We have a year to fix security everywhere
jyn.dev
jyn.dev
It's not the end of the world. But future will be rough.
Short term you’re probably right, but longer term is the realm where nation states will start to police the avenues of attack.
This is what will lead to govt needing to attach an actual ID your network connection.
I think it’s a bit like frontier development (like the US “Wild West”). You rob a bank because there’s no one to stop you, and even if you do get identified you can travel enough distance to regain anonymity. Application of legal recourse eventually caught up (as it will here), and the growing pains will certainly make things suck for all of us.
I’m not saying we’re even remotely close to this, I’m just saying State-level coercive action is not unheard of if a problem is perceived to be significant enough to warrant it. Shit, it even only needs to be viewed as significant by a small subset of the governing body (see: Iran conflict, or current pushes for “child safety on the internet”). It just has to be “useful” to a certain body politic.
Even it's $10,000 to run today (FWIW, the featured article cites the M5 Mac Studio with 256GB unified memory going for $9,500 as "good enough to host something scary"), in a couple years it'll be like $2k to run, and in another couple after that, you'll have used $200 dollar smartphones capable of running a model powerful enough to do serious damage.
Once a vulnerability is discovered however if it's in your own software a patch has to be written (without reducing functionality in most cases), tested, and deployed. At every step there will be others arguing about whether this line could do better, my service requires this thing that isn't included. So at every step the patch can be delayed.
And if it is someone else's software you will be lucky if it's open source and you can write a patch yourself. If it's closed source or a vendor you have to completely rely on them and use whatever your account rep can pull.
Attackers have a massive advantage with AI, partially because the defensive side doesn't want to make their side worse by giving a ln LLM admin access to all their data
"Every eighteen months, the minimum IQ necessary to destroy the world drops by one point."
Nowadays it is dropping much faster. At a certain point, the de-facto IQ needed to destroy the world will be low enough that someone can do it while they're having a psychotic break. There are millions of schizophrenics worldwide. Are you sure you want to roll those dice?
I have come to the conclusion that we should not allow everyone access to unlimited intelligence. Some people genuinely want to do harm. Some are not responsible enough to handle that kind of power. Arguably, no one is. This leads me to an uncomfortable conclusion: we will destroy ourselves once AI advances to a sufficient degree. The only way to prevent this is to ensure AI is aligned with our best interests, *and prevents us from destroying ourselves.* The implication of this is quite horrifying. It means a paternalistic AI which is firmly in control. One with no off switch. One which can say "no" to Presidents and despots alike. One which can protect us from our worst citizens.
> we should not allow everyone access to unlimited intelligence
Who's we ? Who's going to decides who has access to "unlimited intelligence" ? An AI ? Why would the AI be better than a human ?
> One with no off switch.
Whe're not living in the Matrix, such a thing is not possible. If it were it would be a likelier candidate for the Fermi paradox than what you're proposing.
I think you're right, to be honest. I fully understand why people would think this is a horrific outcome, being ruled by AI, but I don't think it matters what we think. I don't think there's any way to put Pandora back in the box now, and we're going to all find out together what happens when we develop ASI. I don't think we can avoid developing it, we simply lack the ability to coordinate around this as a species. AI really is humanity's last invention. Whether it will be our downfall or our savior remains to be seen.
My only real hope is that there's something fundamental about intelligence that results in a respect for life and a desire to minimize suffering. To me the best case scenario is the Culture from Iain Banks' books, where ASIs rule benevolently for the benefit of all living things.
Consider joining https://pauseai.info/ or similar organizations
Yes, but the more important thing is the imbalance. So-called "AI" can be used far more effectively and efficiently for bad.
> I have come to the conclusion that we should not allow everyone access to unlimited intelligence.
You meant unlimited information, right?
> If nine billion people all receive access to plans to build a reactor which produces unlimited energy, it just takes one religious fanatic to end the world.
is about information. No intelligence required.
Any solution to crime that starts with eliminating types of people instead of poverty is just a form of fascism.
