While this may be an interesting theoretical result it almost certainly has zero practical implications for cryptography.

We already know how to build secure random number generators. Pretty much every real world problem with random numbers can be traced back to people not using secure random numbers (or not using random numbers at all due to bugs) or using random number generators before they were properly initialized (early boot time entropy problems).

This random number thing is so clouded in mystery and a lot of stuff gets proposed that solves nothing (like quantum RNGs) and stuff that's more folklore than anything else (depleting entropy and the whole /dev/random story). In the end it's quite simple: You can build a secure RNG out of any secure hash or symmetric cipher. Once you seeded it with a couple of random bytes it's secure forever.