I agree with your point about resource scaling if we assume a classical computing model. But the Von Neumann model is purely classical — it predates quantum computation entirely. So its scaling limits don’t apply to any hypothetical simulator capable of generating a quantum universe.

If our reality is simulated at “Level 0” (fully detailed, quantum-accurate), the simulator’s hardware must be at least quantum-native or beyond-quantum. That means it wouldn’t follow classical memory/clock constraints or the exponential resource blow-ups associated with Von Neumann machines.

In other words, using a 1940s classical architecture to evaluate the feasibility of a universe-scale simulator is like using abacus limitations to argue that supercomputers can’t exist.

Your Level 0 / Level 1 distinction is useful though — the essay is more of a conceptual metaphor than a literal computational model. But if someone did build a literal Level-0 universe simulator, it can’t logically be based on classical Von Neumann architecture.