How is Dropbox detecting duplicate files - not by name, surely? By some hash? It must be unique - how does it know it's safe to duplicate otherwise.
Which, to me, means I don't quite get the security concerns. If you've got a file that you don't want duplicating because of sensitive content, isn't that going to be a file you've created yourself, therefore with a unique hash. So, it won't be duplicated.
The only things duplicated are common files. Ones that won't have been edited from their original source.
(I use Dropbox so I may just be kidding myself and not seeing the bigger picture.)