Yeah, exactly. What if there're no problems to solve? I like the metronome just fine and I want to write a Tetris in Haskell just to see if I can do it, or something like that?
We need a lot of "bricklayers" that don't care about the problem but can be elastically moved around between projects to implement the ever increasing backlog of "solutions". Entire companies are created that solve that exact problem.