Papers with a large number of coauthors

This is a matter of opinion, but whatever the policy is, it should be consistently applied. If D.H.J. Polymath is allowed to post papers without listing the members, then other collaborations should be allowed to do so as well. D.H.J. Polymath shouldn't get special treatment just because it is famous.

I'm gradually becoming convinced that the mathematical community as a whole needs to pay more attention to arXiv policies. Seemingly small questions of policy can have a huge impact on mathematical research, simply because such a huge fraction of all mathematical research gets posted to the arXiv—far more than is submitted to any single mathematics journal. There is a myth that there are no barriers to publishing on the arXiv, but this is certainly not true. Figuring out the right policies is not a trivial problem; the founders of viXra stated some legitimate concerns about the arXiv, but I believe that viXra has failed to solve the problems it was intended to solve. Perhaps the American Mathematical Society and other professional societies should investigate some of the concerns and complaints about the arXiv that are being raised by mathematical researchers, and try to change some things about the arXiv if doing so would serve the community better.