Vitalik Buterin backs crypto ‘bunker mode’ amid rapid AI math advances(cointelegraph.com)
42 pointsby firstcomm2 hours ago8 comments
  • simonw38 minutes ago
    Crypto's greatest weakness has always been security. Even if the algorithms themselves hold, humans need to keep their keys safe. We have decades of experience now showing that humans can't do that.

    I wouldn't be worrying about the algorithms so much as I'd worry about all of the end-users who are going to get their keys stolen and their wallets drained thanks to the deluge of new vulnerabilities in operating systems, browsers, wallets, personal agents and other software.

  • int32_6419 minutes ago
    The guy that made the 'bunker mode' tweet is an Ethereum guy, the chain that they reorganized when the wrong people lost money in the DAO hack.

    What good are strong cryptography primitives for cryptocurrency if your "CEO" can successfully make calls to reorganize the chain?

  • notnullorvoidan hour ago
    My gut is telling me pushing security model to hashes seems like a far far worse bet if you are worried about advances in maths.
    • palmotea37 minutes ago
      > My gut is telling me pushing security model to hashes seems like a far far worse bet if you are worried about advances in maths.

      Aren't hashes far less reliant on math tricks?

      My understanding is the problem with public-key crypto is it extremely reliant on a single math trick (e.g. the factoring problem), so if it turns out that trick was weaker than was assumed, the whole thing crumbles.

      • tptacek27 minutes ago
        Yes. Symmetric cryptography doesn't generally rely on the existence of "trap doors", so there's no direct relationship in any sense between the inputs and the outputs. Huge portions of the cryptographic attack surface are foreclosed in these constructions.
    • MattPalmer108637 minutes ago
      The point is they have less algebraic structure than things like elliptic curves. If you want "math resistant" properties, this is a good thing.
    • jMyles30 minutes ago
      That's the opposite of what my gut is telling me.

      Hashes are so simple. There's no place for these crazy math attacks, which rely on rethinking long-standing and otherwise reasonable assumptions, to hide.

      Surely we can all agree that there's no way to reverse a modulus.

      • dist-epoch9 minutes ago
        Isn't it more secure mostly because of the mixing than the modulus.
  • j2kun34 minutes ago
    FWIW, there has already been a lot of effort thrown at AI attacking lattice problems underlying PQC without anything to show for it. I'm not sure why Vitalik is suddenly worried about lattice problems in particular, unless he has some insider knowledge.
    • dist-epoch16 minutes ago
      The AI unit distance conjecture proof used high-dimensional lattices and algebraic number fields, adjacent to the kind of stuff used in lattice PQC. It shows skill in that area.
  • search_facilityan hour ago
    May be Ethereum should fund 10000 agents to pursue new crypto math with provable strongness/hardness
    • Lerc31 minutes ago
      I think much of the the doomer propaganda is paid for by by the half billion or so that Vitalik Butering gave to the Future of Life institute. It sounds like he wanted it spent on research more than advocacy and is now distancing himself from them.
  • whatan hour ago
    Why should we believe that LLMs are going to break ECC?
    • tptacek25 minutes ago
      LLMs probably won't break curve cryptography. Cryptography researchers armed with LLMs are a different story.
      • contingencies14 minutes ago
        Don't forget LLMs can now design efficient high speed hardware systems, a traditional slow-down for many attackers which represents a significant barrier to scalable next-gen attacks falling. Unfortunately memory prices are through the roof, but that isn't always significant.
    • plesiv38 minutes ago
      Your probability distribution should be flatter at least.
    • cidd13 minutes ago
      You don't
    • nextaccountic15 minutes ago
      It's a risk
  • EA-3167an hour ago
    It’s always tragic when the new hype train runs over the old one. It pains my soul to imagine those gentle folks with millions in crypto being subjected to the incredibly foreseeable consequences of their limitless greed.
  • spottedmarleyan hour ago
    It may be a plausible concern at some point in the future but, once it's possible, it would manifest as an extremely expensive and time consuming attack against a targeted address, it would require an enormous amount of compute. So, I suppose the first line of defense would be to never keep more funds locked under one key than it would cost an attacker to decrypt it.
    • _ink_an hour ago
      Which leaves Bitcoin between a rock and a hard place. The Satoshi funds are probably worth an attack. At the same time they cannot move to post quantum cryptography, because that would also require to move funds.
      • an hour ago
        undefined
      • spottedmarleyan hour ago
        Yeah the Satoshi treasure would be a prime target. Maybe we see those funds get mysteriously split up at some point? But we wouldn't know if they were hacked or protected..
    • tzonean hour ago
      There are exchanges, custodians, etc that hold insane amounts of BTC or ETH in a single address.

      We are talking billion dollars+ worth in a single address. So if either chainskey security is compromised in a way that it is conceivable to brute force it, there will definitely be plenty of targets.

      • dist-epoch6 minutes ago
        Stealing a billion dollar address is kind of worthless. Try cashing it out. Even if you are NK.
      • spottedmarleyan hour ago
        Yes, it would require those exchanges to move their funds around. I'm pretty sure they can manage it.
    • warkdarrioran hour ago
      Cost of running an attack is super cheap if one buys botted computers (it is/used to be $25/1000 machines).