The computer on your desk is blind and deaf and dumb. It does not have the life experience of a two-year-old child, you use all of your senses to make sense of the world, without them you could not function… we are finding in many different areas connecting simple pieces together lets systems behave in surprisingly intelligent ways… machines start to become smart simply when we connect them up. Now in retrospect it is not that surprising - that's really what biology does, if you look at how the brain functions, its a lot of little sub systems, each doing their job, working together.

(Neil Gershenfeld, 1999)

In the past 30 years computers have been steadily shrinking. Mainframes that once occupied whole rooms have been replaced by laptops. In his provocative book When Things Start To Think, physics professor Neil Gerhsenfeld, from the acclaimed Media Lab at MIT, goes one step further. He looks at what will happen when computers disappear altogether and embed themselves in our clothing and home appliances...and then start talking to each other.

As he explains in this interview this process is already well underway:

In your book you talk about how computers should be used to free people, can you explain what you mean by that?

We have been working for a few years on shoe computers and people laugh when they first hear about that and think of Maxwell Smart. But think about the specifications of your laptop and your shoe and compare them. To use a laptop: first of all you need to carry it with you, it needs a power supply, if you travel you need adapters for many sorts of different suppliers, you need the modem cables to plug it in where you go. You need a lap, you have to sit down and open it up to use it. It's really a constraining device.

Now, if you put the computer in your shoe, when you leave the house in the morning you have it with you, you don't need to remember to bring your shoes. We have learned how to recover energy from walking so we can power the shoe so you don't need a power supply. You might need to take a walk to clear your shoe's memory, but you don't need a power supply. Then the shoe can transmit data through your body so we can distribute to the peripherals around you so your wristwatch or eyeglasses could become interfaces around you. Most interestingly of all we can send the data through your body itself. The act of shaking hands could also exchange business cards

So has the shoe become the best platform you have discovered for the laptop?

It is a very nice platform for putting the computing. This point about the technology freeing us; one of the places where it dawned on me was at a beautiful beach in Corsica. I was at the pay phone at the beach with my laptop with a very clever acoustic coupler for the strange phone system and I was logged into MIT to do some perfectly mundane departmental business. I looked down at the laptop and out at the beach and I turned the computer off and realised that the goal isn't simply to provide technology that lets people communicate, but in part the goal is to provide technology to help free people from communicating. Much of what I was doing was mundane business that didn't really need my attention. So the question is to what extent can the machines handle this low-level communication that doesn't really need us.

Think about your telephone; as the tools get smarter they can get more and more intrusive. My telephone does not have any simple social graces. It does not know not to ring when I'm having dinner, it doesn't know not to ring when I'm in the shower. A person who interrupted at those times would be seen as very antisocial and tossed out of the house in short order. But we don't bring that critical judgement to the telephone and all that is needed to cure this is for the telephone to be able to communicate with its environment and to know when I'm in the shower or when I am having dinner. Solving this problem does not really require a deep insight into artificial intelligence it just needs a telephone that knows what's happening in its environment.

So what you're after really is a phone that thinks?

For decades people have tried to make machines that think, that are smart and that's been a very disappointing effort; the machines are still relatively stupid. The computer on your desk is blind and deaf and dumb. It does not have the life experience of a two-year-old child. You use all of your senses to make sense of the world, without them you could not function, so it's not surprising that people who have spent decades typing at computers trying to program them to be smart have failed because the computer does not have access to the world.

For the telephone to know not to ring when I'm in the shower, it does not require a penetrating insight into human cognition, it requires the phone to know what the shower's doing. We are finding in many different areas connecting simple pieces together lets systems behave in surprisingly intelligent ways. We're finding that the machines start to become smart simply when we connect them up. Now in retrospect it 's not that surprising, that's really what biology does. If you look at how the brain functions a lot of what it does isn't mysterious, it's a lot of little sub systems, each doing their job, working together and then the emergent properties of the whole system are what we call intelligence.

In your book you talk about a personal area network, can you describe what that is?

We discovered how objects near you, like your shoes can change the average voltage of your body by a tiny minuscule amount and we can do it intentionally and by doing that we could send data through the body. So this is a network like any network, but your body is the network. The most interesting implication of that is the connectivity. Let's say I have a personal digital assistant and I come to a cocktail party and I want to give you my business card. Right now I can pull it out, punch-up some buttons, aim it at you and transmit it. It is very intrusive but we've spent millennia learning the protocol of how to approach each other physically and shake hands.

With one of these personal area networks the physical act of shaking hands can also be a logical act of a data exchange. So we don't have to change modes from interacting with the physical world to do the digital exchange, we can do that through the familiar physical action. Or when you touch a doorknob it can authenticate you to the house and let you in, if you pick up the telephone it can be downloading your text messages while you're listening to your voice messages. Instead of disconnecting the bits of the digital world from the atoms of the physical world they come together in the same place at the same time. It's not a virtual, or a real reality, it is an enhanced reality.

Part of your mission at the Media Lab is to create unobtrusive computers, and I understand some of your own students use "wearable" computers. Can you describe how they work?

The notion of wearing a computer is just that, wearing a computer, instead of you having to come to the computer on its terms, the computer tags along with you. Now the first wearables are horribly intrusive, they are worn by either people who really believe in these capabilities, or people who desperately need them. So in the latter category there is enormous pull, for example, from various industries, like the driver for a Federal Express delivery vehicle really needs to route an aeroplane as soon as a package is picked up and so they need a wearable computer to route that data out instantly.