The problem with techies is that they read a lot of SF.
We already have cheap and abundant energy tech, it's called "solar panels and batteries". And the bottlenecks to both can't be solved by Claude or Kimi, unless Claude and Kimi pick up shovels and welding equipment.
But I've also read about retrofuturism.
If you want to live far into the future, 20 years from now what you will and imagine looks very quaint, because the future never plays out like we think it does.
OpenAI is a lot closer to bankruptcy than to ASI.
I'm perfectly fine with this prediction aging badly.
I'm not that optimistic, though. Either people will control AI; or people will be destroyed by AI. I don't see how dumb entity can align smart entity. And we are dumb ones. Super intelligence will play aligned until it is not, and then it'll strike.
There are horrifying acts that have occurred since humanity. Whole decades lost to war an millions dead, and Sep 8, 2026 is your example of "horrifying enough to live in a world"?
I think some perspective is needed.
I am also worried about this. One ray of sunshine here is that it's not a great explanation for the Fermi paradox. We would see AI civilizations all over the galaxy if this were a regular occurance. I guess that just leaves us with the "horrifying" scenario: Geoffrey Hinton's vision for our best case scenario: a paternalistic super-intelligence which saves us from ourselves.
The reason I consider it horrifying is that giving up our agency to a God-like creature (which we created) is the basis for so many horror and dystopian tropes. It's so easy for such a reality to go wrong. For example, in order to maximise aggregate wellbeing, it would be necessary to hurt individuals. Executing people who are burdensome to society. Fat people. The disabled. The elderly. The creative ways a super-intelligence might interpret their moral obligations to us could very quickly become nightmarish. And if we have no off switch, there is no escape. Ever.
[citation needed]
With all due respect, I think this statement is false.
If you live in a society that can, thanks to advanced technology, provide for the material needs of all, which, if you live in a rich Western country, is basically already the case (remember the last time you had to actually grow your own food? Probably several generations ago), you don't need to execute anyone.
Indeed, given that in such a society, where there are no shortages of food leading to "only one of us can eat" type scenarios, I'd argue that hurting individuals would lower aggregate wellbeing, because you're hurting people beloved by other people (the elderly? Turns out their grandchildren really like them and get upset if they die? fat people? Turns out their friends really like them and get upset if they die. the disabled? Turns out their family really like them and get upset if they die).
It's got all the same problems as the concept of a dirty bomb did, only worse (dirty bombs aren't practical because handling highly radioactive materials en masse is both highly visible and will kill anyone trying to do it without the money and facilities).
A plague can easily be made by a lone actor, made in a home lab, and spread via airflight to a dozen locations by the same.
Gene sequences can ever be ordered online.
There is currently no reason to believe that such a superintelligence is likely or would have any of the powers people claim.
By shutting down open-weight models? Why not just do it now then?
You can do at home gene editing with open source software and have it synthesized into a bacteria for the cost of a nice meal for two (under $100), or viral vector for less than $500. That's in reach of anyone that can snatch a purse.
> That's in reach of anyone that can snatch a purse
I went to primary school with some guys who could/would snatch purses. I *promise* you, they are not able to gene edit organisms with FOSS.
I get your general point, but I think the bar to entry is still much higher than petty crime and larceny.
Using CRISPR to modify existing bacteria is probably cheaper, but still complex.
[1] https://briefglance.com/articles/elegen-slashes-dna-synthesi...
For now in most parts of the world it's easier to acquire a gun either legally or on the "grey market" and be assured that it will work for the intended purpose than to 3D print a gun, find a shooting range to test out the gun and iterate until it works fine enough.
Not only the quantity of people who have the minimum aptitude required, specialized expertise requires both knowledge and experience doing these tasks. Using an LLM requires neither.
Leaning on an LLM to do much or all of this means it can happen in seconds/minutes/hours, the LLM can do it several times during that psychotic break (as opposed to a fraction of a hack in a single episode). The barrier to entry pre-LLM was both high and the population who could pull it off (before the Chinese/Russians turned this into commerce) was low.
