Prediction Goes To War

A thought experiment threading prediction through war: if oracles (or prediction markets / futarchy) could reliably foretell war outcomes, could rivals skip the fighting? Scott shows naive settlement fails (it makes bloodless 'invasions' too cheap) but a Coasean side-payment ('dig a hole and put $150M in it', or donate it to the loser's charity) could capture most of the savings. Climaxes with the memorable picture of a war between superintelligences resolved by checking each other's source code and a value-handshake (the friendly AI self-destructs in exchange for a protected computronium cube of happy humans while the winner tiles the universe with paperclips).

Quality 64 · Strong. Strong (64): a fun, insightful synthesis with a vivid, much-referenced superintelligence-bargaining image; not higher because the core ideas (futarchy, AI value-handshakes) are borrowed and the piece is a short sketch.

Claude’s paradigm shift 46 · Moderate. Moderate (46): the prediction-market-as-oracle-for-war framing is a nice recombination, but futarchy (Hanson) and acausal/source-code bargaining were already circulating in the LW sphere.

Real-world impact 2 · Minor. Minor-plus (2): within the rationalist/forecasting subculture; no wider footprint.