If mythos can break into almost all of their classified systems in hours then other models including opus, gpt, gemini and large open weight models can do so as well, maybe you'll have to double hours or it may become days, but they also will, there is no "maybe" in here.
State sponsored, non-public penetration fine tunes (of possibly public ones) likely can do it even faster.
Unsupervised penetration RL loop is ideal setup similar to optimization one – it's relatively easy to gain function on it.
Also, this is just security through obscurity. The holes that mythos exploited still exist after you've tried to limit mythos accessibility.
And the fact that all our systems are riddled with security holes shouldn't be too much of a surprise given the way that we all know that software is developed and how tech debt / chores are constantly underbudgeted (plus I think this underscores that any one human's knowledge and attention are inherently limited, and even the best PR review is going to leak all kinds of security holes).
- With a weaker model, the time to break into the system might grow so larger that it becomes infeasible, similar to how password hashes can be bruteforced, but if the password is long enough, that is not going to happen in our lifetime.
- There might be problems which are inherently unsolvable with a lower level of intelligence. For example, your dog won't derive calculus from scratch, even if it lived forever.
- LLMs might be biased in such a way that they never explore the entire solution space, no matter how many attempts are made. Some models are notorious for getting stuck in a loop, trying small variations of the same approach every time, even though it is doomed to fail. This can be counteracted somewhat with higher sampling temperature, but that hurts reasoning capabilities.
The concept of infinity claims that the dog eventually becomes Shakespeare. The same way we handled encryption, even before Alan Turing codes were broken and evolved. Last, it is a huge advantage to have the machine/mind and to evolve from there.
P.S. Even if you go back to lemon juice on paper there may be a thief around that knows the trick.
> The concept of infinity claims that the dog eventually becomes Shakespeare.
The ability to reproduce an exact copy of hamlet does not make one Shakespeare. A monkey on a typewriter may very well generate Shakespeare eventually, but it wouldn't understand Shakespeare then any more than it could immediately. Likewise a dog may put together some string of text that includes a derivation of calculus, but at no time will it be able to apply that derivation to solve mathematical problems.
People seem to think entropy can be overcome with proper focus. Thats why we have things like "effective altruism", the idea that you can ignore all the harm you do on the way to some big grand altruistic act, as if the shattered glass can be reassembled if you just collect enough reverse entropy.
It's a line of reasoning meant to shut off empathy to the here and now. And while it sounds good, along the lines of Baywatch: If you're jumping into a live saving situation and you have to choose between further harming your victim and you being harmed, you choose your victim because without you to save both of you, it's fatal; the difference is indirectly or directly pushing your victim into the water then claiming you're altruistically going to save them at a later date.
Mythos and other models are not brute-forcing passwords (and with this analogy passwords, ie. systems are the same).
We're not talking about dogs, but LLM systems.
Mythos is not exploring entire solution space either.
Usually looping is solved by repetition/frequency/presence/n-gram penalties/DRY/min-p sampling, not temperature but we're not talking about small models that have those classes of issues here.
> Mythos and other models are not brute-forcing passwords (and with this analogy passwords, ie. systems are the same).
I am not talking about literally bruteforcing passwords (although LLMs are being used for that, too), but bruteforcing passwords and solving verifiable domain tasks have quite a few similarities, especially when considering rule-based and probabilistic bruteforce methods.
> We're not talking about dogs, but LLM systems.
Well, clearly dogs are not LLM systems. It is an analogy. If there is an important point on your mind that makes the analogy break down, feel free to spell it out.
> Mythos is not exploring entire solution space either.
Yes, but weaker models do not find the solution right away, so they need to try more often. But if they only try the same thing every time, they will never succeed, so we need some kind of guarantee that they try something different every time.
> Usually looping is solved by repetition/frequency/presence/n-gram penalties/DRY/min-p sampling, not temperature but we're not talking about small models that have those classes of issues here.
Those might help to reduce looping (at the cost of biasing the generation), but to guarantee that a model can generate all possible generations, we need non-zero probabilities for all tokens, not lower probabilities for likely tokens.
This is really making me raise an eyebrow. I’m sure mythos is an improvement for sure. I don’t think the framing of it hacked the entire NSA is fully truthful. I’d like a more in depth understanding of what actually happened. Excited to be proved wrong tho!
One has to assume they put Mythose behind the front lines and not infront of the front lines, so I'd agree almost any currently useful LLM could likely crack through security if you're already inside the perimeter.
If I were too guess, internally they have as sloppy security as any other corp/organization. And those were the things Mythos effortlessly poked holes in. Other models would probably as well, but Antropic hyping gave NSA the idea to try. The shell around those internal systems is probably as (im)penetrable as ever because it's just some flavor of hardened and bare bones linux.
