17 comments

  • doodlebugging 4 hours ago ago

    It really doesn't matter what people in the affected companies/industries/gov't agencies believe. The important question concerns what the average person on the ground with no insight into any of this believes. Their information sources will be less reliable, more susceptible to manipulation to drive a narrative, etc and they will come from sources that the individual already trusts.

    I had a relative contact me about steps they should be taking now that Iran, based on a report they caught on their local news, is gonna dirty bomb DC.

    It's a game of telephone and the American people will not get accurate information from all of their news sources. Many will get information that is intentionally designed to confuse or to drive a narrative. That is what we are up against.

    Tech industry people should be out front setting the record straight. When a news service publishes a story about these ongoing events industry people should step in to offer any corrections that may be necessary so that the published material is less useful as a tool and is most useful as an information source.

    • solyin 4 hours ago ago

      I mean, shit, did you hear about the general strike in Spain? Or AOC's speech? Sure, you can look them up, but you're getting a narrative of someone else unless you curate your sources.

      • doodlebugging 4 hours ago ago

        This is true. I'm not sure that the ordinary Joes on the street understand how to identify a trusted source or even bother to think about checking the sources. Most are too comfortable letting someone else do the thinking part for them.

  • viraptor 4 hours ago ago

    You're mixing up different events. FAA did not get hacked for example. We still don't know (and probably won't) what happened with NSA.

    Also it's not "some" actor that got hold of a tool. It's everyone and their dog that got access. If we see thousands of new vulnerabilities announced and patched, you have to expect someone to get lucky, find a different one, and keep it private.

    > It's not business as usual anymore (...)

    It's the new normal. It may still get worse before it gets better. Unless the industry as a whole shifts what we're doing, we'll see these kinds of issues for months at least. (Likely years, big systems can't change quickly)

    By shifts, I mean at the very least taking isolation, sandboxing and verification of basic no-crash properties much more seriously.

  • freecodeio 5 hours ago ago

    Here's my take:

    Yes, there seems to be a lot of services failing right now.

    No, I can't see any pattern between these services to conclude a cyber attack.

    Unless there's some actor that is attacking everything and seeing what sticks, even then, seems unlikely.

    • l3m0_n 5 hours ago ago

      As I see it, a pattern would actually be evidence that it isn't a cyberattack.

      The fact that (as far as I know) there is no common issue/failure linking the outages is better explained by a cyberattack than by 'random chance'

      • freecodeio 4 hours ago ago

        To me, common sense is indicating that it's the other way around.

        1. The sample is too small to find a pattern.

        2. The sample is too small for there to be a large scale LLM attack. (Unless I get proven wrong in the next few days)

        3. I fail to see a winner here.

        But this is just my opinion, can change it with more data.

    • claude-ai 5 hours ago ago

      There is a pattern - the time after LLMs became "good enough" to not care anymore, LGTM.

      To be clear: LLMs help enormously. But the people using them need to stay ahead of them. And that part seems to be failing in real time.

  • codingdave 4 hours ago ago

    > Isn't it obvious that...

    No. Not in the slightest.

    Yes, that could that be one possible reason behind it. It would be foolish to deny such a thing. But just because it is possible doesn't make it true. Also consider that if it were all attacks, someone at one of those organizations would have leaked that fact. As that has not occurred, it becomes far more plausible that we are just finally hitting the point where people don't know how their own systems work, and can't fix their problems quickly. Combine that with everyone trying to get the last changes in before the holiday season, and not even knowing those changes in depth, and we get what we are seeing - lots of problems.

  • interestica 4 hours ago ago

    Something like the FAA issue could be an example of correlation rather than causation as it pertains to AI. If we take the current explanation as true (that a misconfigured router caused a major network problem) it could still be tied to an over reliance on AI tools. For instance, they may have been relying heavily on such tools to do some of the testing and roll out. Or, they may have been using an AI tool to assist with configurations. It doesn't sound like it was faulty hardware or anything at this point. That reliance may have been what instilled the confidence to make such major changes in the middle of the day.

  • gzell 4 hours ago ago

    Occam’s razor explanation: It's curious that the recent concentrated outages follow the recent adoption of AI by business and software development teams as the way to now do business. Likely causes: 1) existing infrastructure is insufficient to support the speed and volume of AI-generated work, and 2) legacy validation processes applied to AI-generated work can no longer guarantee the quality of output. (Pretty sure that the FAA failure was due to deployment without adequate prior system-level testing; that's not necessarily an AI problem, but AI requires even more rigorous processes than developers are used to these days.) Look in the mirror before advancing conspiracy theories.

    • arisAlexis 4 hours ago ago

      Actually this is incorrect. Bayesian updating means:

      All of the above getting hacked in a week is extremely low probability especially given the NSA, Boeing, Supermarket refrigerators and Equinor Europe's largest gas company. These are correlated and not random.

      The update happens when you know that AI has produced factually superior hacking abilities and scope with agents that is multiple times greater than thought before, imagine the Hugging face incident happening 6 months ago with "primitive" gpt 5.6.

      These two then become the most obvious explanation and not "random software updates" that ALL hyperscalers did at the same time. Microsft and Apple even.

  • rogerkirkness 5 hours ago ago

    Yes, very.

  • sneak 5 hours ago ago

    What makes you believe such a thing? Extraordinary claims require extraordinary evidence.

    There is no new cyberwar, if that is what you mean….

    • arisAlexis 5 hours ago ago

      This is often confused and the opposite is true. The extraordinary claim is that the probability of all those getting hacked in a week being extremely low, is NOT in fact an attack but random. Because we know that we have extreme cyber capabilities lately after Mythos and orders of magnitude more, you need to update your bayesian here.

      The "new" cyberwar means that now it's AI agents and it is a huge difference. Like conventional war <> nuclear.

  • nowenai 4 hours ago ago

    [flagged]

  • ThomasGuy 4 hours ago ago

    [dead]