Hi, I'm the speaker. The presentations motivation and conclusions are indeed very high level. The body of the presentation presented several different applicat

Hi, I'm the speaker. The presentations motivation and conclusions are indeed very high level. The body of the presentation presented several different applications, including Zcash and several applications of Proof-Carrying Data (the extensions of SNARKs), to exemplify this. Unfortunately not could be described in detail, but they all have corresponding technical detailed papers (see http://www.cs.tau.ac.il/~tromer/publications.html http://www.cs.tau.ac.il/~tromer/publications.html). Regarding the conclusion, patcon's interpretation is accurate, and more generally: There are numerous situation where parties cannot cooperate due to mistrust, and modern cryptography, SNARK included, allows them to use digital cryptographic schemes instread of certifications, accountants etc. About the hash functions: collision resistance is discussed in the paper, but note that such algebraic/combinatoric/geometric hash functions are a totally different animal from SHA etc. The latter are much more efficient, but completely heuristic, whereas the former often have security proofs that relate them to well-established computational hardness assumptions.