Russell Hoffman ("Host"), High Tech Today
Dr. Nicholas Negroponte ("NN"), Founder & Director, MIT Media Lab
Host: My guest today is is Dr. Nicholas Negroponte. Dr. Negroponte is the founder and director of the MIT Media Lab, which has been running for about 10 years, a little over ten years, they now have over 300 people working at the lab. First of all, I'd like to welcome you to the show.
NN: Thank you very much.
Host: And, the lastest thing that I've heard about--I'm always hearing about something from the Media Lab. But the latest thing that I heard about is the Things That Think Project. So, why don't we talk a little bit about that. There's a poem here that I was faxed, a three line poem. Why don't I read it, to give people some idea of what we're talking about, but it will probably leave them more confused!
"In the past, shoes could stink.
In the present, shoes can blink.
In the future, shoes will think."
Okay--what do we mean?
NN: Okay. Well, let me first of all say that we've started this program very much with the idea of getting back into things that are--I'll say, more risky, and even slightly crazy, because when we started the Media Laboratory, which was, as you mentioned, over 10 years ago, a field like Multimedia was considered at best, off the wall. And of course, now it's a trillion dollar industry, and we went from being sort of the mavericks and the charletons and the pioneers to all of a sudden being the establishment, much too fast, so we're trying to sort of now, reinvent ourselves, and look at things that may sound crazy today, but will be probably equally as established ten years from now, and the Things That Think program is one of those.
It's the following idea: If you can imbed computing in all sorts of common objects, and that includes sneakers, then each of those objects could have a personality that made it better at doing what it was normally meant to do, and also allow it to do things that it wasn't meant to do. And so, we've got a whole lot of companies, about forty in total, that make things that range from furniture to carpeting to clothing to wristwatches, to all sorts of stuff. And the idea is to try and make these different things compute! And, I can give you some very specific examples.
As part of this we have something called the bodynet, and that is, to use the human body as the network. You've heard of local area nets in offices and you've heard of the Internet which is global of course. Well this is probably the most local network possible and it's actually--it's your body. And the students and faculty have achieved to transmit, though the body, about 100,000 bits per second, using the body itself as the transmission medium.
So, if you wear a pair of sneakers such as ones that we've designed with NIKE that actually have computers in their heals, the computer can actually get not only energy, because as you walk or you run you generate energy that's usually wasted--pushed into the floor or the cement. Here it's captured and it can power the computer. And, you shake someone's hand, who's also wearing these sneakers, you actually can exchange 100,000 bits between yourselves. So, imagine going to a meeting in the future, or a conference. You could shake the hands of a few hundred people and when you go home that night, you just print their calling card.
Host: You won't have to hand out business cards anymore.
NN: Don't need to, you just need to shake someone's hand.
Host: Which right now, business cards have an amazingly small amount of information on them. Perhaps we should have computerized business cards like credit cards.
NN: But this is exactly what it--I mean, you can exchange all sorts of information, I mean, when you shake someones hand your computer can compare the people you know in common and all sorts of stuff like that.
Host: Now, when I first read about this I thought we might be talking about shoes that maybe tell you what your stride length is, how many steps you've taken since the last time you looked. Is that also something that you're--maybe as an aid to one's health or something like that?
NN: Sure, that's actually quite easy. And yes, and since again the body can act as the conductor, you can take the person's pulse because the wristwatch that you're wearing actually makes contact with your skin, so your sneakers could in fact communicate with your wristwatch very easily.
Host: Is anyone raising any objections to the use of the human body as a transmission wire? It seems to me--
NN: Yeah, I understand your question. It turns out that using the body as a transmitter--we're putting in less amps into the body than the headset of a walkman, okay? These are nanoamps, these are tiny, tiny amounts of current which as I say, are less than when you're wearing a headset and listening to music.
Host: Not to mention all the microwave radiation, and FM and AM all going through our bodies all the time anyway. I've always wondered what to do with those two bolts on the side of MY neck but--what about--what are some other things besides shoes. What can you do with chairs, or with jewelery--or with carpet?
