We can not ignore that signals were initially analog
Of course not.
Although up sampling does not have more information, practical implementations in real world can sound better than no up sampling.
Yes, but in that case the upsampling is used as a technique in the digital-to-analog conversion process. The upsampled material as such is not in any way "better" than the original, un-upsampled material.
But only listening will tell you which gets better results.
Absolutely - as long as account is taken of perceptual bias and subjective preference factors. Just because something sounds better to *you* doesn't mean it is better in an absolute, universal sense.
And although opinions are just opinions, we weight opinions according to people experience and shown public evidence of the quality of their work.
IMHO it is impossible to separate quality of a format from implementation. But when we have some systematic in the opinions emitted by several people we can trust we consider this opinion as plausible and it becomes accepted by many people. Surely skeptics can have different views, but unless they present something new, other than the usual "what matters most is the recording" and "all formats have quality enough for me" I do not loose my time with it. YMMV.
How you spend your time is of course your own decision, but you seem to present a world view where everything is just "opinions", but some opinions carry more weight than others based on who is presenting that opinion. It is, unfortunately, a very common world view these days. My background is mostly in engineering and applied science, where opinions don't matter much unless backed by verifiable, repeatable evidence. I don't care how famous or revered some person is, if he or she is making some claim, he/she better be able to present evidence to support the claim.