Visions of tomorrow

Posted by Peter Cochrane on August 18, 2004

The Independent, 18-08-2004
By Charles Arthur

Futurologists have a reputation for grand (and untestable) predictions. But Peter Cochrane is different. Charles Arthur is impressed

The second time I met Peter Cochrane was only a few days after the first. And of course before the first meeting, years had passed without our paths crossing. "It's because things happen in clumps!" he explained breezily. "Nothing is spread out evenly. Everything clumps together. Galaxies are clumps of matter in the universe. Cities are clumps of people in a country. And so on."

The more you think about this statement, the more like a universal axiom it appears to be, working at any level, from the subatomic to the galactic. Not bad for someone who claims not to have done well at school.

Insights like that, though usually more attuned to a technological arena, are frequent in Cochrane's new book, Uncommon Sense. Although it's essentially a collection of some 50 columns he wrote for the online site (the column still continues), it's a far more satisfying read in print than on screen.

Cochrane had a distinguished career as the head of BT's research and development labs, a company for which he worked for more than 38 years before spinning himself off to become an all-round consultant and guru.

But, you're probably thinking, people who predict the future are 10-a-penny. They always pull the same trick: predict something conveniently far off (10 years, a favourite minimum distance, sounds near enough to be plausible, but is also far away enough that they won't be called to account) or simply extrapolate from a rather specialised set of existing data to produce forecasts that make mathematical sense but none at all socially. Thus George Gilder could in the Nineties put out newsletters for which people paid big money that extrapolated a rosy future for internet stocks, which all fell to earth in the bust.

However, Professor Cochrane (the title is from the Collier chair for public understanding of science and technology at the University of Bristol) has always struck me as different from both crowds. As the example above shows, he has an apt way of digging underneath our expectations and beliefs to extract the nuggets of truth.

Such as this: at school, you're taught that the world is generally well-ordered and completely understood by "linear" systems, with just little bits that are chaotic and hard to follow. Instead, once you're thrust into business and life, you realise that chaos is the norm - just think of the unpredictability of traffic jams on roads you travel regularly - and that there are only small parts we have well under control.

Makes sense? OK, now try some technological ones. Satellites have never made much difference to telecoms, except where installing cables was hard, and for broadcasts; and they're not the answer to the "last-mile" problem of getting rural homes wired up with high-speed internet. There, the solution is line-of-sight radio, which can manage up to 10 megabits per second.

Or this: power lines won't be able to deliver high-speed internet access, because all the appliances around the house keep ruining the data stream, the copper used in mains cables is wrong for data transmission, and the signals are transmitted by the wires. It's cheaper, he concludes, to put in those line-of-sight radios. (It'll be interesting to see whether Scottish and Southern Electricity, which trialled this technology earlier this year, will indeed roll it out more widely - and how customer satisfaction goes.)

More widely, he also points out our weakness - as a species - in dealing with global problems, because those are multi-dimensional: there are lots of parameters, and tweaking any particular one might make a big or a tiny difference. But mathematicians are at ease with multi-dimensional problems, and model them frequently on computers (the weather is a multi-dimensional problem). For problems such as Aids, water supplies, drugs, he comments, "without the assistance of computer models that relate data and decisions to actual outcomes, we are going to be lost in a sea of meaningless heated debate". For example: imagine you produced a cheap Aids vaccine and cure for malaria. Millions more would survive in Africa - but without birth control, famine would quickly follow, because the system is not set up to deal with so many people. "The whole system has to be addressed," he says, "otherwise the solution could be far worse than the intended cure."

There are plenty of other targets: managers who are happy to indulge their children at home, and yet are control freaks to staff who need to be creative; why you should never delete anything, but just buy more hard disks (because it makes more sense to store everything, because you'll be able to search it increasingly effectively in future); the challenge of pricing software correctly ("too high and it won't sell, too low and it won't sell, give it away and people won't value it"); and, as a sort of rallentando at the end, a prediction for some time between 2010 and 2025 of a "Cambrian explosion" of smart machines. "If we factor in the long-term impact of exponential increases in chip performance and bit density, we are looking at more than one billion times [performance improvement] over the next 30 years."

Beyond 2020, he expects "devices at least one hundred billion times more powerful than today" with location awareness, cognition and contextualisation, and the ability to continuously and automatically configure to a dynamic world.

It sounds like we'll need them - because we certainly have problems doing it. I'd recommend arming yourself with a helping of Cochrane's wisdom. He'll probably still be around in 2025 (a sprightly 77-year-old, I don't doubt), but the more of us who can clump together and think like him, the better.