If you’re sufficiently paranoid, the only thing you can say for certain is that there isn’t much you can truly trust.
There's (mostly) no reason for Joe Bloggs to want to protect themselves from a well funded state actor but they may want to prevent themselves being locked out of Google services because some bot triggered a threshold.
What's permitted is to build something that targets a particular person in such a way that it cannot possibly affect another person's security.
The example I use (though IANAL) is that a request to backdoor WhatsApp's encryption would not be permitted under the law. However I think that pushing an update that checks for a particular person's hard-coded phone number and forwards messages to law enforcement would be permitted.
The law in question: http://www5.austlii.edu.au/au/legis/cth/consol_act/ta1997214...
It doesn't matter what service you use. No email is private, I don't care who they are.
p.s. - I'm a paid fastmail user.
I wonder if the Monero Message System is still being developed. It would solve the metadata problem.
Russia is also famous for its cadre of crack Salisbury cathedral appreciators.
This was not always the case. It might not be the case now that Russian companies scan for this stuff, but why risk it?