The whole thesis falls apart though. You can't be on your way to "power over everything" and get distilled into free Chinese models within months. Pick one.
The bottleneck is compute and data, not the model. That's why they could only gate it for a bit. The ITAR thing proves it: no nationality controls in place, so the only option was killing the whole thing. Not exactly what an all-powerful gatekeeper does.
> The whole thesis falls apart though. You can't be on your way to "power over everything" and get distilled into free Chinese models within months. Pick one.
But is that last part actually true though? Sure, there might be 600B+ models available for download and local inference if you have the hardware, but does the users who use Anthropic switch over to those even if they're available even as hosted models? Seems like some do, most don't, Anthropic and Claude remains very popular among the people who use LLMs, there is no denying that.
The hotness we are seeing is smaller 'expert' models with an 'orchestrator' model in front that evaulates the prompts and routes to the appropiate small models and then synthesizes the collected answer. Easier to split across many smaller, cheaper servers and more efficient than a huge monolithic model.
Do you have more info about this? I can't tell if you're being misled by the unfortunate "Mixture of Experts" terminology (which don't work the way you're describing), or alluding to something different.
Or, maybe I'm wrong, but my understanding is: MoE is just an architecture to keep the activated weights smaller per token. The experts get routed basically token-by-token, and the "experts" themselves don't have a semantic domain so the "expert" word was maybe a poor choice.
> does the users who use Anthropic switch over to those even if they're available even as hosted models?
I'm currently spending $200 for Claude. That's around my maximum that I can afford. I could stretch that to $500 I guess. But I saw reports of people spending tens of thousands of dollars with Claude API. That's certainly outside of my budget.
So if/when Anthropic decides to stop subsidizing subscription (if they ever do that thing, I still not sure about that), I'll certainly look at the other options. And available "open weights" LLMs hosted by someone will be my first pick. Right now Claude 4.8 feels very advanced, but things move very fast...
The ai labs would be very dumb to get rid of subscriptions. First, I don’t even think the subscriptions are losing money, I suspect they’re around break even, maybe small loses. More importantly, the subscriptions are how they lock in users and convince companies to pay api rates. Without user loyalty that they cultivate with subscriptions businesses will just use the cheapest model on open router or maybe local models.
I don't think you're appropriately understanding the full gamut. The individuals who only spent $200/months will be stuck. But the pie is increasing in size, it's not stagnant. There are a lot of orgs who can afford to run a 1T model and even more that can run a 600B model. These newcomers are what's being fought over
I have a question that perhaps you or someone else here has an answer for: I enjoy using Opus via Google Antigravity (usually agy) for perhaps 90 minutes a week. For Google’s subsidized $20/month plan they seem to give out a reasonably generous amount of Claude tokens. How does this compare with Anthropic’s $20/month plan using Claude Code?
BTW, I also use DeepSeek v4 Flash very frequently: fast and so cheap it is almost free.
I disagree. It is not the model alone. It needs a system which capitalizes on it. And this is very complex. Hardware, software, architecture - it takes a lot to get it right.
Try running the latest OS models on a normal Mac or PC. Claude Fable and Mythos are systems not just pure models.
And of course marketing. Don't believe the hype.
I think Claude is often times underwhelming. Security concerns are also a concern companies have a blond spot for. The really toughest pro security (Yes, pro! Totally different framing!) company I know is Google after all.
What I can companies advise to do is, really having more than just bug bounties but a professional hacker team that does nothing else but attacking them the whole day and night 24/7. This needs to be coordinated with the government otherwise you might sound an alarm and will be SWATed for doing good. And I would pay them huge sums since the risk and fallout warrant such a treatment, not the standard wage.
Hackers are the real deal, not AI. Proof: Hackers using AI.
> Try running the latest OS models on a normal Mac or PC.
It can be done through the magic of SSD offload. The worst case involves seconds-per-token speeds, but that's OK if you only care about low volumes of slow unattended inference, which maximizes utilization for the hardware.
(The real worst case, where you're streaming the whole model from the cheapest storage you could feasibly think of, involves multiple minutes per token for a single inference, or even hours per token batch if you're doing many inferences in bulk. That's a lot less helpful, so there's a space for smaller models at the edge, even for unattended workloads.)
> I disagree. It is not the model alone. It needs a system which capitalizes on it. And this is very complex.
AFAICT … despite saying you “disagree”, you appear to be agreeing with the parent comment that the model is less important and compute (all that complex infra) and data (also complex infra) are more important.
> > The bottleneck is compute and data, not the model.
