Import AI 440: Red Queen AI, AI regulating AI, and o-ring automation
Import AI 440: Red queen AI; AI regulating AI; o-ring automation
Import AI 440 highlights two threads: Sakana used GPT-4 mini to evolve Core War programs, and specialized warriors beat 89.1% of human-designed warriors. The post says DRQ uses MAP-Elites plus matches against prior champions; a separate policy proposal ties AI rules to automatability triggers, with example thresholds of <=1% false positives, <=1% false negatives, and <=$10,000 per model evaluation.
Why it matters: Sakana had GPT-4 mini keep playing and evolving against past champions in Core War, raising the share of programs that beat human players from 1.7% for single-shot generation to 89.1%.