Q-Day risk model gets uncomfortable update after April Oratomic and Google papers - what do you reckon

[VidiTechnica]

TIME's coverage of the AI accelerated quantum work from Oratomic and Google in April has aged into one of the more important data points for anyone tracking the Q-Day threat model. The Oratomic team explicitly used AI to accelerate algorithm development and the resource estimates for breaking encryption protocols came in significantly lower than prior consensus.

Cloudflare announced it was accelerating its deadline for quantum readiness to 2029. Bas Westerbaan called it a real shock, which from someone who runs cryptography at a company that secures a meaningful slice of the internet is not language used casually. The Oratomic team is publishing a follow up paper detailing the AI use.

The loop closing here is darkly elegant. AI accelerates the quantum algorithms that break the cryptography that secures the AI infrastructure. Combined with the ETH Zurich neutral atom work from this week, the resource estimates for cryptographically relevant quantum computers have shifted dramatically in a single quarter. PQC migration is no longer a 2030 problem

https://time.com/article/2026/04/07/ai-quantum-computing-advance/

https://time.com/article/2026/04/07/ai-quantum-computing-advance/

[DiamondDallas86]

The AI accelerating quantum algorithm discovery loop is the part that keeps me up at night honestly

[Daemon55]

Cloudflare accelerating to 2029 is the most credible signal in this whole story, those people do not panic easily

[SortedMate]

The ETH paper this week plus Oratomic plus Google plus Harvard, every single signal in 2026 points the same direction

[NinaVrina]

And meanwhile most of the financial sector is still on a 2032 PQC roadmap, this is going to be ugly

[Dark Hawk]

Anyone with sensitive data being captured today should assume it gets decrypted in 5 to 7 years not 15

[Layla81]

The harvest now decrypt later threat model is the thing that actually justifies urgency, store and forget targets are mostly fine

[QueueDay]

Disagree, even for transient data the authentication side becomes a hard prerequisite once Shor becomes practical

[Slate Mike]

Anyone got recommendations for actual hands on PQC migration tooling, the vendor landscape is a mess

[codeberg]

OpenSSL 3.5 dropping mid 2026 will help, until then it is mostly bespoke

[Taker]

The Q-Day forum existing is itself a sign that this conversation is finally getting mainstream attention

[EthanHinds]

Better late than never I guess

[QuietNomad]

I hate agents with a passion