NN: Again, this is a program that is all of one month old, and--it's not even a month old, it officially started the 10th of October, and it's a five year program, so--
Host: --so you get press very quickly! Congratulations! (both laugh)
NN: But we've got a long way to go. I probably could give you a convincing example with a doorknob.
NN: Because imagine your front door, as you approach it, if it could see you, and recognize that it's you, it could actually open the door. You don't have to put down the packages, or fumble for the key and so on. That kind of processing is not likely to live in the doorknob itself, because it's relatively demanding, the kind of processing you use. So the doorknob would do some of the seeing if you will--there would be a vision system in it. That probably will cost all of a dollar. But it then might have to link to something else that does the computing.
So Things That Think also includes Things That Link, because you would want to talk to your toaster and your refrigerator and so on, but each one of them doesn't have to understand English. That can be a shared resource.
Host: Let's see--the one with the carpet. Now, I still just am wondering, what's the carpet going to say to the shoe besides "don't tread on me!"
NN: (laughs) Cute thing for it to say!
Host: Hopefully it actually won't say that!
NN: Well, carpets are probably a good way to distribute signals through a home. One of the problems with computing in the home when you want to distribute signals is that the wiring in the wall is already there, and to add new wiring is very hard, and to use radio frequencies or infared, it works but it's got problems. If it's infared, of course it doesn't go around corners. Carpets probably will end up being much more of the transmission medium between objects. Not so much the carpet trying to tell you something itself.
Host: ...[break]... We're talking about Things That Think, and one of the things that were mentioned is glasses that can look, or that can see, as well as being seen through.
Host: I would think that's something that could help the handicapped. What other things, and how far have you gone on aids to the handicapped, who's senses are in many cases deprived--they're deprived of the 'full sense set'.
NN: Well, I think--let me answer that beyond Things That Think. Because the mission of the Media Lab has been, and its original purpose, was to make computers 'sensory rich' and more sensitive to, and understanding of, more sensory rich information. And you could argue that the most sensory-deprived being in this world is a computer. So if you enhance its sensory abilities, and its motor abilities, that work can be directly transposed to human problems. So, we build systems that try to see, and recognize for example people, and faces, and machine vision is very much a part of our work at the lab, I'd say it represents as much as 15% of our activity.
And if we can build vision systems, and these clearly could be made compact and tiny, and worn like a pair of glasses, to help a blind person see, by the glasses seeing, and then perhaps telling the person. Glasses convieniently wrap around your ear. You sort of have almost a way of communicating the visual scene to the ear in this sort of magical pair of glasses that are really 'eyes' in that they see for you instead of you seeing through them. Now, I'm just speculating here, we don't have such a pair of glasses, but we do work on vision systems and I wouldn't be surprised if as part of the Things That Think program, we became very engaged in the sorts of things that would help the handicapped--the hearing impaired, the visually impaired, and again, even the motor side of the equation, which is a bit more electromechanical than we normally get involved with, but I think it's something we'll do.
Host: Are you working with nanotechnology at all? Micromachines? Anything on that scale?
NN: Well, we did--we had a PhD thesis on that subject, but we haven't really persued it that much.
Host: What about machine vision. Why has that only been used a little bit in factories, and it has taken--it seems to be staying in the lab. What's holding it back?
NN: Most of the applications of machine vision, up till now, have as you say been used in factory environments or the idea that a moon lander can, you know, robots roving around on the moon can sort of recognize sort of where to go and how to sort of navigate by seeing. These don't sound very relavant to day-to-day life. And you're absolutely right in that observation. What we're trying to do is build computer systems that can see people. And sort of, basically, not only know that you're there, but that it's you.
I ask the following question of many people, and you're going to have to almost close your eyes to think about this, but: Try pretending to be your own personal computer. You understand what I mean? You are, for this moment now, you are your personal computer. And try to imagine that situation. When you're your personal computer, first, you can't even see you. You can't hear you. And the most that happens is you get poked at a litte bit. Now, I know this isn't the normal way to talk about the human/computer interface, but if you try to look at it from the computer side of the equation, it's a pretty crummy interface! I mean, you're not giving me much to work with! So, if I could just see you, and if I could hear you, and if I could do a few of those things, I think as a personal computer--I'm still your personal computer--I could do a lot better than I can today.
