Yes, you are right. However, this "design choice" makes things less efficient and more cumbersome to compute, not the other way around. There are easy ways to g

Yes, you are right. However, this "design choice" makes things less efficient and more cumbersome to compute, not the other way around. There are easy ways to get random numbers from continuous dynamics that do not involve something as painful as quantum fields. There are also some consistency arguments (contradictions arising in classical theories that do not exist in quantum theories, so maybe it was not even a "choice", but the only way to make a self-consistent simulation). And there is the whole interpreation-of-quantum-mechanics issue: it is a separate can of worms to even claim that measurements of quantum observables produce random outcomes (but that question is too philosophical for me).