I agree arXiv and Sci-Hub are useful. In the case of arXiv it does indeed cost money to run but not much as they are taking in PDFs and storing them without rev

I agree arXiv and Sci-Hub are useful. In the case of arXiv it does indeed cost money to run but not much as they are taking in PDFs and storing them without review, copy editing or curation. And that is fine. But arXiv has not replaced journals still. Getting work disseminated has not been an issue since the dawn of the internet. And that is not expensive either. The service Elsevier provides albeit at a high margin and predatory business tactics is more than just archiving and displaying a PDF. And Sci-Hub is stealing the extra value they provide by curation, and peer review facilitation. We might agree with it since they aren't a nice company but the costs of copy editing, curating and reviewing those papers is not being done by Sci-Hub they are just doing the cheapest part of the process by being a repository for paywalled papers. They aren't being more efficient publishers by getting the value add part for free. Someone has to pay for that part. And if Sci-Hub just hosted non pirated work that was not reviewed or copy-edited it would be just as laudable as arXiv.