I'm not going to dig into the paper, but I note that the abstract claims a constructive proof. I sort of assume that anyone with such a proof could organize a convincing demonstration by letting a disinterested observer provide a series of random numbers while clicking a stopwatch.