Three Minutes: MIT Lab Director

Photograph: Webb Chappell
For decades, the Massachusetts Institute of Technology's Media Lab has been the ultimate electronics toy shop. Michael Bove, an expert in visual media and self-described Media Lab "lifer," describes what's cooking in its new Consumer Electronics Lab, the emergence of electronic ecosystems, and just exactly what will be left for PCs to do.

PC World: What are your new lab's main themes?

Michael Bove: One is power. How do you use less of it, and if you can't use less of it, how do you get it from new places? Because, increasingly, we're making these smart things that are power-hungry, and we've got lots of them.

A second one has to do with sensors, actuators, and displays: How do we sense things about people and how do we feed information back to people, and what good technologies are available for those?

A third is cooperative wireless: How do you make wireless communications systems that scale differently from the traditional ones?

Fourth, and overlapping, we're trying to create self-organizing ecosystems of smart devices. You can put arbitrary amounts of intelligence and frankly arbitrary amounts of communication into something that fits in your pocket now. But you can't necessarily work seamlessly with everything else that might be around it, even though they have a need to exchange information.

And fifth, we have an interest in new materials, which potentially include smart materials or electronically active materials, and design and fabrication methods for them.

PCW: What kinds of prototypes are you building?

Bove: One example of our work is the smart architectural surfaces project that we're doing in collaboration with the Information and Communications University in Seoul. They've had an interest in building smart rooms.

Smart rooms are spaces that are equipped with a lot of input and output and communication. Smart rooms generally consist of a room full of video cameras and projectors and microphones and speakers, and somewhere there's a closet with a rack in it full of servers, with big bundles of cables running back to it.

The idea that the ICU researchers had was, "You guys at the Media Lab have a good sense of how to build modular stuff; could you think about how to build a room like that?" And so, in conjunction with them, we created these [smart architectural surface] tiles.

Each one of these tiles is a computer, but it's also a display, camera, speaker, microphone, and a variety of sensors, and they just snap into the wall studs, which provide the power. They can talk to one other using wireless, and they can either work as a group or they can run applications individually.

Imagine that the whole room was covered with these things. The system could pick out my voice from your voice, by having the different microphones compare what they're picking up. It also would know what I'm looking at, what I'm pointing at, and so forth.

Each of these tiles is basically a high-end pocket PC, so they're cheap and they don't use a lot of power. And if you have a pocket PC in your hand and it has Wi-Fi, and it's running Linux, and it's running a particular piece of software--then as far as the tiles are concerned, it's another one of them. So you don't have to talk to the wall, you can talk to the microphone in your hand, and they hear it. And you can interact with those tiles by tapping on the [pocket PC] in your hand with the stylus. It becomes, on an ad hoc basis, part of the infrastructure of the room.

Similarly, it may well be the case that in the future we wouldn't have a videoconferencing system. Everyone comes in with a camera phone, and you flip open the lid and you put it on the table in front of you. The camera phones talk to one another, they figure out who's speaking, they get the picture of the person who's speaking, and the microphones get together and find the best copy of the voice of whoever's speaking, or in fact collaboratively process the voice they're picking up and produce a better version of the voice than any one of them could do. And that's what gets transmitted somewhere else. There isn't a computer that's managing it. It's all being done by these individual devices acting in concert.

PCW: What will it be like living in an electronic ecosystem?

Bove: One thing to think about is seamlessness. Say you're having a conversation with someone using your handheld device. And you're moving from a place where there's a Wi-Fi network to a place where there's a phone network, but the conversation keeps happening and you don't notice it. You get in your car, and the conversation switches from the thing in your pocket to the phone that's part of the car. Or you're listening to a satellite radio program in your car, and when you get out of the car it continues happening on the thing in your pocket. And when you get to your office, it switches to the speakers in your office. You don't lose a service because you switch from one context to another.

There's also the notion of the personal VPN [virtual private network]. If I've got bits somewhere on the network, I've got all of them everywhere. And it may be that there are some environments in which it's better to access them than others, but there isn't a fundamental technical barrier to my getting access to my data from any place.

The situation right now is becoming unmanageable. I don't know how many disk drives I have, I just have disk drives all over the place. And whatever I want is on the one that's not in front of me right now. It would be really sort of nice if, wherever I was and whatever access means I had to the network, all of my information just appeared to be there. And whatever security issues and digital rights management issues and everything else would disappear in that context.

If I've got a piece of music somewhere, I should always have it accessible to me, wherever I am, with whatever device I've got. If I've got a particular picture I need, it shouldn't matter that it's sitting on a piece of equipment 2000 miles away.

PCW: How does your lab look at overcoming power constraints?

Bove: Ideally, you'd like to generate power as needed, rather than have to carry around something bulky just in case. It's also possible to connect power from a variety of unusual places. For instance, you can parasitize enough power out of a Wi-Fi network to power a sensor, which then can retransmit information.

PCW: What about fuel cells?

Bove: It's probably going to be a while before they're exactly what you need for very small things. You're probably better off looking at both improving traditional rechargeable battery technology and figuring out how to use a lot less electricity. Because even if eventually you go to fuel cells, you get to make them smaller and cheaper if you solve those other problems as well.

PCW: What's your take on Microsoft's upcoming Personal Media Center devices?

Bove: It's like the little pocket-sized portable DVD players, which are actually kind of fun. But I don't think anybody would be willing to consume all of their video on a device like that. It's not a new idea. In the early 1960s people had transistor radios in their pockets and big stereos in their living room, and you didn't listen to your transistor radio in your living room.

PCW: Down the road, do you expect to carry one handheld device that does it all?

Bove: It's not really necessary that you think of devices in isolation. You really need to have a collection of devices that are talking with each other. Sometimes you do need a Swiss army knife because you can't carry a toolbox. And so it's important to have all that functionality in there. But the Swiss army knife is a terrible Phillips screwdriver.

Increasingly we're doing work in collections of smart stuff that gets together to solve some particular problem.

PCW: Such as?

Bove: I spend more time sometimes chasing down people than I spend dealing with them once I've finally found them. But your [electronics] stuff knows where you are.

But it doesn't know it knows, and it doesn't provide a way for other people to get at that knowledge.

Think about it. My phone knows where I am. I shouldn't have to call somebody with whom I am meeting and say, "I'm two blocks away in a taxi, I'll be right there."

That's a perfect example of a situation in which all of the necessary [devices] have this knowledge. But the problem is a combination of them not knowing that they have it and not having a way of sharing it with one another. A lot of those issues will be resolved.

PCW: As new crops of well-connected smart devices appear, where will the PC fit in?

Bove: It's not going to disappear, but it's going to be a partner in a bigger enterprise as opposed to being the enterprise. There's a reason to have a whole chunk of storage in one place, but it's not going to be your only storage or your only processing or your only access to broadband. That's really the point.

There will be reasons to sit down with a keyboard and a screen in a fairly traditional computing environment for a long time. As long as we have to do our IRS forms, we're probably going to do them on a thing like that.

Subscribe to the Power Tips Newsletter

Comments