• reksas@sopuli.xyz
    link
    fedilink
    English
    arrow-up
    1
    ·
    3 days ago

    Aah… that is indeed a problem since the threats have to be dealt with fast once they are reported since they are now basically public…

    • fruitycoder@sh.itjust.works
      link
      fedilink
      English
      arrow-up
      1
      ·
      2 days ago

      If a public tool can find a CVE in minutes to hours, it doesn’t matter if some of the people using signed an NDA.

      All it takes is someone how isn’t going to report it to also find and exploit it

      So the exploitation window doesn’t start when it is reported it started at when the tool could have found it

        • fruitycoder@sh.itjust.works
          link
          fedilink
          English
          arrow-up
          1
          ·
          2 days ago

          I mean the alternitive, in this case, was security through obscurity, in which these exploits existed, could be reversed for years, and no one else would know .

          • reksas@sopuli.xyz
            link
            fedilink
            English
            arrow-up
            2
            ·
            2 days ago

            yes, but with llm they can be uncovered so fast they cant be processed fast enough. And you have to check every report carefully since llm come up with false information all the time, but malicious actors dont have to care about that. Also, now attackers don’t have to stockpile the zeroday vulnerabilities, infact they have to use them fast before they are potentially patched.

            • fruitycoder@sh.itjust.works
              link
              fedilink
              English
              arrow-up
              1
              ·
              4 hours ago

              Right, the next step is how do we filter BS faster for devs and fix/mitigate bugs faster or even better prevent this category of bugs from being implimented.