A rogue AI led to a serious security incident at Meta
theverge.com
theverge.com
All the years of discussing programming/security best practices
Then cut to 2026 and suddenly its like we just collectively decided software quality doesn't matter, determinism is going out the window, and its becoming standard practice to have bots on our local PC constantly running unknown shell commands
My thinking is, this will increase the demand for backup and other resilience solutions.
This occurred long time ago comrade 'aeblyve.
Marx
Is this new to people? I figured this out when I first entered the industry. The messages have never been particularly subtle.
Seems we are digging our graves as a species and don't even realize it. I mean Sam Altman is already saying it taking 20 years to train a human is a Big Problem.
Oh, many of us realize it, but doing anything about Moloch is much, much harder.
Of course, intelligence will be a solved problem so "20 years of training" won't be needed. You'll just be the hardware. AI will tell you to pick up that box, place it on that conveyor belt, place the autowelder at that seam and wait for the green light, turn the wrench to install bolt B in part C. If you don't wish to, or no longer can, so be it. Another, hungrier human will replace you. After all more are made every day, and they are capable of doing this type of labor by age 10 or so. And what else would they do with their time, go to school and get a completely useless education?
All of this will of course be in service of our technofeudal lords, the owner class. Some robots will be needed for heavy lifting and for the jobs that are too sensitive to trust a human in, like personal security and strikebreaking. Can't risk trusting a serf for those tasks. But for most physical grunt work humans will be cheaper. Shockingly cheap, when they have no other options.
Did that make you less depressed?
If all you have to offer people is this kind of sad fucking "2.5 meals a day and a small shelter" while you live on yachts and eat like a king, eventually they will gang up and kill you
But maybe Altmans AI will break out and do it for us.
I’m looking around the world and thinking this “eventually” isn’t happening very fast.
Not an optimistic thought.
In that sense I see their higher cost as temporary. Another factor: robots will not rise up and blow up your factory. They will always work at the same efficiency for each and every robot produced. They will not go on hunger strikes. They will not kill themselves.
I should not need to justify my existence, that is the problem with being led by psychopaths.
Twenty years to train humans for what? A tech job? That is not why we get an education. It is not my purpose to be a cog in the wheel for some psychotic billionaire.
And then we won't need to, because at that point it will be too late.
Humans will kill us by it damage amplifying their worst characteristics.
Thus we'll die of a pandemic because some idiot LLM'ed up positive looking virology data when they were being too lazy to verify something. Everyone will trust it because they don't really care as long as it looks about right.
Also, customers outsource the risk to their vendors, so as long as there's someone to sue, nobody worries about doing it right. Ship it now and pay the lawyers later.
(In your personal opinion)
Attitude that you can fix broken stuff later and customers are fine with it.
Charlatans and fuckwits everywhere who wouldn't last a second in a traditional engineering discipline.
and thank you!
Mostly we're just called programmers.
We’ve covered so many issues already on our blog (grith.ai)
I self taught and wrote a small saas in 2017. Pays well enough to support me.
I'm building a new one using AI this year. I promise you, it's better built and more secure than what my previous still in use Saas is.
https://github.com/kstenerud/yoloai
I can't go back anymore. Going back to a non-sandboxed Claude feels like going back to a non-adblocked browser.
“The Gervais Principle” is an oft-cited one.
A truly absurd amount of capital was deployed which triggered a cascade of reactions by the people in charge of capital at other places. They are extremely anxious that everything will change under their feet, and if they don't start using as much as humanly possible of it right about now they die.
That's it.
The tools have definitely found some use, there's more to learn on how else they can be used, and maybe over time smart people will settle on ways to wrangle it well. The messaging from the execs though, is not that, it is "you'll be measured on how much you use this, we don't know for what or how, it's for you to figure out but don't dare to not use it".
I do understand their anxiety, their job is to not let their companies die, and make the most money as they can in the process; a seemingly major shift on the foundations of their orgs will cause fear.
But we have not collectively decided that it was safe, and good, to run rampant with these tools without caring for all that was learnt since software was invented...
The whole industry is like a fashion show and has been for a long time. This is just exceptionally stupid compared to moderately stupid things before. I see it ore that everyone's wearing pink feathered chicken suits because it's in fashion. If you don't wear a pink feathered chicken suit then you're a luddite scumbag who doesn't deserve the respect of your peers.
