If you want to do exactly what some evolved thing does best, you probably cannot come close to matching it with technology.

Not necessarily true - evolution (and simulating evolution) is great at finding local maxima/minima, but not as great at moving out of those in the case where the local min/max is not the global min/max. So, for example, birds might not be the optimal way to do flight efficiency, but between birds and optimal flight efficiency if there's a region of worse flight efficiency of any real size (more than you could vault in a couple generations of lucky mutations) then evolution will never find it because the intermediate steps to get there will be selected against too heavily to jump the gap.