While one could be wrong about the basics of physics and math, you can encrypt things, cheaply and quickly, that would take a computer made of all the matter and all the energy in the universe to break in more time than the universe has existed.
You have to take some care. But an NSA magic code breaking machine is in the same category as flying saucers at Area 51. Exceedingly unlikely.
That's about as good a guarantee as nature can offer.
There's a very specific reason I included B and C in my listed cases. I agree in many or most cases it could be said that it's unlikely certain encryption methods would be bypassed, but to me that only gives you a fairly good prospect of being secure, not an absolute certainty. That may or may not affect your original argument, I'm not sure.
Encryption protects infromation as long as A) the computation power, time and desire available to break your encryption does not exceed the level you encrypted, B) the underlying math principles on which the encryption was based do not see a change in some manner reducing the effectiveness of the algorithm for this task, or C) The implementation of the encryption algorithm did not suffer some flaw reducing it's effectiveness to the level if can be beaten.
I know A and C have happened in the past. I'm unsure whether B has occurred. It's worth noting that in the case of B or C, the strength you choose now may not have any effect on the outcome.