His usual answer is "vote on values, bet on beliefs." So we would just vote on what to include, and then bet (through prediction markets) on how to achieve these goals.
It seems to me that many of these "rationalize all the things" schemes tend to bottom out at and build onto an "incorruptible kernel of truth" which, if corrupted, causes the whole thing to collapse (or worse: have a veneer of impartial truth while being secretly corrupt). It's a sort of microkernel approach to government.
With Hanson's futarchy, it's the hypothetical mechanism for "voting on values" in order to choose what does and doesn't go in the almighty bundle of metrics (and who gets to measure the result: they have the real power). If that can be corrupted then the whole thing falls apart. Yarvin's "neocameralism" has a "cryptographic decision and command chain" which everything else rides on (https://www.unqualified-reservations.org/2008/05/ol6-lost-th...).
This "all the eggs, one basket" approach seems fragile and ripe for subversion, especially when having fallible humans administer it. It seems more resilient to avoid concentrations of this kind of power.
Who determines what is on the ballot, or how unstructured ballot items get classified? Are "make college free" and "make university free" the same or different policy proposals?
