That would quite possibly be the worst optimization ever.
"Yeah, I could have tracked and updated n values pretty cheaply, but instead I decided to exponentiate an exponentially huge matrix and use that to update a vector containing 2^n complex numbers associated with the possible assignments of the original n bits. Also, you should use bogosort. It's the best."
Of course. I used that assumption because it was also implicit in the original claim. If we don't have any idea what the top-level rules are, then claiming quantum mechanics is an optimization artifact falls a bit flat.
If the top-level rules are classical-ish, then quantum mechanics is expensive instead of cheap. If the top-level rules are quantum-ish, then that begs the question of why we thought quantum implied simulated. If the top-level rules are totally different... then there's not really much to be concluded.
"Yeah, I could have tracked and updated n values pretty cheaply, but instead I decided to exponentiate an exponentially huge matrix and use that to update a vector containing 2^n complex numbers associated with the possible assignments of the original n bits. Also, you should use bogosort. It's the best."