The proof accounts for that. You can even allow for the ANN being totally wrong on "almost all" seeds except for a non-negligible fraction, the definition of a

The proof accounts for that. You can even allow for the ANN being totally wrong on "almost all" seeds except for a non-negligible fraction, the definition of a CSPRNG allows for that. If it is wrong for all but a negligible fraction of seeds, then you haven't gained much over random guessing. So the only way that an ANN could help is if the ANN approximation is sufficiently bad that it doesn't give you any significant speedup in cracking the CSPRNG. Of course, this is assuming that practical CSPRNGs conform to the theory, but we're just debating theory here.