I like this concept, now how does one tell the difference between a tenacious insider with a crazy idea that just might work, and a tenacious insider with a horrible idea that they just won't let go of?
Good crazy ideas revolve around solving one or a handful of extremely hard engineering tasks, which make everything else easier "assuming the crazy thing works."
Bad crazy ideas break any part of that description. The single hard thing is new science/research instead of engineering, or the new hard thing won't make the rest easier, or the new hard thing is supposed to get rid of all the other problems.
I don't know, but I agree with you that this is a good question.