Does Grete Hermann's thought experiment concerning the relative nature of QM stand valid? In the 1930's, Grete Hermann argued of the relative nature of quantum systems. In her words (after translation),

Quantum-mechanical characterization is not, like the classical one,
attributed to the  physical system, as it were, "in itself', i.e.,
independently of the  observations  through which one acquires
knowledge of it.

https://www.hcs.harvard.edu/~hrp/issues/1999/Hermann.pdf
In the paper, she formulates a thought experiment, in which two cases are examined where the past wavefunction of a particular electron is calculated retrodictively after observation using different measurement methods, and the result is that two different wavefunctions are derived for the same electron at the same past moment of time.
Does this thought experiment, and what it purports to show, stand valid, and what is the significance today?
 A: In the thought experiment in the article the initial wave function is the same, given the initial setup is the same, the measurement process is different: in one case it goes by identifying position in the other identifying momentum. The wave function evolves casually down to the measurement, but without including the instant of measurement, the wave function contains a certain amount of uncertainty that is specialized at the moment of the  measurement, but that uncertainty does not change across the causal evolution is always there and is the same.
It's not possible to retroactively assign a meaning from the measured position or momentum, to the initial wave function as the article attempts to do, because the measurement "instant" itself is not part of the causal evolution, i.e. the  wave function collapse if you want, wave collapse considered here instantaneous, ​is not causal and looses information (the generic uncertainty becomes a number).
What is relevant in the article is the attempt to decouple causality, from the completeness of the description of state and from the measurement process itself, historically speaking the three things were intermixed at that time and it was not obvious which part was going wrong.
For example the linear equations of quantum mechanics imply causality, orthochronus Lorentz transformations imply causality. But there is no such finite or infinite set of variables that can fully describe an electron as you would do with a classical point particle or a classical wave. There is at best an infinite set  of statistical variables, which represent in essence a more complex underlying phenomena which is a combination of fields of photons (whether virtual or not) underneath.
In this interpretation there is one single causal evolution for all states together.
Another relevant thing of this paper nowadays is the Beobachtung (in the intro) i.e. is not just about the apparatus as Bohr was stating, but the measurement result itself, which means every experimental result is perfectly legitimate (as in Everett's interpretation), and there is to the extreme opposite one causal evolution for every single experimental result, in this scenario the effect of the measurement process itself is considered "negligible" given you just "live" in one of the possible causal evolutions.
Then the description is extensively quasi-classical and let's say inconsistency prone. Ray tracing is just an approximation, and using the huygens-fresnel principle is also approximate even from a pure classical optics standpoint (i.e. the phases are not taken into account and you would need at least a kirchoff integral which takes into account phases and therefore interferences between different wavelets).
There is no such thing as a single photon and there can be always random interactions with virtual particles, in a short enough amount of time.
If you look at coherent states for example, if you describe these photon propagations as a bunch of coherent states of photons, there is no real notion of numbers of photons there either, i.e. there is no constant number for the "bunch", and even if there is a shape of experimental approximation of a very dim light in a photomultiplier, the theory itself strictly has no notion of
"one photon", and the wave function collapse can be interpreted in this case as a step from a coherent bunch of photons to an uncoherent one (where again some information is lost), and in this case wave function collapse may even take an enough small finite amount of time.
If you compare this article with the EPR article for example you see that Einstein does not attack the problem of how many variables are describing the state, and whether this description is complete, because in essence there are many thermodynamics systems with a lot underlying unmeasurable variables. He does not attack either the wave particle duality, because one may argue about a deeper/smarter theory with somewhat of this underlying information or the causality given the  linearity of equations.
What Einstein attacks is entanglement, i.e. the correlations across subparts of the system, and this is actually a fairly modern issue still and which is not mentioned in this article instead
PS: Take care that such historical articles often needs to be read in context and it may not be evident how to understand them.
A: Her interpretation holds an intermediate position between Bohr and Everret's relativive ontology:

*

*Bohr relativises with respect to a classical measuring apparatus


*whilst Everett insists the entire universe follows the deterministic quantum evolutionary law without any non-deterministic evolution as usually given by the measurement postulate. Hence his theory is radically relative.
Hermann does not go as far as Everett, but she goes further than Bohr. She keeps the measurement postulate and says the quantum is relative not only with respect to the classical measuring apparatus but also the 'concrete particular' of the outcome.
She also points out in this essay that QM does not overthrow causality. In fact, her analysis was concerned with the Kantian notion of causality which she affirmed still held despite the changes that QM required to our understanding of nature. For her, classical thought (ie after Newton) assumed predictivity with causality. With the advent of the quantum, predictivity is abandoned but causality still holds. She also notes that QM is remains retrodictive and hence we have causal retrodictivity.
A relational ontology of QM is still viable, and Rovelli outlines one in the SEP (Stanford Enc. of Philosophy).
It's worth adding the focus on phenomena being fundamentally deterministic is after Newton and before the indeterministic physics of QM (or even after, as one can argue that Everrets ontology is deterministic - at a huge price). However, Aristotle notes that some natural philosophers asked "is chance a cause?", showing that some natural philosophers were open to nature being indeterministic in effect/cause until the success of Newtonian physics foreclosed that option.
