

Nah. They try to defend “value indefiniteness,” a position which is largely indefensible but has become very popular in academia despite having been all but ruled out, and should go the way of the aether. A physical theory needs observables with well-defined conditions under which they obtain well-defined properties or else it is not possible to make a single empirical prediction with the theory.
If you claim particles do not obtain definite values for their observables at all, this is the Everettian interpretation, the “Many Worlds” interpretation. But this trivially doesn’t work because, again, if particles never obtain definite values at all, then you cannot make a single empirical predictions with it. The theory has no connection to empirical reality. See Tim Maudlin’s paper “Can the World be Only Wavefunction?”
Indeed, whenever we have a statistical distribution, we presuppose that there exists an underlying real state of the system, but we are just ignorant of it. In order to derive the Born rule distribution, then Everettian mechanics must, at some point, admit to there being definite values to the observables. But to do that would be equivalent to a “collapse” approach, which they want to avoid, so they try to use arguments involving decision theory and such, but, as Adrian Kent showed in his paper “One world versus many: the inadequacy of Everettian accounts of evolution, probability, and scientific confirmation,” these explanations are always circular, as well as the paper “Epistemic Separability and Everettian Branches: A Critique of Sebens and Carroll” by R. Dawid and S. Friederich.
Indeed, Everettians also love to claim that their view is “local,” but if their viewpoint really is mathematically consistent with quantum mechanics, then, at some point, it must reproduce Born rule probabilities, meaning it must reproduce violations of Bell inequalities, and so it cannot be local, as shown in Aurélien Drezet’s paper "An Elementary Proof That Everett’s Quantum Multiverse Is Nonlocal: Bell-Locality and Branch-Symmetry in the Many-Worlds Interpretation". They often get around this by just redefining locality to be in terms of linearity or no-signaling, but any interpretation can be local if we just change the meaning of locality.
Of course, there are also “collapse” interpretations. The collapse, obviously, cannot just occur “when you look at it,” or else you end up devolving into crackpot solipsism, as per the “Wigner’s friend” thought experiment in Wigner’s “Remarks on the Mind Body Problem.” The “collapse” must occur before then, and it also must be an invariant collapse, or else the minds of other observers would depend upon how you personally look at them, and their own minds would not have independent existence. “Collapse” thus can only be a consistent view of physical reality if the collapse both occurs under well-defined conditions and is invariant.
But these are pretty much ruled out by John Bell’s paper “Against ‘Measurement’” which points out that the “collapse” approach cannot constitute a physical interpretation of quantum mechanics because orthodox quantum mechanics does not tell you when this “collapse” should occur, under what well-defined conditions, and so it does not give you an unambiguous ontology.
If this “collapse” really occurs, it is a non-reversible process, yet all unitary evolution operators are reversible. That means if I build a measuring device, and you give a complete physical description of the measuring device in terms of quantum mechanics, then an interaction with that measuring device would be described via unitary operators, and thus would be reversible, and so orthodox quantum mechanics would predict that an interaction with the measuring device is reversible, whereas a “collapse” approach would not.
This would in principle lead to different empirical predictions, as we would have something interact with a measuring device and then attempt to reverse the interaction, and the predictions between a “collapse” theory and orthodox quantum mechanics would deviate from one another. Theories like GRW and the Diosi-Penrose model are thus separate theories, not interpretations of the same theory. A physical collapse model can only be consistent if you believe orthodox quantum mechanics is simply wrong.
The “measurement problem” within orthodox quantum mechanics stems from the assumption of value indefiniteness. Nobody has proved it is possible to make quantum mechanics consistent with value indefiniteness without running into the measurement problem, and it is my position that it is not logically possible to do so. The measurement problem is a proof-by-contradiction that value indefiniteness is just an untenable position, an outdated position that has largely been ruled out but people still cling to their outdated ways due to preconceptions.
