I think there is a simple reason why app.net did not take off: There was no significant differentiation from twitter other than paid / no ads argument.
Indeed many people including Dalton were saying that from 2007-2011ish when the Twitter API was wide open and speculation about it's potential as plumbing ran rampant. However over the last few years, it became increasing clear that the contingent of people inside Twitter who saw it as new and amazing plumbing for the next generation of applications were losing, and the contingent who saw Twitter as a media company selling advertising with a fully controller and branded user experience were the ones who won out.
Anyone building an app on top of Twitter for anything other than personal or experimental use is a fool. Unfortunately for anyone building on top of App.net, it doesn't have the users to make it broadly interesting or significantly profitable.
However the 256-character limit in the micro-blogging service and its support for embedded links (so you didn't need to use ugly URL shorteners that hid what you were about to click-through to) and images (which did not require yet another visible URL) are very nice features.