“Donald Trump’s blocking of Anthropic is capricious and chaotic” - current title
I don’t understand the posted title quote and assume it’s missing a lot of context or was misinterpreted as it’s a secondary attribution. “Mythos broke into almost all of our classified systems in hours”.
When you put it on those networks already and gave it compute?
(see other comment about HN titles). I think expecting an HN post title to match the article title is an overzealous interpretation of the fieldname ‘title’ in the HN submission form. Happy to be corrected if it’s right in the HN forum rules, but I’ve found highly upvoted posts to have an accurately descriptive title that is other than the source article’s title.
In other words, ontologically speaking,
post.title -= article.title
I used to treat it as post.title = article.title, but the community taught me by example to cease being a purist.
Anyway article’s flagged so this is just pedantic at this point.
I made a point about this in relation to anthropic last week: nobody inside the strategic information spaces is worried about AGI they're worried about core strategic information leaking out. Either it's in the model, or the model exposes pathways to finding it in the core strategic systems.
Those "tapes" DOGE took away? Nothing on them can be considered private any more. That's how brute force risk happens. Mythos' risks are showing doorways to exfiltration surely? Why bother when you can walk out the door with a data dump?
The NSA is just a highly specific subclass of the problem. Their traditional publicly stated approach to security is "nothing electronic which enters our domain leaves" and yet somehow they have assessed these systems as capable of breaching their walls? That's super bad.
I suspect they ran an analogue/instance inside their protection rings. I doubt they ran a test outside in the global internet. If they have actually lost control of their boundary, that's a bigger story (which I doubt) and contextually he could have been referring to information systems in NSAs duty of care, not things inside Ft Meade.
I dont seem to understand why no one is talking about this obvious fact? I mean suddenyl everyone is banning .. ok .. well how many months behind are the open source models?
If you assume that open models catch up in 6-12-36 months, then you either assume exponential growth destroying the global economy and probably the world in a few years, or you assume it plateaus and commoditises.
Even if your country prevents access to compute to protect the trillion dollar companies, it’s not going to apply for every country, and as models get better it becomes easier to compete. There’s no way an AI non proliferation treaty will be passed or even enforceable.
If I was going to pay for a news subscription, it would probably be the economist. Or maybe the financial times. They both seem to still have solid journalism.
I’ve found that high community-upvoted posts don’t bury the lede by parroting the headline. I used to be a headline title scribe until the HN community showed me the light.
> On June 11th Mark Warner, the vice-chair of the Senate Intelligence Committee, said that General Joshua Rudd, who leads the National Security Agency and the Pentagon’s Cyber Command, had told him that Mythos “broke into almost all of our classified systems, not in weeks, but in hours”.
Have to give it to Cyber Command. This is cheap and effective propAIganda.
Of course, America is now the only nation on the planet with advanced weaponised AI models that are so good they beat billions of dollars and decades of IT security experience with some of the brightest minds in their fields within hours.
If this were true, you’d see the president yapping and bragging about it on Truth before the NSA director even gets a chance to publicly talk about it. Probably doing a live stream about how he personally prompts his way into an unconditional Iranian surrender. You know it, I know it.
Nice try, William, but unless I see the Senate Intelligence Committee freaking out with you sweating black goo like Giuliani, I ain’t believing it.
This is the same kind of bullshit that was showing a gun on TV that could apparently give people heart attacks with some frozen, untraceable darts.
If the US really was in possession of a technology that could hack into the most secure environments on the planet autonomously within hours, you would see all their partners pulling their access from shared IT systems and blocking all traffic coming from the US immediately.
Especially considering they have been caught spying on allies before:
If mythos can break into almost all of their classified systems in hours then other models including opus, gpt, gemini and large open weight models can do so as well, maybe you'll have to double hours or it may become days, but they also will, there is no "maybe" in here.
State sponsored, non-public penetration fine tunes (of possibly public ones) likely can do it even faster.
Unsupervised penetration RL loop is ideal setup similar to optimization one – it's relatively easy to gain function on it.
Also, this is just security through obscurity. The holes that mythos exploited still exist after you've tried to limit mythos accessibility.
And the fact that all our systems are riddled with security holes shouldn't be too much of a surprise given the way that we all know that software is developed and how tech debt / chores are constantly underbudgeted (plus I think this underscores that any one human's knowledge and attention are inherently limited, and even the best PR review is going to leak all kinds of security holes).
I don't think that is necessarily true.
- With a weaker model, the time to break into the system might grow so larger that it becomes infeasible, similar to how password hashes can be bruteforced, but if the password is long enough, that is not going to happen in our lifetime.
