Can you point me towards some of the best comments?
There’s some around this thread (responses to Luke’s comment). Also I think that QM sequence has responses from physicists.
What do you mean by ‘fragility over’ potential theories of everything?
The MWI is concluded from exactly linear quantum mechanics. Because we know QM to be only an approximation, we lack any strong reasons to expect exact linearity in the final TOE. Furthermore even though exact linearity is arguably favoured by the Occam’s razor over any purely speculative non-linear theory, that does not imply that it is more probable than all of the nonlinear theories together (which would have same linear approximation).
In my opinion, things like multitude of potential worlds allow for e.g. elegantly (and compactly) expressing some conservation laws as survivor bias (via some sort of instability destroying observers in the world where said laws do not hold). Whenever that is significant to TOEs is, of course, purely speculative.
Whats the standard response by MW enthusiasts to your point on Solomonof induction?
As far as I know, the arguments that Solomonoff induction supports MWI never progressed beyond mere allusions to such support.
My understanding would then suggest that neither MW nor Copenhagen can give an exact picture of photon noise
In raw form, yes, neither interpretation fits and it’s unclear how to compare complexities of them formally.
in which case the problem would seem to be with Solomonoff induction as a formalization.
I explored some on how S.I. would work on data from quantum experiments here. Basically, the task is to represent said photon noise with the minimum amount of code and data, which can be done in two steps by calculating probabilities as per QM and Born rule, and using the probability density function to decode photon coordinates from the subsequent input bits. (analogous to collapse), or perhaps more compactly in one step by doing QM with some sort of very clever bit manipulation on strings of random noise as to obtain desired probability distribution in the end.
There’s some around this thread (responses to Luke’s comment). Also I think that QM sequence has responses from physicists.
The MWI is concluded from exactly linear quantum mechanics. Because we know QM to be only an approximation, we lack any strong reasons to expect exact linearity in the final TOE. Furthermore even though exact linearity is arguably favoured by the Occam’s razor over any purely speculative non-linear theory, that does not imply that it is more probable than all of the nonlinear theories together (which would have same linear approximation).
In my opinion, things like multitude of potential worlds allow for e.g. elegantly (and compactly) expressing some conservation laws as survivor bias (via some sort of instability destroying observers in the world where said laws do not hold). Whenever that is significant to TOEs is, of course, purely speculative.
As far as I know, the arguments that Solomonoff induction supports MWI never progressed beyond mere allusions to such support.
In raw form, yes, neither interpretation fits and it’s unclear how to compare complexities of them formally.
I explored some on how S.I. would work on data from quantum experiments here. Basically, the task is to represent said photon noise with the minimum amount of code and data, which can be done in two steps by calculating probabilities as per QM and Born rule, and using the probability density function to decode photon coordinates from the subsequent input bits. (analogous to collapse), or perhaps more compactly in one step by doing QM with some sort of very clever bit manipulation on strings of random noise as to obtain desired probability distribution in the end.