Do you know if there are any arguments in the literature to the effect that actually useful quantum supremacy may never be possible in principle (not merely in

Do you know if there are any arguments in the literature to the effect that actually useful quantum supremacy may never be possible in principle (not merely in practice)? What I mean is: decoherence with environment must be held at bay long enough for the final answer to appear in the qubits with high probability. Based on examples from thermodynamics (such as impossibility of extracting useful work from a heat reservoir without a second, lower-temperature reservoir into which to dump heat), I could imagine a scenario where decoherence would always foil the computation at some point, perhaps at some intermediate time or in the final measurement. In particular examples, this would show up as irksome experimental limitations, but in fact these limitations would all stem from a deeper prohibition. And to the extent that decoherence could be managed (such as in the cited experiment), the resulting computation would not be "useful".