Abstract
Two years ago, I proposed that near-term, sampling-based quantum supremacy experiments, such as Google's, could be repurposed to generate bits that are certifiably random under computational assumptions. In this talk, I'll discuss an aspect of that proposal that I haven't had time to go into in earlier talks: namely, how does one actually derive the conclusion that fresh entropy is being generated, from relatively "standard" computational assumptions? I'll also discuss the major challenges that remain in turning this proposal into a genuine near-term application of quantum computers.