I'm most excited about the platform, so forgive me if I rave about it for a bit.
Like Diaspora, splicious is going to be secure, distributed, and open-source. Unlike Diaspora, the back end is built on a distributed secure object platform called KVDB, so it's trivially extensible. There's a simple, direct translation from a sequence diagram into code for the platform. We also intend to use Caja in the UI to expose client-side presences of the server-side persistent objects; this means that people will be able to share stateful objects rather than just static data.
KVDB stands for "key/value database". You can think of it as a map (in the computer science sense) on steroids. Rather than merely supporting strings as keys, it uses prolog terms. To query the database, you provide another prolog term; any term in the database that unifies with the query gets returned.
Also, there's a guarantee that at a particular key you'll find either a) nothing, b) one piece of data, or c) a set of continuations. If there is no data to return, the query can get stored as a continuation; the database can also forward the request to its peers. If there is data, the query can consume it or leave it in place. By varying these options, you get messaging, pub/sub, nosql, or distributed transactions all with the same programming model.
Data and continuations at a pattern correspond directly to sending and receiving on a channel in the pi calculus, which is an object-capability-secure language. We've shown that we can write security policy---including deniability, which lies at the heart of the ocaps security model---using Caires' behavioral-spatial types for the pi calculus. We intend to port the spatial logic model checker to Scala and integrate it into the compiler. Given such a compiler plugin, we can write security policy as a type annotation and have the compiler verify the implementation.
The splicious app we've built on the platform so far has been built around maintaining your privacy; you can only be introduced to someone else through someone who knows you both, and only if you both agree. We're looking now at how to allow people to share content with people they don't know in order to grow their audience; we want to be very careful with that.
Bitcoin integration, we hope, will allow people to support their favorite content creators directly; also, by giving a cut of revenues to people who reshare, we hope to encourage wide distribution of content that's meant to be public.
There's no way to prevent resharing of content that's meant to stay private; the best that can be done is to allow the creator to note their wish about the extent they want it distributed and then notify someone if an action they're about to take violates that policy.