I'm surprised to hear that, because from my perspective (neuro), things are changing fairly fast. Bioarxiv is seven years old, and when it started, it was tiny

I'm surprised to hear that, because from my perspective (neuro), things are changing fairly fast. Bioarxiv is seven years old, and when it started, it was tiny niche service with a few dozen new papers each month. People were skeptical of the contents of preprints and journals were reluctant to accept (or cite) work posted there. Now, it gets thousands of submissions a month and is a de rigueur part of many researchers' (and some journals') workflow. In parallel, governments and funders have put some weight behind open access. An open version needs to be available within a year for all US-funded work from 2013 onward; Canada's similar tri-council policy took effect in 2016(?). It's true that there aren't draconian penalties, but there are some small nudges towards compliance. The NIH wants a PubMed Central ID for any paper you're claiming credit for on a grant. Google Scholar nags about non open-access papers on your profile. Many journals (even initially-closed ones) even handle the whole open access deposition thing for you now. As a result, I find I need the library VPN a lot less these days. On the other hand, I agree that supplanting the publishers is tough. A Cell, Science, or Nature paper still carries a ton of cachet and its associated career impact; ELife isn't quite there yet.