science also is telling us that reality is a strange miasma of superpositions and that we actively participate in the creation of reality by simply existing/observing.

It doesn't tell us that at all. This is just bizarre metaphysics invented out of someone's ass one day and became popular among academics, despite it having no empirical basis for it and not even being logically consistent if you take it seriously for more than five seconds.

Quantum mechanics is just a statistical theory. You literally superimpose states in classical statistical mechanics as well. The only difference is quantum mechanics has an extra degree of freedom in the state description of the system that includes phases, and those phases evolve deterministically and influence the stochastic dynamics of the system. This gives a kind of "memory" effect whereby the same operator can have different behavior if the history is different, such as, a photon having 50%/50% chance of being reflected/transmitted by a beam splitter, unless its immediate previous interaction was of a beam splitter as well, then it is 100%/0% because the state of the phases are different.

No, Sean Carroll is just wrong and he presents nothing to justify his position. The cat doesn't stop existing when you're not looking, nor is there is a multiverse, nor do things spread out as infinite-dimensional vectors in configuration space when you aren't looking. You just do not know its state because it is statistical as quantum mechanics is a statistical theory. Multiverse believers love to put their idea side-by-side another idea which is even more absurd in order to make it look more viable, but they never bother to defend their ideas on their own merit, without a comparison. Any time you ever encounter a multiverse believer, they will constantly bring up Copenhagen even if you never mention it.

Carroll responds to a variant of Copenhagen that believes in a "spreading out" axiom that things diverge into a multiverse of every possibility represented by a vector in configuration space when you aren't looking, but then suddenly "collapses" back down into a definite configuration in state space when you look. He then attacks the "collapse" as silly, and therefore we should believe things spread out as a multiverse forever. But nowhere does he ever give any convincing justification for the "spreading out" axiom to begin with. That axiom is not grounded in any empirical evidence or in the mathematics at all, and so multiverse believers can only make their position look coherent by putting it beside another silly belief which also presupposes that axiom, and thus they make it appear reasonable that they never justify it.

Just look at the awful slide 24:35. Someone can make this same argument in a perfectly classical universe. If we could not track the definite states of particles because they behaved randomly, but in a classical sense which did not violate Bell inequalities, we would also only be able to track the states of systems as vectors evolved by matrices. Someone could also come along and claim that particles do not have real states when you are not looking at them because they are not there in the mathematics, and that they are being the "reasonable" one for believing that the universe just evolves as a big deterministic vector.

We would all look at them as if they are silly. Yet, somehow, this is stated unironically among multiverse believers as if it is somehow made less silly by quantum mechanics, when absolutely nothing in the theory makes this a less silly position.