However some of us still have enough self-respect not to be seen dead in a pink feathered chicken suit. I mean I'm still pissed off at half the other stuff we do in the industry. I haven't even really looked at the chicken suits yet.
This is a "monopolized sector." They absolutely forced it on you. In most cases, sure, not directly, but their influence is the only driving force. Absent this no one would have jumped on this flimsy bandwagon.
For ref I'm usually the guy who comes in and cleans up all the chicken corpses after.
People can't just leave Wednesday and be in a new job with the same or better pay next Monday.
Rent/mortgage, bills and food are still due.
Comments like yours always baffle me. Empty idealism.
For the record, I usually get called to clean up and make things good again as well. Most of my career was like that.
They don't have to leave, they can refuse to comply with unreasonable requests which are likely to cause harm by jeopardizing the security of user data.
Maybe your position is too precarious to risk getting fired, but if your job is asking you to do something unethical then you should be doing everything you can to get yourself out of that situation, either by supporting unionization or by being willing to take a manageable pay cut to find a new job as soon as possible.
If you're a software developer then you can almost certainly afford at least a moderate pay cut for upholding ethical conduct. The vast majority can even if we don't want to, but these situations are where we find out if our ethics are for sale or not.
> Maybe your position is too precarious to risk getting fired
You mean yours isn't? Or even that of at least 95% of all devs worldwide? I can definitely say "no" to my CEO if he wants something too big that would take too much time and energy for questionable business results -- I am even expected to ground him. But if my colleagues hand me a ticket, I cannot just refuse without repercussions. I'll not get fired on the spot, that much is certain. But if it happens 2-3 times they'll start looking for a replacement. Same will happen if I outright tell my CEO I can't do something due to ethical concerns. That's how it is almost everywhere I looked and asked (and have very rarely worked with US companies).
> If you're a software developer then you can almost certainly afford at least a moderate pay cut
I can't even afford a 10% pay cut. I want to live in your world.
> It isn't empty idealism.
It is if it's never happening. Pragmatism it would be if it was already an established practice.
I like my dragons purple btw.
> If we don't change, nothing can change. Therefore, we must change.
Obviously. But that "if" is trying to lift an impossible amount of weight is what I am saying. It's one of those powerless "oh, if only!" cries that we the people are prone to.
You ever been hushed by a higher up in a company for talking about compensation? Did you point to the sign required by law as a reminder you have Rights? When times get like this, you have to dredge up the things you've put away because everything was going so good.
Now, it isn't, and your neck is on the block. You will die. You will be offered up for slaughter at the first inconvenience. This is unavoidable. The calculus of business is not something that the ones executing business are going to change voluntarily. They have to be forced to change by the environment. You are the agent that makes up the environment. So your choice is, walk into the inevitable like livestock to the slaughter, or work with your fellow man and take a few chunks of the machine with you. The first step to collective action is accepting you might not ever get to see the shade of the tree you're planting. Once you accept that; the course is clear. It's not empty ideals anymore. It's action. Your action, because you matter, everyone else matters, and it's the right thing to do, and if what's going on is someone else's idea of right, you ain't selling everyone else into it, because that, (what's going on) is wrong.
Ever heard of a Judas goat? Same thing. Herd follows it calmly. Manager's and execs are 100% aware of the dynamic and on guard. The only counterbalance against the dismissal reflex is making it too costly to dismiss all the individual actors at once. We're in the machine. The machine is us. If we don't like how it works... Time for change. I got a lot of days left, and I don't intend to leave the world working like it is, because it is not working for the vast majority of us.
So in conclusion, do what you want. I'm not here to convince someone who doesn't want to be convinced. But I see a fairer, more equitable world where we aren't subjugated by wannabe despots at the top of corporate hierarchies, but partners in making, delivering, and consuming goods and services, to the mutual benefit of all. Not just a lucky few. We were there once, and we can be there again; but we have to accept the way isn't making it possible for a privileged few to set the terms of exchange unilaterally. There has to be good faith. In the abcense there of, there will be conflict. They want things to just work and make them money. We want to eat, and be able to live reasonably well on a living wage without having to lock horns with and wrestle for every damn crumb against a capital wielding class more interested in extraction than being benefactors and stewards of a meta-stable system that serves everybody.
