Stefan - I don't think anyone has said anything contrary to that.
However - it does not quite get at the nub of the matter - which is the spooky way that non-locality surfaces. True, it (provably) can't be used for FTL signalling. But equally, it does exist.
The way to think of this is that two entangled photons - even if 1000s of light-years apart - behave as a single object when measured. Complementarity then means that the apparently random results from measurements at far distant locations can be correlated.
This spookiness is subtle and emerges directly from the maths of QM. It shows something deep about the structure of the universe - that non-local things are everywhere. It is consistent with the modern (Raamsdorf et al) ideas that the structure of spacetime (and the speed of light as an information propagation limit) emerge from entanglement as a fundamental process.
So we have here a 70 years old issue that tantalisingly hints at something new.
How you interpret the spookiness is up for grabs: the different interpretations cannot be experimentally resolved (there might be some possibles, with some interpretations, but none to my knowledge that have panned out).
There is a meta-observation - which many don't like - that might push you towards a many-world type answer. That is if it turns out vanishingly unlikely that the universe could evolve to generate complex structures and life then the anthropic principle makes that quite OK as long as we have many worlds.
On the issue of wavefunction collapse: it is non-measurable, because decoherence produces the same results. The fact that measured microscopic systems can be processed coherently in a way that restored the "multiple results" speaks against any form of waveform collapse, because there is in principle noting to stop complex processing that restores coherence after a measurement. Thus I don't like anything that requires collapse. But pilot wave etc ideas can be compatible with a many worlds no collapse interpretation.
It is a mark against Mills that he seems not to have engaged with this whole issue: the physical world certainly does engage with it!
It's spooky if you interpret it as such. Read the quote again, you can have an interpretation, and it is an acceptable one, of QM, where it is no effect at all, no spookiness with more that there is a correlation that you get from initiating the photons at the source in a certain way. You don't need some kind of communication between the photons at the measurements. Kind of strange that you dismiss LENR and Mills, but still prefer a spooky version of QM before an acceptable interpretation where there is no spooky action at distance. You don't need a many world interpretation. In a probabilitic interpretation, something that is natural you will find that you need non locality to explain what you see due to Bell. But again if you view QM as something unique of it's own e.g. just another way of generating the expectations than taking a probabilistic mean, you can satisfy Bell and have a locality view. What's interesting is that Mills shows another system that is not a probabilistic one
which is not a local hidden variable theory.