The interpretation I currently have is this: If an observer O measures the state S of a system, in doing so they themselves become entangled with that system.
To another observer O', the state of both S and O are still completely probabilistic until he measures them himself. However the state of S to O is now deterministic as they have become perfectly correlated with one another (through the measurement/entaglement). This is why to observer O it seems the wavefunction of S has collapsed as there is no longer any randomness in S relative to O (but this is not to say there is no randomness in the system (S+O) to an outside observer).
Further if I were not measuring a system (my friend let's say) they could be in one of a very large number of states with each having a certain probability. If I were to now look at them and 'measure' them, in effect one of these states is being sampled and I have become entangled/perfectly correlated with it.
My question is, have I completely missed the point or is this the general idea behind "Relational Quantum Mechanics"?