So what is the value of a paper journal over a website? That the pictures look better?
I love how you compare a scientific journal to a cinema. It really brings your point home. But you're wrong: Science isn't entertainment. (I have experienced both NPG and AAAS treat it as if it was, though.)
Sci-Hub and arXiv are merely distributors. Like a movie theatre for a movie. Elsevier is like the movie studio that actually produces the movie after authors write it. The question of marginal value is missing that point. It's looking at the fact that, under this analogy, Elsevier is the movie theatre and is blind to the fact that it is also the movie studio.
One way or another the work has to get done. Shifting it from one pile or person to another doesn't change that. You can live without some of the steps I outlined, but quality can suffer then. Not in all cases, but some. I've seen plenty of HN links to arXiv posts where it's clear the authors had little sense of the best way of presenting their material and therefore have a harder time communicating their work.
What marginal value do you place on having new scientific knowledge more clearly and consistently communicated to other researchers? That's what publishers do.
Subscription prices are certainly way too high compared to the cost, but saying the add little or no value just tells me that a person probably doesn't understand what publishers do, or that "researcher" skill sets don't always overlap with the skill sets to perform the work I outlined above.
Good points. I'm not implying that publishers add no value; I agree that they add some (but charge far too much). I'm just trying to find a way to determine the values of their different inputs.
Then they did not articulate that viewpoint very well. They also completely neglected to provide any justification. I stated a claim and provided extensive detail to support it. Replying with a comment that boils down to "nope" doesn't add anything to a conversation.
Is the cost added by the journals justified by the ultimate difference in output? It ultimately depends on your goals, probably. In my opinion, a good competing system would more or less completely eschew physical copies and save that process for anyone who is willing to take on the costs associated with that. And ideally, the peer review process would be independent from publishing, since that seems to be the invaluable service that publishing companies provide.