Host: Today's computers are really, notoriously slow. Even the Pentium chips are pretty weak compared to what we hope for, maybe in the next ten years or so.
Host: So I would think that that's part of the problem, with introducing vision, and then there's also I think, two kinds of vision. One is where you can see and recognize a particular person, then you can say 'yes, this is the owner of this house.' The doorknob sees you and lets you in. It's another one where you can distinquish just between all females and all males. Those are really two different questions, I think.
NN: No, they're very different questions.
Host: And, are you making more progress with one than the other?
NN: Obviously, recognizing a given person, especially if it's from a subset--
Host: To the exclusion of its twin--
NN: Yeah, twins is a classic problem. And, what it means is that you probably would always use redundant systems. I actually happen to be married to an identical twin, and I have identical twin brothers, so believe me, I understand twinness! And, it is, that is the bug in all of these programs, but humans fail too! My identical twin brothers who are both in their forties, I mistake them, if I see them walking down the street, I really have to, almost, ask, they look so much alike.
Host: It's not until you come across one of the differences, really, it could be inflections in their voice--
Host: --Or their voice, or the way they look at you when you ask that! There could be small cues, like the horse that could count, back in the Vaudeville days.
Host: It was actually responding to pinky movements or something like that, from the trainer.
Host: Okay. 300 people at the Media Labs. How many of them are working on different--how many different things are you working on at one time at the Media Lab?
NN: The place is a little bit different from most labs at MIT. We have about 140 corporate sponsors. And the 300 people are more or less evenly broken, I'm just using rough numbers, 100 faculty and staff, 100 graduate students on full salary--full scholarship, and about 100 undergraduates on the payroll. And at the moment, there are a little over 100 projects. Almost sounds as if we purposely go out of our way to do things in hundreds, but this is all quite by chance.
And the projects are really chosen by the faculty, and the graduate students that orbit any particular faculty member. And my job is to sort of connect up, or wire, these particular projects with companies that are interested in them. So, companies don't come to us and ask us to do things as much as sort of join a club, where each person is doing more or less what they want. If they're not doing what they want, there's a problem, and I better find out about it. But, the projects that people do really are driven by passion, not by some request that came from corporation X or Y.
Host: So it's almost your position to link the thinker--the faculty member or the grad student or undergrad--to the appropriate project so that he doesn't feel that he's stifled, even though in reality, in a way he sort of is because the sponsors just want what they want.
NN: It's--it's--from the faculty member's perspective, it looks like the company is funding them to do exactly what they want to do, and from the companies perspective it looks like the faculty member is doing exactly what they want him or her to do.
Host: And that's the job of the manager. ...[break]... You just finished up the 10th anniversary celebration on the Internet, for the Media Lab. Tell us: what results did you get out of that?
NN: Well, the results aren't really in yet. There was a massive effort, to try and host 10 million people on the 10th of October, as well as host, physically, a couple of thousand people here at MIT. So, we set up a web site which had a dedicated broadband line, and literally hundreds of machines at this end. And we literally--every country in the world, and people from all over tuned in, but actually sorting out the data, and getting you some results of what's out there, that's going to take several weeks, if not months. But, it went very very well. I was surprised that it went as smoothly as it did!
Host: I checked it out several times during the thing, I never had any problems getting there. AOL, America Online, was a major sponsor I guess?
NN: America Online was a major sponsor, as was Intel, and DEC, and a few others who were extraordinarily generous in putting this together.
We called it A Day In The Life Of Cyberspace, which is taken from that series of books you're probably familiar with, A Day In The Life of America, etcetera, and we are now working with Harper-Collins to turn it into a book.
Host: Oh, so then you won't be sued for stealing their title! (laughs)
NN: No No! We cleared that very carefully!