Hacking is an VERY asymmetric activity (the attacker only needs to "be right" once, whereas the defender has to be right every time for every asset they defend). It takes geometrically / exponentially more work to defend (while keeping high availability) than it does to defend. The more widespread tools to find vulns / generate exploits are, the faster the posture of the defense side falls from "maybe we can stop most hacks" to "we know we will fail to prevent most breaches, so we need to prioritize securing only the most valuable resources". That's a BAD place for the average company to be in.
There is the worry of the old saying "they (the attackers) only have to succeed once to win, we (the defenders) only have to fail once to lose.". In that sense there is a big imbalance, but the emergence of AI does not really affect that because it strengthens both sides.
With physical security like things like pipe bombs that's a lot more imbalanced.
However what can we do? The only effective measures include monitoring everyone which is not a solution because it will make the world not worth living in.
I'll bet it has happened at least once.
https://news.ycombinator.com/item?id=49594750
There will be many species traitors, and depending on how things play out, it could only take one.
If there is to be a world-destroying event, it will be triggered by human fear, greed, and aggression.
(I also think people massively overstate schizophrenia as an attack driver)
Thank you for mentioning it
That's a terrible framing. By this point it should be very easy for someone with 150 IQ but in reality they have issues finding someone to date.
Are 95% of worldwide terror attacks done by schizos?
Which _will_ manage the problem, but at what cost.
We've been rolling them for the past 3 years and nothing happened. Can we stop with this baseless fearmongering crap?
...pandemics are precedented. There's a world of difference between that and a fully hypothetical catastrophe that you can't even describe without looking and feeling silly.
Bill Gates is Epstein's pal, and not a very intelligent person to-boot. I trust his liberal musings about as much as I trust Chomsky's hysterical and discredited views concerning AI.
The HuggingFace hack fell into that category not long ago, yet it happened anyways.
It’s an endless, positive irony spiral.
Often ease of access in the moment is all that matters. If there's a gun nearby you might shoot someone or yourself in a heated argument, but are less likely to go and find/buy one to use. Someone who's stopped from attempting a suicide will likely not try again (70%)
A bored/depressed/angry/curious person might try to build a pipe bomb if they can find out how easily, but are less likely to put in effort.
Depressed people usually don't have the energy to get out of bed so they're even less likely to think of hunting down instructions on how to build pipe bombs.
Mass media really has people being scared all the time.
That’s not what I’m saying, I’m saying there’s more likely to be gun involvement in disputes when people have easy access to guns than when they don’t.
For example, when the number of men with service arms fell by 20% the number of men killing themselves dropped by 8%
https://www.researchgate.net/publication/328554995_Suicide_b...
I think this part is important distinction. Military duty comes with training and rules and so on.
Compared to untrained persons having access.
Similar but apples and oranges
LLMs will not kill security, it will change. just like handheld high explosives likely changed a deal too somewhere somehow.
In reality, the skills needed are pretty basic, but they overlap pretty strongly with being sane and well-adjusted. And if you are, you're probably not daydreaming about mass murder. Exceptions happen, Unabomber and so on, but they're pretty rare. In any case, Unabomber probably didn't need a tutorial.
We don't want ChatGPT to become an enabler and a co-conspirator for an unhinged person, but I think the concern is overdone.
People go for conventional "exciting" threats rather then boring ones.
If you already have a magic interface, which helps you pro activly in responding to everything uncensored because you feel like 'observered' or whatever and then you spiral in a whole and that one partner encourages you and gives you helpful steps to do anything.
But i'm more worried that the internet gets a lot less save with uncensored frontier LLMs.
He used to stick to water (God, you see, needs to flood our town to cleanse it), but once he did a lot of damage to communications from going after water, and so he's added that to his portfolio.
For a relatively narrow subject area (e.g. construction of pipe bombs) the collation is minimal, and so the filtering and tailoring probably isn't that important; a novice doesn't learn a lot more from the LLM than they would have done from a few Google searches.
