> might indicate lattice based crypto (ML-KEM etc) breaks with quantum computers That is doing way too much work for one paper. At best it says a particular attack model might get cheaper than people hoped, not that ML-KEM is broken, and those are very different claims when you have to ship something on mainnet.

The awkward part is not whether this paper is right, it’s what it does to procurement and rollout timing. If every headline turns into "maybe ML-KEM is toast," a lot of teams will sit on their hands and keep shipping classical crypto a little longer, which is exactly the sort of inertia an ecosystem hates. That also hands vendors an easy out, "buy our hybrid stack, come back later," and those products have a habit of becoming permanent. If Simon’s work really changes the conversation, the first-order casualty is probably adoption momentum, not Kyber itself.