659 points | 4d ago | Discuss on Hacker News | Back to Radar
And the cost to fix this is gonna be $12-billion. No wait, $30-billion. Fuuck it, make it $50-billion. And give it to me now or else!!!
Ultimately, yeah (with the help of GWB) https://www.nsa.gov/Press-Room/Press-Releases-Statements/Pre...
Not unlike Bloomberg terminals…dos era UI but does the job well
Definitely just a big coincidence due to firing people, upgrades, and AI.
...maybe.
Seriously though at what point does this become justifiably suspicious?
The truth lies somewhere in the middle. My theory: someone figured out how to bring down each network. And is progressively breaking each system to help developers fix their security holes. Given enough breaks, we will figure out their modus operandi. And then figure out who they are.
Or something even more conspiratorial: us govt is halting each system to install back doors.
But give it another year, trust me bro.
It does? Can you share an objective metrics that can highlight this that doesn't focus on single, isolated events?
I know you'll provide this because of your strong anti-"trust me bro" stance and dedication to scientific fact.
The problem is much more systemic than having 1 decent org.
Who decides when a DDOS of Trumpcare.gov that is traced back to Akamai servers was an attempt to extort government contracts or “those pesky Chinese state actors”?
Compare to that, most people in industry at this point are either business disguised as technologist or in some cases experts in one vertical.https://google.github.io/building-secure-and-reliable-system...
GSA's 18F played a major role in applying Site Reliability Engineering (SRE) and modern DevOps practices to federal government infrastructure.
https://www.gsaig.gov/sites/default/files/management-challen...
The only references to 18F are in a footnote citing two reports about 18F done in 2016 and 2017...so very much during Trump.
if only hunger wasn't an issue
On a serious note, I see it as a paradox. Competency is built through struggles, and the non-competent do not want someone that seemed lesser at some point to outrank them, be the go to person.
> according to six people
Meanwhile, 24 hours later, we know the headline being questioned here was indeed accurate. Your heuristic will presumably now adjust somewhat, as to misinform people less in future.
Also consider that the media absolutely loves to run any story it can find about a network going down right now because it causes fear, and fear gets clicks. Now think about how many major networks have a problem in a given week. You just never hear about them until now.
No and that was the purpose of DOGE
(yes that quote was separated from its context with some malice)
No I don't. Why would you risk your career and potentially your freedom to leak a report about enemy action, in which doing so could potentially harm your country too?
https://www.cnn.com/2026/11/16/politics/us-intelligence-iran...
[0] https://harvardlawreview.org/print/vol-127/the-leaky-leviath...
I doubt this is the case, but since we're talking about conspiracy theories: this looks more like things breaking by accident as things are shimmed for control by a third party and less like a focused attack.
It's exactly like the media reporting every aviation notice after an airplane crash. Leading people to speculate "why is everything broken right now!" even though it's the normal course of business, they're only just noticing it now because they're being told to notice it.
https://www.nsa.gov/portals/75/documents/news-features/decla...
If the network was one of those no-fail important ones (think NORAD stuff) then nobody would be talking about it as doing so would be treasonous. So, at most, this is a less-important internal network. It will probably be back up by end of business.
Anything the catches the public's collective conscious, the media is there to stoke it into piles of ad dollars.
It's extremely profitable to write stories telling social media pitchfork mobs how correct they are. Influencers have been catching onto this game in the last decade too, and went from "here is what is actually going on" contrarians to "here is what is actually going on" agenda towing. It just pays so much more, and people trust you more (because people naturally trust those who tell them they are right).
> The frequency illusion, also known as the Baader–Meinhof phenomenon, is a cognitive bias in which a person notices a specific concept, word, or product more frequently after recently becoming aware of it.
Or at least it is very similar.
https://en.wikipedia.org/wiki/Gatwick_Airport_drone_incident
>> Flap, originally a military term to describe a "period of panic or chaos", is used by ufologists to describe surges of sightings in one geographical area. Longer surges spanning multiple countries may be called waves. https://en.wikipedia.org/wiki/UFO_flap
Including NSA Net? How often does it fall down?
> We are probably only hearing about this because of the coincidence of the other recent network failures
This seems to be an explanation for why the press is covering it - this time. Which also hints that NSA Net failures happen with some kind of regularity and that that news orgs had been getting wind of it.