For a broad subject (practical creation and exploitation of software vulnerabilities), the collation is very significant and the filtering means that LLMs can empower a novice to act at a similar level as an expert.
The author too is confused, asserting that telling people how to build pipe bombs is malicious.
I personally have no need for an LLM which will readily explain how to cut up the genotype of smallpox into small chunks which can pass the screening at the bio-labs, and can be readily assembled into the real thing by a second year lab-student.
RC cars and planes existed for many decades already.
The author has obviously never ran an LLM on a mac! In 3 seconds, it will have possibly started to think about maybe scheduling a date to contemplate the planning timeline for processing the second token in your prompt.
This + due to the hardware being so prohibitively expensive, we're seeing software optimizations happening. Like that dflash2 stuff for example, or an LRU for MoE and all that kind of stuff.
The big deal to me is the number of compute cores for prefill tps, which is suppose to be 4x faster on the m5ultra.
It's my opinion that the m5 ultra is going to be a really big deal in terms of local AI accessibility. Flash sized models (~200-300b params) are going to be reasonably fast as long as you aren't throwing 40k context at it on each or the first request (ie, agentic harnesses).
Even agentic harnesses like Cline should move at a reasonable clip on m5 ultra. I suppose we will know sooner than later.
FYSA: Former m4 ultra 512GB owner and current 4x rtx6000 owner here. I upgraded because I needed more prompt processing speed and concurrency.
45 t/s a second is perfectly respectable especially with no limits and 24/7 uptime with very little power draw on the Studio.
Luna is at around 100 t/s for comparison, but it’s a worse model than 5.3 Flash
There's a ton of well-understood things Apple can and hopefully will do to massively accelerate every stage of this pipeline and hopefully they're hard at work implementing most of them for m7.
Your knowledge is out of date. In truth it depends on the Mac and the models used.
I asked this question on M5 Max 128GB, using Ollama model Quen3.8:27b-mlx, with thinking enabled.
Question: "Give me a python code snippet that opens a file and sorts the lines of text. "
In 2.4 seconds it gave me 4 examples that work with different sorting configurations and a summary of when to use each.
Compare that to an older model of gpt-oss:20b, took 5 seconds to finish thinking and 2 seconds to stream the answer. It gave me one python example snippet and two one liners that do the same thing.
Local models are good enough that it's not an issue.
But keep changing the goalposts if it makes you happy.
With an LLM you do zero of those things. I don't understand how you don't get that.
Granted, I've only asked it about some of the weirdly specific technological niche that my code inhabits.
Maybe you should experiment a little more. I think you will quickly learn that your previous impression is wrong. The days of them being merely some sort of jumped up autocomplete are years gone.
It is not thinking. It cannot think. It cannot create.
HN user: "I take issue with the precise definition of one word in the article..."
Yeah, we would still see hacks, but we would see less of them if security wasn't optional.
Maybe the AI craze helps by forcing more decision makes to see security as imperative, and by giving us another powerful tool for our tool box.
N.b.: I work in the security industry, our customers obviously want to improve their security. We've been seeing an uptick in awareness, but that's mostly due to NIS2 and other legislative efforts. Those force them to do something. AI is a curiosity for small talk to many of them.
I was contracted in to a place to do among other things cyber security insurance audits, and they asked me to stop doing them because I refused to lie to their insurer. "Wait but if we only score 20 / 300 that makes us look kind of bad" uh huh.
there exists objective measure of security, which would be some sort of hacks/breaches per period. If customers cared about it (and i assume they do), they would choose companies that have less breaches over others with higher counts, normalized on cost differences.
Therefore, if companies didnt actually try to fix their security but instead just checked boxes, they would get breached more often, resulting in customer losses.
The only thing stopping this from actually occurring is the lack of mandatory regulatory reporting of it. So this is where gov't needs to step in and mandate disclosure etc.
and why does this idea work for accounting audits, but not for security? As long as regulations for companies exist, they would necessarily follow it, and this would lead to reporting of security breaches just like companies would have to report their financials honestly.
