Maybe I've been in computer graphics land for too long, but I'm somewhat surprized by the author's initial surprize. Isn't it obvious that you get a fractal surface if you sum up frequencies with 1/f amplitude?
I suspect perhaps the author is surprised because squinting/defocussing your eyes at the original noise doesn't much look like the final result.
Thats because as well as removing the high frequency components (like squinting), this algorithm also is rescaling the amplitude.