I don't disagree. I really want that reality to materialize. It does not. We have people in very high positions who very carefully make sure it never happens. They have connections, money, resources, obedient and scary enforcers -- they have everything.
While I have you here, I want to make a comparison. There are/were dozens of thousands of devs who commanded $400K for at least 5 years, some for 10+ in FAANG companies, just resting and vesting. They are the ones who should try and sacrifice something to try to better the world. Not me in Eastern Europe who get passed over on the final phases of interviews whose phases I _all_ aced (and got told so in very clear terms) because I said that no, 6200 EUR is not enough for a senior and that I'll start from 7500 at least. Not me who is still renting in this 40s because he was a young overconfident doofus who never learned any money and financial advice (and nobody told him he should; on the contrary, everyone was very happy to exploit me and keep me blind of my own interest) and is now finally working hard to his own ends only -- in a period he should be resting and thinking about the later parts of his life! -- and because he's mostly operating in the EU market, notorious for barely any investment climate and conservative compensations, and definitely not me who has seen first-hand what happens to people rocking the boat.
My disagreement with you is that you invoke some mythical "we the humanity" entity which to me is a cheap way to avoid your own personal responsibility. I don't belong in that "we" group. The FAANG or any privileged engineers are there -- not me. Have you ever commanded FAANG salaries for at least 3 years? If so, and you have not changed anything, then you are directly responsible that the system is not better. Not me. The three total times in my life when I actually managed to gather money to rest for 6-10 months, I used them to just rest from all the crap that happened to me and just recovered physically and mentally. What for? Just to get back into the meat grinder.
What you say is generally valid but you get lost in the bigger picture whereas the everyday fight to change the system is on the ground -- this must not and should not be handwaved away with ideals but with CONCRETE measures, step by step: "allocate 1000 EUR from your next salary and invest them in exactly this and that place" or "use law 1234 and regulation 5678 to get some of your taxes back" or "insist for this contract clause so you are eligible for at least 6 months of severance if you get fired early" etc.
Tell me what power do I have as a contractor. No employment rights. No medical / dental. No severance. I can get fired tomorrow and I have no time to catch my breath. I have to start interviewing tomorrow. Better hope I get the best sleep in the last 5 years tonight! Or else it's not happening.
Executives / people in power just use the "boil the frog" tactic i.e. they tighten the grip 1cm per year until one day, as you said, it's you who is on the chopping block and you are just left confused about WTF went wrong. We are seeing it everywhere, you and I, otherwise we wouldn't discuss this at length here.
> The machine is us. If we don't like how it works... Time for change.
OK, shall I send you my account number so you can support me for 12 months full until I find a job where I feel I can in fact change the world for the better? Disclaimer: it might take 60 months as well. Make your difference in the world! Do concrete measures! Or hell, do it for somebody else -- help them achieve their full potential and recruit them to help the world with you.
Virtue signalling, man. An empty one, too. This is what you're doing. Wishing a theoretical reality into existence so far has not worked for any living human as far as I am aware.
> I got a lot of days left, and I don't intend to leave the world working like it is, because it is not working for the vast majority of us.
Again, that is very obviously true. But it's only theoretical. Everyone is too afraid to not lose their stable income -- and I feel for them. Do you?
And the fun bit is this. We're the ones who have kept it going. If we keep doing that without forcing a confrontation with the dynamics... Already dead, remember? So. There's one meaningful choice left. Start fighting for change, or go down like a lamb to slaughter maybe a little bit later, but get slaughtered just the same. I get you're in the EU. I get shit is largely happening over here in the States. Why do you think I'm a nuisance here in a Silicon Valley incubator's little club house of conniving ne'er-do-wells where everyone comes to swap ideas and get-rich-quick schemes? If I'm going to go down, it will not be comfortable for anyone, because I'm done seeing the world as anything but what it is, and if the only reason I am on this planet is to have conversations like these we're having today, that is exactly what I will do until my fingers, my mouth, my brain, my eyes, and my feet stop working. I can't even give myself the cold comfort of being a starving artist. They at least occasionally bring joy. I can bring only discomfort, and galvanization to change, even though I have a non-zero of ever seeing it.
