The quantum eraser experiment tells us that a photon shot at two slits is a wave, unless you measure which slit is taken and you do not destroy the measurement result.
I've found this very similar to the notion of 'lazy evaluation' in computer science. Only evaluate when it is certain the result is required.
Konrad Zuse's theory for 'rechnender raum' or a computing universe was always just a theory.
But could the behaviour of photons not be described as the 'lazy evaluation' of a computed universe. If the 'which slit information' is not required, the universe never bothers to compute this information. Somehow, the computed universe is optimized?