I'd love a not-more-often-than-quarterly (preferably annual) summary of important science (natural and social) papers, ideally including mathematics. I looked for something like this a while back, or for some kind of meta-journal that selected the best (or at least most important) from all sources, but couldn't find such a thing.