What non-trivial threat model? Requires one side to have already been compromised and for a new client to be written to not throw away the message. For any secu

What non-trivial threat model? Requires one side to have already been compromised and for a new client to be written to not throw away the message. For any secure messaging scheme or encryption protocol how do you actually protect against that? Without authentication and keeping remote party secure NOTHING will protect you (besides the threat of violence I guess). <science-fiction> There is currently no way to encrypt a message using a secure DNA-based public-key encryption that makes it so only that person can read the message. </science-fiction> You have to show it to whoever is "authorized" at some point. We hope the keys are kept safe to assure us of that.