I suspect that the opposite is true. That it's rare to unheard of for NSA Net to fail like this and rare for a news org to get wind of something so deeply protected.
I rather doubt they have a public status page.
> This seems to be an explanation for why the press is covering it - this time.
That's one possibility. Another is it goes down sometimes, but that morale in the administration has sunk to where it gets leaked.
If more people misspell than not, then what do the predictive statistics result in?
It peaked at early mature swype, and has been downhill since then.
Just like good crypto must survive robust cryptanalysis, good network security needs to be robust even when adversaries know the network exists (and perhaps even where in physical space key nodes are located).
That being said: the network being down isn't something that immediately makes me worried in the near term. "Turn it off" is a better first response than "observe if anything seems weird" for communications channels that prioritize secrecy over availability.
NSANet is one of the most locked down, observed and monitored networks on the planet
https://www.whitehouse.gov/presidential-actions/2025/11/laun...
Coincidence or is there some common explanation that I missed?
I think the opportunity right now where we have powerful AI hacking systems but not enough protections is to establish long-term presence in all networks of interest. Not bring them down.
For example, we updated a GCP managed base image of Ubuntu that was released with 350 CVEs patched, some were as high as 9.8. In doing so, they (GCP) changed the DEFAULT setting for hardware GRO from off by default to on. That's a massive networking change packed into a security update.
We only caught it AFTER it started causing issues in production. We didn't have the OS pinned because (decently valid reasons) or catch it in non-prod (decent reason + human mistake).
This is on a modern highly automated tested resilient well architected platform.
I can only imagine these other platforms are dramatically more exposed that we are, hence these common large scale outages.
More plausible to me is that spy agencies had previously found and used some of these hundreds or thousands of vulnerabilities to hack into those networks just fine. But now other people like cyber criminals are using these vulnerabilities via AI. That is messing up both the attackers who had already gotten in quietly and do not want the attention and the defenders who now have to respond to a noisy intrusion they cannot ignore.
Outages could be grey-hat testers and emergency updates.
A temporary period like we are experiencing is to be expected when moving to a time where a single engineer can perform the equivalent work of a team of excellent engineers working for weeks. Security has always been a game of who has the capabilities, means, and requirement for any given attack surface.
The site has pretty robust defenses, so nothing has broken through yet, but a penetration or an effective DDoS feels inevitable as the probing steadily escalates. I expect that juicier targets are seeing a lot more of this kind of activity, possibly leading to some of these outages.
I run rate limiting on my nginx config instead which helps a lot more to prevent site issues from too many requests.
Granted I am on the free cloudflare plan, maybe the paid ones do a better job.
"If you don't read the news you're uninformed, if you do read the news you're misinformed"
The sentiment is quite old, however. Thomas Jefferson wrote in 1807: ...the man who never looks into a newspaper is better informed than he who reads them; inasmuch as he who knows nothing is nearer to truth than he whose mind is filled with falsehoods & errors.
Where it gets interesting is the "Exclusive" part of the headline. Like somebody else said, any other source would either have to have access to people similar to CNN's "six people familiar with the matter" or else reference CNN.
The fact that Reuters amplified the story (and attributed it to CNN) gives it a little more weight.
Though at the same time, the public isn't privy to even 10% of the incidents going on right now.
The information I have is consistent with general cybercrime and script kiddie patterns, just a much higher volume, and much larger diversity of targets.
https://www.bleepingcomputer.com/news/security/optimizely-su...
Is it possible that the false of productivity is leading to worse products and massive loss in valuable headcount? Instead of AIs directly writing bad buggy code? Or is "everything rage-bait, and AI is all sunshine"
It would be interesting if someone could do a proper analysis, though!
There's a group of people who think AI is perfect and will solve all of humanity's problems. That's annoying. There's another group of people who think AI is the worst thing to ever happen to humanity and is the downfall of society. These people are more vocal in my experience and they're everywhere. If you ever use AI you're a bad person who they'll disown like you're racist. It's somehow become the most popular 'social justice' issue of our time even as ICE exists alongside it. These people see a problem and immediately try to figure out how AI is involved. That's far more annoying.
Best example I can give that came up in my circle recently is the video game "Big Walk". It was recently denied for a silly award nomination because the awards have a strict no-AI policy. Which to me is absolutely fine, go nuts. But the vitriol towards this game, that everybody absolutely ADORED before the developer disclosed that they fixed ~10 lines (!) of code with AI because they couldn't figure it out otherwise, is absolutely bonkers. People I respect treated it like they were mourning something that WAS great until this catastrophe happened.
