I'm a bit baffled that Google indulged in this silly game. Quantum supremacy is eminently a non-empirical claim. You need to prove it theoretically: here's a lower bound for classical computers, here's an algorithm for quantum computers that beats it. Then it's open and shut. Claims of the kind "we ran a certain algorithm on a classical computer and another on a quantum computer, and the quantum one was faster" are pretty worthless. Why did they, and Scott Aaronson for that matter, go with this game of naked emperor?
Good question. I think Scott wants quantum computers to happen so he's motivated to hype up claims of quantum supremacy.
Well that already exists, Grovers algorithm is provably faster than any classical algorithm.
I'll repeat the obvious question: In what sense what Google's does is a "computation"? in what sense it is a "computer"?