The converse is also true -- I use the journal's screening to figure out what to read because I don't have time to read every single thing, even preliminarily. ... IMO, what will actually happen is that a free/open system is that the loss of the imprimatur of journal publication will mean increased reliance on other ways to quickly evaluate works.
I have an idea for this. I sometimes hang out on this website where lots of people are submitting things to be published. I don't feel like reading all of the submissions, but the great part is that other users can score up some of the submitted content so I can filter through the chaff and just look at the stuff that is most likely interesting or worthwhile. And I can just go to subsections of content that I'm primarily interested in. Then - and this is the cool part - every submission is debated in a moderated, open forum - where again, I can filter through the comments using a number of criteria and read only the "top rated" information, if I so choose.
I think the website is called slashdot. Have you ever checked out how it works?
Without name-brand journals, name-recognition will become even more important, which will lead to even more of the sort of "superstar" science in which funding and interest is ever more concentrated in a few research groups.
I would argue that without name-brand journals, name recognition will rely on more realistic metrics (like the quality of the work, not necessarily where stuff gets published) and therefore may be a better measure than the current system.