Login
You're viewing the functional.cafe public feed.

Replies

  • 💬 0🔄 0⭐ 0
  • 💬 0🔄 0⭐ 0
  • Sep 4, 2026, 1:54 AM

    @malwaretech

    A masterpiece, THE analysis everyone involved in this area should read. Not only for your credibility, the breadth and the strength of your arguments. The clarity and sharpness of your writing is fantastic.

    I will make this mandatory reading for my students.

    Thanks a lot for your work.

    💬 0🔄 0⭐ 0
  • Sep 4, 2026, 2:47 AM

    @malwaretech agreed. I crank through meetings with hundreds of companies on these points, as I hear from ciso who get a lot of bad pressure. maybe your post will do us all some good.

    we have to ask why is the marketing so much more effective than the anti marketing? my presentations since 2012 warned it's becoming our industry Loch Ness, and yet here we are. the people selling Nessie don't want us to prove the lie.

    should also point out generative code tools came well before the start you say. GitHub Copilot was public preview June 2021 and GPT-3 API dropped waitlist in November 2021. full year before ChatGPT. DALL-E 2, Midjourney and Stable Diffusion also were 2022. so that start date reads off by a year or more.

    💬 2🔄 0⭐ 0
  • Sep 4, 2026, 4:36 AM

    @flyingpenguin @malwaretech
    😬

    In some (or many) ways marketing utilises the same psychological effect as phishing/social engineering

    It shows you something you want (or want to avoid) e.g. "not getting hacked".
    Then tells a story with at least a tiny connection to (a perceived) reality e.g. "ai is everywhere"
    Distracting you with a lot of irrelevant details to cause mental (over)load to limit the cognitive capacity available for rational thought.

    Add time pressure, "if you don't act now it's too late"

    And play with biases and psychology
    Authority "Microsoft, Google said it or do it" "leading researchers say" ...
    Group pressure, "everybody does it"

    And then scam you out of your money

    💬 1🔄 1⭐ 1
  • 💬 0🔄 0⭐ 2
  • 💬 0🔄 0⭐ 0
  • 💬 1🔄 0⭐ 0
  • Sep 4, 2026, 10:02 PM

    @malwaretech oh viability! better, but you said availability, and that's why I said it's off by a year. if it's viability then your own privesc tree automates with a Python script. can we agree Copilot, running Codex, wrote those from June 2021? then GPT-3 spear-phish was outclicking humans at Black Hat in August 2021. Check Point's January 2023 sightings were people self-attributing to ChatGPT, but that's as bad a vendors self-crediting. and even then it's the observed, not the whole. I guess I have more evidence to dig up.

    💬 1🔄 0⭐ 0
  • Sep 4, 2026, 10:32 PM

    @flyingpenguin I feel like you're trying too hard to make widespread AI cyberattacks a thing. I deliberately structured my entire argument around the fact that people would still attempt to resurrect their AI-boogeyman no matter what I said.

    "For argument’s sake, let’s take CrowdStrike’s AI-enabled claims at face value, assume AI is driving the drop in breakout times, and that trend does continue as a result of it. The core argument of this piece still stands, regardless of opinions on why breakout times are dropping."

    💬 1🔄 0⭐ 0
  • Sep 4, 2026, 10:44 PM

    @malwaretech other way around. i've spent twenty-plus posts since April arguing that the AI attacker boogeyman is a Nessie sighting, the AI Loch Ness thing I've been on about for over a decade. my objection is the method being inconsistent. your conclusion doesn't depend on who's right about adoption. and your piece is stronger if you avoid the dates and keep the basics. that's what I'd like people to hear.

    💬 1🔄 0⭐ 0
  • Sep 4, 2026, 11:29 PM

    @flyingpenguin No, because a reasonable reader is not going to pretend stable diffusion models or GitHub CoPilot preview were launching cyber attacks just to push some dates back. That is the behavior of a professional nit picker whose views I am not going to move one bit.

    The fact that AI is not responsible for the reduction in attack times is an important case to make, or the people making it are just going to continue poisoning the discourse with their nonsense.

    💬 1🔄 0⭐ 0
  • Sep 5, 2026, 7:25 AM

    @malwaretech Fair enough. Your first reply said the cause doesn't matter, conclusion stands regardless. Your second says proving the cause matters. I'm only asking which one you wanted to write.

    💬 1🔄 0⭐ 0
  • Sep 5, 2026, 11:14 PM

    @flyingpenguin the cause doesn't matter to the argument, but you're also wrong about the cause. Two things can independently be true at the same time. Those arguments don't overlap in any way, one does not refute the other.

    💬 0🔄 0⭐ 0
  • Sep 4, 2026, 4:18 AM

    @malwaretech
    Thanks for the great article. It contains a lot of material for an internal post in writing 😁

    I totally agree on the technical side.

