Cassandrich22h ago
Folks do realize the people telling you QC will be a thing and break RSA and EC are exactly the same folks who were telling you AGI will be a thing, who were the exact same folks telling you the metaverse is the next big thing, who are the exact same folks telling you NFTs would take off, who were the exact same folks telling you bitcoin would replace Visa, who were....
Show threadLaurent Bercot19h ago

@dalias There's a difference though; the theory is not bullshit, the obstacles to QC are practical. Given infinite time to work on this (which we don't have, fossil fuel and rare metal reserves will depleted in ~20 years and we'll be lucky to still have a civilization in 50) a real quantum computer could be built; whereas the foundation for the other bubbles was pure hype built on wind and astroturfing.

But it doesn't change anything in practice, QC won't happen. We'll reach cold fusion first. 😔

Show threadCassandrich17h ago
@ska There is no reason to believe that a working QC could be built even given infinite resources. As an abstract mathematical construct it exists, yes, but that's very different. So does a Turing machine with an infinite tape.
Show threadLaurent Bercot17h ago

@dalias If we had infinite resources, we could build a Turing machine with an arbitrary long tape. I don't think that's qualitatively different here.

But that's pure hypotheticals.

Show threadCassandrich17h ago
@ska But you still wouldn't be able to do any meaningful computations with that in human-scale time, thanks to relativistic limits and Planck limits and Bekenstein bound and stuff.
Show threadCassandrich17h ago
@ska What I'm in particular suggesting is likely impossible, regardless of how much money you throw at it, is a QC that can factor arbitrary numbers without either space or time blowing up exponentially.
Show threadAriadne Conill 🐰17h ago

@dalias @ska i think the thing people need to understand is that you need enough qubits to hold enough state to factor the key.

right now they are factoring 22-bit RSA keys with like 150 qubits or whatever.

i'm not going to lose any sleep over this, we can just migrate to 4096-bit RSA keys and be fine for the next decade at least.

Show threadCassandrich17h ago
@ariadne @ska Even with the worst-case predictions about QC, you could migrate to 64k RSA and be fine until the heat death of the universe.
Show threadAriadne Conill 🐰
@dalias @ska also, quite a few of these elliptic curves have been found to be problematic
Mar 28 at 4:45pmWeb
Show threadCassandrich17h ago
@ariadne @ska I'm pretty confident about curve 25519, but skeptical of anything NIST has touched.
Show threadAriadne Conill 🐰17h ago
@dalias @ska yes, same, and the brainpool curves.