Funny you mention Swartz. He was entirely on board with "information wants to be free". He explicitly condoned piracy in pretty much every form. Also there's many ongoing legal problems for these AI companies that address all this. Anthropic is already on the hook to pay 1.5 billion for pirated books alone.
But this is going to be the issue of our time. If we make it past this point in history (I'm skeptical) I have a feeling that the "stealing" of information in such a way will seem comical to future generations. We'll gape in awe at how everyone was in a rabid race to hoard THEIR information to get pieces of paper and that we basically destroyed the only planet we have to live on in service of having more of an imaginary "value" (whatever value means on a dead planet).
But I guess I'm of the opinion humanity is ultimately doomed because of our greed. We've been warned and are watching in real time as climate change does exactly what we were told it was going to do decades ago and our answer to it is to put out even more emissions (drill baby drill!). IF AI can cause a technological boom that can maybe fix some of that damage perhaps all the "theft" was worth it? If not oh well the theft probably won't matter for much longer. And yes I know AI is causing even more emissions, but compared to the overall problem that's built up over 200+ years it's basically negligible.
In regards to mass surveillance that was happening well before AI. We all have phones in our pocket that track what we do at all times. You haven't been able to buy a car that doesn't track you for a long while now. Unless you're using cash you purchase things with tracking devices. We ask all of our questions to Google, an international behavioral tracking company. Does AI make that worse? Yep. And people are finally paying attention and getting Flock the hell out of their cities. Hopefully that will go far and people like Facebook, Google, etc. will finally be prosecuted for their decades of crimes against humanity (doubt).
You can say all the things you just said about AI about the internet as well. I can go to anna's archive and steal all the books I want. I can get any movie, piece of software, whatever else that I please. It enabled all the crimes you mentioned and much more. The difference with AI is that it might just be the most important thing since the invention of fire.
When you steal a movie from archive.is and watch it what is the result? You stole .00012c from copyrightholders? Are you stopping them from making new movies or taking new revenue streams or setting up your own theatre to profit? no.
When an AI companies steals decades of information, uses it to produce copies, and sells those copies, it is monumentally different.
Aaron Schwartz was supremely aware of class issues in this country, and did not believe meritocracy was real. I sincerely doubt that he would be in support of the upper classes stealing the wealth of information from the lower, and using it to make themselves even more tremendously wealthy while skirting the law. The point of my illustration is that this is a class issue, corporations are viewed as ultra-persons capable of skirting the law all the way to profitability to serve their masters. Meanwhile we are all little peasants that get to watch it happen, but if one of us tries to do the same we will see the inside of a prison cell.
It really is another form of theft the elite have managed to repackage as innovation and improvement.
The fact that SWEs fall over themselves flocking to this technology because of the so-called productivity gains while simultaneously fuelling an industry that will not only undermine their future career prospects, but is actively profiting from en-masse dragnet theft of IP and public information is mindbogglingly and monumentally stupid.
It would hardly be shocking that hostile nations are able to cause some problems with Mythos level tools.
If anything, it is more interesting how we can invade another country and the idea that the country or a collection of hostile countries could hit us back in any way, shape or form is just an impossibility. An insane conspiracy theory.
Maybe it is not that but surely the probablity is not zero.
Naturally, when you tell an entire occupation that they are no longer necessary employees and that AI writes all your code, when something breaks in a major way those people are probably going to ask the question.
(https://www.cnbc.com/2025/04/29/satya-nadella-says-as-much-a...)
Summer of 2019 was also a pretty good one [2].
I fully expect AI will cause insane disruptions at some point, but all evidence points to this just being a series of normal outages. Worst timing ever to have a patch Tuesday.
[1] https://www.csmonitor.com/USA/USA-Update/2015/0708/NYSE-Unit... [2] https://www.thousandeyes.com/blog/looking-back-biggest-inter...
As a developer, I know very well that systems break, sometimes frequently (AI or not). It has always been the case, and likely always will be the case.
But, I am not surprised that people are mentioning AI - feels like most posts on here for the last year have involved AI in some capacity.
I'm sure most of us here have been on the receiving end of managers suddenly obsessed with a failure mode because it happened to occur twice in one month.
