The problem with "assume the maximum worst everything" case is that it places constraints on you that are essentially certain to be overly broad. Building a wor

The problem with "assume the maximum worst everything" case is that it places constraints on you that are essentially certain to be overly broad. Building a workable threat model & defending against it isn't practical (or necessary) if you assume worst case everything. For example, have they recorded all IP traffic ever? It's possible, but absolutely overwhelmingly unlikely. But all IP traffic originating from $small_anti-US_country_x in the last year? That's a whole lot more plausible, and thus probably worthwhile to defend against. Also, does that mean non-US CAs are trustworthy? Are non-US comms also subject to tapping, modernday Ivy Bells[1] style? Also, you have no real idea of exactly if/how they're analysing their captured data in order to extract "threat metrics" or whatever. Posting in a reddit thread about kittens would seem utterly innocuous to virtually everyone; but suppose their steganography detection throws a false positive on your image and by random coincidence they also have hits from some IP once used by a 'terror organisation' in the same thread. Should threats of that nature be considered, purely because they're possible? How do you clear your name? Why won't you give up the password to the bomb plans claimed to be embedded in your image file? Is it because you're protecting other commie mutants[2]? How much paranoia is reasonable? [1] https://en.wikipedia.org/wiki/Operation_Ivy_Bells https://en.wikipedia.org/wiki/Operation_Ivy_Bells [2] https://en.wikipedia.org/wiki/Paranoia_%28role-playing_game%29 https://en.wikipedia.org/wiki/Paranoia_%28role-playing_game%...