Comments by "Jeff Huffman" (@tejing2001) on "Parallel Worlds Probably Exist. Here’s Why" video.

  1. Sorry for the wall of text, I've had this on the mind for a while and just had to say it. Why postulate many worlds at all, when you can think in terms of knowledge, probability, and correlation? The density operator (wave functions only represent pure states, and thus aren't a sufficiently complete description in my view) represents your knowledge about a system, much like a classical probability distribution (and in fact diagonal density operators are classical probability distributions over the basis states. Also, by choosing an appropriate state basis, any density operator can be made diagonal). The difference from classical probability distributions is that there is no assumption that there are any density operators that make every property of the system definite simultaneously (this is the key idea, take some time to think it over), which also means it's no longer possible to do away with the probability stuff and deal with a single fully definite state instead, since those don't exist. You still get determinism though, in the sense that no information is created (or destroyed, for that matter) through time evolution. Entanglement is just correlation (so no action at a distance), superposition is a result of your knowledge about the system not being applicable to the question at hand (rather than your lack of knowledge, like classical uncertainty), wave function collapse is the change to your knowledge that results from interaction (much like the classical probability distribution for what hand of cards you were just dealt "collapses" when you look at the hand). It's also worth noting that both decoherence and wave function collapse are a result of interaction of the system with things that aren't being modeled, namely the environment and yourself, respectively, so the apparent creation/destruction of information in how we think about those processes isn't real (no "dice" involved, even in measurement). The No Cloning theorem just says that when you run an object through a perfect copier, the outputs will be perfectly correlated (nawww, really? copiers do that?) (and you'll need raw materials about which you have maximal information in order to make the copy too). The No Communication theorem is no surprise since nothing about this whole perspective involves action at a distance. And so on. Once you accept that there are no probability distributions that make every property of a system definite, everything else falls into place quite intuitively. You lose the expectation of definiteness of reality, sometimes called "realism" (that your questions have definite answers, even though you don't know them), but this is a much smaller sacrifice than it might at first seem, since you still have the expectation of coherence of reality (that if you and I have incompatible knowledge about reality, one of us is wrong).
    1