-
phryk
moparisthebest, skye: The attacker wouldn't be able to read anything from *before* they joined tho.
-
phryk
so, if anything from such a chat is leaked you have some assurance that it was someone who was in the room at the time. Not a scenario that will show up in most threat models, but I wouldn't say it's useless *per se*.
-
phryk
jesus, that was in *march*? 😀
-
moparisthebest
I think you just proved the point that forbidding future joiners from reading past history in public rooms would be bad, so thanks 😁
-
phryk
still, depends on your threat model.
-
phryk
might add some real world value with semi-ephemeral recruitment channels for surreptitious groups. recruit in blocks, delete channel after a block is complete, create new channel. reduces risk for new inductees.
-
moparisthebest
It certainly doesn't have *no* use ever, just no use for fully public channels like this one
-
phryk
no, that definitely not. but i think the use-case of moving someone over into a more secure communication environment shouldn't be overlooked. 🙂
-
moparisthebest
systems like slack let you choose when you invite people to the channel "show all/none/hour of history" you don't need e2e for this, just mam filtering basically
-
phryk
i'm thinking about state-level actors, so that might not help if they can just get the data from slack directly. also, the partial attribution can help identify infiltrators.
-
skye
Ok