AI is tuned to speak confidently and ship independently; it is rare that it will suggest that one go back to a broad survey of stakeholders to ensure no Chesterton's Fence is being torn down (https://fs.blog/chestertons-fence/). "I've read the documentation and logs, and nothing suggests this system needs these legacy capabilities/components/capacity" is something we hear all day, and it is a siren's song if there ever was one.
Couple that with a general (and valid!) consensus that modernization of legacy systems must be accelerated for a variety of reasons - to mitigate this new and continuing "Black October" of hyper-agile agent-driven cyberattacks (whether the network is public or private); to have more flexible infrastructure for new features to stay ahead; to allow human teams to show their efficacy in a world where technology jobs are increasingly at risk - and you have a perfect storm where change management is no longer what it once was.
I think you went off somewhere else.
Which is to say, a parent post here on HN doesn't narrow the topic to it's focus. That's what makes discussion here interesting.
I suspect the situation is that AI's caveats are just easier to ignore than a person's warnings. Plus all the other factors you list. Also, AI tends to give a person the kind of communication they ask for. I always ask for options but I suspect other people ask for just answer and so they get them.
Asking the coin-tossing machine if it's safe until it says it's safe. :)
Doesn't stop so many people from believing it wholesale, and doesn't prevent so many programmers from saying "I don't review code anymore"
Also the amount of code that is being produced (at least in some places) is almost unreviewable
Some weeks ago, our Claude subscription failed - most people stopped working and switched to tasks they had postponed.
There was this tv program in the 90's about a fully automated japanese warehouse. The staff was one maintenance guy. They ask him if it was a lot of work. He said, everything is new then he made the face like a puppy to fill an awkward silence.
I've often joked about the mechatronics fortune in our train yard, they need to study hard, get the diplomas, get experience. Then the only tool they get is a flashlight and they have to walk around indefinitely not finding anything. Hundreds of employees like that. A good day is when they all find nothing, there are many good days, they are not suppose to find anything. If anyone does find something it most probably wont be you. They apparently still find enough things to prevent the full switch to AI and those pretty line scan pictures.
https://upload.wikimedia.org/wikipedia/commons/7/74/Line_sca...
The new (modernized) trains have IT guys, they usually have red/purple faces and run up and down the train with their laptop. They plug it in all over and the color gets brighter. Bright enough that I don't make jokes about it.
The FAA's recent outage for example, at its core, seems like people trying to rush needed upgrades that were slow for a specific reason (e.g. reducing the testing window for 90 to 30-days).
The Microsoft outage seems like three different things linked together (and poor "full failure" recovery prep, lengthening the outage).
The bank "outage" may not have even existed, 2/3 of the banks claimed they didn't have an outage at all, and the other said it was a limited technical issue. The articles were based on DownDetector spikes.
This has reached a Witch-hunt levels of fever pitch.
You'll commonly get regional failures, when failure happens.
You're out here categorically saying none of these had an AI or cyber-attack angle with no verifiable proof and having a tantrum that other people are doing the same thing in reverse.
The list of unanswered questions is long. Maybe focus on pushing for more details instead of immediately saying "nothing to see here" regarding a series of important failures within a week's time.
BTW here's JPMorganChase's statement: > JPMorganChase has experienced a temporary issue related to a software bug in on-premise hardware systems – there is no evidence this is cyber-related. Where we were party to a transaction involving other banks, those banks may have been impacted, but this is not a systemic issue. Customers still have access to their accounts and funds.
If you were trying to send your rent payment via Zelle from your Citibank account to your landlord's Chase account and it continually failed yesterday with no clear reason immediately provided, of course any complaint is a "witch hunt" since your Citibank account "works."
This is just a media hype train. Zelle outages have happened many times before but the stories didn't blow up they way they are now.
So let me see if I understand what I should be:
- Be more skeptical of official explanations, including where it includes a technical breakdown of how it occurred (e.g. Microsoft/Outlook outage).
- Be less skeptical of random commenters claiming, without proof, that it is AI or Cyber Attacks.
Because that's the implied point you're making. Be more skeptical of A and less skeptical of B, even though A has someone placing their reputation behind it, and B is anonymous commenters throwing made up theories out into the ether.
it wouldve been easy for the faa to blame their new AI vendor or these companies to blame some new ai process set up by consultants etc...
The explanation for this one still doesn't ring true for me and that's exacerbated by how many times the story has changed after telling the public that they didn't need to know the reasons.