- There might be problems which are inherently unsolvable with a lower level of intelligence. For example, your dog won't derive calculus from scratch, even if it lived forever.
- LLMs might be biased in such a way that they never explore the entire solution space, no matter how many attempts are made. Some models are notorious for getting stuck in a loop, trying small variations of the same approach every time, even though it is doomed to fail. This can be counteracted somewhat with higher sampling temperature, but that hurts reasoning capabilities.
The concept of infinity claims that the dog eventually becomes Shakespeare. The same way we handled encryption, even before Alan Turing codes were broken and evolved. Last, it is a huge advantage to have the machine/mind and to evolve from there. P.S. Even if you go back to lemon juice on paper there may be a thief around that knows the trick.
> The concept of infinity claims that the dog eventually becomes Shakespeare.
The ability to reproduce an exact copy of hamlet does not make one Shakespeare. A monkey on a typewriter may very well generate Shakespeare eventually, but it wouldn't understand Shakespeare then any more than it could immediately. Likewise a dog may put together some string of text that includes a derivation of calculus, but at no time will it be able to apply that derivation to solve mathematical problems.
People seem to think entropy can be overcome with proper focus. Thats why we have things like "effective altruism", the idea that you can ignore all the harm you do on the way to some big grand altruistic act, as if the shattered glass can be reassembled if you just collect enough reverse entropy.
It's a line of reasoning meant to shut off empathy to the here and now. And while it sounds good, along the lines of Baywatch: If you're jumping into a live saving situation and you have to choose between further harming your victim and you being harmed, you choose your victim because without you to save both of you, it's fatal; the difference is indirectly or directly pushing your victim into the water then claiming you're altruistically going to save them at a later date.
It's just delusions to keep moving forware.
Mythos and other models are not brute-forcing passwords (and with this analogy passwords, ie. systems are the same).
We're not talking about dogs, but LLM systems.
Mythos is not exploring entire solution space either.
Usually looping is solved by repetition/frequency/presence/n-gram penalties/DRY/min-p sampling, not temperature but we're not talking about small models that have those classes of issues here.
> Mythos and other models are not brute-forcing passwords (and with this analogy passwords, ie. systems are the same).
I am not talking about literally bruteforcing passwords (although LLMs are being used for that, too), but bruteforcing passwords and solving verifiable domain tasks have quite a few similarities, especially when considering rule-based and probabilistic bruteforce methods.
> We're not talking about dogs, but LLM systems.
Well, clearly dogs are not LLM systems. It is an analogy. If there is an important point on your mind that makes the analogy break down, feel free to spell it out.
> Mythos is not exploring entire solution space either.
Yes, but weaker models do not find the solution right away, so they need to try more often. But if they only try the same thing every time, they will never succeed, so we need some kind of guarantee that they try something different every time.
> Usually looping is solved by repetition/frequency/presence/n-gram penalties/DRY/min-p sampling, not temperature but we're not talking about small models that have those classes of issues here.
Those might help to reduce looping (at the cost of biasing the generation), but to guarantee that a model can generate all possible generations, we need non-zero probabilities for all tokens, not lower probabilities for likely tokens.
> I am not talking about literally bruteforcing passwords (although LLMs are being used for that, too)
They are? Seems like a much worse way to brute force that a tight loop written in a compiled language.
PassGPT: Password Modeling and (Guided) Generation with Large Language Models
https://huggingface.co/papers/2306.01545
Although most activity is likely hidden (blackhat or state)
This is really making me raise an eyebrow. I’m sure mythos is an improvement for sure. I don’t think the framing of it hacked the entire NSA is fully truthful. I’d like a more in depth understanding of what actually happened. Excited to be proved wrong tho!
Yeah, this article cites someone saying that someone else said something. Maybe it was said, maybe not. Maybe it was a exaggeration, maybe not.
Very insightful.
One has to assume they put Mythose behind the front lines and not infront of the front lines, so I'd agree almost any currently useful LLM could likely crack through security if you're already inside the perimeter.
From the outset, Mythos’s PR has been rather dodgy.
https://archive.is/aA1dB
The link does not seem to be working
Works for me though, even when using a proxy that is usually blocked everywhere.
If I were too guess, internally they have as sloppy security as any other corp/organization. And those were the things Mythos effortlessly poked holes in. Other models would probably as well, but Antropic hyping gave NSA the idea to try. The shell around those internal systems is probably as (im)penetrable as ever because it's just some flavor of hardened and bare bones linux.
https://archive.ph/dXddV
“Donald Trump’s blocking of Anthropic is capricious and chaotic” - current title
I don’t understand the posted title quote and assume it’s missing a lot of context or was misinterpreted as it’s a secondary attribution. “Mythos broke into almost all of our classified systems in hours”.
