1:31 PM
@ray That piece demonstrates some apparent misunderstandings of how genetic algorithms or natural selection work, and the bridge analogy doesn't make any sense. A genetic algorithm for making a chess-playing machine wouldn't use, as its generation, individual moves; it would use entire games. And it wouldn't look any steps into the future; it would just make new generations of machines based on machines that were more successful than others at winning chess games in previous generations.
1:41 PM
... There are many, many naturally-occurring bridges. And - oh! - a project in 1999 that designed one using a genetic algorithm. But fundamentally, if you're looking trying to evaluate the soundness of natural selection by picking a goal and then trying to design a genetic algorithm to meet it, then you're not getting what natural selection does. The question isn't whether natural selection could make a bridge; it's whether it could have ...
« first day (1 day earlier) next day → last day (16 days later) »
Transcript for
Mar25
Mar '1527
Mar28
Discussion on answer by ray: Chovos H…
Imported from a comment discussion on judaism.stackexchange.co...