At the same time there is a generation of students growing up who are unwilling to leave their capabilities of their desktop. The early cyborgs will have cameras that service their eyes, but they can put them anywhere they want. So you can have an eye behind your head when you're riding a bicycle to look backwards, or an eye down in your feet if you are navigating through complex terrain. It gets connected to displays you look through.

You can put your senses on the network because these cyborg students have radio frequency network connections, so other people can look out through your eyes and you can talk and hear and see what they're seeing. Your notion of community extends not just to people adjacent to you but to people anywhere. Right now, we're talking to each other through a special studio, with these sorts of capabilities you and I can talk to each other where and when we choose to. Then the computer you wear can carry all of the information you have about the world, so if you meet somebody you can call up the history of when you last got together and what you know about them.

We wouldn't have the problem of trying to remember people's names and where we have met them before?

That's right. When I first started teaching to these kinds of cyborg students I thought it would be horribly intrusive, I wouldn't have their direct attention. What I discovered was I had more of their attention because they made me put my lecture notes on the net so that in the glasses they wear which are also computer displays they could see me and floating next to me they could see what I'm lecturing, so they would annotate me with my context so I had more of their attention than the conventionally equipped students who had to look down at their papers and up again I had more, rather than less, of their attention.

Now what's happening is after a first wave of creating these cyborgs and wearables there has been a tremendous transformation. While there is still kind of a 'cyber chic" in looking like this, now everybody involved wants to make them go away. They want to wear what looks like ordinary comfortable clothing but just has these capabilities. So we're been discovering how do things like sew circuitry using electrically conducting threads, so we can put the chips right into fabric and it looks and feels like ordinary fabric, but it actually is a computer. We can embed displays in what look like ordinary glasses. So you can keep your familiar clothing, but enhance it with these capabilities.

In When Things Start To Think, you make the interesting point that if the book had been invented after the laptop, it would be seen as a very clever invention. Can you elaborate on that?

It is considered retro to say "I don't like computers I'd rather read a book" and there are very polarised debates between people arguing for digital books and people arguing for printed books who consider the digital book advocates illiterate heathens who don't know the value of learning, reading and writing. But just go through the specifications, they are really quite remarkable; you can open a book instantly, unlike a computer, you can rapidly flip the pages, you can drop it and it keeps working, it doesn't take maintenance, and you can view it from all almost any angle and almost any light. Those really are specifications that go far beyond what a laptop can do. The one thing a laptop can do is change, so that leads to the question of how do we bring that to the book.

Now the display in a book is very interesting and it's only recently we have come to understand the physics. There is a deep reason why a sheet of paper doesn't need the bulky power hungry back light that a laptop does. It's because the fibres in a sheet of paper are translucent, so a photon of light comes in and it bounces many times and then it comes back out. Mathematically it is described as diffusion, like ink spreading in the page, but this is light spreading and that is the mechanism that lets light from any colour and angle illuminate the sheet of paper. So we use the room light and turn it around to light up the sheet of paper, rather than putting a light bulb in it. That's such a nice idea we figured why don't we keep the paper and make the ink change.

A colleague, Joe Jacobsen, developed a process called "micro-encapsulation" to grow clear shells that are about the size of laser printer toner. In them are small particles that are black and white and have a charge difference. With an electric field we can move them relative to each other, so what this means is it looks and acts like ordinary ink, because it really is, but we can change it after you put it down.

With this electronic ink we can do many things; the first thing is we can make reusable paper that can go back in the printer and be re-imaged without throwing anything out. Even more interesting is to make a book that has sheets of this paper and electronics in the spine, so you can download many books into one book and change what the pages display. The most exciting effort is a notion of radio paper; the idea is to print a solar cell and a radio receiver right on the sheet of paper so you can have for example a sheet of newspaper that always stays current.

How would you describe the actual gadget for downloading the day's newspaper or book that you want to read?

For the book the goal, (we can't quite do this yet, we are working towards it in the Lab) would be for it to look like an ordinary book, the pages would look like ordinary printing and the spine would look like an ordinary spine. The pages would actually be made out of electronic ink and the spine would have tucked in it electronics. So for example the book could have an electronic address on the network so from a computer server you could pick out a text you want to read and over the network send out a message to your book. The book could receive the message, the text and then the pages would change into what looks again like ordinary printing in an ordinary book but it has just turned into a new book.

Are implants the next natural step after wearable computers for the Media Lab?

What we have been discussing is how to bring the "bits" and the "atoms" back together. How to weave the capable bits of the digital world with capable atoms of the physical world and this is the work for next decade. If we look ahead we've got a decade's work to do everything I'm describing; to make electronic ink books, make a bathroom that looks after your health care needs, play rooms that help children to learn, tools for developing countries, things like that.

The next step after that is to start to stick these things in people, rather than having to put it in your clothing, even better I could put this in your optic nerve and put the circuitry in your ear. Some people are beginning to do it. There are two communities; one are people studying this for medical reasons, like helping the blind to see, the other are lunatics who do it because it because it's great fun to explore.

Technologically we could start to put things in people, the problem with that is I don't trust us enough. Computers crash too often, I don't want to have to take a nap when my computer goes down. I think we need to go much further in learning to trust ourselves as technologists to provide technologies that are so valuable and so capable we're willing to live with them in that intimacy.

platform #5

© 1999 Australian Broadcasting Corporation