I was fortunate enough to have coffee on Sunday morning at TAM London with David Deutsch. We talked briefly about QM. Now I know some of his theories are contentious but his work is built on Everett's work which Deutsch and his supporters deduce to be essential. It completely redefines what Probability means.
Yes, I agree with that.
According to them the Copenhagen Interpretation's use of 'Probability' is logically flawed.
I wouldn't go that far. It's not on a firm footing, it's probably incomplete, but I wouldn't say it's "logically flawed". After all, it works extremely well, so it's at least almost right. It's pretty rare that something with flawed logic at its root works that well.
I no longer know which interpretation I prefer. I think MW interpretations have many things going for them but the sheer impracticality of creating an entire set of universes for every quantum state for the entirety of time seems very very wrong to me.
Many people have that reaction. But think about the enormous complexity inherent in
all macroscopic physical systems, even classical ones. A classically chaotic system is completely impossible to simulate on any computer beyond a certain point, because the precision required grows exponentially with time step. The same goes for vanilla QM (irrespective of the issue of measurement). For example, it's possible to simulate a QM computer on a standard classical computer, but the number of classical operations required grows exponentially with the number of qbits the QM computer has (that's a basic fact about QM that's independent of MW versus CI).
All large systems are incredibly complicated when you study them in detail, and MW only differs from the CI when macroscopic systems (measuring devices) are involved. So it's really not clear at all that MW is one bit more complex than Copenhagen (and in fact it's impossible to be sure, since Copenhagen never actually defines measurement or says what actually happens during one).
With a few hints from Sol, I've begun to think of the universe as a complex probability wave function, and when it 'splits', it gains a new set of probability density ripples corresponding to each of the outcomes - including superposed representations of us as observers.
Dropping pebbles in a 4D (nD?) pond...
Yes, not a bad way to say it. And really, the pond contains all its own pebbles and their dynamics, because at least in principle the entire universe should have a wavefunction which simply evolves according to the total Hamiltonian. Such a concept is incomprehensible using CI, because measurement is some kind of vaguely described deus ex machina.