> I disagree. It is not the model alone. It needs a system which capitalizes on it. And this is very complex. Hardware, software, architecture - it takes a lot to get it right.
For now I suspect however that the gigantic models are not needed and you will be able to do pretty much what you need in a specific domain with 120b or lower. There is so much trash in the frontier models. I don't need all the world's slam poetry for my coding tasks for example.
Model capability is a function of model size. Raising the bar raises model performance in every domain.
An "idiot savant" model that's overtrained for a specific domain would beat a generalist model of the same size. But scale the generalist up enough, and it'll trounce the specialist. Removing poetry data from a model training mix doesn't give you much - it might even cost you some performance - and "idiot savant" approach of overtraining for a domain has a hard ceiling.
So far, it seems like there's some equivalent of "g factor" in LLMs - a broad "intelligence" value that performance across many diverse domains correlates with. And, as a rule, larger models have more of it.
But he already got it, no? Claude Fable can only be made available to US citizens, which implies that every user who wants to use Claude Fable must provide proof of citizenship in some way, basically KYC.
what Dario wants is to retain any influence whatsover on how the research progresses before the inevitable nationalization of the frontier. he gets to keep the N-2 tech and maybe influence the N-1 tech, but the only influence on the frontier he has is today; whatever he imprints in the pipeline the government takes over.
IOW I don't think he thinks in the same categories as most folks here.
Best-possible-model (N) - Two Generations (2), same with N-1, N is the SOTA in this example. I'm not sure that actually clarifies what the comment is trying to say other than they think the models will be nationalized (can't even imagine what that would look like).
Regulatory capture is the OpenAI and Anthropic end goal, for certain.
But I also think they exist in a sort of un-designed corporate narcissism, which is a common trait in bubble economies — I am not judging them particularly severely.
Netscape under Clark and Andreessen and Sun under McNealy both fell into corporate narcissism: the belief that only they really mattered, that they were chosen, and that the world needed to rearrange itself to just let them shine. They arguably let themselves get played by Oracle (a corporate psychopath) and others as a result.
OpenAI's position is profoundly corporate-narcissistic: all we need is all the money in the economy and not to have to do anything upsetting like think about turning a profit for the next four years. Like rich kids. It would be nice if you believed we were so important that we should get an enormous stipend for just being us.
Anthropic's position is: we think we're so unique and ominous that government needs to make us both essential and terrifying. We have to exist otherwise worse people will.
HN is the builder side of the conversation, and in my experience, few safety people congregate here.
The safety side of tech is a PTSD inducing shit show. Governments are more than happy to champion age verification laws, because parents, around the world, are clamoring for anything to pump the breaks on the social media experiment.
Society outside of HN is quite tired of Tech, and I despair of figuring out a way to make this clear to the commentariat.
As someone on the "safety side of tech", social media is being exploited to increase surveillance and government control precisely because its actual social influence is heavily on the wane, and capital is happy to sacrifice what's left to increase the profits of the expanding public/private tech surveillance industry (with "protect the children" controls on social media like age verification being the usual backdoor route it always is).
Society may be growing tired of Tech, but governments aren't, and in fact they're heavily expanding their back channel reliance on not-traditionally-military Tech as an extension of their Defense spending.
Cyber security has the maturity that trust and safety hopes to achieve at some point.
Social media was being exploited from inception. Palantir had sales documents for sock puppet management software back in the PHP era.
I don’t disagree that Government is interested in tech, but I will push back on the dismissal of child safety that is inherent in your comment, intended or not.
For all that some people in the firm may have tried to do the right thing, Social media firms have created bad outcomes for children, and executives were briefed on the harms they were going to cause.
This is the dismissal that concerns me, because it ends up miscalculating the level of anger and unhappiness amongst the voting populace, and therefore the political will to pass regulation to reign tech in.
> Society outside of HN is quite tired of Tech, and I despair of figuring out a way to make this clear to the commentariat.
I don't think anyone in tech is really truly engaging with how quickly the shine has come off the tech industry. Except maybe Apple, who even so still have some work to do.
Technology and science is the intersection that is supposed to make our lives better, easier, more prosperous. The last decade or two what marvelous technology has came from silicon valley that hasn't served primarily the billionaire class and made life worse for the common people.
The yoke of silicon valley is feeling heavy. People might just throw it off.
Spot on. There's a certain level of drinking the kool-aid or getting high on their own supply. Anthropic is a lot worse than OpenAI but OpenAI had to go through rounds of shedding.
