Some selection pressures should be too weak to produce phenotype changes. They should be lower than the noise, inconvenient for the genetic programming language or protein building elements, or blocked by the spandrels of competing adaptations. This does not appear to be the case. A selection pressure for just about anything will, in the fullness of time, produce behaviour or anatomy perfectly fitting that pressure.
E.g. frogs get upset by earthquakes, and thus frogs have adapted to detect earthquakes early so they aren't too surprised. They just sort of panic in response, but it's not like the earthquake poses any threat to them aside from the cortisol spike.
E.g. border collies are adapted to herding sheep. You hardly need to train a good border collie, they see sheep, they herd sheep.
Evolution is also too fast, but never mind.