There can always be tons of "could have" examples but his decision to push CUDA in my opinion was risk and innovation that paid off