Resonant Intelligence: Born Rule Fusion of Large Language Model Ensembles

A theoretical framework proposing Born Rule Fusion (BRF) for combining large language model ensemble outputs using quantum-inspired interference. BRF introduces bilinear cross-terms modulated by phase cosines — a structural operation absent from all surveyed ensemble methods. Three potentially fatal problems are presented with equal rigour. The framework is falsifiable: a concrete phase coherence test using the Rayleigh statistic is specified. All conceptual contributions are the sole responsibility of the author.

v2 (2026-03-03): Corrected Council of AIs accuracy figures to match primary source; fixed citation numbering in Section 7; added normalisation assumption parenthetical in Section 3.1; de-duplicated Weinberg discussion across Sections 2.2 and 5.5; resolved orphan lines at three page boundaries.