Suppose that fast quantum random number generator (QRNG, https://qrng.physik.hu-berlin.de/) is placed in a subsystem which has limited interaction with the rest of the world. What would happen if decoherence rate for this subsystem were not fast enough to extract all the random numbers being created by QRNG?
It seems that then the subsystem has to evolve into a state with macroscopic coherent properties (Schrödinger cat state). At least this occurs if one simulate QRNG using measurement of a rotated qubit and copying a result in a vector of qubits representing a (partly) closed subsystem.
If so, why the question about feasibility of macroscopic coherent states is still not settled?