Like humans, manuscripts are supposed to be monogamous. That is: one manuscript, one journal. However, also like humans this is not always the case and perhaps more reflective of traditions then reality. Authors submit their manuscript serially to numerous publishers, a process satirized in the short, "down the impact factor ladder."
This funny "mating" dance scholars play with journals is in the intent of publishing with the "best journal." Some might say, it's Darwinian-- the authors do this for their survival (i.e. they get grants). Authorea makes this mating dance easier and more efficient as we recognize it won't disappear over night. You can easily shift between 8,000 different publisher styles-- no more rewriting and revising to go to another journal.
Beyond typically serially submitting our manuscripts to numerous journals, authors have adopted another form of publication, so-called "preprints." Preprints are un-reviewed versions of a manuscript and they appear on sites like Authorea, arXiv, bioRxiv, SSRN, and others. The rise of preprints in new disciplines and dominance of them in certain communities has led some to question the way we treat our manuscripts. To continue the poor metaphor, are preprints simply dating before finding the one? The Ingelfinger rule
would argue that they are one in the same, and many journals will not publish a preprint. While this may boost the exclusivity of a journal, it slows down communication of results to the scholarly community as well as to the public.
Similar to preprints, depositing a manuscript into a repository is generally allowed but it must not be the version of record. Again, a policy that is aimed to push eyeballs to the main journal--a place where the content typically resides behind an expensive paywall.
So, what do we propose? Publish your work openly, share it openly, communicate your ideas to the world how you think they will best be received. Don't worry about how it has always been done; there are better ways now.