Yet another question about Schrödinger's cat

I had always believed that the point of the (in)famous Schrödinger’s cat thought experiment was pretty obvious: people, back then, were puzzled with the state space being (basically) a vector space. Something like: "Quantum mechanics predicts superpositions. But then, by this mechanism, you would have to concede that also a cat can be in such a weird state, which clearly never is. That’s a paradox!". Period.

But, if that was the point, there would’ve been no need for the part of the joke when the observer opens the box and becomes aware of the "actual" macroscopic status of the cat: the first part with the radioactive atom, the poison vial, and the cat, woul’d been enough to get to a paradox.

What was the meaning, expressed in modern terms, of the Schrödinger’s cat thought experiment for the scientists that first discussed it?

For example, does it correspond to one or many of the following points?

1. Whether quantum mechanics is a merely epistemic theory or a realist one.

According to the translation of a couple of passages from the Schrödinger–Einstein correspondence that we can find on wikipedia, it seems the heuristic went along the lines:

• Quantum mechanics does work for microsystems.
• Here is a mechanism by which you could transfer indefinitness from a microsystem to a macrosystem.
• But we never see indefiniteness in the macro world, including when the observer opens the box.
• And surely the observer, upon opening the box, gets to know the real state of affairs.
• Therefore indefiniteness cannot reflect the real state of affairs, even for the microworld, but only our knowledge/ignorance about reality.
• In other words, we say quantum mechanics is not a "complete theory".
• This sucks.

So, it seems, they weren’t puzzled by superpositions by themselves but by the fact that those superpositions didn’t (seem to) possibly describe the "real state of affairs" of quantum systems but seemed to be "just" an epistemic device.

2. The quantum-to-classical transition.

Regardless from whether QM describes reality or our knowledge of it, there’s something to explain: why do macroscopic objects never behave such as they are in superposition of macroscopic variables like position or momentum? The mechanism by which superpositions are transferred from micro to macro in Schrödinger’s cat must break down somewhere. But where? (Please notice I don’t want an answer to those questions: I already know the answer would involve decoherence and einselection)
But this doesn’t need the observer opening the box to have a paradox: the cat would be enough.

3. The measurement problem.

Conventional QM is formulated in such a way that epistemic content seems to explicitely influence the dynamics: unitary evolution has a hiccup when a "measurement" or an "observation" takes place. "Measurement" or "observation" aren’t defined in terms of more fundamental physical processes: they’re just primitive terms in the theory. This wouldn’t be so bad by itself. The problem is that observers (whatever they are) are themselves physical systems, and measurement (whatever it is) is a physical process so, under reasonable assumptions of reductionism, measurement should be governed by SE instead of having Born hiccups.
Even substituting the word "interaction" in place of "measurement" doesn’t help, because it would similarly lead to an unexplained breakdown of reductionism somewhere along the measurement chain.
(Again, I don’t want a solution of the measurement problem explained to me: I already have a couple of interpretations of QM that solve this and that I happen to like)
But this, in turn, wouldn’t need the cat in the first place, to get a paradox: the radioactive atom by itself would be enough.

So, am I correct if I say the original point of the Schrödinger’s cat thought experiment was only point 1. and not points 2. or 3.?

In modern times, do people typically illustrate Schrödinger’s cat paradox in order to also express points 2. and 3.?