    When I suspect an impact of AI on the attacks it's on the social/human side e.g. phishing.
    (I intentionally write "suspect" because there is no "exponential" increase and nearly everything i see could also be done without AI)

    We had our first deep fake phishing attempts, there are context based phishing attacks after external accounts were taken over, and generally the (non English) language of phishing "improved"

    💬 1🔄 0⭐ 0
  • 💬 0🔄 0⭐ 0
  • 💬 1🔄 0⭐ 0
  • 💬 0🔄 0⭐ 0
  • 💬 1🔄 0⭐ 0
  • 💬 1🔄 0⭐ 0
  • 💬 0🔄 0⭐ 0
  • Sep 4, 2026, 7:16 AM

    @malwaretech Wannacry still spreading is interesting. Didn't ever consider it. So, are we in a situation when, at some point in the future, someone will forget to renew a registration for the sinkhole or accidentally delete it, and then it starts ransoming systems again?

    💬 1🔄 1⭐ 0
  • 💬 0🔄 0⭐ 0
  • 💬 0🔄 1⭐ 0
  • Sep 4, 2026, 7:55 AM

    @malwaretech "Jumping the gun towards worrying about zero-day exploits and AI-attackers is not in anyone’s best interest."

    arguably that's in the best interest of the attacker who can then still compromise the entire network with a single phishing mail :^)

    💬 0🔄 0⭐ 0
  • 💬 0🔄 0⭐ 0
  • Sep 4, 2026, 8:36 AM

    @malwaretech I agree with the message of your post - but I think you underestimate how well an LLM (even a local one, no need for latest cloud models) can use an initial vector (from a regularly automated scan) to gain persistence and exploit. While nation state actors have personnel doing this for their 9-5, this ability is now part of any script kiddie's toolbox as well.

    LLMs are also super useful to analyse centralized loki flows for "anything that seems suspicious or out of the ordinary". Keep all rule sets in place for known things, they're much more effective at that, but an LLM can pick up on unknown things remarkably well.

    💬 2🔄 0⭐ 0
  • Sep 4, 2026, 10:54 AM

    @troed @malwaretech I’m not an expert (in fact I know very little), but the cost of the attack is also a limitation you have to consider IMHO. The random guy would still need to spend a few thousand dollars of tokens for each attempted attack. At that point there isn’t much difference between using the LLM and hiring someone for doing it, paying some existing service or run scripts you find online. All things that have been available for decades.

    💬 1🔄 0⭐ 0
  • 💬 1🔄 0⭐ 0
  • Sep 4, 2026, 12:19 PM

    @troed @malwaretech If we speak about local models then yes. I was mostly thinking about the edge models AI companies are asking a lot of money for. But it does indeed make more sense to use local models as you suggested.

    Is there however any proof this is actually a problem? Is there any statistics/study showing that attacks from this kind of people have significantly increased and that they are causing new problems we are somehow unprepared to face?

    💬 1🔄 0⭐ 0
  • Sep 4, 2026, 12:32 PM

    @antopatriarca

    My post listed the exact two things I thought useful to know in relation to the blog post. It did not support various media driven emergency claims.

    I worry that people have some sort of general Anti AI sentiment (prevalent here on Mastodon) that cause them to underestimate how good the models are at cybersec in general though. The time it takes me to exploit an IoT device is ridiculously low nowadays compared to before, and that was sort of my thing ...

    @malwaretech

    💬 1🔄 0⭐ 0
  • Sep 4, 2026, 2:33 PM

    @troed Yes, I’m not saying gen AI is useless for this. It certainly seems to work well for attackers. However, I don’t think it is necessarily the paradigm shift for infosec the marketing claims. I may be biased because almost every single claim I have ever checked from AI companies was either misleading or completely unproven (and often ended up being false), but I’m not going to make decisions on what seems to be more science fiction/wishful thinking than reality. Particularly on things that don’t make sense to me. So I would not bet against LLMs making attacks harder to defend or more common (I think it is probably true even if maybe not exponentially more), but AI reactive systems seem to instead only make sense after the network is properly proactively protected and it is probably going to be the least important part of the defence.

    💬 0🔄 0⭐ 0
  • 💬 0🔄 0⭐ 0
  • Sep 4, 2026, 12:59 PM

    @malwaretech

    💯

    "Historically, organizations have chosen to prioritize business function and availability over security. As attacks and attackers get faster, this becomes increasingly untenable."

    💬 0🔄 0⭐ 0
  • 💬 0🔄 0⭐ 0
  • Sep 11, 2026, 11:27 PM

    @malwaretech The talk those two openAI devs gave at the blackhat had the same conclusion: We build the tools that attack you and you must buy the defense tools (or tokens) from us because this is the only way. They did say this without any shame. Its like selling land mines to the attacker, buying a health company, and sending the sales team to the target country to sell wheelchairs and artificial limbs for their people

    💬 0🔄 0⭐ 0