Agree, though, that the rest do have reasonable explanations
So this outage seems a bit overdue to me.
Is "Claude says it looks good so let's skip the usual testing window and push to prod on a Friday" really not a sentence you can imagine being said in this current environment?
My point is that people are taking a potential future problem and applying to actual things that have recently happened, that didn't fit that mold. Which is incorrect.
When AI or Cyber Attacks cause an outage, people are absolutely right to say so. But claiming - without proof, that every outage is related to the two is noise.
In my line of work I deal with a lot of banks. They have been absolutely panicked for the last few months as they have been using AI to find and exploit bugs in their own software (A huge number of banks are members of project glasswing and other AI cyber testing with less limits than the public ones we get). This has forced them to drop almost everything else and immediately address updates in their systems, especially to a lot of legacy software that's sat around forever.
Just this many updates is going to cause instability in itself. Old systems that aren't touched can run forever without many problems, but when you can RCE said app the normal 3+ months of testing goes out the window. Now imagine this with tons of bugs in hundreds or thousands of applications at a single bank and you can see why things are starting to get unstable.
You have aides and resources and money. Not having enough information or understanding is a choice.
It's criminal negligence.
I can't put my finger on it, but something sinister feels afoot.
https://beready.campaign.gov.uk/
It's nowhere near the kind of readiness and preparedness campaigns of the cold-war of course, but it's a step up in alertness that's higher than anything I can remember.
[To be clear: I mean exactly what I said, which is "there could be" which is not the same as "I think there is" -- so call it a 5%-10% chance of real cyberwarfare -- which means I'm not dismissing those claims out of hand, and find the claims at least plausible, even if I don't currently believe them]
It’s not nothing
But my gut says incompetence and rushing upgrades out of fear of cyberwarfare, not an actual attack.
Too many people are addicted to their smartphones and don't feel the same level of pride they used to in their work. All systems are shared, instead of one person being the ruler of a certain part of the stack, it's a revolving door and it's rare that anyone stays somewhere for even 5-10yrs. It leaves a wake of modifications that nobody else can keep up with. Also, the tech industry is moving fast just to move fast, a lot of old infra is still C++ and COBOL but good luck hiring anyone to use that, so they have to convert the systems to use the Latest Tech of the Day. The old, stable people don't have any interest in learning the new thing, and would rather retire with their knowledge. The new people want to rebuild from scratch in the New Thing and are convinced they need to job hop immediately after.
There are now post mortems for both outages. The FAA outage was caused by a botched upgrade to an ancient system. The MSFT outage was caused by a latent concurrency bug that was surfaced by a later incident. My experience in this industry tells me that this is plausible because these are exactly the type of things that tend to cause outages. My industry experience also tells me that hundreds to thousands of people were involved in the responses to these incidents, so if the companies were lying about the cause it would be leaked to the press and posted about all over Blind.
The other outages all seem to be normal nothing burger outages that happen all the time and nobody ever hears about because they don't make the news or even the front page of HN. The reason they make the news now is most likely because we just had the MSFT / FAA outages and that's in the news cycle now, not because they are somehow linked.
I can put my finger on it, and nothing sinister is afoot.
Frankly, for the past few days, my gut is pushing a hypothesis to explain why nothing happened yet: bad actors are holding off until the winter starts.
The combination is a cycle
1. AI finds bugs
2. Engineers us AI to try and fix
3. AI testing approves, minimal engineering review
4. Engineers push release rapidly, it's critical after all
5. Bugs introduced
6. Cycle starts again
I also think generally that AI is probably both the cause and solution to these problems. We're certainly getting more code churn than otherwise, and in lots of cases that's really good! But you're going to have to deal with inevitable instability too.
But I'd note, these "named after a thing" phenomena (eponnyms) are kind of terrible for exact description since their definition tends to be fuzzy, there's a temptation to fudge their borders and they create an insider jargon (if only I could a device that could reach of the computer and slap everyone who says "steelman" when they mean strengthen an argument, etc).
With all the recent asshattery of leaking kernel exploits, I would presume trust is gone, and it's cover the big boys first.
(a linux kernel bug vulnerability could be patched without lkml involvement or distros even knowing)
So certainly there is a big social-psychological phenomenon. Or it's gonna be a cold winter! :)
(This of course is on top of whatever unusual things might actually be happening)
I think the real ones were the Microsoft and FAA due to deploying the AI slop.
