The flash hider doesn't touch the bullet in that scenario.
My brain is giving me this scenario and it is most likely wrong, it's my brain after all.
Think of the face of the action as a clock, say this face is .001" proud at 6, barrel is proud .001" out at 3 and shim is perfectly parallel. In my brain, I going to bottom out at those 2 surfaces early, pushing the muzzle away, right and up in our scenario. The tension is located in those two areas, over time, the potential of failure of the shim is highest there because of crush and the harmonic oscillation during firing.
Now, you say, but, but, but, thicker recoil lug and such, a shim has no where the load capability of a precision ground recoil lug, it's impossible to heat treat a shim that thin to the same degree as a recoil lug, at least to my knowledge.
They also make shims and people shim AR barrels all the time, yes they do, but I don't and it's because I don't want one more point of potential failure. Either I swap receivers or barrels or both or replace the barrel extension, but never a shim.