I truly hope this is a successful project. If it is, other areas of academia/professionals might take note and adopt this system.
I wonder why they limited it only to mathematics? Are there other free publishers out there for the areas? I cant imagine what the publishers reciprocation might be.
It replaces the preprint servers every department ran/runs, but it's not like a journal in that papers there are in various states of revision rather than having already gone through peer review (and, as a corollary, there are sometimes cranks). Browsing it is akin to browsing github, you can see all the cool stuff your peers are working on but you want to treat them with caution. The intent of this project seems to be to organize the review process like an ordinary journal does, then link to papers that are ready for prime time. It's a really obvious step, since we all have to do our own typesetting anyway (thanks for that, Knuth) and very few people read new papers in print journals anymore.