In Physics and many other fields Arxiv.org has basically replaced journals as the primary way of getting the latest research results. Unfortunately many high-impact-factor journals like Nature forbid you to publish a preprint of your article there (at least that was the case in 2012) so that they can defend their claim of publishing "cutting-edge" results only (which due to their long review processes are often 12 months old by the time they're actually published).
The most ridiculous thing is that the journals today do little more than act as an intermediary between the authors of a paper and the reviewers, which work for free and are colleagues from the same field. That and maybe styling the papers a bit, which is (IMHO) unnecessary in most cases as the source LaTeX version usually provides very good design already.