The first question posed by our leader librariems is one which occurred to me as well. As pointed out there are now a number of large-scale digital collections available and being built, HathiTrust being a main one. I believe the main advantage to HathiTrust is one they advocate as well: non-profit/non-corporate. Google being the main example here, there is no guarantee they will be around forever. And even if they are here to stay companies change their focus. Google is not a library, if they one day decide to stop supporting GoogleBooks no one can fault them for taking it offline. HathiTrust as a non-profit has a huge advantage over profit driven digital libraries.
HathiTrust also has the advantage of being a group of of many different libraries with records stored at multiple locations. This provides a level of safety for the whole system and a means for continuance. Schools can join and leave the system but HathiTrust still remains because is it not solely dependent on one entity.
When it comes to the DPLA and HathiTrust I might be too positive, but I do not see a need for much infighting or overlap. I view DPLA as an extension of our national public libraries. The most difficult aspect of HathiTrust is the large number of items under copyright while I think DPLA could focus on items within the public sector.
As part of this discussion on overlap I found the article by York to be most interesting. It was more focused on overlap of materials between libraries and how HathiTrust could affect collection development. Over the next ten years or so as more schools join HathiTrust and more items are placed in the system it could truly begin to change how libraries develop collections and make budget decisions.