First I’ll outline the problem. It’s well
within the capabilities of computers to program other computers, and robots can
do things like build computers. The machines of the future may be able to build
more and better machines without us getting involved at all. And if, as Descartes
and even Dawkins sometimes seem to, you think that a designer could not design
something more powerful than itself, you are wrong.
Now, there might come a point when the
machines are better at designing machines than humans are. After that point,
not only will each generation of designers have more to build on than their
predecessors, but they’ll also be better designers. This will have
two effects. First, technology will advance much faster. Second, the robots’
designs will be better than anything we can come up with, since they are better
designers than us. Our puny weapons will be no match for their superior intellect, and the robocalypse will be upon us.
Here’s the solution. The problem was robots becoming
better at designing new robots. To keep up with them, we need to become better
at designing new humans. We need to use our knowledge of biology to produce
better humans who will in turn be better at designing the subsequent
generation, and so on. The robots’ powers will increase exponentially, but so
will ours. This will give us a fighting chance in the Robagnarok.
Of course, this plan of mine evokes the twin
spectres of eugenics and designer babies, and the subsequent generations of
bioengineering geniuses will presumably evoke brave new spectres of their own. The
reasons for not going down those roads are as strong as ever, and I don’t like the
sound of growing old in that world at all. But unless we can think of something
else, it may be our only chance. And since the sooner we start, the more likely we
are to beat them, I think we'd better come up with another idea pretty quickly.
No comments:
Post a Comment