Hacker News new | ask | show | jobs
by fc417fc802 70 days ago
You seem to just be rehashing what we already clearly agree on. Obviously if you view classically breaking PQ algorithms as higher likelihood than QC breaking classical then you are going to disagree with the premise.

Can you actually back up your prediction that crypto related QC will remain either relatively ineffective or extremely expensive in the medium term?
1 comments
adgjlsfhk1 70 days ago
The requirement for favoring hybrid isn't that "you view classically breaking PQ algorithms as higher likelihood than QC breaking classical", but you think that the likelihood than QC breaking classical is less than a billion times more than the likelyhood of classically breaking PQ.

Hybrid has essentially no cost, so we should favor it as long as it has a greater than negligible chance of providing protection. IMO the likelihood of CRQCs breaking ECC is pretty high (>50% by 2040) and the odds of classically breaking lattices is low (<1% by 2050), but creating a 0.5% chance of breaking cryptography for the entire world seems way to high when we have a free mitigation right here.

link
fc417fc802 69 days ago
Not so. One of the core premises of the article that we're discussing here is that hybrid is proving to be quite difficult for entirely nontechnical reasons.

I agree that my previous wording was sloppy to the point of error. The point I was trying to communicate was that we already had agreement that an elevated assessment of the chance of a classical attack against a given PQ algorithm would lead to one disagreeing with the aforementioned premise that we should switch to a PQ only scheme making use of said algorithm. Rehashing that is just stating the obvious.

What wasn't presented was any reasoning to back an elevated risk assessment for any particular PQ algorithm, of which there are several. So at that point the "argument" amounts to little more than "nuh-uh, that risk assessment is wrong" which isn't exactly convincing or insightful.

link
adgjlsfhk1 69 days ago
> hybrid is proving to be quite difficult for entirely nontechnical reasons.

This is hard to square with the reality that hybrid systems are already widely deployed while pure PQC aren't/

link