Although human engineers can learn a lot from the way that organisms have solved problems, organisms have not always evolved the most practical way of doing things. The evolutionary history of an organism constrains adaptations; it is not always possible to reach the most practical form of an adaptation without going through a large number of intermediate steps that reduce the fitness of organisms.
This leads to the notion that there are landscapes of genes. And the landscape is hilly. On the tops of the hills are ΓÇÿlocal optimaΓÇÖ: this is where adaptations (eye shape, say) are pretty good given the starting point. But look over to another, higher hill and the same problem has been solved ΓÇÿbetterΓÇÖ. The trouble is, to get over to that other summit requires crossing a valley. In the valley, fitness is lower than on the top of either hill. So populations get stuck with pretty good, but not always great, adaptations. For instance, human eyes are pretty good. But they have a blind spot. Octopus eyes, which are constructed similarly, are ΓÇÿbetterΓÇÖ because they have no blind spot. But for humans to evolve eyes without blind spots would require crossing a crevasse of blindness.