The other ones seem to be usual outages that normally no one reports on.
https://en.wikipedia.org/wiki/Frequency_illusion
Combine this with news reflecting both supply and demand for stories, it's impossible to tell. There's certainly a constant supply of outages.
I guess this is what the run up to Y2K was like but with less time and orders of magnitude more systems?
It's not quite Baader-Meinhof because there is deliberate action involved, but it can also make a trend appear out of what is actually random noise.
However, AI works like a speedy intern who hasn't yet got to grips with the whole codebase, and sometimes makes assumptions. That's mostly because the codebase is usually bigger than the context window, so it is 'writing blind'.
1. All flights grounded
2. Microsoft outage
3. Optimizely malware
4. Classified NSA network down
I don't know what, but something is clearly happening.
News professionals would never be granted "need to know", and therefore would never be granted any level of secure access at all.
There are a couple of things to keep in mind with nsanet though:
- It can handle anywhere between confidential to TS//SCI information
- It is largely (completely?) an off-infra network, meaning that it doesn't run over the same infrastructure that the open internet uses. It uses an entirely(?) discrete backbone. SIPRNET uses similar architecture.
- You cannot easily cross pollinate data from unclassified networks to nsanet. Everything from removable media to EMSEC is considered prohibited without proper channels being informed and paperwork being performed.
- Lastly, we're unaware of the timing of this supposed source. This could have been over the last week or a month ago or <n>. Scant detail is given, so I'm unwilling to believe "lolz nsa pwnd" just yet.
I would not be surprised if something similar was involved here. It worked with stuxnet, too.
Unspecified computer parts were interdicted by US intelligence while being shipped to Neda Industrial and infected with Stuxnet. Four other external vendors working on the enrichment plant were also compromised, likely by Israeli intelligence physically entering buildings.
Stuxnet then spread across each companies internal networks and onto USB drives that were then carried into the facility and plugged into the air gapped network.
Beyond that it's also typically blocked by GPO prohibiting USB, so even if that were the case, the net result would be no real data transfer and a nice email to the DA mailgroup about the workstation and logged in user who triggered the alert.
https://gdmissionsystems.com/encryption/taclane-network-encr...
none of these systems are even vaguely similar
AI-accelerated attacks, I think most of the cybersecurity ones are attributable to that exploiting previous incompetences. But the flight grounding and microsoft outages? Looks more like direct incompetence to me
Either way I think the takeaway is that we've been slacking for a long time and now we have something that doesn't need man-hours, caffeine, and obsessive human focus to take advantage of that fact. At least for those of us who have been in the field for a while there's lots of work cleaning up these messes for the foreseeable future but I been on calls with peers of mine in infosec who have definitely been losing valuable sleep and health over it
Of what?
It's easy to spot patterns when paranoid.
If four politicians die on the same day of heart attacks, maybe they're unrelated. Maybe they're chemical assassinations.
Same here.
> Because we don't know they're unrelated.
They might be related! But the burden of proof is not on the null case.
Aren't these being treated as "normal" outages by the media and powers that be?
We collectively are very much in arm-chair speculation at this point. If this is a coordinated cyber attack the US government right now wouldn't want to demonstrate weakness or fault. Iran would be the obvious culprit... but going full arm-chair on speculation... There was a major and disappointing result in our elections to the current powers that be...
If this was a staged set of attacks carefully done to minimize casualties... I bet it would be enough to declare a state of emergency and get additional executive powers "legitimately". If that was the case I would expect more propaganda and less bury it under the rug if that was going on.
Be careful out there the world is scary. Don't take any of this as fact, this is all personal speculation based on my own observations. Let me stand-up from my armchair and go for a walk...
- Epstein/SV network within the US flexing muscles in response to mid term losses
- Iran, obviously
- China, obviously
- Russia, obviously
- CIA attacking other parts of the deep state in some internecine struggle or other
Literally all of those are completely plausible scenarios where the actors are motivated and capable
Based on my personal experience and to a certain extent sitting back down in my arm chair. I don't believe this is either China or Russia, they both have priorities that don't immediately involve the US and (as a government) the US is actively supporting both right now in very desirable (to them) ways.
Internal US action and Iran both seem equally likely in my head right now, but this is purely speculation and attribution is the hardest part to actually nail down when a state actor is involved.
