We use principles from many body quantum physics to create algorithmic primitives for model counting, a core component of probabilistic inference, knowledge representation, and planning. The aim is to improve efficiency for artificial intelligence workloads at scale by deriving algorithms from physical insights and information flow behaviors.