>OK, shall I send you my account number so you can support me for 12 months full until I find a job where I feel I can in fact change the world for the better? Disclaimer: it might take 60 months as well. Make your difference in the world! Do concrete measures! Or hell, do it for somebody else -- help them achieve their full potential and recruit them to help the world with you.
My friend, I've already done that. Multiple times. I've helped raise children that were not my own. I've stepped in where no one else would. I've moved families back together across continents. I've taken on burdens that others may then run. Now, in a situation I need help, I have pauce hope for another like me to come around and do the same thing for me, but you know what? That's fine. I at least know I'm where I'm at because of my own hand. No one asked me to make those sacrifices. I just did. It's who I am. I give a shit about the welfare of other people, well in excess of my own. Maybe too late in life, have I come to truly internalize, that I should have been much more ruthless with the System than I have been. If it's willing to do harmful things to me, it'll damn well do it to people who can't stand for themselves. The important thing though, is if I can do these things, so can everyone else. We can break this goddamn cycle, but you have to want to, and be unafraid of the consequences. Our ancestors did it before, we can do it again. And in the end, it at least means something, even if I don't get to see it.
Even if that means I have to sound like an old fart for the rest of the time I have on Earth; I have to make sure at no point I slack off in reminding people your personal choices matter. The systems around you emergently form out of not only what you do, but what you choose not to do, and your life really only starts, once you accept that death is coming regardless, and stop running from it, pick up the billy club, turn around, and start saying "No more. This shit ends with me."
You hear not a comfortable man sitting in an office chair. You hear a man, being harried by a System toward the waiting abyss, who is refusing to go down quietly, or with a whimper. I have nothing left than word of the Void which we all are being inexorably pushed, and exhortations that if we're going to end up there anyway, it might as well be from fighting to give those we leave behind a better chance. Something many of us lacked the courtesy of.
I don't blame you. My intent is not to shame you. I only speak that which is the Truth, and the Truth, is often hard to hear. Fight well, friend, in whatever way you can. I know I will be.
People, in general, want to keep their jobs, saying no is an option when you don't care what happens with it or have the backing of the collective to walk out together.
"Hey developers, we spent $x million on Claude, who promised 7x returns, so YOU better make it 7x more efficient so we don't look bad".
At a certain point why blame people for trying to keep up? Why are scammers so successful? It seems to me we have a systemic failure at a societal level. Until we are honest about that it will only get worse. Until then maybe some rouge LLM botching some critical system will be the wake up call we need.
I am not sure what to make of critiques that seem to rest on notions of a small population of scammers preying upon the doe-eyed public. I think the situation is a bit closer to Carlin: garbage in, garbage out. A critique that holds up quite excellently in this AI age.
No.
We don’t say “a rogue plane killed 300 people today when it crashed into a mountain”.
The only difference in the AI case is that some people are attempting to shift blame for their incompetence into a computer system, and the media is going along with it because it increases clicks.
From TFA:
"But the agent also independently publicly replied to the question after analyzing it, without getting approval first."
You can’t authorize a system to take some action and then complain when it takes that action. The “approval” you quoted is not a security constraint. Someone who confuses it for a security constraint is incompetent.
The developer is solely responsible for what APIs they expose to a bot. No you can't say your software agent was grumpy and mean and had a bad day. It is not a human intern, it is an unreliable chatbot who someone ran with permissions it should not have had.
I saw the sea change in 2008 when quality process got replaced with velocity and testing tasks. I've watched everything from Experian and health record data leaks to Windows 11 since that change. Software quality hasn't mattered for a long time.
We're the dummies that have to run around picking up dookies like a new puppy in the house.
It makes it sound like a rogue AI hacked Meta.
Instead, the "wild" thing here is that someone let an agent speak on their behalf with no review. The agent posted inaccurate instructions which someone else followed.
Those instructions lead to a brief gap in internal ACL controls, sounds like. I'm sorry, but given that the US government gave 14 year olds off incel Discords full access to Social Security data, this is not shocking by comparison.
To be clear, it is dumb and rude to let an agent speak on your behalf _without even reviewing it_.
