Quantum mechanics is just a statistical theory. There is no magical role for observers.
In classical statistics, we also describe systems as a vector that is formed by superimposing basis states, and in classical statistics, the definite configuration of the system is also not tracked in the mathematics. The whole point of statistics is that you assume that there are limitations that disallow you from tracking the definite configuration, so it is not there in the mathematics, but that does not prove it is not there in the physical world.
There is just a huge problem among physicists in that many of them never take a class in statistics and have no idea what statistics even are. I remember once trying to talk about probability vectors with a physicist and they legitimately had never heard that term before but could not admit there was something they do not know so they accused me of making it up and that it is not a real concept. This is just how dire the situation is with how little physicists actually engage with statistics.
In statistics, you do not track the definite value of the system, only likelihoods of different possible configurations, and so the definite configuration does not exist in the mathematics, even if it exists in the real world. Physicists, who don't engage with statistics, see that their mathematics does not typically contain a definite configuration, and then declare this must mean that there is no definite configuration in the real world!
Of course, if we go observe the particle in the real world, we find it has a definite configuration, and so they then introduce an additional postulate that says it suddenly "collapses" down such that it acquires a definite configuration the moment an "observer" or a "measuring apparatus" looks at it. Yet, if you look at the mathematics of this supposed "collapse," it is literally mathematically equivalent to just applying Bayes' theorem to the degree of freedom of the quantum state containing the probability vector!
That's just straight out of classical statistics! No mystery or magic, but of course, if you know nothing about statistics, you won't understand what I'm talking about at all. What the hell is Bayes' theorem? What the hell is a probability vector?
All the supposed "weirdness" stems from this completely unjustified lunacy. Nothing in the academic literature justifies this nonsensical dogma. I often see people trying to justify it by referencing a lecture from Feynman or Deutsch, who both argue you can "prove" that particles do not have real values until you look at them from the double-slit experiment.
Their argument is as follows. Cover slit A and collect the statistics of where the photon lands on the screen when it goes through slit B. Cover slit B and collect the statistics of where the photon lands on the screen when it goes through slit A. Superimpose those two statistical distributions. If particles have definite values at all times, Feynman and Deutsch argue, then the statistical distribution when you open both slits should be that superimposed distribution. Yet, we know it isn't in practice, therefore, they conclude particles have no properties until you look at them.
Why must it be? What law of logic says that? What law of statistics says that? People who regurgitate this argument just insist it must be so, and yet, that is all they can say about it. It must be so. Why? There is no argument beyond that. It is an entirely arbitrary premise. There is no statistical law or law of logic that says it must be so.
We can prove anything if we're allowed to just make up whatever rules we want and insist it must be so.
And don't even get me started on the huge industry centered around blatantly lying about Bell's theorem.
It is trivial to represent any arbitrary quantum system in terms of a Markov chain.