It is too late for a chronology of what turned up when, so maybe by age of article, and in reverse:
In January this year Kathryn Greenhill reiterated a recognition that Scribd-like services have got so much right, while specifying why Scribd (and one guesses any for-profit cloud system) ought not be relied upon - this warning was in reply to Brian Kelly's enthusiasm for how Scribd has enhanced access to papers.
No, if I try to make coherence it is going to take too long. So, a list of items that contain points of interest I'll just have to get back to, unless someone else can point me to a summary of all issues?
In July 2009 Kerim Friedman asked whether Scribd would serve as an Edupunk repository.
Christopher Kelty weighed in there with:
"archival persistence? How would these tools allow for permanent findability and a certain sense that one can be sure it will stay available for a long time? DOI numbers require an institutional home... COiNs data are easy to add to a blog post... Zotero can find things with this data... so maybe part of the blog post should be best practices for eduPunk future-proofing... "
Archival persistence and ownership issues seemed to be the major argument of commenters against Joseph Esposito's June 2009 proposition that libraries "should begin to close their IRs" to save money, in favour of Scribd served repository.
Also in June 2009, and same venue, Michael Clarke described the profit-potential of Scribd for publishers - might it also serve institutions in the same way?
Interesting that many articles about Institutional Repositories are shared by writers through Scribd.
Before I leave this sidepath, I must also keep this link to all of Brian Kelly's interesting discussions of Institutional Repositories, top of which (at present) is an article about measuring the effectiveness of institutional repositories.