I've been thinking about this game for a while. It seems to me that there is a fundamental problem with translating evolution to games. In the real world, evolution creates new things. You start with a single cell (or something even simpler), and you eventually arrive at fishes and ants and elephants and humans. Games can't do that; you'll always be limited by the parameters the game designers envisioned.
In the pitch, Devin mentioned that they were aiming for two species, and three traits. I think that approach - having a fixed set of species - might not be the best way of approaching the game, given that you'll already end up with a very limited system, compared to real evolution. Instead, a different approach could be to do something that's closer to real evolution. Don't have species (in the real world, "species" don't exist; they're a human concept that attempts to describe how nature works, but fails when you start looking a bit more closely). Instead, during mating, compare the genetic difference (i.e. the difference in traits) between individual animals, and then allow them to mate if they're not too different.
That way, you'll get "automatic" speciation without implementing the concept of species explicitly!
If you hunt for slow creatures in one place, and for fast creatures in a different place, you'll eventually create two different, incompatible species. This also allows for weird concepts like ring species (neighbouring individuals can mate with each other, but individuals further away can't mate anymore).
This would require increasing the number of traits to include purely cosmetic traits like eye color, and maybe even introducing some entirely hidden traits that aren't visible to the player, and don't have a direct gameplay impact, but do impact speciation. After all, that's how the real world works, too.
Since some of these traits wouldn't have any direct gameplay impact, and since this approach would remove the concept of "species" as something that's implemented explicitly in the game, it might not be harder to implement than the concept shown during the pitch, but might lead to a more interesting, deeper game that emulates real evolution a bit more closely.