niplav comments on Mistakes with Conservation of Expected Evidence

[niplav]

Part 2 (and the dream algorithm) remind me of semi-decidability.