- cross-posted to:
- technology@lemmy.world
- cross-posted to:
- technology@lemmy.world
Anthropic AI discovered the vulnerabilities, the issue is all the sudden patching all at once and at an unprecedented level. Also heavily reliant on AI coding to produce and roll out. Impossible not to introduce bugs. We are painted into an AI corner now. Bc the AI companies rolled it out too fast with no guardrails. So we’re all on roller skates and the roller rink is covered in oil. But to walk on shoes is too slow. So everyone is dragged into it. Move fast and break things isn’t just for unscrupulous tech CEOs. It’s now business as usual and essentially mandatory. Alternative is to be too slow to even patch vulnerabilities. Greed did this to us. Nearly zero attempt at regulation allowed it to run amok. Like gang of toddlers hopped up on speed, drunk with power, and determined to out-toddle the rest.

knew it. now stop that shit before it gets to the nuclear codes
Huh, I literally didn’t notice any of that. Guess it’s a sign that I’ve sufficiently federated my independence from this big dumb monopolies
I mean 99.99% of people would not immediately be aware of NSA outages unless there was a another navy at our shores haha. But banks, yeah. If it went after one of the backbones of fintech like Plaid, it would certainly affect almost everyone in the world.
Futurism is stupid
I’m sure this will be hated here but no where in this article does it state AI caused the outage. It said the vulnerabilities were found using AI, but never specifies how the patches were created/reviewed. So we are just assuming since the patches were large that they used AI to generate them without significant review where someone said “this might be to large to push during business hours”
It’s possible that the weakness was AI discovered but human remediated, but so far if an org is enthusiastic enough about AI identifying vulnerabilities in their systems, they are likely to be inclined to take in the suggested remediation.
You generally don’t run these tools without getting a suggested remediation to go with the discovery.
There’s gonna be stories decades later about a relatively small group of underpaid, overworked techs that thanklessly kept everyone safe from AI vulnerabilities just like they all raced to patch the Y2K stuff and no one really noticed.
It seems that it’s not about the patches themselves but about the act of patching. All of this constant patching seems to bring other problems and that’s what we’re seeing.
Yeah, in my opinion we are seeing companies not testing the patches in test environments seeing how they act and releasing beta test versions to the companies that rely on the software to be able to test and fix their own software to ensure it works. That to me is all oversight and management issues. Whether you have 1 patch or 12, they all should have the same approval processes, if anything grouping the 12 together and implementing in a test environment, then making that beta available to the companies you sell the product too.
If the security flaw was so bad that they need to implement something now without prior testing they should have made a public announcement that the systems were being taken down for emergency maintenance to ensure the security of their, and their customers data. Which no company would want to do, thus they would have to decide which patches are actual high level risks, and how to get them tested outside the live version.
The frightening thing is that scenario is the most plausible one, given the short but catastrophic history of AI.
My thoughts are that it’s poor management and no oversight. As an example if we look at say the PHP vulnerabilities found. Those flaws were found using AI from what we’ve been told. Companies like Cloudflare went out and developed patches and advised companies/users on how they could implement them to ensure the vulnerabilities were covered. A company/user then would need to review/run those patches in a test environment and see how it acts within their environment. What breaks, how much load it puts on the system… what steps will be needed to update it to work with their environment, if none then verifying it works, setting a schedule and deploying a beta version of their companies products to the companies that relies heavily on them telling them when they will go live.
Those companies/users are then responsible for doing the same.
If someone vibe coding just pushed patches that took services across the nation down for a large number of users without notice, we should see JPMorgan and such filing lawsuits for breaching contacts/terms of use.
I’m not a coder, but after working IT related fields for 15+ years, someone writing a patch doesn’t just get a greenlight by their manager, then their managers VP, or however they have there company structured.
Now if this was your home based projects you use for yourself or maybe family/friends… Sure. Someone could easily just push patches and have no one to answer too.
Well, look, people’s actually that stupid, and AI boosters are too rich to market people into doing that. If people are giving their agents full control even in production cases/environments, then that’s more than expected to happen. They’re already vibecoding everything. Look at Proton i.e.
Worst part is that they’re not shameful about doing so.
Al found “vulnerabilities”…vibe coder who replaced senior dev. Patches with vibe code from AI.
Yes, AI caused this shit.
I’ve been getting wary of futurism articles, they seem to really distort the truth to say what we want to hear, not dissimilar to conservative misinformation sources
Nearly every news site that gets posted on social media is like this. Including lemmy. Most of the accounts posting links here are bots anyway.
It’s in the same idea as citogenesis.
#1 Social media users have a fact that they want hear.
#2 A news site publishes a distorted article about that fact.
#3 Social media users check news for that fact and repost it.
#4 Now other news sites have sources. They publish their own articles about that fact, citing the article and social media.. It’s especially problematic about elections and AI. With elections it’s about polling and voter sentiment. None of them are trying to report on reality. They’re publishing whatever they think viewers want to see. That may have some overlap with reality, but it’s not reality.
Same here. This is one of several articles where the headline was not only missleading, but outright wrong. I am critical of AI and all, but i don’t want my worldview reinforced for the price of truth. I do not consider futurism a valid source any longer.
Because it originally started with doge and there are no smart people in the room to not use AI. Do you seriously think this admin could do anything on their own? They of course, used AI.
FAA union president says DOGE fired essential employees without assessment
A new report looks at what the DOGE cuts actually cost the public
A call with over 70 staff who remained at education could not identify the problem, and it was only when staff called members of a team that had been fully eliminated on their personal cell phones that the problem could be fixed. And many from this team were then asked to return to full employment. The second is a loss of institutional memory, and that’s about the loss of expertise. The third is oversight and accountability that has been weakened. And the fourth is independence of federal expertise that has been strained. And lastly, there have been major disruptions to agency operations and organizational stability. So when you look at this pattern across 19 different policy areas, it really shows how much federal workers touch every port of four lives. Communities across the country, from rural farmers to students to families, are paying the price for these cuts.
As the Washington Sun reports, Wall Street and National Security Agency sources have already told Congress that the outages are being triggered due to system updates designed to thwart future attacks by AI.
‘It wasn’t caused by US using AI, it was caused by OTHER PEOPLE using AI!’ Is a meaningless distinction if you’re trying to determine whether the rise in the use of AI (LLMs) is to blame.
Also, nowhere in this article, nor in the thin article linked, says anything about the Chase outage, AI or otherwise.
All ya gotta post is “AI BAD!” and lemmy stands up and barks like trained seals. 🦭⚽
It wasn’t AI, IT WAS ME!
YES! I DID IT LIKE THIS!

