What bugs me about this article is that we have 'half silvered mirrors'. By definition they divert half and allow half through. Like the one at 'A'. But then suddenly, with the one at 'D' we get "And what D does to a photon, depends on the angle at which the photon arrives" - so not a half silvered mirror, but something else, with no explanation of how or why the angle affects the outcome.
As a layperson whose understanding changed from billiard balls to waves to probabilities I suspect there is no 'reality' that everything can be reduced to - and I certainly don't think 'amplitudes of configurations' will be it. Even if the description is useful, they do not actually exist, just as billiard balls and the rest are just useful-at-times descriptions.