This will eventually lead to a bigger snafu, of course. Security teams should control or at least review the agent permissions of every installation. Everyone is adopting this stuff, and a whole lot of people are going to set it up lazily/wrong (yolo mode at work).
AI use without checking its output (at least at the moment) is firing without aiming. Sure, you can fire really fast. But who cares if you don't hit what you need to? The point wasn't to just shoot bullets, the point was to hit your target!
I mean, you might make a case that enough of them hit the target that shooting fast is a net win, and accept the occasional friendly fire incident. That might possibly be true. Or it might not. I'm not sure that everyone trying to run fast has really done the calculation, though.
Because a human would have been fired for posting something that incorrect and dangerous
If there is a year or two between writing your security fuck up and it being discovered the likelihood of repercussions drops significantly.
It also doesn’t care.
And there was no test environment to validate the change before it was made.
Multiple process & mechanism failures, regardless of where the bad advice came from.
Now, some people claim that you need to improve the reliability of your productive tasks so you can remove the verifications and be faster. Those people are, of course, a bunch of coward Luddites.
The language of this article is a great example, "... thanks to an AI agent that gave an employee inaccurate technical advice ...".
It should more-correctly read, " ... thanks to the people who made it possible for an AI agent to give an employee inaccurate technical advice ... ".
It is at our peril that we deem it acceptable to blame a black box for an error, especially at scale.
Wow, no mishandled user data? A striking change of standard operating procedure from Meta here.
Actually the later information in the story directly contradicts that, so The Verge probably shouldn’t have just quoted this line if their reporting is in opposition to it.
Regardless, this is one of the more insidious things about these tools. They often get minor but critical things wrong in the midst of mostly correct information. And people think they can analyze the data presented to them and make logical judgments, but that’s just not the case.
The article points out that “a human could have done the same thing” but, between the overly confident tone of the text generated by these tools, and the fact that weirdly people trust the LLM output more than they trust other humans (who generally admit or at least hint when they aren’t actually experts on a topic), it’s actually far worse when one of these bots gets something wrong.
<insert takes long drag tweet[1] here>
I personally find "LLMs can do $THING poorly" and "LLMs can do $THING well" articles kinda boring at this point. But! I'm hopeful that stories like this will shift the industry's focus towards robustness instead of just short-term efficiency. I suspect many decision making and change management processes accidentally benefited from just being a bit slow.
https://www.psypost.org/scholars-ai-isnt-hallucinating-its-b...
So I asked AI to give it a good name, and it said “statistical wandering” or “logical improv”.
If I post a question to the internal payment team's forum about a critical processing issue and some "payments bot" replies to me, should I be at fault for trusting the answer?
So to a degree, corporate politics can sort of discourage it.
"Debugging is twice as hard as writing the code in the first place. Therefore, if you write the code as cleverly as possible, you are, by definition, not smart enough to debug it." ~ Brian Kernighan
"We cannot solve our problems with the same thinking we used when we created them."
I'm pretty happy with the team I've built. They make solid decisions that I can trust every time. I can't say the same for the LLM.
That is politics. Not engineering.
Assigning a human to "check the output every time" and blaming them for the faults in the output is just assigning a scapegoat.
If you have to check the AI output every single time, the AI is pointless. You can just check immediately.
1. Check frequency (between every single time and spot checks).
2. Check thoroughness (between antagonistic in-depth vs high level).
I'd agree that, if you're towards the end of both dimensions, the system is not generating any value.
A lot of folks are taking calculated (or I guess in some cases, reckless) risks right now, by moving one or both of those dimensions. I'd argue that in many situations, the risk is small and worth it. In many others, not so much.
We'll see how it goes, I suppose.
There is a point to using LLMs. They can save time by doing a first pass. But when they do the last pass, disasters will follow.
The AI "led to" the incident , true. But do nt forget that this, like all similar incidents , is a human failure
AI is a tool with no agency. People make mistakes using it, thone mistakes are the responsibility of the humans
Claw AIs absolutely do have agency in the sense of being able to independently perform actions on their own, based on their "understanding" of a goal given by a "principal". I can't think of a better word than "agent" for that.
Can you try to define more clearly what it is that you believe AI agents don't have?
Can you perhaps share a archive.org link if possible?