Re: Artificial life likely in 3 to 10 years

Guy A Hoelzer <hoelzer@xxxxxxx> wrote:
in article fb49p3$fbt$1@xxxxxxxxxxxxxxxxxxx, DK at
dk@xxxxxxxxxxxxxxxxxxxxxxxxx wrote on 8/29/07 10:18 AM:

Anything "artificial" that makes any sense
to create implies something else - custom built stuff that is either
built in known and predictable ways or possesses desirable traits
and displays predictable behavior. I think we are infinitely far
from that.

I didn't have this kind of engineering goal in mind, but I'm not so sure
that this would be so much further off. The engineering of self-organizing
systems is happening now in several contexts, and I think we are coming to
understand how to do this. It is not much different from directed molecular
evolution, or agricultural breeding practices, in the sense that you control
the fitness criterion and allow the self-organizing system to solve the
fitness problem as it emerges. The "engineer" in this setting does not have
to determine the mechanism of the solution.

OK, you are right. I wasn't thinking along those lines. Of course
it's *possible*. In theory. The directed evolution is the dumbest of
all dumb approaches (we do it here :-)) but if the probability is high
enough it sure will work eventually. So the question really is how
high that probability really is. I guess you are in the optimist camp
then and I am not.

DK



.

Quantcast