Ultradian studies what makes AI systems reliable — the structural conditions that hold regardless of how a system is built.

Preprint · 2026

When a Single Agent Provably Cannot: An Elimination Criterion and a Cross-Domain Taxonomy of Coordination-Forcing Conditions

July 2026