When the 10x developer topic comes up, I think about people like Joe Armstrong or Peter Norvig.

They come up to solutions to problems that very few others would even think of. How many times more productive does a solution like this make Armstrong, compared to the median developer?