I’d guess it goes something like this: the answer being computed is what set of genes is best adapted to the environment (a search problem over the space of reachable organism genomes); each organism is a possible answer; every generation, an organism producing more or fewer than the average # of offspring represents a computed 1 or 0; after enough generations… Not a Universal Turing Machine, no, but still computation.
Yes, what gwern said. Evolution produces (very slowly and wastefully) things that are well adapted to their environments. It seems reasonable to call this an instance of computational power. If you (PK) prefer not to, though, fair enough; I think we would only be disagreeing about words, not about things.
I’d guess it goes something like this: the answer being computed is what set of genes is best adapted to the environment (a search problem over the space of reachable organism genomes); each organism is a possible answer; every generation, an organism producing more or fewer than the average # of offspring represents a computed 1 or 0; after enough generations… Not a Universal Turing Machine, no, but still computation.
Eliezer gives a few examples of this kind of thinking in http://www.scribd.com/doc/2327578/Worlds-Most-Important-Math-Problem-Eliezer-Yudkowsky-Future-Salon and I gather it’s a reasonably well-established way of mathematically approaching evolution.
Yes, what gwern said. Evolution produces (very slowly and wastefully) things that are well adapted to their environments. It seems reasonable to call this an instance of computational power. If you (PK) prefer not to, though, fair enough; I think we would only be disagreeing about words, not about things.