Axiomtheoreticalv1
The performance of an agent is bounded by the accuracy of
The performance of an agent is bounded by the accuracy of its world model regardless of the sophistication of its decision-making process.
Why This Is an Axiom
This is stated as a principle from model-based reinforcement learning (Dreamer architecture research). While empirically demonstrated in AI, the lesson treats it as a foundational theoretical principle that applies universally to agents (human or artificial). It's the 'garbage in, garbage out' principle elevated to an axiom about agent architecture.