In Scott Aaronson's book Quantum Computing Since Democritus, he off-handedly mentions the need for quantum-hard crypto and that perhaps something based on Wolfram's cellular automata would do the trick. Wolfram says he used one of his own cellular automata rules as the PRNG in Mathematica. Which makes me wonder, why don't we already use something like this in modern crypto? And, how exactly would we implement this if we wanted to do so?
I'm not nearly well versed enough to speak intelligently on the subject, but my understanding is that only quantum computers can do a "perfect" key exchange, because the detection of interference by a third party can only be detected by a quantum computer.
Kind of, quantum key exchange is based on the uncertainty principle. By encoding the key exchange in quantum properties of particles, any mitm observer will change those properties, thus breaking the exchange. You don't need a quantum computer to do QKE, just the right kind of detector. There's already a bunch of people that do this without quantum computers.
