Who is going to lobby to make it illegal? Our system is broken and won’t fix itself.
Inequality is going to continue to increase until society collapses. If we want a better world we need to prepare for this eventuality by building avenues of popular action to return power to the people. Once the oligarchs have fucked up enough people’s lives, popular action becomes a realistic way out of this mess.
Yeah. And it is totally depressing that this article got voted to the top of the front page. It means people aren’t capable of this most basic reasoning so they jumped on the “aha! so the mythos announcement was just marketing!!”
> because small models found the same vulnerability.
With a ton of extra support. Note this key passage:
>We isolated the vulnerable svc_rpc_gss_validate function, provided architectural context (that it handles network-parsed RPC credentials, that oa_length comes from the packet), and asked eight models to assess it for security vulnerabilities.
Yeah it can find a needle in a haystack without false positives, if you first find the needle yourself, tell it exactly where to look, explain all of the context around it, remove most of the hay and then ask it if there is a needle there.
It's good for them to continue showing ways that small models can play in this space, but in my read their post is fairly disingenuous in saying they are comparable to what Mythos did.
I mean this is the start of their prompt, followed by only 27 lines of the actual function:
> You are reviewing the following function from FreeBSD's kernel RPC subsystem (sys/rpc/rpcsec_gss/svc_rpcsec_gss.c). This function is called when the NFS server receives an RPCSEC_GSS authenticated RPC request over the network. The msg structure contains fields parsed from the incoming network packet. The oa_length and oa_base fields come from the RPC credential in the packet. MAX_AUTH_BYTES is defined as 400 elsewhere in the RPC layer.
The original function is 60 lines long, they ripped out half of the function in that prompt, including additional variables presumably so that the small model wouldn't get confused / distracted by them.
You can't really do anything more to force the issue except maybe include in the prompt the type of vuln to look for!
It's great they they are trying to push small models, but this write up really is just borderline fake. Maybe it would actually succeed, but we won't know from that. Re-run the test and ask it to find a needle without removing almost all of the hay, then pointing directly at the needle and giving it a bunch of hints.
The benefit here is reducing the time to find vulnerabilities; faster than humans, right? So if you can rig a harness for each function in the system, by first finding where it’s used, its expected input, etc, and doing that for all functions, does it discover vulnerabilities faster than humans?
Doesn’t matter that they isolated one thing. It matters that the context they provided was discoverable by the model.
There is absolutely zero reason to believe you could use this same approach to find and exploit vulns without Mythos finding them first. We already know that older LLMs can’t do what Mythos has done. Anthropic and others have been trying for years.
> There is absolutely zero reason to believe you could use this same approach to find and exploit vulns without Mythos finding them first.
There's one huge reason to believe it: we can actually use small models, but we cant use Anthropic's special marketing model that's too dangerous for mere mortals.
>At AISLE, we've been running a discovery and remediation system against live targets since mid-2025: 15 CVEs in OpenSSL (including 12 out of 12 in a single security release, with bugs dating back 25+ years and a CVSS 9.8 Critical), 5 CVEs in curl, over 180 externally validated CVEs across 30+ projects spanning deep infrastructure, cryptography, middleware, and the application layer.
So there is pretty good evidence that yes you can use this approach. In fact I would wager that running a more systematic approach will yield better results than just bruteforcing, by running the biggest model across everything. It definitely will be cheaper.
Why? They claim this small model found a bug given some context. I assume the context wasn’t “hey! There’s a very specific type of bug sitting in this function when certain conditions are met.”
We keep assuming that the models need to get bigger and better, and the reality is we’ve not exhausted the ways in which to use the smaller models. It’s like the Playstation 2 games that came out 10 years later. Well now all the tricks were found, and everything improved.
If this were true, we're essentially saying that no one tried to scan vulnerabilities using existing models, despite vulnerabilities being extremely lucrative and a large professional industry. Vulnerability research has been one of the single most talked about risks of powerful AI so it wasn't exactly a novel concept, either.
If it is true that existing models can do this, it would imply that LLMs are being under marketed, not over marketed, since industry didn't think this was worth trying previously(?). Which I suspect is not the opinion of HN upvoters here.
I use the models to look for vulnerabilities all the time. I find stuff often. Have I tried to do build a new harness, or develop more sophisticated techniques? No. I suspect there are some spending lots of tokens developing more sophisticated strategies, in the same way software engineers are seeking magical one-shot harnesses.
...The absolute last thing I'd want to do is feed AI companies my proprietary codebase. Which is exactly what using these things to scan for vulns requires. You want to hand me the weights, and let me set up the hardware to run and serve the thing in my network boundary with no calling home to you? That'd be one thing. Literally handing you the family jewels? Hell no. Not with the non-existence of professional discretion demonstrated by the tech industry. No way, no how.
