Close Menu
    Technik

    KI findet Fehler, die Hacker ausnutzen können. Machen Sie sich bereit für Bugmageddon.

    Share.
    View 28 Comments

    28 Kommentare

    1. Kinexity on

      Assuming that this is actually the case this „Bugmageddon“ will end not long after it started. It’s not like the devs are not going to use AI to find those bugs too.

    3. grumpy_autist on

      As someone with 20 years of experience in software engineering of critical systems (telecom, banking, etc) – we usually have a backlog of few hundreds or thousands bugs to fix but management does not give a fuck because it does not impact revenue or security compliance.

      Many „critical“ security bugs are not being fixed because they have like mission-impossible scope (i.e you need to break into datacenter and connect your laptop to a specific chassis, then wait 5h for a right signal, etc).

      So AI would do shit except add few thousand new backlog entries no one gives a fuck anyway. And also add extra work for human engineers who eventually need to analyze and reproduce the bug.

      We already have had for years source code bug scanning tools and it’s pain in the ass to work with with a lot of noise and false positives. It works and there is value to it but you need to plan and do it right – something that vibe managers will never grasp.

      Sure, they may or will be few high profile bugs found in popular software but nothing that will transform the industry.

    4. origanalsameasiwas on

      I guess it’s time to move away from the internet and disconnect all devices from the internet. It’s going to be a ride to where I don’t want to be a part of. Let me know when the smoke clears.

    6. Bubbly_Extreme4986 on

      So long as everyone has access to them at the same time it’s an even game

    7. dope_sheet on

      Sneakers was a fun movie, just a couple decades ahead of its time… „Too many secrets.“

    8. EliteG77 on

      Don’t wanna be that guy, but paid/subscription based articles shouldn’t be posted here.

    9. rio_sk on

      If someone is using AI to spot holes in a system they’ll probably also leave their home IP on the logs

    10. lokey_convo on

      I’m sure this wont be a problem with all the overly complicated poorly engineered vibe coded software out there…. right?

    11. hitsujiTMO on

      It’s forgetting one very important point in the article. Mythos had access to the source code.

      It’s not finding these bugs looking at raw machine code.

    12. sam_my_friend on

      My security team has been sitting on the same bugs backlog for YEARS. Most are never a priority, and AI is not going to change that.

      Man, many companies still use root user in their Dockerfiles and it’s the first no-no in the book of no-nos.

      You wouldn’t IMAGINE how many tokens we have pushed into repos, just for security team to come a *year* later to let us know we have to rotate said token.

    20. marmaviscount on

      The exact opposite result will happen, get ready for all the annoying bugs to be solved before release and a decline in hacking stories in the news.

      This is a tool to find and remove bugs

    22. Jjerot on

      The only specific examples I’ve seen of this finding vulnerabilities were things already known for 10+ years. Everything else had been vague gesturing, sounds a lot like hype marketing which the AI industry is built on.

      We’re already living through bugmageddon, look at what AI code has done to microsoft. No hackers required, it breaks itself.

    23. MaybeTheDoctor on

      The AI is also proposing fixes, so maybe people should just implement the damn fixes and deploy the fixed code.

    26. Intelligent_Ice_113 on

      that’s why you should write software with a prompt which ends with „… Make no mistakes.“

    27. wintrmt3 on

      Fuzzers were a much bigger breakthrough in finding security bugs, and the world did not collapse.

    28. ebfortin on

      We’re still in the hype with this new model from Anthropic. The only info we have is from them stating how incredible it is. Allow me to be a bit cautious when a company says its product is just incredible and so good it can do harm to the entire universe.

    Leave A Reply