That helps, but it doesn't mask the data groups like the NSA are really after (who is talking to who, and when).
The way in which it will work is having all users of the system constantly send encrypted streams to each other ( mostly gibberish ). When you want to send a real message you simply replace the encrypted gibberish with some real content.
It will implement "onion" style repeated encryption to prevent direct links from being statistically analyzed as well.
Look for it in 6 months or so. Details will be up at circleofdistrust.com in about a month or so.
Essentially, set up a new block-chain that allows transactions to contain 'large' messages (IE. large enough for long plaintext, but probably not enough to embed videos). Then, when you want to send a message, encrypt it using the intended recipients public key, then publish the cipher-text to the blockchain.
Edit: OTOH, this this does just mean they attack one or both nodes if they care when content is being sent.
Dear NSA: I can and will be designing a fully anonymous open source distributed messaging system that is entirely secure even if you tap every line in the world. At the very best, you will be able to accuse people just of running the software. If you don't like this, I'd advise you that your only option is to hire me and/or pay be enough money to become disinterested in finishing the creation of the software.
What if I don't trust server X to really be doing that?
What if the government contacted owner of server X and altered the software not to do that?
The system I am designing keeps anonymity without any need to trust the server. Hence the name. Don't trust anyone.
