It looks like you wrote up a correct but abstract&approximate description of the problems with that approach. I claim that as you make your model of these problems more detailed and careful, refine the vague concepts and reduce the model uncertainties, the way that this looks like it just might work will disappear.
I know this isn’t very helpful to hear, and isn’t convincing. But that’s what my thoughts are.
It looks like you wrote up a correct but abstract&approximate description of the problems with that approach. I claim that as you make your model of these problems more detailed and careful, refine the vague concepts and reduce the model uncertainties, the way that this looks like it just might work will disappear.
I know this isn’t very helpful to hear, and isn’t convincing. But that’s what my thoughts are.