Heavy is dead!?
That’s a joke lads. It was *hiccups*
HIM!
Agh!
How did you know?
I didnt. That was a joke too
“outages are being triggered by system updates to fight AI hackers.”
is NOT
“was AI that caused massive outages”
It is if the patches that knocked the system offline were vibe-patches
But it kind of is because they used AI to write the code to fix the vulnerabilities and pushed out so much so fast that they couldn’t check if it actually worked
And where in the article does it say that?
Or are you just assuming based on just vibes?
Without a very explicit statement, we can’t be sure, but a lot of us have actual real world experience guiding us on exactly how likely AI discovery gets paired with AI remediation. It is very likely and it does have outcomes pretty much like this. It is supremely familiar to our own experience.
Well yes, and if it does come out that it was indeed the case, I will fully concede.
But I also have seen many instances of software breaking because some human just made serious errors. This was before the age of LLMs too.
I’m assuming it beause I am currently doing the same thing at my job 🤷♂️
and what do you think they are using to do these updates?
Strange. I was going to post that but then told myself “nah, that was in May. Surely they weren’t that slow“
Of fucking course it was. I’m in a creative field using profession specific software and they started offering AI tools a couple versions back.
Thankfully you can keep the full AI stuff disabled, but the amount of broken, non-functional or defective features in the base program have multiplied. It wasn’t the most stable to begin with but it’s so much worse now.
They released an update that supposedly fixed something about a new feature, but the bug is still there. No doubt they asked AI to fix the bug, believed it when it said it was fixed, and happily rolled out an update that fixed nothing.
I’m so fucking tired of wasting working hours trouble shooting and finding ways around broken shit.


Honestly this doesn’t surprise me at all. Too bad the companies won’t want to blame their ai usage so they’ll just keep up with the cyberattack bullshit.
It technically was a cyber attack, though
Was it, though?
My money is still on cyber attacks.
Could be both, though.
And they still want to make it more powerful.

It’s not enough to simply “fuck ai”. The lifestyle and culture that incubated this possible course is an enemy too.
When do we leave? Like a spouse in a shitty relationship or kid leaving abusive parents. By staying we becoming part of the enemy.
I never thought I would become religious in my life but I hear Islamic people talk and what they sounds like truth sometimes. Our lifestyles are not compatible with truth of the universe. And the universe is constantly working against the falsity we create in order to rebalance truth and nature. It’s like …that’s why we have and fund wars, so that we can patch our ego driven foundation instead of living in sync with truth and nature. We exist now for a false “God” so to speak and the universe (actual “God”) is constantly trying to rebalance.
I never in my life thought I would start thinking like this. That’s how bad shit has gotten. Like this culture isn’t respectable anymore, it’s actually almost real “evil”.
I feel like an enemy of humanity just for participating in the western lifestyle.
Just make your own culture. We have the technology. Lemmy is a great example
Even before ‘AI’ we were enemies of not just humanity - but all life. To quote a beloved live album title ‘every day, things are getting worse’.
How do we leave?
Truth and courage to stop participating and funding it? But then think about that. …If enough people started detaching we will just get treated like every other culture they abuse that doesn’t assimilate.
It’s like how Jesus taught lessons of how to exist in the chaos that life is, and then he was crucified and his lessons were used as a method of control. That’s how far it goes back.
Whatever the how is, it’s going to a centuries long task.
















