This is a very cool and exciting idea but my feeling is that email clients are hard to code because they're full desktop applications. Even standard email clients are hard to get right - building a social network version (with encryption) would be quite a challenge.
The rewards could be nice though - such as being able to use mailinator or mixmaster for anonymity.
http://catb.org/jargon/html/Z/Zawinskis-Law.html
Gmail is an email client. Facebook is now, too. So are the default 'mail' applications for Android and iOS. And let's not leave out Emacs :)
The challenge of email (which comes with its advantage of being ubiquitous today) is its plethora of standards documents one must read and respect if one wants to make a serious go of developing software that will work well with most of its corner cases.
But if you disagree, please write one - I'd be a very willing beta tester and would even be keen to help in a limited way.
Also, I think it's implicit that you trust your server. It's other people's servers you've got to worry about.
Email did this badly from the start, because it started in a much more trusting world of people who did not spam one another. Authentication features were added later, but there was no requirement to use them, which limits their effectiveness.
Perhaps for a project starting fresh now, this can be handled better?
Email is a distributed social network. It just doesn't define very many types of actions or objects.
You can create "messages" and then you can "send" them. The user sees the guts of the message -- the "from", "to", "subject" and "body" portions, mainly.
As far as I know, nobody has extended the standard email formats and protocols to support new message types that have caught on lately. Actions like "Post photo", "Tag photo", "Take a special action defined by this third party which should be displayed like so when rendered in a 'feed'".
This seems like mostly a matter of agreeing on some new types of messages and some suggested ways of handling and displaying them.
I'm not saying this has to be built on top of existing email protocols -- it's really just independent sites exchanging messages of some sort. But email is a good working example to point at today.
The hard part is getting adoption, and getting different implementations to interoperate. Once you dive into this, you find a huge soup of different projects working on different pieces of the puzzle. Some of these projects have a future, others do not. It's going to take some time for this to shake out, and for the developers to resolve issues with bringing this stuff together.
A lot of this work has been going on without a lot of fanfare. See for example http://indiewebcamp.com/ or http://www.internetidentityworkshop.com/
Naturally, it couldn't hurt to bring a lot more help into these efforts.