I'm sure this has been posted already in the massive 146-page backlog of this thread, but for me one of the more informative posts about this was this page here, where you can listen to a sample of a song with a good USB cable and a bad USB cable:
https://archimago.blogspot.com/2014/01/demo-measurements-what-does-bad-usb-or.html
At least for the digital stuff, the signal either makes it through properly or it doesn't, and when things go wrong, the changes are very nontrivial (pops, static-sound, silent parts from bad decoding, etc), and can't change overall "tone" or color of the song (in order to do that, you'd have to make very consistent changes to a variety of bits which is no longer error but something so specific that the probability is near-zero). To my mind, a cable is either good enough to transmit the signal, or it isn't - in other words you either get the correct sound or you get garbage. You don't get a "better" sound somehow past the bare minimum, is the way I see it.
I think it doesn't help either that you can look at pictures of musicians/DJs/etc at big concert venues who can be seen using very high-end cables - but I think if anything, in a big environment like that (where you may have rain, or a huge crowd with lots of cell phones, lots of equipment, large distances between components, etc), at least there it would seem more defensible to me to get something with beefier shielding, or spring for the gold-plating to resist corrosion better, or the better tech for transmitting a clear signal over longer distances that is less prone to integrity issues, whatever.
But for the average joe with a modest in-home setup with things mere feet apart, it's overkill -- and not even in a "diminishing returns" sense, but just straight overkill, zero added benefit.
I suspect that the placebo effect is just extraordinarily powerful. At least when I've looked around on Google, most of the objections tend to center on AB/blind testing as flawed, but I have not really heard a good reason behind that defense. If I can't tell something apart even with the ability to move seamlessly between them, as far as I'm concerned they're perceptually identical, and I don't see why this is a flawed way to approach it.