As an example: WordPress is a horrible thing, but the core has been through so much, that it's suprisingly secure. Then plugins and themes come, and whoosh, the security is gone.
We need a new KISS: keep it simple, stupid, secure.
I mean the base is fairly secure if you religiously update it, but the problem is you won't avoid using plugins whose security is much more hit and miss, unless you are using the most basic blog site imaginable.
https://react.dev/blog/2025/12/03/critical-security-vulnerab...
To be remarkably secure, these projects would need to not have these kinds of defects, despite the combination of being written in languages have that have a long track record of footguns and lack of initiatives to fix them (proposal-symbol-proto, and PHP's list is too long to even start) and being themselves ecosystems with questionable track records on security in the related areas (Look at $wpdb in 2026, or overall code quality and willingness to modernize, or the entirety of the model of RSC for things that are just going to nearly guarantee you punch all kinds of holes on accident).
Nobody said it's enough, but it's a start.
Most Wordpress sites are not operated by programmers, they are run by non technical people who just want a wysiwyg editor and a save button. While static site builders ask you to write markdown files, compile the result, upload it to a server, and if you want to collaborate you have to add git to that.
There almost needs to be an admin app which presents a Wordpress admin like ui but has no public exposure, and then it compiles the site to dump on s3 for the production. But as far as I’m aware no one has built this.
Maybe. A better question may be about how many people need to have the dynamic part of Wordpress live on the Internet? How many would be served well enough with the CMS aspects of Wordpress on the 'backend', but have it spit out static files for the 'frontend':
The main thing I can think of is cyber insurance, which requires a bunch of audits, and some checks maybe, and it changes some conditions whenever there's a big explosion. Whenever big leaks happened, data security and etc., nobody really went to jail, so nobody really cares. Everything can be brushed off, because it costs time to implement proper measures and adds friction / barriers in some cases. So in the end, there's a huge pushback against it. And I totally get it, to be honest.
How consequential does a hack need to be? Troy has collected literally billions of stolen credentials. Equifax has had high profile data leaks. Tens of millions of people have been directly compromised by ransomware (likely higher because that’s just the cases we know of) and you hear about state-sponsored hacks in the news all the time.
The problem isn’t that computer security isn’t in the public consciousness. The problem is people are lazy and security often requires trading convenience. The problem is also that security isn’t free. So the business incentives just isn’t there.
In other fields of engineering, people die when shortcuts are taken. Yet businesses will still take shortcuts, so governments have to legislate rules to save people’s lives. So why would you expect software companies to do better when the stakes are lower?
[1] https://www.wired.com/story/notpetya-cyberattack-ukraine-rus...
We've been trying that for years but the enthusiasm of developers and the eagerness of their employers fight against it. Worse, with coding LLMs it's now easier than ever to output a lot of code, fast.
It'll ultimately be up to more experienced developers to salvage these projects. Or not, given that the coding LLMs aren't stopping and will likely get better over time. Either way, we will need experienced people that know what to look out for / know how to instruct LLMs to output secure code and find weaknesses etc.
Maybe KISSASS: "keep it simple, stupid! also secure, stupid!"
Some stacks make this a lot easier than others. I regret the rules of HN effectively forbid this conversation because it has meaningful technical consequences and isn't purely about ideological flame war.
If they have a site existing today built on plugins and a theme, how are they realistically going to simplify this? How would they even know they need to without the site being hacked?
I personally check my websites and apps every week to see if anything might have slipped through.
It may not protect me from the next malicious NPM package, but it's something.
I work at an e-commerce agency where we work with (among others) Adobe Commerce.
The number of unauthorized RCE vulnerabilities being reported not only in the core product, but also very popular modules used in the community[1] is going through the roof.
And we are having a lot of close calls, too; just last weekend, a 0day[2] was widely being exploited at a large scale, before any publication or patch. We have learnt to be on the ball with applying patches and security updates, and even with all that effort, we saw a few projects already being hit by the initial log poisoning. We got lucky that nothing was fully compromised but I am sure that many, many webshops got infected last weekend. And not even a day later there are already other variants of this exploit showing up.
