JulianFlux Electrodynamic Vector Retrieval and Causal Routing

JulianFlux replaces Euclidean-based vector similarity with a continuous electrodynamic field to reduce hallucinations in AI agents. It uses a Julian-Gauss Fast Transform to project 768D semantic spaces into a 128D manifold, lowering arithmetic costs while adapting bandwidth via nearest neighbor distances.