The March 29, 2007
Any sufficiently advanced technology is indistinguishable from magic.
- Arthur C Clarke
Ever since Ada programmed Babbage's calculator, we've been trying to get
machines to think like humans, and they're getting pretty good at it. You
might think that because so many of us are trying so hard to make machines
think, that everyone wants machines to think; or that machines thinking is an
inevitability. Not necessarily, on either count. A lot of people are bothered
by the concept of 'artificial intelligence', and savor the citations of its
failures. For the record, The Almanack believes in the inevitability of a
machine passing a Turing test: in the foreseeable future, it will be possible
for a machine to convince at least a casual interrogator that a human is
responding. But that still doesn't mean that a machine will think 'like a
human'. There's no percentage in building a machine that's silly, whimsical,
and irrational - that would take away the fun from humans, and make the machine
Ray Kurzweil has a lot more to say on the subject of machine intelligence, with
a lot more erudition than I can bring to it, although he doesn't waste a lot of
ink on machine irrationality. Click on the 'Singularitarian' link, at right.
If artificial intelligence eschews whimsy, does natural intelligence embrace it?
Of what possible evolutionary utility might silliness have been?
The Times today reports that the Mass Extinction
didn't result in the immediate
ascendence of modern mammals. Some species rushed to fill the gap created by
the death of the dinosaurs, but were superceded by the 'long fuse' of others
which evolved more slowly.
PBS ran a show called Boomer Century
last evening. Boomers, of course, were
the children of the 50s, an era of unprecedented growth and prosperity for
US industry. We grew up with the assumption that our technology and expertise
were just naturally superior and more evolved than the rest of the world's.
It came as a huge shock to find in the 70s that German and Japanese
companies were developing, and they could make steel and cars, the staples of
our economy, as good as ours. No, better. The reason American industry was
so effortlessly superior in the 50s was that the competition had been bombed
Now, call it whimsy if you like, but there's an analogy in these two stories.
Consider that pre-war American industry was like some dinosaur-era mammalia.
When the competition went extinct, it leapt into the breach. Other 'species',
if you will, took longer to evolve - it required 40 years of rebuilding
and the Marshall plan to seed the genes, but competition has cast Toyota in
the role of dominant mammal of the post-extinction era.
The imperfections in the analogy are enormous, which makes it highly unlikely
that a machine would ever come up with such a non-rational idea. On the
other hand, a machine might be very useful for analyzing the idea to find both
the imperfections, and the potential applications.
Please add a comment.