OK, I agree that finding a non-CA way of computing physics doesn't necessarily rule out a future CA model being better. Current physics models (integrating the

OK, I agree that finding a non-CA way of computing physics doesn't necessarily rule out a future CA model being better. Current physics models (integrating the Schrödinger equation) aren't computable even in principle, because they involve infinite continuous vector spaces. So a CA model is a theoretical breakthrough because it's computable, even if it's not efficient. Numerically integrating the Schrödinger equation scales exponentially in the number of quantum states, while the CAs proposed in NKS only scale with the volume of space, so there is some size of system where it'll beat numerical integration.