> of course this sort of analysis has to be made in a lab and by smart
> people, a rare asset among policemen. yet I don't believe an arms race
> can be won this way.
Important point you made about it being possible to make assertions
about an supposedly encrypted volume with the appropriate resource and
this should not be belittled.

But I believe making it harder for them to break our autonomy is way
more important than any political means to win this arms race... As your
also noted, using encryption schemes like this does increase the need
for scarce resources and knowledge about the situation.

Thus, this works well either way... The key point is "scarce resources",
I think:
If just a few rebels use this kind of encryption, the default routines
(e.g. airport security) tend not to spend so much on using specialized
resources looking for the needle in a haystack.
If many do, it cripples the bad actors' power to do it in a mass
surveillance fashion because of the costs involved (most states are
borderline broken, btw).

Of course, they will probably target some high profile individuals and
catch them as examples to make the propaganda about hidden encryption
being useless against governments... but it's would not be a good
argument because, well... They already use this same technique again and
again.
And, finally, when they really want to make a point to scare people,
they just resort to direct violence anyways (e.g. the famous $5 wrench
or the less famous and a bit more expensive 3 shots in an alley).

Plausible deniability (even if not 100%) is an important weapon to
suffocate their resources.