Post #949760

2026-03-26 15:43 UTC

Last year, my position was that we still had time to design PQ authentication mechanisms. Now, based on the pace of progress and on statements like Google&#39;s, I believe: 1. we need to finish rolling out PQ key exchange yesterday 2. we need to start rolling out PQ auth now 3. it&#39;s too late to ship any new non-PQ design or system https://blog.google/innovation-and-ai/technology/safety-security/cryptography-migration-timeline/

Replies (6)

@freddy@social.security.plumbing 2026-03-26 15:46

@filippo I like their aggressive timeline, but I'm not sure there's any specific argument or reason that explains why it should be expedited. Am I missing something?

Open ##1074195
@fay59@tech.lgbt 2026-03-26 15:53

@filippo too late forever or too late for initial deployments?

Open ##1074196
@ikke@ipv6.social 2026-03-26 16:07

@filippo Interesting, I just cam across https://infosec.exchange/@mttaggart/116163107290977793 the other day, basically saying that it won't be feasible any time soon.

Open ##1074197
@Ciantic@twit.social 2026-03-26 16:21

@filippo You got me interested to know what it would look like in authorized_keys, and can it be this short! Looks neat. ssh-mldsa44-ed25519 434f4d505349472d4d4c44534134342d456432353531392d534841353132 https://datatracker.ietf.org/doc/draft-sun-ssh-composite-sigs/02/

Open ##1074199
@alxndr@tech.lgbt 2026-03-27 16:45

@filippo Oh, did Google announce their intent to buy 40% of the world's 8-bit home computers and/or dogs?

Open ##1074200
@flux@mastodon.unwi.re 2026-04-06 17:46

@filippo @cryptohagen are you following this? :)

Open ##1074205