Today, let's talk about distributed intelligence. The
University of Houston's College of Engineering presents
this series about the machines that make our civilization
run, and the people whose ingenuity created them.
My new issue of
Science magazine has a fine editorial by
Vice President Al Gore. Gore puts forth his metaphor of
distributed intelligence. He begins by talking about
computers. Twenty-five years ago, computer pundits
expected a few huge machines to handle all our computing
needs. No one saw the personal computer coming. No one
saw how machine intelligence would be distributed today.
Gore reminds us that we live by metaphors. The factory
was once our metaphor for making things. Our livers were
blood factories; our minds were idea factories. The
metaphor of the factory, the large central production
unit, has been breaking down ever since Henry Ford. Where
is a computer made today, or an automobile? Pieces are
made all over the world: China, Mexico, Taiwan, the US.
My computer was assembled at a little shop down on the
Not just computer manufacture, but computer use as well,
has been decentralized. Ten years ago the last vestige of
the large central computer was the so-called
super-computer which we used only for huge computations.
But each new generation of PCs has taken over more
super-computer turf. At the same time, super- computers
themselves become smaller and more affordable.
Besides, we can now do many large calculations with
blinding speed using arrays of cheap PCs wired to work in
parallel with one another. The upshot is that machine
intelligence actually gains in effectiveness as it
distributes itself into the population.
Now, Gore says, that's the way human potential also
works. Large central organizations can't think. Only
individuals can think. We must capitalize on individuals
thinking in parallel.
Yet we cling to the metaphor of the factory. The purpose
of a large central factory is to maximize the production
of a standard product at the lowest unit cost -- just as
the purpose of a super-computer is to maximize the speed
of a very long calculation. By the metaphor of the
factory, research is a waste of time.
But society is faced with problems that cannot be met
with the output of factories. They must be met with human
intelligence, and intelligence must be applied by many of
us working in parallel.
Henry Ford's cheap Model-T's had far less road life than
today's cars. They were fuel-inefficient and needed
constant maintenance. The factory itself was ill-equipped
to improve the product. Modern automobiles had to emerge
from the distributed intelligence of competitors and
users -- not from the Ford factory.
So Gore says we need to feed intelligence at the
grass-roots level. We need to fund education and
research. The worst thing we can to is let an educated
over-class develop. The worst thing for all of us is a
super-computer model of intellectual elitism.
Gore finishes with another wonderful term. Our survival,
he concludes, depends on our ability to create -- a
I'm John Lienhard, at the University of Houston, where
we're interested in the way inventive minds work.