Evolutionary changes occur by random mutation and, in sexual reproduction, by random genetic mixing. Whether or not those changes are successful is governed by survival, but it's not directed by survival. Survival doesn't choose specific evolutionary changes with a goal of increasing survivability.

This has become a completely semantic argument, but HN is a great place to argue semantics, so here we go.

> Evolutionary changes occur by random mutation and, in sexual reproduction, by random genetic mixing

The genetic changes of evolution are due to random mutation, but the ones that don't survive are not "evolution". Evolution is the subset of random mutations that increase reproduction/survival rates. If, at some point, a chimp was born with two heads, we wouldn't say that "chimps evolved to have two heads" because that two-headed chimp didn't pass on its genes. We'd say "a chimp randomly had two heads one time" and wouldn't bring it up in a discussion of evolution at all.

> governed by survival, but it's not directed by survival

To be directed by something does not always imply intentionality. Of course the most common usage of "direct" indicates something intentional, but that wasn't how I meant it. I meant "direct" in the sense that a particular landscape will direct the course of a river.

I think what he meant was that each individual mutation is random. There isn't someone intelligently deciding what the next batch of mutations should look like.