He is information theoretic minded, and, with respect to comp, has a coherent view of physics. Of course he does not seem aware of the necessity of that view once we postulate comp, and the fact that this necessitates to take all computations into account, (not just the quantum one) and justify the quantum interference from inside any self-justifying universal number.
Comp shows that the bit ---> qubit road is two sided.
Technically, due to diagonalization used to make the self-reference, you get the split between truth and justifiable, which provides a tool to distinguish the qualia and the quanta, as different but related mode of information.
I think Ron Garrett explains (very shortly but rightly ) the qubit -> bit justification. Comp provides a reverse of that justification, and this doubled by the communicable/non-communicable (G/G*) splitting: the bit -> quantum-bit, and the bit -> quale-bit*, with the explanation of the fact that the quale bit* can't be quantified and described (provably so in the ideal case of arithmetically self-referentially correct machine)
Comp forces, just to remain coherent, to extend Everett's way of embedding the observer into the physical wave, to his embedding in all arithmetical relations, by first person indeterminacy, with the advantage of explaining a fundamental role to the (universal) person points of view, and hopefully so, to justify QM or refuting comp, or weakening it or constraining it.
Ron Garret might think that classical is a bit trivial, but comp makes it into two type of weird qubits. So to speak.
Bruno