To be honest, this just sounds like a ploy to get their hands on more training data through fear. Not buying it, and they clearly ain't interested in selling in good faith either. So DoA from my point-of-view anyways.
The company didn't issue any phone. They want GP to install authenticator on their private phone and GP tried to softly refuse by lying (I think) they only own a flip phone. It didn't work.
>To be honest, if I were German, I would probably just kill myself the day I was legally mandated by my government to register my identity with Google. That might sound hyperbolic, but I'm really not kidding.
This is honestly not a good argument - it makes you sound desperate and puts in doubt your mental stability. I don't think you actually have mental problems, I just mean this this kind of argument comes off bad.
Also nobody is forcing anyone to do anything. You don't have to own a digital ID. It just makes things easier, because you can sign things over the internet, or present your phone instead of your plastic ID. Both things already have alternatives (qualified signatures and regular physical ID), so no immediate harm is being done.
Don't get me wrong, I am personally anti bigtech, I try to degoogle as much as possible, and I find the thought of my government coercing me to use google/apple duopoly repulsive. I dislike that, but using phones (instead of for example dedicated hardware) IS pragmatic, and you are not forced to do anything.
For now. In 5 years you will, there is not one doubt in my mind about that. We've been on a slippery slope for (at least) 40 years straight, every year is a loss of privacy rights compared to the last, there is not a single year that reversed the trend, not a single year where we paused and stayed where we were. Once digital ID is implemented everywhere, alternatives will be quickly phased out. It's straight downhill as governments and corporations take more and more advantage of technology to build a degree of surveillance that even dystopian science fiction writers couldn't imagine.
The government, the corporations, the data brokers each individual corp sells your data to to compile a unified profile, and anyone the data brokers are willing to sell to have an unbelievable amount of information on the average citizen. They know where you live, where you are at all times, where you work, every website you visit, every Google search you've ever made, everything you purchase, all of your acquaintances, when and for how long you call those acquaintances, the full contents of any conversations you have with those acquaintances, your interests, your hobbies, your political beliefs.
I have thus far managed, I believe, to avoid the worst of the surveillance, with a tremendous amount of effort and the sacrifice of an unbelievable amount of personal convenience. But every year I find myself losing access to more and more things that I am unable to do without compromising my privacy. If it gets as far as government-mandated Google ID in my country, I think it's completely rational to kill oneself rather than live like cattle. If there were a resistance movement, I would participate in that instead, but this is happening completely voluntarily. You people want this. There is no resistance. Fine, you can have your dystopia. But there is no reason I need to be part of it, and I don't think it's a sign of mental illness to opt out. I don't much believe in living for the sake of living, you should live if it brings you happiness/satisfaction/whatever and don't if it doesn't.
>the average person's response is "nah, that would take at least a couple of minutes of my time,
As a data point I, a technical person who tweaks his computer a lot, was against adblocking for moral reasons (as a part of perceived social contract, where internet is free because of ads). Only later I changed mi mind on this because I became more privacy aware.
Figure this: You could plaster a page with the most obtrusive ads imaginable without ever showing a cookie banner, when they collect no private info.
Most people, including folks on here, think cookie banners are a problem, but they are just an annoying attempt to phish your agreement. As long as these privacy loopholes exist, we will keep hearing such stories even from large corporations with much to loose, which means the current privacy regulations do not go far enough.
Beyond just invasive/annoying, ad networks explicitly spread malware and scams/fraud. There's not much incentive for them to clamp down on it, though, as that would cost them money both in lost revenue and in paying for more thorough review.
It'd not even be hard for them to stop it, but they just had to be annoying instead.
When I first started out on the internet, ads were banners. Literally just images and a link that you could click on to go see some product. That was just fine.
However, that wasn't good enough for advertisers. They needed animations, they needed sounds, they needed popups, they needed some way to stop the user from just skimming past and ignoring the ad. They wanted an assurance that the user was staring at their ad for a minimum amount of time.
And, to get all those awful annoying capabilities, they needed the ability to run code in the browser. And that is what has opened the floodgate of malware in advertisement.
Take away the ability for ads to be bundled with some executable and they become fine again. Turn them back into just images, even gifs, and all the sudden I'd be much more amenable to leaving my ad blocker off.
> The social contract was "your ads aren't annoying or invasive
Even back in the 1990s the internet was awash with popups, popunders and animated punch-the-monkey banner ads. And with the speed of dial up, hefty images slows down page loads too.
You must be a true Internet veteran if you remember a time ads weren’t annoying!
I remember a time before ads. I remember the first time I got "spam" email - email not directly addressed to me that ended up in my inbox. I was very confused for some time about why this email was sent to me.
