Response to "the microprediction domain is empty or very small"

[microprediction]

Response. Not so sure about this one. If you think microprediction oracles are useless you are coming up against an applied mathematical truism, an operations research truism, and so forth. Purely in the interest of thoroughness, my book tries to reduce this position to absurdity if it was not already absurd, by demonstrating ad nauseam that repeated quantitative tasks are useful - both by explicit example and through the power of abstraction (value functions, etc). Clearly, Chapter 2 could have been ten or a hundred times as long, surely.