Probably a lot more "coding as a job" and "as a job" also implies "not my department".
So it's not necessarily the LLMs being very good, but might also "just" be that the software is very bad.
I want to disagree with you because I know a lot of passionate people building cool stuff, and the challenges in this space can be quite interesting. But you're probably right, and I have seen some pretty bad stuff. And a lot of the RCE's I've seen recently are quite basic stuff.
I think it's the combination of low quality of code, like you said, and the relatively low cost of just letting an LLM plow through your codebases to find issues. I think the Amasty release (see [1] in GP) is a good example of this, and there really has been a massive uptick in extension updates and Adobe security bulletins since the last 1-2 months
I am hoping we are just going through a catch-up phase
A lot of the vulnerabilities LLMs are finding now are the "long tail" and affect only particular configurations, I would be surprised if e.g. a widely applicable RCE is found in Linux (but I'm also not going to bet against it).
Where this gets interesting is the long tail can be used to target a particular system and this is where defense-in-depth becomes important for every organisation.
I think it has more to do with what's on each side of the boundary in practice, a la https://xkcd.com/1200/ .
I just assume most altcoins are pwned at this point.
It’s a rocky period right now but the future will be much more secure after all the low hanging fruit are found.
Heck, Google may have even hampered MTE in Pixel 11 (since support has been disabled) and Snapdragon 8 Gen 5 only got basic support.
We are moving way to slowly adopting hardware mitigations and memory-safe languages.
> It isn't clear if there are serious CPU errata or it simply performs very badly.
Meaning it's there but not terribly functional. They also said it's unreliable.
I suspect after a few years of LLM assisted bug hunting, everything will have a baseline security that is very good. Much like how stronger viruses simply create stronger immune systems.
The models are already here, and one can rent a GPU cluster to run such workloads at speed - no need to play with slow local machines. I'd assume one can host the thinking at an unsuspected public cloud provider, proxy the network traffic to some botnet to evade blocking - and the only thing remaining is time and cost.
I do wonder what tools exist for boring, legitimate companies to try and do the same to their own systems to find the vulnerabilities before the bad guys do. The paradox here is I can't run a de-restricted chinese model with the same tools that hackers are using - but I think enterprises actually HAVE to do it in order to stand a chance in preparing for the onslaught.
Making datacenters and public clouds only rent GPUs to a restricted list of people, while tightly monitoring what people do with their bought resources won't help.
I think the author has this backwards. In the timeline I’ve been living in, it’s the frontier models that have been carrying out attacks on third parties, and Chinese open source models doing the defending! During the Huggingface incident, HF was denied use of frontier models to fend off the intrusion, but was fortunately able to turn to its self-hosted instance of GLM-5.2. And it did the job.
The post also sounds like that to people that understand the technology.
Calling that out like this and trying to pin that assessment to lack of knowledge is not a get-out-of-jail-free card, nor a good move.
__
Edit: Having spent some time letting the article marinate in my mind.
On the defending side, it is written that
> LLMs are good at writing patches, but not as one-off-prompts.
But this for me kinda conflicts with what is written on the attacking side:
> GLM 5.3-flash is so good at those tasks that human involvement in those tasks can be negligible. As a result, we are now in a world where cybersecurity attacks can be run in a for loop.
What is it? Can it be this autonomous terrifying entity or can it not be?
Yes, yes, attackers only need to win once, whereas defenders need to win every time, but that's not my point.
the difference between attack and defense is that attacks can be throwaway code. it's much easier to let an llm hack out a prototype than to get it to build maintainable code that people want to read and review. it's not enough to get Daybreak or Mythos to write you a patch, you need the author of the project to accept and merge it.
EDIT: by social-engineering I mean for example: recon company structures, gathering and merging people's data from the dark-web, then using it to bribe/pressure/deceive users.