To be maximally fair to them, I think it is difficult to be one of the key businesses in a market bubble and not fall victim to this kind of thinking, especially when the continued inflation of the bubble depends on you — lots of people lose their shirts if you don't push hard to be "special".
But as you say, there is a measure of getting high on one's own supply now.
And there's the curious solipsistic energy of Sam Altman whimsically musing in public that it turns out his product is too expensive for people and they complain when you make the price realistic (when it possibly needs to be more expensive for OpenAI to survive).
They seem to believe that the ordinary rules either will not or somehow must not apply to them; it's increasingly bizarre to watch.
Maybe the people around pets.com were this bizarre; we didn't have so much livestreamed interview content to show us.
I struggle with the practicality of the whole thing.
The amount of tokens required to properly distill a frontier model is so large that by the time you could consume the # of tokens you would either be banned for extremely obvious abuse or a new model would be released, rendering your efforts less and less valuable over time. Intelligence is not a linear thing. Being behind just a little bit can have exponential consequences.
> Being behind just a little bit can have exponential consequences.
That seems to be the argument of Dario, Sam et. al., but I'm not ready to believe it. Time will tell, but this can be a marathon and Anthropic and OpenAI is in getting ready to sprint the last lap of the first mile.
I'm uneducated on how distillation works at more than a basic level so forgive me if this is a stupid question.
Isn't "distillation" of another provider's model exactly how these models got training date in the first place: Massive amounts of the written word + Prompt -> Answer. Why wouldn't distillation produce similar "reasoning" in the new model? It's just inputs and outputs.
What you're describing is (pre-)training. Distillation requires richer labels, the probability distribution over tokens (it would be logits rather than probabilities but that's not important). From a chat transcript you can only understand the argmax/most likely token of that distribution (and only if the API allows you to set the temperature to 0). It's not impossible for an API to give you that but they won't if they don't want you distilling their models.
The intuition is that distillation exploits not only the "right" answer but the relationship between answers (what's the second most right answer? the third? etc).
Among other things, because you simply can't get those "massive amounts" of text from a SOTA model at reasonable cost. And complex reasoning cannot possibly be trained in a pure one-shot fashion, real post-training takes massive resources. The whole story doesn't add up.
This is totally inaccurate, the APIs provide the reasoning logs. You ABSOLUTELY can distill from APIs, in fact, that's the primary way distillation is done currently.
> Anthropic’s cautious roll-out was justified. The problem with publicly releasing models, however, is that guardrails can be jailbroken, and apparently that is exactly what happened shortly after the release
The future is unevenly distributed. Anthropic, and Amodie in particular, seem to be of the mind they can control a bit of the unknown using words. They are likely being guided by the very product they built. *AI CAN MAKE MISTAKES
That Project Glasswing bullshit reeks of it. Corporations have take control of our attention, our Internet, and now our thinking.
We took the specific vulnerabilities Anthropic showcases in their announcement, isolated the relevant code, and ran them through small, cheap, open-weights models.
Is not
We sent open weight models against a codebase to find vulnerabilities.
As I understand it, ITAR regulations for export controls have just been applied to any form of Mythos.
These are overseen by U.S. Departments of State and Commerce, and forbid foreign nationals from access to any form of Mythos, either within or outside the U.S.
Only U.S. citizens and immigrants that are holders of a "green card" may now access Mythos.
It appears that Anthropic does not have internal controls to implement these restrictions in any form, so the only option was to shut Mythos down.
Penalties for ITAR violation can reach ten years in prison and a million dollars per violation. (I can post a link to those details if there is any interest.)
As long as Anthropic is a U.S. company, there is no escaping this.
This is how the US gov does business now, capricious and vengeful.
Textbook retaliation for not letting them use an abliterated version of Claude in weapons systems.
This effectively renders any US closed model useless for any foreign company. Could happen to OpenAI, Google, etc. Too much of a risk to implement something that can be yanked out because the company didn’t behave the way they want.
Looks like it’s time for Kimi, Z, Deepseek to take the front row. They’ll catch up in a few months anyway. Kimi code 2.6 is crazy good
This is a suicide shot for the American economy. The numbers only lined up for AI to rescue the USA from its debt if it captured a significant portion of the world's AI spend, and while it was a longshot before, there's basically zero percent chance the world trusts American AI when the government is pulling strings.
It was a zero percent chance anyway. Look at Europe working to leave American software behind in recent years. And it was greatly accelerated by leveraging American AI to build the exit.
You can read it all over HN. It's about weakening American influence and building Eurocentric economies and influence. And exercising the same level of choice that Americans prefer as well. Americans also want to escape Google, Microsoft and Apple and more. They've all been caught investing too heavily in government influence and thought control (aka marketing).
