The main issue with this line of argument is that you don't see people who like other interpretations claiming quantum computers won't work. Saying quantum computers imply many worlds is the classic mistake of wanting to show A=>B and arguing B=>A instead of ~B=>~A. If quantum computers were inconsistent with collapse interpretations, you'd expect people who think collapse interpretations are correct to be confidently predicting quantum computers will never work. But they don't; they just think the state inside the computer isn't collapsing.
I'm often tempted to argue quantum computers at least clearly favor ontic interpretations (ones where quantum states are real). Because good luck computing things by subjectively having a computer instead of actually having a computer. But, as far as I know, you don't see quantum-bayesianism-ists having any qualms with quantum computers. I think because if you're already biting the bullet of interpreting diagonally polarized light as subjective, and Bell tests as subjective, then interpreting a computation as subjective isn't fundamentally different. It's just more in your face about it.
Quantum computers do defeat hidden-variable models, which would have to hide the computation somewhere. But yeah, Bell and EPR ruled those out years ago anyway.
I don't see people who believe hidden variable models, like pilot wave, claiming quantum computers won't work. So I don't think quantum computers disprove hidden variable models.
I do agree quantum computers disprove local hidden variable models. Because they can run Bell tests, and local hidden variable models can't violate the Bell inequalities.
Last time I checked pilot wave theory was not developed enough to permit modeling of quantum computers, or at least no one had done the work yet to apply the theory to that use case. It is very undeveloped.
I don't think anyone has managed to write down a super deterministic model that can violate bell inequalities and contain computers. The computer bit is key here because it's what let's me create ludicrously difficult to solve setups, like picking measurement bases using sha1 hashes of starlight.
I'm often tempted to argue quantum computers at least clearly favor ontic interpretations (ones where quantum states are real). Because good luck computing things by subjectively having a computer instead of actually having a computer. But, as far as I know, you don't see quantum-bayesianism-ists having any qualms with quantum computers. I think because if you're already biting the bullet of interpreting diagonally polarized light as subjective, and Bell tests as subjective, then interpreting a computation as subjective isn't fundamentally different. It's just more in your face about it.