If your threat modelling includes Mossad, FSB and CIA, by all means do take E2E encryption and metadata into consideration every time you communicate with someone in any way.
For most of us it's just selecting a group of people who know what is a joke or hyperbole in the current context. I've typed out some weird stuff in Telegram and Discord chats I wouldn't want my family or employer to see out of context.
But my threat model doesn't include my employer or my suing Telegram or Discord for my chat logs, so I don't worry about it.
Are you certain these groups are private? E2E and no meta data leak?
> isn't a significant backlog to trawl through and identify me.
Are you that confident in your OpSec? No browser fingerprinting, GUID, IP address, location data?
Surely the alphabet soup agencies profile HN users.