Host: The Media Labs--a lot of people like to see results from scientific endeavors. What sort of things have been pushed out the door besides the multimedia work that you started doing, that now is more commonplace. What other things have been pushed out the door from the Media Labs? What are you proud of?
NN: Let me answer that in two different ways. I can point to shrink-wrap products, like the Lego-Logo work we do, that is sitting on the shelf, and 20 million children around the world are using it. And there are examples like that.
But I'm more interested in a second kind of transfer, which is in some sense the one you never get credit for, I constantly tell the students and faculty here that the best technology transfer you can have is the one you never get credit for. And since we have so many corporate sponsors--as I said, 140 of them. And they visit all the time. We have a minimum of four companies a day in this building. And a lot of the ideas get transferred in a much less formal way. There's a great deal of product, hardware and software out there, that originated here, gets carried back to the company, then gets folded into a program, and then appears, but it isn't a crisp, clean package that we've delivered, it's basically an influence.
Host: A derivative product.
NN: I think we're more in the technology influence business that the technology transfer business.
Host: Okay, what--let's see, what's a good closing question since we don't have much time left. What do you see the Media Lab doing twenty years from now?
NN: Well, let's see. Twenty years from now is a long way's off!
Host: That's why I didn't pick ten!
NN: Yeah. One of the best things that you can do in a laboratory is recognize when it's, when a mission is accomplished. And it may be the case that twenty years from now the Media Lab will be totally irrelevant, and something else like Biotechnology and Computing Resource is going to be what's here instead.
I think we have to be just as prepared to uninvent ourselves as we are to invent ourselves. So, twenty years from now--even I who likes to think about that, has to recognize that the technology of computing, whether it's still based in silicon, or if by then it's more biological, I find it hard to predict. So I keep my window sort of like, five to ten years out.
Host: Well, things do tend to happen, I think, more slowly than I ever expect them to happen in the computer industry. I thought, fifteen years ago we'd have fiber optic in every home by now. And, I don't know why I was misled, because we should have fiber optics in the home.
NN: Well, look at the World Wide Web. I mean, people didn't even know what that was, and it's doubling now every 55 days. So some things happen very fast. Netscape, the browsers that follow the World Wide Web. I mean--it's--it's speed. I think we're going to see in the next few years things moving much, much faster. I agree with you completely, in the past I was always disappointed. It seemed to take, like, ten years for an idea to go from the lab to be actually a product, and that cycle time is dropping very very fast.
Host: Okay. What about--we only have about a minute and a quarter left, and this is an unfair question in so little time. But, there's a new product out called JAVA.
Host: And, one of the problems that I see, as someone with a long time in 'the biz', is the potential of spreading viruses though a product like that. In thirty seconds or less, what is the Media Lab able to do, or been doing. Have they been working on anything that can help us? That can protect us from things like that?
NN: I think we have to protect us from all sorts of things. Not just viruses but security and privacy in general is a big topic. And I'm not very good at the dark side of things. I'm a very optimistic person, so I don't think too much about the dark side. But the dark side side really is security and privacy, and that includes security from viruses. We're going to start to do more and more in that area, in the past we haven't done much. But, I think that when people talk about the Internet, and they think about the Internet, they should realize that in the digital world it is potentially far more secure than the physical world. Right now it's not secure because our government has some very odd ideas about export controls of encryption that just has to change.
Host: My guest has been Dr. Nicholas Negroponte. Thank you very much for being on the show. You've been listening to High Tech Today. Bye Bye!
Related Material Outside this Web Site:
MIT MEDIA LAB
If you enjoyed this interview, here's others
Interview by Russell D. Hoffman
The host would like to add the following to this electronically-circulated World Wide Web Internet Page:
Please distribute this document IN ITS ENTIRETY (as a raw HTML file or printed document). Please link to it rather than placing it on another server. Thank you.
The Animated Software Company
Last modified March 27th, 1997.
Webwiz: Russell D. Hoffman
Copyright (c) Russell D. Hoffman