The more original a discovery, the more obvious it seems afterwards. --Arthur Koestler
And the more obvious a discovery, the EVEN WAY MORE OBVIOUS it seems afterwards. --me
People who claim that no patent can be judged as obvious after it's been invented (such as Arthur Koestler) are missing something. There are two kinds of inventions: obvious ones, and original ones. The original ones come from asking an obvious question, and deriving a unique solution. One that has never been thought of. One whose details are not obvious given the question. This is a true invention -- something from nothing.
On the other hand, obvious inventions can come from asking an original question. An inventive question. And some answers are completely obvious once you ask the question. I claim that these are the kinds of discoveries that Arthur Koestler is referring to. You can even argue that the patent system should cover these kinds of inventions, because a public purpose is served by encouraging people to invent new questions by giving them a monopoly on the answer.
The trouble arises when people ask obvious questions, derive the obvious answer, and then think they can get a patent. And right now, they do. There are innumerable examples of obvious inventions deriving from obvious questions, and the public harm in allowing an unproductive monopoly on an idea should be obvious but the USPTO continues to grant these kinds of patents.
The solution is to look for obviousness in both the question and the answer, and only give patents for originality in either the question or the answer.