Besides, why would it matter if the observer is conscious or not?
Our psychology is biasing our reasoning about these thing, leading us to attribute something magical to the process that make us "us". It's understandably utterly disconcerting to lose the ground under the sense of self that we inhabit, but that's not a reason to discard, a priori, a line of questioning reality. I know it's hard to shake this uneasy feeling off, even for non-religious people.
But let's imagine for a moment that whatever the substance of our consciousness actually is, it can be forked/cloned into copies of itself that no longer share the identity after having been cloned, yet retain the memory of past events (external or internal to itself). I realize this is a bold assumption, but arguably less contentious to accept than whether machines can ever become conscious (perhaps because it sidesteps our innate aversion for outgroups, non-biological entities being extremely outgroup).
Now, imagine you fall asleep and get cloned in two, and the two resulting persons are put, still asleep, in two rooms (room 0, and room 1). After an hour they wake up. They both have copies of your memories and both have a copy of whatever-the-thing-that-holds-consciousness is (soul, matter, different kind of "stuff", not really relevant, as long as it's not incompatible with the "cloning postulate"), so in a way, they both think they are _you_; they feel a direct continuity with what they were before falling asleep and getting cloned. They'd have no way to tell which of the two entities they are though.
Imagine now being one of them. You get out of the room and peek at the room number on the outside of the door. Which number will you see?
The sequence of thoughts from the moment each clone realizes it woke up in room X, depends on the value of X, which potentially has lasting psychological effects that boost the divergence of the two clones. For example: imagine you've been told that cloning doesn't "split" you in two equally "genuine" clones, buy instead that the first one is the "original" whole the one in room 1 is the "clone". When you wake up and realize you are the clone, your thought process will probably diverge from the original.
As soon as the copy was created, it would diverge. But both would have ~equal claim on being "me".
> They'd have no way to tell which of the two entities they are though.
Not until they learned it from someone authoritative.
From your answer it seems you're also ruling out, a priori, the possibility that cloning a conscious being is not only impossible in reality, but absurd, regardless of any concession that would allow consciousness to still transcend matter (provided it allows for cloning).
I'd be interested to hear your thoughts on how you derive such strong conviction about the non-cloneability of conscious beings.
Of course, that's different from people being cloned all the time implicitly, aka many-worlds.
However, my instincts also tell me that reproducing quantum states for quantum computing must be basically the same thing as cloning objects, difficulty-wise, so I expect large quantum computers to become exponentially difficult to construct - but smart people who actually understand the math don't seem to believe that. So I don't have firm grounds for knowing the world works the way I think it does.
That doesn't require us to clone an object; the argument only requires us to imagine how would it feel if it's happening all the time to us and conclude that we wouldn't feel a thing subjectively.
For example, in your scenario, neither of the consciousnesses is ever going to experience what the other experiences. Thus they must be different, and not truly clones.
On the other hand, if I copy a piece of software, it is in some sense the same software in both instances, though their histories are different.
Apparently we don't mean the same thing by "cloning". Let's call it "forking".
Btw, the same "uniqueness" issue applies to running software as well. The instance of a running software is not the same thing as the code of the software. If you take a snapshot of a software running on machine A and resume it on machine B, they no longer _are_ the same instance. It's clear that they exist independently because you can act on them independently and their state and future reactions will diverge.
That said, it's trivial for us to run n copies of the same state transitions with software (in lock step), something that might just not be possible for us to do with arbitrary precision on biological systems. That's not what I meant by cloning: I'm not saying having an exact "lock-step" reproduction of whatever the internal mechanism of consciousness is (that would be another debate).
But let's get back to the main point of my argument in this thread:
Wigner's friend thought experiment is absurd/paradoxical only if we cling to the notion that we'd somehow be subjectively affected by the fact that nature effectively forks off and entertain parallel histories for complex systems such as conscious beings in the same way as it seems to fork off and entertain parallel histories for simpler quantum systems when they are in superposition state. We don't have to invoke the many-worlds interpretation, which is just an interpretation; we can jus look at the math that describes how quantum states evolve: our most accurate description of what we observe in the world tells us that the behaviour of particles is consistent with viewing them as if they fork off and evolve in parallel over time until, at some point, when we add ourselves in the mix as "observers" then suddenly puff this superposition collapses and we observe the particle having a concrete state, at random.
Common ways to explain this collapse are:
1. The act of observing a system, done by a conscious being, collapses the quantum state, because consciousness is special
2. A conscious being it's just a very complex system that obeys the same rules governing the system it obseves, and thus gets involved in the superposition while performing the measurement, literally becoming part of the superposition. Thus the observed randomness is just an artefact of subjectively "waking up in room 1"
Wigner's setup exaggerates this problem by adding another observer (the scientist friend). It poses absolutely no problems of you espouse the second explaination above. It poses all sorts of problems if you absolutely cannot stomach the possibility that all of us people might not live in a uniform linear objective reality that works according to our intuition.
Simple != Intuitive
Certainly Occam's Razor implies that we are.
Do you assume that it's absurd to consider that a conscious being is really no different than any other memory equipped machine and thus can be in a superposition state like anything else without necessarily "feel" anything special about it?