that it sort of gives you a hint, and you wish it were there. But we have a name for a little bit of intelligence. It's called stupidity. The thing that makes me feel that it's not the interface is that computers are pretty good when we use them as communications intermediaries with other people. So for instance, the Internet is very useful for talking to other people. The telephone is very useful, and it has a terrible interface, but we don't complain about it, because the thing on the other end is intelligent.
BILL JOY: It's hard to imagine having an emotional attachment to a device with a complete lack of awareness of where we are in space or time, and with no real metaphorical understanding of the way we understand the world. So we have to make the devices aware of space, aware of time. To the extent that we're really nomadic creatures and wander around, they have to get small enough and unobtrusive enough that they can be with us.
BRAN FERREN: Why do we have this expectation that we should have relationships with computers? If I had said "your relationship with your toaster" or "your relationship with your adding machine," you wouldn't engage me in discussion. Why are we willing to consider this notion with a computer, especially when it is essentially an adding machine?
DON NORMAN: Bran, I've seen you. You have real relationships with some of the exquisite mechanical objects you own. You love great lamps and cameras. I disagree with Danny: It is the interface. The physical form is everything. Danny and I once spent an afternoon going to Japanese department stores. Most of the stuff in them is just like the stuff in American department stores, except for household appliances and gardening tools. That's where the fun is, and the shape is all the difference. It shows not just how you interact with it, but even the differences between cultures. The Japanese do their gardening differently than Americans. When you get to that level of intimacy, that's when you have a real relationship with your tools. That's why computers will disappear. You won't see them, and the form factor, the way we hold them and interact with them, will become everything.
BRAN FERREN: Let's push this a little further. I think it's entirely possible that in the not-too-distant future, we might be connecting to the Internet by way of an implant somewhere in our own bodies. Don, what do you think?
DON NORMAN: I actually don't think that's such a big step. We can implant a telephone today in the little empty space right under the ear and use facial bones for conduction. The hard part is dialing, but we could use voice for that. I could imagine implanting a calculator or a memory enhancer.
BILL JOY: A computer that monitors your health would be something a lot of people would probably be willing to have implanted.
BRAN FERREN: Dean, you design replacement body parts. In 25 years, instead of getting pierced or tattooed, will teens terrorize their parents by getting a "Netplant"?
DEAN KAMEN: Well, nobody I know is seriously working on a replacement brain, but mostly I think that's because they think it's rather trivial compared to their other organs. Most people operate without one now. But on this business of feeling some emotional attachment: There's a real paradox in society that has nothing to do with technology at all. For instance, we know more about astronomy than ever before, but there are also more people who believe in astrology today than ever before. We have a culture that has gotten so complex that most people can't understand it. And people are not comfortable with what they can't understand. Computers fall into that group. So there has been an attempt to imbue them with emotional characteristics, and we've given them some anthropomorphic properties that really dazzle people, like you can talk to them and people actually think: It listened to me.
BILL JOY: One of the truly stunning transformations that we haven't touched on yet is the rise of the virtual world. Most of us use this big virtual bookstore called Amazon.com. Lots of interactions in the economy are becoming virtual. And learning is becoming a virtual process.
CHUCK HOUSE: This virtual world notion that Bill outlines will be transforming in many arenas, not just education. We will cease to think of computers as physical things, but more importantly, we will start to think of our own physical presence differently. We will be able to be virtually present in the classroom or anywhere else for that matter, and the boundary will blur as computer, communication, and persona are blended. I call it the vivid experience--voice integrated with video integrated with data. Talk about virtual education!
BRAN FERREN: That's one of the things that concern me: the transition to virtual education. It seems to me we're losing a type of learning that has to do with the kinesthetic interaction of real things in real spaces with real people. For example, some of the best electrical or mechanical engineers I know are ones who grew up taking military surplus apart. We're transforming into a software-based view of the world because increasingly the hardware can't be taken apart, which means you can't learn and understand by taking it apart and putting it back together again. Story, do you think this is important, this idea of learning by feeling, touching, and experiencing?
STORY MUSGRAVE: It's very important, and I think we're swinging too hard the other way in dealing with abstract intellectual ideas. We're taking the body out of the equation; we're taking Earth out of the equation. Look at how we spent our spare time in the 1950s, and compare that to what we do now: We're staring at a monitor and shooting it with an electron gun. We work on computers all day, then we get on the Internet at night, then we watch television. What has that replaced? It's certainly replaced reading a book. How much of human relationships is that replacing? How much of our relationship with nature and Earth is that replacing? How much of the kinesthetic sense of space and time in the real world is that replacing? We need to get more balance.
BRAN FERREN: Is it possible to swing back?
DANNY HILLIS: Probably not. I think we're heading for some altogether new relationship with technology, which seems to be becoming fundamentally incomprehensible and fundamentally self-generating. I think our relationship is going to be more like the one we have with nature. Namely, we can influence it in certain ways, but we won't be able to really control it in the way we are used to controlling machines. All we may be able to do is try to keep the weeds out of the garden.
Danny Hillis's clock projectBill Joy's work, www.sun.com and www.java.sun.com Dean Kamen's work, www.dekaresearch.com and www.usfirst.orgMarvin MinskyJoy Mountford's workDon Norman, or e-mail him
