> We show that if you have two low-quality random sources—lower quality sources are much easier to come by—two sources that are independent and have no correlations between them, you can combine them in a way to produce a high-quality random number
"Independent and no correlations" sounds like a crippling assumption if you want to use any two deterministic PSRNGs. How can you possibly guarantee they're completely un-correlated and independent without seeding them with collectively more bits of entropy than you can get out of the combined system?
I'm not sure what "independent" is even supposed to mean for a deterministic sequence, which by definition is recursively dependent.
"Independent and no correlations" sounds like a crippling assumption if you want to use any two deterministic PSRNGs. How can you possibly guarantee they're completely un-correlated and independent without seeding them with collectively more bits of entropy than you can get out of the combined system?
I'm not sure what "independent" is even supposed to mean for a deterministic sequence, which by definition is recursively dependent.