It might be that the connection between information entropy and thermodynamic entropy aka Landauer's limit, is because classical computers have to work in a heat bath. Quantum computers work at low temperatures and with small ensembles, each in thermodynamic isolation from each other, these are called qubits. --note that photons are ab initio thermodynamically isolated from other photons, so it's a given they can be qubits. Susskind says that quantum information and its entropy are no less and no more than the entanglement measure, between those same small ensembles of quantum particles. Once entanglement exists it exists until a measurement is made either locally or by the environment, classically a heat bath. It's when you try to leave the bathroom that entropy gets complicated . . .
What all the ah, confused thinking and argument might be about is what the relation between information and energy is. Relations are more general than functions. Relations are axiomatic and need not refer to anything with physical logic. You can relate meaningless information to other meaningless information. However, we prefer computers that do have meaningful relations. We want to be able to give energy a meaningful relation to information; for instance efficient machines don't waste more than a "reasonable" amount of energy if they can compute "fast" results. That's a meaningful relation between energy and computational time.
It's fairly obvious an observer doesn't need to be intelligent; the meaning of observe is somewhat fluid. I think it's just an example of a receiver of information. The process of successfully receiving some information, some data, implies the information wasn't discarded or lost during transmission. How or why we want to take this observer paradigm, or give it extra capabilities after receiving some input data, and point it and ourselves at the universe is obviously restricted by the way the observer receives transmissions. Processing received information is another, more complex area, whose only purpose has to be understanding what the information is saying about where it came from and how was it sent. It needs a context and that context is physical logic.
If anyone hasn't watched this Susskind lecture, they should. It nails the idea of information, in its physical guise to the wall of modern theories. I think he is more or less trying to explain how information appears, or maybe how you can store it by creating a pair of entangled black holes, or . . . something.
Information doesent exixts by itself. It needs some controversial "force" that contradict it to become "information". Without... it remains "presence". (To be or not to, be, thats the question".)
