Seems like it is assuming all of those ancestors had unique DNA when in fact they were probably all 99+% similar if you look at it from the perspective of the whole genome. I'm no biologist though, would be interesting to see a real explanation.

I don't know either, but I read on HN some time ago that the size of a CD were comparable to ... I don't even know, a human genome. So, assuming that's a lower bound, 1% of 600MB is still 48 million bits. 2^48 would be the number of different strings representable with 48MB, but maybe not all of those will be viable.