When you put it on those networks already and gave it compute?
(see other comment about HN titles). I think expecting an HN post title to match the article title is an overzealous interpretation of the fieldname ‘title’ in the HN submission form. Happy to be corrected if it’s right in the HN forum rules, but I’ve found highly upvoted posts to have an accurately descriptive title that is other than the source article’s title.
In other words, ontologically speaking, post.title -= article.title
I used to treat it as post.title = article.title, but the community taught me by example to cease being a purist.
Anyway article’s flagged so this is just pedantic at this point.
I made a point about this in relation to anthropic last week: nobody inside the strategic information spaces is worried about AGI they're worried about core strategic information leaking out. Either it's in the model, or the model exposes pathways to finding it in the core strategic systems.
Those "tapes" DOGE took away? Nothing on them can be considered private any more. That's how brute force risk happens. Mythos' risks are showing doorways to exfiltration surely? Why bother when you can walk out the door with a data dump?
The NSA is just a highly specific subclass of the problem. Their traditional publicly stated approach to security is "nothing electronic which enters our domain leaves" and yet somehow they have assessed these systems as capable of breaching their walls? That's super bad.
I suspect they ran an analogue/instance inside their protection rings. I doubt they ran a test outside in the global internet. If they have actually lost control of their boundary, that's a bigger story (which I doubt) and contextually he could have been referring to information systems in NSAs duty of care, not things inside Ft Meade.
What happens when open source models achieve Mythos level capabilities in six months' time?
I dont seem to understand why no one is talking about this obvious fact? I mean suddenyl everyone is banning .. ok .. well how many months behind are the open source models?
If you assume that open models catch up in 6-12-36 months, then you either assume exponential growth destroying the global economy and probably the world in a few years, or you assume it plateaus and commoditises.
Even if your country prevents access to compute to protect the trillion dollar companies, it’s not going to apply for every country, and as models get better it becomes easier to compete. There’s no way an AI non proliferation treaty will be passed or even enforceable.
Not being funny but does most of HN subscribe to the economist? I dont think ive ever paid for an online newspaper ( and Im not trying to be edgy )
If I was going to pay for a news subscription, it would probably be the economist. Or maybe the financial times. They both seem to still have solid journalism.
more likely, most of HN who care about reading this article use something like archive.is
HN post title does not match link title
> NSA director: 'Mythos "broke into almost all of our classified systems in hours"
> Donald Trump’s blocking of Anthropic is capricious and chaotic
I’ve found that high community-upvoted posts don’t bury the lede by parroting the headline. I used to be a headline title scribe until the HN community showed me the light.
The article has nothing about mythos breaking into classified systems in hours.
So you either posted the wrong link or are just spreading FUD.
Yes it does.
Third paragraph:
> On June 11th Mark Warner, the vice-chair of the Senate Intelligence Committee, said that General Joshua Rudd, who leads the National Security Agency and the Pentagon’s Cyber Command, had told him that Mythos “broke into almost all of our classified systems, not in weeks, but in hours”.
Indeed. I missed that part.
Have to give it to Cyber Command. This is cheap and effective propAIganda.
Of course, America is now the only nation on the planet with advanced weaponised AI models that are so good they beat billions of dollars and decades of IT security experience with some of the brightest minds in their fields within hours.
If this were true, you’d see the president yapping and bragging about it on Truth before the NSA director even gets a chance to publicly talk about it. Probably doing a live stream about how he personally prompts his way into an unconditional Iranian surrender. You know it, I know it.
Nice try, William, but unless I see the Senate Intelligence Committee freaking out with you sweating black goo like Giuliani, I ain’t believing it.
This is the same kind of bullshit that was showing a gun on TV that could apparently give people heart attacks with some frozen, untraceable darts.
If the US really was in possession of a technology that could hack into the most secure environments on the planet autonomously within hours, you would see all their partners pulling their access from shared IT systems and blocking all traffic coming from the US immediately.
Especially considering they have been caught spying on allies before:
https://www.spiegel.de/international/germany/cover-story-how...
You know what they say in intelligence circles.
Fool us once, shame on you. Fool us twice, it's open windows season.
None of the partners or adversaries seem to give a fuck about Mythos, so there is a good chance this is just another lying NSA director as usual.
Come on, people. You don’t run the NSA if you’re an honest man. It’s a spy agency.
Actual title "Donald Trump’s blocking of Anthropic is capricious and chaotic"
flagged because reasons. @dang