Is there every a point that a photon is not a particle? Sure it is wibbly-wobbly and can have all sorts of wave-function state but a change of state doesn't change the nature of what the particle is.
It never was a particle. See later.
In the video the mysticism starts when the explanation for the wave function collapses was attributed a now choice(made by the experimentor) affecting the choice? the photon made after it passed the slits. Not being able to observe the photon in a wave state is problematic. The photon is scared to get caught with it's pants down?
Wavefunction collapse isn't really a coherent thing, in terms of logic. It's the result of trying to hang on to the macroscopic (and rather comforting) notion of there being only one world where things happen in a linear order. If you instead work from the logic of quantum mechanics upwards to the macroscopic, you wind up with multiple-world QM.
What the wavefunction actually describes is the probability that the quantum thing interacts at that particular place. (Except the values themselves aren't probabilities, and so add together strangely. You have to do a little simplification to get the probabilities out.) In some cases, this'll be near 1, and the quantum thing will behave very like a localized particle. In others, it'll be much smaller, and consequently more spread out, and behave very much like a wave. (It'll be spread out because the entire probability-volume of the wavefunction has to add to 1; the particle has to be
somewhere.)
In Copenhagen QM, we get some ontological wibbliness because the quantum thing has to choose from these probabilities "randomly" (in a philosophically ill-defined way) and then it interacts there, and we carry on our merry, classical way as though it was a Newtonian particle all along.
In MWQM, on the other hand, there is no choosing. Instead, the future continues exactly as the wavefunction did. Now, that's quite tricky to think of, so here's a concrete example: in the double-slit experiment, the wavefunction shows a variety of places the electron is likely to land on the screen. In MW, all of these are in some way "real" - and consequently, there is a variety of
you, all of whom see the electron land somewhere slightly different. Then, all of the different yous react, and all of them think they saw the electron land in only one place, and so all of them think the wavefunction has collapsed.
But really, they've just been entangled in it, and turned into a superposition themselves.
(About the linear order thing: in quantum mechanics, you can do the maths by selecting a starting position, selecting an ending position, and calculating the probability that your start will turn into your end. One issue this produces is that you can swap your beginning and end and normally end up with the same value, and they can be anything you like. So transitioning between
any two states of the universe has some non-zero probability. Including completely nonsensical ones.)