In MinutePhysics’ video, he does not defend any of the absurdities of the worldview he is proposing. He just attacks the alternative because it would have to not be spatiotemporal and calls that “crazy.” That’s not an argument, that’s an appeal to incredulity. There is no law of logic that says nature must necessarily be interpreted as spatiotemporal. If my two options are (1) believe the empirical evidence that demonstrates reality is spatiotemporal, or (2) adopt an entirely incoherent position that descends into irreconcilable contradictions in order to delude myself into believing it is spatiotemporal, then I am going to pick #1 every time, and saying it is “crazy” is a bit of an absurdity.
Indeed what is even more ridiculous about MinutePhysics’ and 3blue1brown’s dismissal of such a view as “crazy” is that it was also the view of John Bell, the guy who developed the theorem. Apparently, John Bell was “crazy,” yet, it was Bell who had the deep insight into the theory in order to develop this theorem in the first place, and so clearly he understood it better than a couple of YouTubers.










The Harvard physicist Jacob Barandes proved that quantum mechanics is mathematically equivalent to a statistical theory with history dependence in his paper here. There is thus always a rather simple and intuitive explanation for most quantum mechanical phenomena without resorting to things like multiverses or collapsing wavefunctions, but that it is just a statistical theory + history dependence.
I will use this simulator to illustrate: https://ophysics.com/l3.html
Rotating only the first 90 degrees blocks the light.
Rotating only the second 90 degrees blocks the light.
Rotating both 90 individually blocks the light.
Rotating the first at 45 degrees and the second at 90 degrees allows light to pass through.
Rotating the first at 90 degrees and the second at 45 degrees does NOT allow light to pass through.
To say that there is history dependence means the behavior of the particles is a function over its history, and so the behavior of the particle during an interaction can change if its history is different. If not all the light is blocked by the time it reaches the second one, then the behavior of the photon at the second one can be different if, in its history, it had interacted with the first one rotated at 45 degrees, and thus the second one may not block all the light as we would normally expect it to because you have changed the history of the particle.
In classical statistics, you can represent the statistical outcome of an interaction according to p(t)=f(U(t),p(t-1)) whereby U is the definition of the operator describing the interaction and p is the probability distribution. Mathematically, you can decompose the quantum state into two real-valued vectors according to its two degrees of freedom where one is a probability vector, common in classical statistics, and the other is the phase, and the way the probability vector evolves with an interaction can be defined by p(t)=f(U(t),p(t-1),h(t-1)) where h is the phase.
The phase can be interpreted as a sufficient statistic over the system’s historical trajectory, because you can get rid of the phase entirely by expanding out h(t-1) over its whole history such that you get:
p(t)=f(U(t),p(t-1),g(U(t-1),p(t-2),g(U(t-2),p(t-3),…)))
This extension would stop at a base case. It is quite trivial to prove that a degenerate distribution would be a base case, and so if there is a point in the system’s history where you know its value with certainty, then you can stop the expansion there, and thus the phase disappears from the evolution rule. Of course, that doesn’t mean you shouldn’t use the phase mathematically, you just don’t interpret it as a physical entity, but as a sufficient statistic over the system’s statistical history.
You can think of each physical interaction like a black box the particle enters, the box alters its state, and then it leaves the black box. The black box is described by a function, which takes the particle’s pre-interaction state as an input, and then outputs its post-interaction state. You can conceive of quantum mechanics as just a form of statistical mechanics whereby the statistical behavior of the particle is a function not just of the particle’s statistical state right now and the current operator, but of all its previous statistical states and previous operators.
Each black box is, in a sense, “aware” of the whole history of the particle which enters it. The second polarizer “knows” that the particle had just interacted with the previous one at a 45 degree angle, so it changes its behavior accordingly. If we express it like this:
p(t)=f(U(t),p(t-1),g(U(t-1),p(t-2)))
Then the behavior of the second polarizer given by this function can change if U(t-1), the previous polarizer’s orientation, is changed.