> I think the difficulty of creating cryptosystems is a bit exaggerated in this piece; I think we haven't seen much diversity in the past, because diversity was

> I think the difficulty of creating cryptosystems is a bit exaggerated in this piece; I think we haven't seen much diversity in the past, because diversity was bad for public cryptanalysis, but there are six different approaches listed on the wiki for post-quantum crypto The difficulty is not really exaggerated, and diversity is still bad for public cryptanalysis. The main approaches of post-quantum cryptography utilize very advanced mathematics and very different algebraic structure. We can use basic probability theory and game theory for basic modeling, but it's very hard for us to compare and contrast relative security for algorithms variously built over polynomial rings, modules, Galois groups, lattices, supersingular elliptic curves and error correcting codes. There is some linear algebra featured in all of this, but very few people are sufficiently well trained in all of these mathematics to consider them on equal footing. Researchers don't like this diversity. It's hard to competently analyze the advantages and disadvantages of systems with wildly different mathematical underpinnings, and you have to keep up with a lot of new material just to stay competitive. In contrast the literature for hashes, block ciphers and stream ciphers is relatively sane and unified; and at least for classical public-key cryptography you only really have to wade into number theory. The result is that the research landscape of post-quantum cryptography is pretty fractured into tribes of mathematical lineage.