4 ms·
“Module-LWE is not breakable within a Python script” —-Ptacek, 2025 “Apache chunked encoding is not exploitable” —- Dowd, 2002
by philodeon 1y ago
“Module-LWE is not breakable within a Python script” —-Ptacek, 2025
“Apache chunked encoding is not exploitable” —- Dowd, 2002
- tptacek 1y agoI mean, if you're putting me in the same camp as Mark Dowd, I'm flattered. What I think you're not seeing is that this isn't a SIKE vs. Lattice kind of debate; it's a Curve25519 vs. P-256 kind of debate. P-256 was never broken. Curve25519 made smart engineering decisions that for years foreclosed on some things that were common in-the-real-world implementation pitfalls. P-256 has closed that gap now, but for the whole run of the experience they were both sane choices. That's a generous interpretation. Another parallel would be Rijndael vs. Serpent, where the Serpent advocates were all "I don't know about this Rijndael stuff seems dicy". Turned out: Rijndael was great. But Bernstein wants you think that rather than a curve-selection type debate, this is more akin to a "discrete log vs. knapsack" debate. It isn't.
- philodeon 1y agoI thought the whole point of DJB’s “cache timing attacks in AES” is that Rijndael is not particularly great.