Reconstructing Physical AI Decisions: Why State Matters More Than Models

Most physical AI failures are system failures rooted in missing state. Logs, vectors, and explainability cannot reconstruct the exact conditions under which decisions were made. This paper outlines a shift toward structured, connected, and version-pinned state models that enable deterministic replay and true auditability.