Views
Back to Personal Views > Publications & Opinion > Back to Homepage


Why the physicists clubbed a chicken
PODIUM, Peter Cochrane

From a speech by the Head of Research at BT to the Royal Society for the Arts, Manufacturers & Commerce

A FEW years ago, I had a rather haunting experience. I stood in front of a large column of formaldehyde containing a splendid human brain complete with the eyes, the spinal cord and all the spinal feeds to the rest of the body trailing down. It suddenly occurred to me that this was the complete human being, the complete system, and we can use only a small part of it. Do we really understand it - and do we use it effectively? No.

My ambition as a human is threefold. I would like to be eclectic, holistic and omnipotent. The third one is always a bit of a problem, but I would really like to try to understand everything, I would like to try and influence everything, and I would like to be a positive force for change. But it is becoming increasingly difficult, and the only way I can see of achieving a fraction of what I would like is to somehow to join forces with technology.

The amount of information now being created and published is phenomenal. It's like lying underneath a waterfall of data. The half-life of information in my industry is now about six months. If you were to read the research reports on neurology alone 24 hours a day non-stop for a year, after one year you'd be only two years out of date. This is a stunning problem for our species. Perversely I now spend 15 days a year looking for something to read. For an hour every Sunday morning I go through the newspapers desperately trying to find something worth reading. Only one in four Sundays do I find something - which is incredibly time expensive. When I was 20 years old, 15 days a year was no big deal, but now I'm coming up to 53, 15 days a year is a big % of what I have left. So I really would like technology to bring stuff to me, but I don't want technology that is cold and sterile. I want technology that is serendipitous, something of the quality of someone coming into your office and saying: "Seen this?"

The library at my laboratory used to occupy a huge room full of shelves of books. It was ripped out nearly five years ago and replaced by what has turned out to be the world's second-biggest digital library, with over 500 Gbytes. The biggest library is currently at Los Alamos and is about 5x bigger at 2.4Tbytes. The digital library has transformed the way we do work, and has become an essential part of our lives. What we do now cannot be done on paper, and our work output has gone up about 10 fold in the space of 10 years.

Some very interesting things are starting to happen in the way engineers and scientists are now trying to look at what machinery is doing, as opposed to what evolved systems have done. By 2020 we expect machines to be writing better software that humans - which is a little earlier than we thought just 10 years ago. My latest laptop has much more processing power than an ant, so why isn't it more powerful? For one thing, it suffers sensory deprivation on a scale that is hard to understand or comprehend, so no wonder it's dumb.

Some people are now creating little machines that have an interesting mix of brainpower and sensory capability. One of the most remarkable I have seen is a tiny, six-legged robot with only three transistors - like three crude neurons - and it can walk. As an engineer, I couldn't conceive of how you get something with just three transistors to walk. The clever thing is that it's got a lot of sensors, and it's the sensory feedback that creates the intelligence.

A virtual shark has been created that can swim inside a computer. No one programmed it to swim; it evolved the ability to move in its viscous world and, remarkably, it does so in a way that mirrors the carbon life form that it is emulating. We now have to think in terms of electronic agents swimming through a network of fibres, able to search and find and bring information to us.

For thousands of years philosophers sat round tables with a chicken in the middle trying to understand how chickens work. But they didn't make a lot of progress. The guys who did the first smart thing were the physicists. They clubbed the chicken to death and took it to pieces. That way you get a first inkling of how it works. However, engineers have concentrated on building chickens, and it is this route that will probably see our first glimmer of true understanding. But the guys that really worry me are the software engineers; they just want to specify chickens!

We may not be able to create a butterfly yet. We may never be able to do everything that Mother Nature can, but there are things that we can do. Complementary technologies help us to win in a sea of activity that is becoming very difficult, very fast and very complex.