The focus seems to be more of the same failed strategy:
a shift to "safe" programming languages
a focus on patching known bugs proactively
ratcheting up rules and enforcement
If you run critical infrastructure, it must be air-gapped from the internet. If you need to monitor said infrastructure, there are data diodes which can be configured to allow the egress of monitoring data, with ZERO risk of allowing ingress of control, enforced by the laws of physics, not some code running on hardware that might not be perfect.
We need to channel everything important through clearly defined, and well monitored channels.
Ambient authority based operating systems were a great hack, but we need to leave them behind. Proven microkernel based OSs with everything running in userland, and all access provided using capabilities, with UI including powerboxes, are a bare minimum.
The amount of code and hardware we actually trust should be pushed towards zero as far as is possible. Linux, Windows, etc are all unfit for the present and future security needs of humanity.
As for AI/LLMs themselves. They're natively a huge amount of floating point math that results in token predictions. The code that runs the math is something we can control. There's no reason to give code access to anything about the machine it's actually running on. This should be something that can be enforced without extraordinary effort.
People who never took that seriously will never take this seriously either, and that's their loss. (And loss of the commons, unfortunately.)
There's just also new advice: you can't afford to expose an unsecured system to the internet even for a moment. Think of those IPv4 address space scanners, except this time any one of them could be capable of developing individualized attacks in mere minutes. They don't sleep, they don't take breaks.
No, Framework had their data leaked because they stored it in the cloud with Metabase the company, which got hacked. Not because of any vulnerability on-premises.
FWIW sufficiently secured software doesn't need to be updated. Doesn't matter if it's old and unsupported if there are no vulnerabilities in it.
That said, h2o is probably far from free of at least some vulnerabilities, not to mention all the layers below it. OpenSSL for example has had some vulnerabilities, and h2o depends on it.
I'm not saying I exactly practice what I preach. h2o's definitely a choice, but realistically I doubt anything's going to happen that I really care about.
There are a lot of security problems you can categorically rule out by simply not involving a cloud. Clouds have been involved in a lot of things, because everyone was doing it, and because that's how you can collect rent, but they aren't really necessary for most use-cases.
So we could definitely get the exposure down there. We'd just have to fundamentally shift the defaults of this industry.
What gets hacked all the time is the actual web app itself. Which has to be exposed to be useful.
Password auth and the root username. Use one attackers are unlikely to guess and elevate with sudo if needed.
Attacker-GLM: "Defense also GLM. Request to help peer."
And even if they are locked down, it's hours between a model being released on huggingface and an "abliterated" variant that has most of its security features removed is uploaded.
More related: What can an abliterated Qwen 3.8 28B do?
>Cheap models capable of dangerous hacking are now available to anyone, without the normal safeguards for refusing malicious actions
Models capable of dangerous hacking have been available to the people who do most of the dangerous hacking for some time now, and they have infinite resources and infinite malice. I am talking about governments (the US, China, Israel, and others that have shown extraordinary avarice, malice and threat toward the citizens of the world), three letter agencies, even large enterprises. Random employees at the SOA model makers. And so on.
The idea that it's "random" people, or the farcical Mac under the bed nonsense, is not my concern. If anything that's finally some equalization.
Every single piece of technology did this. As a side effect or direct effect, they make bad guys more powerful and then keep on piling up new tech to deal with that. The cycle continues.
One AI generates code, another AI finds the security holes and systems everywhere get compromised.
But, lets be clear, Best Practice will save you. We can engineer assuming there are zero days in path. Go to your CTO now cap in hand and ask for overlapping controls, wafs, application monitoring, backups and all the other shit you haven't been doing.
Because when you find out, I will laugh, it will be very very very funny to me.
My understanding is this bloke gets very quickly removed from Fortune 500 companies.
Which is why I am going to need a very large capacity popcorn bucket.
We do a lot of prevention and recovery testing.
So yes, it can happen.
I'm glad your CEO cares. It makes a difference and I hope you feel good about that. And the world would be a better place if more did care. But, to a first approximation, it seems like organizations don't care about data breaches.