This and the continued outages, such as the new Apple services outage, all collectively feel to close together without some evidence of a shared infrastructure causing it or a coordinated attack.
I don't personally believe this is "AI agents hacking the planet". The companies involved (well at least half of the outages I've seen) have enough compliance and protection obligations they are regularly audited against that my Occam's Razor gut feeling tells me that this is either long established persistent and maintained access being detonated one-by-one OR its a focused attack by one or more teams under the same directive taking down significant US focused institutions for attack or reprisal.
Have you considered the possibility that this isn't even the detonation/attack? That this is testing and/or a modest demonstration of capability.
If these are attacks, they don't at all feel like they're at the hands of people who have shot their wad.
For instance, I have heard that Iranians have developed rather sophisticated cyber-warfare capabilities. I don't know if that is true, and if true, I don't know how likely it would be that they would have reached capabilities to cause the observed outages. But even if you assume that the probability (or rather degree of belief) is quite small but still realistic, say between 0.1 and 10 percents, than considering the current situation it might be absolutely rational to see it likely that they are behind these outages. The full analysis requires estimation of the probability that this is all just a coincidence, and of other major "non-coincidental" branches. But one can do this at various levels of precision and explicitly observe how the output depends on the priors and the estimated probabilities (and therefore, on the assumptions).
First of all, I have not said that just assigning arbitrary probabilities sheds much light on the matter. Rather, what I alluded to, is that even rough estimates might often shed a lot of light on the topic. The Bayesian formulation helps to separate the problem into relevant parts, make the relative importance and interdependence of various factors that enter the reasoning clearer, and makes the assumptions more explicit. But in my post I have not really explained anything; I have just pointed to a tool somebody interested might use to critically assess similar issues. So yes, if you took it away, the post would lack an important part of its message, namely a pointer to a framework that the OP (or anyone interested) might find useful. However, I do not plan to give a treatise on Bayesian analysis in a HN post.
As for the other angles of your post: 1. Yes, the usage of the framework absolutely can shed light on many matters of this kind. What is your background? How much experience do you have with statistics or inference in general? There are many good sources where you can learn about these methods. A somewhat controversial one, but freely accessible, is E.T. Jaynes' Probability Theory, which actually directly addresses some aspects of what I spoke about. 2."it just seems to make explicit what everyone already does implicitly" People apply very bad judgement to matters like these all the time, this forum included. And even if people all used those things implicitly (which often they don't) it would still be worthwhile since it allows one to inspect the argument and find flaws or observe what are the dependencies on various assumptions.
It has come to my attention that the political elites and the Epstein class simply do as they please with virtually no consequences.
They don't. This administration thought they could Venezuela Iran from the air.
Even more importantly, when these pretextual actions occur, everyone interprets them the same way and agrees that the secretly-planned action is now the one we must take.
It's a big change so hiccups are expected.
I don't think it's a stretch to say that some number of services go down every day and we don't hear or care about them. Three or four services going down and/or being attacked over a few days is only a pattern if it's actually different from before vs. you're just more aware about them.
While this rule doesn't apply to the NSA, speculation that recent down time from these U.S. companies being caused by state-sponsored cyber warfare seems unlikely.
After the very deadly crash in Spain a few weeks ago...
You don't need a spy if the target self-sabotages.
AI cyberattacks, foreign interference, and Ghost in the Shell type stuff sounds cooler though.
That's basically what's happening to the US right now. It is corruption from the Oval Office all the way down. Pardons are being openly sold. The news was manipulated ("a ceasefire is imminent") to profit off the oil markets. Stock trades on insider information abound. All of this is an inevitable consequence from a Supreme Court who manufactured out of thin air substantial criminal immunity for the president. This built on other decisions where this Court in particular just could not see corruption (eg Snyder v. United States [1]).
In the private sector literally nobody is ever punished for cost-cutting on systems to the point that they break. In fact, they're rewarded for it. It's entshittification from the CEOs all the way down.
We'll probably find as more of these incidents happen that AI code was written and reviewed by another AI, never tested and deployed to production. Foreign actors will be blamed, Congress will be blamed, etc. As an aside, the DoT is asking for $50 billion after the FAA outage (the worst in 25 years). Then Representative Mullin voted against $25 billion for the FAA in the Infrastructure Investment and Jobs Act in 2021 [2].
Now the administration wants to sell this narrative that they've been hacked. No, they're just corrupt and everything sucks as an inevitable result. Everything is shit. Nobody is hacking us.
