[onemoreact]

Productivity in the shortest time scales is hard to measure.

But, when someone spends a few days and makes an optimization that literally prevents well over 1 million dollars of new hardware from being purchased that's real value. (This actually happened.) Create that type of value a few times in 5 years and yes someone can really create 10x the value of the average programmer.

PS: Not a programming example but http://en.wikipedia.org/wiki/Polymerase_chain_reaction is one of the more extreme examples. Kary Mullis literally created billions of dollars of value paying him 10x what the average researcher makes for 40 years and it's still a net win.