I saw that counter coming. A little bit different, better argument. Several things in here. So, let's look at them.
re paper encryption
That was defeated regularly in the Cold War in a number of ways. Easy or not, the mathematical proof didn't translate directly into the real world due to human issues and physical ones like intercept or observation. FBI's crypo unit has been defeating custom pencil and paper ciphers of criminals for a long time, too. So, we can say the best, provable encryption makes the job more difficult if no observation of the act of encryption, KEYMAT, or decryption take place. That's a lot more limited than mathematicians pronouncements imply. ;)
re universe
"universe does not favor encryption"
Oh, I think it doesn't. For one, encryption only happened one time in known universe that we know of. When it did, it screwed up more often than it worked. Then, even the best forms are defeated by stuff above thanks to other properties of the universe. Universe seems to favor plain text to me. Its own codes are plain to observe, too. Obfuscated at worst.
re computers
That was a nice dismissal but computers are the whole point, right? We talk encryption that we're going to use on a computer most likely. Then someone says some stuff like how we can trust the math. Then I have to point out we run electrical impulses representing machine instructions, not math. Then the conversation drifts to pencil and paper or arcane stuff.
At least you admitted we can't trust the math on a computer because it doesn't represent what it does. Often not on pencil and paper either or in speech if under surveillance. So, we can't trust the math at all. It's always math + all kinds of circumstances and methods. Even then, we can only trust it with probability C as in odds of Compromise.
re paper encryption
That was defeated regularly in the Cold War in a number of ways. Easy or not, the mathematical proof didn't translate directly into the real world due to human issues and physical ones like intercept or observation. FBI's crypo unit has been defeating custom pencil and paper ciphers of criminals for a long time, too. So, we can say the best, provable encryption makes the job more difficult if no observation of the act of encryption, KEYMAT, or decryption take place. That's a lot more limited than mathematicians pronouncements imply. ;)
re universe
"universe does not favor encryption"
Oh, I think it doesn't. For one, encryption only happened one time in known universe that we know of. When it did, it screwed up more often than it worked. Then, even the best forms are defeated by stuff above thanks to other properties of the universe. Universe seems to favor plain text to me. Its own codes are plain to observe, too. Obfuscated at worst.
re computers
That was a nice dismissal but computers are the whole point, right? We talk encryption that we're going to use on a computer most likely. Then someone says some stuff like how we can trust the math. Then I have to point out we run electrical impulses representing machine instructions, not math. Then the conversation drifts to pencil and paper or arcane stuff.
At least you admitted we can't trust the math on a computer because it doesn't represent what it does. Often not on pencil and paper either or in speech if under surveillance. So, we can't trust the math at all. It's always math + all kinds of circumstances and methods. Even then, we can only trust it with probability C as in odds of Compromise.