So, I am away from computer for four days for a really cool event and I come back and in the mean time they maybe found a polynomial quantum attack against Learning With Errors, a lattice problem? (https://eprint.iacr.org/2024/555) If this paper is correct then this is some serious breaking news shit, because lattices are like the main candidate for quantum-secure public key cryptography. (there are others but they are much less practical and for other types there have also been attacks) I mean, this paper seems to attack just a particular setting, is very impractical and does not work for schemes that are actually proposed, but an existing impractical attack often signals the way for more practical attacks. So, if it is not a false alarm, this is pretty big. It could signal the attackability of lattice schemes and undermine the trust in them. And it takes a long time to move to a new standard. Oh well. I guess we have to wait for experts to check the paper for mistakes before we can say anything.