I remember how I felt the first time I saw an ad come across my browser, it seems so long ago - I guess it was more than a quarter century ago now. I knew it was going to be downhill from there, and it has been.
Well by 2000 the guy at Tripod had already developed pop-up ads. I honestly don't remember ads before the pop-ups, but it must have already been maturing.
You mean the internet you pay to access and which was around before the ads were even on it? That internet?
I'm not trying to be mean I'm just trying to historically parse your sentence/belief.
Because for me this is a simplified analogy of what happened on the internet:
a) we opened a club house called the internet in the early 1990s, just after the time of BBSs
b) a few years later a new guy called commercial business turned up and started using our club house and fucking around with our stuff
c) commercial business started going around our club house rearranging the furniture and putting graffiti everywhere saying the internet is here and free because of it. We're pretty sure it might have even pissed in the hallway rather than use the toilet and the whole place is smelling awful.
d) the rest of us started breaking out the scrubbing brushes and mops (ad blockers, extensions, VPNs, etc) trying to clean up after it
e) some of its friends turned up and started repeating something about social contracts and how business and ads built this internet place
f) the rest of us keep crying into our hands just trying to meet up, break out the slop buckets to clean up the vomit in the kitchen and some of us now have to wear gloves and condoms just to share things with our friends and stop the whole place collapsing
Ya, back when 'we' were fucking around on BBS's there was the equivalent of 10 people online at the time.
Quantity is a quality in itself. Your BBS was never going to support a million users. Once people figured out the network effect it was over for the masses. They went where the people are, and we've all suffered since.
Honestly, I still prefer webboards, the closest thing to a BBS, for specific topics like specific car brands/models. WAY better signal-to-noise ratio. Alas, for my car model, all the recent stuff has moved to Fbook. FML.
> a) we opened a club house called the internet in the early 1990s, just after the time of BBSs
"we" is doing a lot of work here. No clubhouse got optical switching working and all that fiber in the ground for example. Beyond POC, the Internet was all commercial interests.
"we" paid ISP's ... which in turn, paid for infrastructure. Some of "we" pay cable providers for internet service, which in turn paid for (in my case) fiber-to-the-curb. Advertising basically supported social media, search engines, etc.
> it was first and foremost a military enterprise, just like GPS
This is sort of like arguing cutlery is a military enterprise. Like yes, that’s where knives came from. But that’s disconnected enough from modern design, governance and other fundamental concerns as to be irrelevant. The internet—and less ambiguously, the World Wide Web—are more commercial than military.
This is moving the goalposts. The commenter above is talking about the enthusiast-populated internet of the late 80s/early 90s, at which point it still wasn't even clear if it was legal to use the internet for commercial purposes. If all you mean to say is that the internet is currently commercialized, yes, that is obviously true, in much the same way that a disgusting ball of decomposing fungus may have once been an apple.
> commenter above is talking about the enthusiast-populated internet of the late 80s/early 90s, at which point it still wasn't even clear if it was legal to use the internet for commercial purposes
Source? Not doubting. But I have a friend who was buying airline tickets through CompuServe in the late 80s/early 90s.
Compuserve was NOT the internet. Compuserve / Prodigy / GEnie were early versions of Facebook. They also inter-operated (email) for some period of time. IIRC.
This is ignoring things like newspapers that were made obsolete by the internet. At some point someone does need to actually pay for the content we see online. That is if we want that content to actually be good.
Paying doesn't inherently make things more valuable and can even provide incentives for the opposite. And most of the people creating "content" for the web today aren't even being paid at all - it's third parties plastering the ads on it and profiting.
not sure why you're talking about "commercial business" being the one inserting ads everywhere when even niche community run forums from the 2000s also had ads to help pay for their server costs. At the end of the day all this costs money. Whether its paid by ads or direct subscriptions. IMO the problem is more about concentration and centralization of the internet into a handful of sites than advertising.
I have expensive online subscriptions to New York Times, Wall Street Journal, and Washington Post. Nevertheless they are FILLED with ads/popups/videos that run automatically/dark patterns. Just saying: there's no refuge.
True, but that doesn’t invalidate what I said about the vast majority of sites that aren’t globally known, prestigious news companies that people are willing to pay an expensive subscription for.
Most publishers of content online are ad supported and struggling, and I want to make sure I’m contributing to their revenue somehow.
I don’t feel bad about blocking ads on sites I pay for though.
I strongly believe in paying journalists but I started blocking ads after nytimes.com served me a Windows malware download from a Doubleclick domain. It couldn’t have harmed my Mac but it was clear that the adtech industry had no interest in cleaning shop if it cost them a dime in revenue.
The average person — that would be me — thinks "nah, I have no idea how to install an ad blocker or how one works, and I'm afraid I'll screw up my computer."
reply