I honestly don't know how we get out of this but we need a hard reset of some sort. One such hard reset was the French Revolution. There are an awful lot of people who need to spend the rest of their lives in jail, penniless.
[1]: https://us.transparency.org/news/supreme-court-undercuts-fed...
[2]: https://heritageaction.com/scorecard/members/M001190/117
While all flights grounded is significant on it's own, and would have got press, even in a world where the fear of AI or foreign interference wasn't so strong, but there's been more than a few substantial local and regional groundings due to software over the past 4-5 years.
Microsoft outage is not new, Teams and Outlook have had more than a handful of outages way before the current doomer craze. While I had to search these for accurate recall, having worked closely with Microsoft for a decade plus, the recent outage is not isolated and to jump to conclusions is short sighted. For example, pre-doomsday summary of Microsoft's noteworthy outages: (1) Azure Outage -Sep 2018 South Central US (San Antonio DC); (2) Recurring multi-day Teams outage that effected users world wide February - April 2020 (3-4 outages in total); (3) December 2021 - Azure Active Directory outage locked people out of Azure and Outlook for several hours. (4) July 2022 - half-day outage for exchange, teams, sharepoint for most users. Same span outlook and exchange were down much like this last event for several hours. (5) Jan & Feb 2023 several outages for globoal Microsoft Outlook and network events.
That's not even all of the events in the last 5-6 years.
Optimizely, NSA network, and FAA mostly would have been under reported and or ignored by the general public.
I'm not saying this time couldn't be different this time, but what is the thesis? AI, foreign actors, incompetence?
"I don't know what, but something is clearly happening" is fear talking not rational thinking. Not all dots require connections. If you had full awareness of all the wild things that go on in one day, the amount of correlation you could draw from that if you're trying to tell a story would be pretty salacious.
But correlation != ____________
I agree that correlation doesn't imply causation. At the same time, I think an increase in uncommon events over a relatively short period of time does increase the chance that they are correlated.
Does it increase it enough to make it plausible or likely? No clue. But we've seen similar in Europe, and the consensus among intelligence/military analysts seems to be that Russia and Europe are in a non-military conflict of some sort.
It might be just bad luck here, or the start of something similar.
I saw some speculation here [1] that it might be the root cause.
Because then you have to define the "major" and "short" in "major outages in a short timespan", in such a way that
A. Includes all four of those B. When applied to the past, would actually, meaningfully differ from the status quo (How often can something happen before it's alarming? Once a decade? Twice as often as that? Half as often?) C. Isn't a clear overfit, threading the needle to meet A and B, but is actually robust against perturbations.
You will likely be right that this is a meaningful period, simply from the flights being grounded and the Microsoft outage, both meaningfully unusual events, but it's very easy to make a fairly challenging assessment seem simple if you just arbitrarily choose things to count on your fingers.
'Mr Bond, they have a saying in Chicago:
"Once is happenstance. Twice is coincidence. The third time it's enemy action."
[From the original Ian Fleming novel]'
What I however know is that somehow the human brain likes pattern matching. We look into the sky and see patterns made by the clouds. We see some random cloud formation, match it against 100000 patterns and say "This is a car". We look at ink blots on a paper, match it and say "This is a knife" (Rorschach test).
Nothing of this is true ... it's just that we matched a pattern. And that he hold something plausible.
But ... thinking something is plausible is far from being it true. To draw an actual conclusion, we need a rigorous thinking process. Based on logic, not assumption. With facts, not assumptions.
If I would have to describe the current times (e.g. the last 15 times) then it would be the sudden lack of all this virtues in people engaging with other people. We seem to be certainly back in the old times, where rumors are hold to be true ("This red haired lady with a dark cat is certainly a witch. Burn her!"). And, worse, often people base actions on this hearsay, so we are also a bit back in the Lynch justice mindset.
> I don't know what, but something is clearly happening.
And I know, something IS happening. That's an inherent truism. Example: You, dear reader, are breathing. Breathing is "something". It is happening. Clearly even. I mean, if we start to write extremely vague, then another human matching pattern can see anything in it. We shouldn't talk in horoscope language and we shouldn't assume that horoscope language contains anything of value.
Did you try sending it via outlook.com?
Showing 400 of 747. Read the rest on Hacker News
Comments are loaded live from Hacker News and are not stored by Mid or Real.