korrents

korrents · video

Livecoding post-quantum signature ML-DSA in Go on Twitch — Day 2/4

Filippo Valsorda · 3h 15m · youtube.com ↗

6 korrents from this recording

1h2h3h
Filippo Valsorda did not write this page.

Every claim below is a statement made in this recording, quoted word for word and linked to the second it was said, so you can hear it rather than take our word for it. The wording comes from the transcript published alongside the recording; the sentence above each quote is our reading of the claim, not their wording.

  1. 0:02:23 · watch on youtube.com↗

    because indeed the format that NIST uses for uh private keys is both useless as an internal representation and a very poor uh exchange representation
  2. 39 min later
  3. 0:40:59 · watch on youtube.com↗

    Um but we want to get started with certification with FIPS early because it takes like a year or two.
  4. 19 min later
  5. 1:00:03 · watch on youtube.com↗

    this has a space of 2 to the 12 or 2 to the 13 in which things can go wrong, right? Uh which means that if you um if you run a few thousand tests, you are almost guaranteed uh like let's say 100,000 tests, you're almost guaranteed to hit every uh edge case in spaces that are that small
  6. 13 min later
  7. 1:12:48 · watch on youtube.com↗

    I I'm generally allergic to having a lot of um API surface because there are things we we are going to be bound to forever and if for some reason it becomes hard to distinguish errors, we are bound to doing that.
  8. 15 min later
  9. 1:27:19 · watch on youtube.com↗

    Did we have to have two different ways to do the same thing? It's just more code, more opportunity for bugs.
  10. 1.6 hours later
  11. 3:06:19 · watch on youtube.com↗

    but also I feel like we should wait a month or so um before freezing the FIPS module anyway because there's always some stuff that might come up right after the freeze, right? Uh bug fixes and so on.