re: 2, there's been some promising developments with Verifiable Delay
Functions that make me think that the block regulation problems are
solvable without requiring brute-force search proof of work. Are those
inapplicable for some reason?