What I mean is that the idea that reinforcement learning is a 'low level dual' of the problem of consciousness is also a theory about the problem of consciousness. It's part and parcel of philosophical topics that one can't get out of the game...