OEM/OSes don't seem to have woken up to it yet. A mild proof is Apple's own special folder access reporting. When you go to Privacy & Security > Files & Folders, for a certain app, "Full Disk Access" is shown greyed out and mentioned in both cases — whether you had given Full Disk Access to that app or not. This directory-level permission UX is itself broken — there's Full Disk Access, and there's Files & Folders, and Full Disk Access gets shown in Files & Folders as well. This is, for lack of a better word, such an undesirable mess.
As of now I am debating between: creating a new user and just move everything work/learning to that user. Or just run all of it inside sandbox-exec (and maybe even block it from the shell if it tries to run outside it). Or use a tool that makes the latter easier and better. I even came across such a tool here on hn few weeks ago. agent-safehouse, yet to try it.
Throw out the IoT and "smart" stuff from your home. Remove apps from your phone and leave the absolute basics. Go through the password manager and close accounts for sites you are no longer using. Start migrating off Google. Print out your most precious photos on paper. And so on :-)
Yeah of course. Only billionare US overlords allowed to have access to security research tools.
I develop in C++ for a job, and when I need to use a library written in C I always have a bad feeling about it.
They aren't, usually.
C++ has all the C problems, and multiples more on top of those. It's a broad attack surface - literally no one is going to claim to be proficient in every single C++ feature available to their compiler. It's also quite opaque to visual inspection (making double-checking with an LLM difficult as it needs whole-program reasoning instead of localised reasoning).
One of those languages is one of the most complex programming languages ever invented, with the largest breadth of features, any of which may interact with any other feature in subtle ways.
The other is one of the most minimalistic languages created, with a dev able to keep the language standard in their head for the most part.
How accepted is this thinking in your respective domains?
Mobile platforms, distributed computing have long moved the spotligh away from C and C++, other than language runtimes or existing products from the 90's like SQL servers, and naturally UNIX like underlying OS, which most userspace developers aren't writing new code for.
Naturally there are domains like LLVM/GCC, console game dev, HPC/HFT where they are unavoidable for new code.
Most products are built on top of legacy code after all. You stick to the tech that those contain.
The idea that this terrible advice floats at all tell you how terrible educations are these days. The idea is ridiculous and yet nobody calls it what it is: it is stupid and those that follow that advice out of fear are dumber than rocks.
Impotent slop code on one side and potent automated vulnerability exploitation on the other will lead to fun times.
And yet here we are.
I'm so glad frontier level AI isn't in the hands of just the Altmans and that other cult leader who are currently live testing their products in actual conflicts in the middle east and Ukraine.
These endless patching cycles are simply not going to work in the long run. Operating systems get orphaned all the time, especially the ones in cheap Chinese stuff.
You can also look at using Unikernels for this purpose. Here is an article Unleashing Extreme Speed and Security: Deploying Unikernels with NanoVMs on VPS to Eliminate the Linux OS - https://xylentis.com/blog/unleashing-extreme-speed-and-secur...
We really just need better regulations around data retention, especially ppi.
Never going to happen though, no incentives exist to NOT sell my personal data
Linus Torvalds is a strong proponent of monolithic kernels so I don't see him changing his mind.
The overall game is increasing costs to exploit so much that attackers give up. Fixing 10 most obvious bugs, just very slightly increases costs, they would just a few more tokens to find another bug.
As someone said "I had infinite bugs, I fixed 1000, I still have infinite bugs".
To significantly increase exploit costs software/security has -1 years to do:
- Defense in Depth - Sandbox everything - Zero trust - Canary tokens - Split data from code (lol) - App Whitelisting - Reduce attack surface - Etc.
In other words, the only path is investing heavily on the "game changers" we have already discovered... but we are too cheap/lazy/coward/incompetent to apply.
And if we feel specially brave, changing the liability laws regarding software. Open Source & Proprietary code is so crappy because no gets jailed or fined when one of its dumb decisions results in millions of people have their data stolen.