And on the other side of that, an American company that deprives the US of AI for defense, is defacto weakening American defense because competition militaries will gain a technological edge by simply taking control of AI companies in their country which the US hasn't done (yet).
There are very valid arguments on both sides, I think.
"When you further combine this realization with the company’s pronouncements about AI’s ability to conduct all economic activity, you realize that Anthropic’s leadership effectively wants to have power over everything and everyone."
This is fearful stuff on all sides, and none of the people involved might realistically be able to navigate the danger.
Open models will catch up eventually, TOTL models will get distilled into smaller, more efficient versions, it’s not something you can moat indefinitely
China will, but they'll only be useable by hackers torrenting it and running it on small GPU clusters you learn about on IRC. Everything old is new again.
"What this degradation represented was both the capability and willingness of Anthropic to silently alter its models to achieve its policy preferences. In other words, Anthropic willfully validated some of its critics’ worst fears in terms of being a supply chain risk."
I never really understood this "US person" restriction. There are 350M people in US, mostly citizens and green cards holders, surely some of them could be working for a foreign power.
They don't even need to know they are. You can assume that if the model becomes available again, a lot of people will find themselves working for companies distilling these models that just happens to ultimately do work for foreign entities, whether or not the people accessing the models knows or not.
> As long as Anthropic is a U.S. company, there is no escaping this.
Reminds me of the RISC-V Foundation → RISC-V International move to Switzerland. Around the time some dumbass Republicans tried to impose export restrictions on a set of open, world-wide used specifications.
Pandora's box has been opened, and there's no closing it. Capable AI models will be everywhere.
Individuals can leave, but the company cannot transfer restricted intellectual property.
Europe has extradition treaties, so the U.S. can force anyone in Europe back to the U.S. for criminal indictment who demonstrates inappropriate possession of this technology.
Would be very hard to demonstrate that they did that. If all employees move to some country with a slow legal justice system and strong labor laws, they also recreate the training data because that can be transferred, they can train another version in said country which is perfectly legal.
Can you demonstrate beyond any reasonable doubt that the model weights have been transferred? No. Will the EU judges move to extradite said individuals (and many are EU citizens)? Also no, especially in the face of spurious accusations. And even if they were open to, you can stonewall everything and you will probably outlast any US administration pursuing that.
Well, force is a strong word… it’s still just accords, that the US doesn’t seem to be valuing lately… so if they say no, what’s the US going to do? Start a war over a company?
"they by extension think that only they should have final say over AI generally. When you further combine this realization with the company’s pronouncements about AI’s ability to conduct all economic activity, you realize that Anthropic’s leadership effectively wants to have power over everything and everyone."
That might be one of the most important points in the post. Very troubling.
It's questionable whether the current government can even unite the talent required for this project. Seizing it might just push all the talent to Europe or China.
The idea of open-sourcing something that falls into the "national security" category is clearly a non-starter unless there's more powerful, classified models that can outmatch them.
I think Anthropic has clearly demonstrated the most responsibility here: they've been crying for regulations, they were careful about Project Glasswing, and they've got comically over-sensitive filters around numerous topics.
1) It’s safe to assume the US would do its best to prevent it, and even if Anthropic was successful in exfiltrating their data, code, models, and people, I’d imagine the US would immediately block all US companies from working with them. So they’d be blocked from their own US-based compute, plus Google, Amazon, Microsoft, xAI, Meta, etc.
2) Where would they go? China maybe, but as far as we can tell it doesn’t have sufficient compute for Anthropic’s level of need. The EU likely as or more restrictive in different ways to the US - the EU is hardly buzzing with AI innovation. Some Middle Eastern countries might have the money, energy, and interest in carving out such a position, but no compute. Plus I’d imagine the US would act directly against any country or region receiving them, economic or otherwise.
3) Then, as said elsewhere, the US would block GPU sales to wherever they found a safe haven, preventing the buildup of the compute they’d need to continue.
There are fortunately some initiatives and interesting developments in the European market. Take bunny.net for example. We have to start somewhere in Europe, right? Better late than never.
Oh please, the earlier spat with the Trump admin was the best thing that ever happened to Anthropic. Before that, Claude was really only well-known in developer circles, not the wider normie-sphere. After Anthropic got the "Trump hates them, so it MUST be good!" stamp of approval, the company's recognition and popularity took off.
This too, will end up being a good thing for them. The ban will end up getting lifted due to some "amazing deal" in the coming weeks and Anthropic will now have the "Trump tried to ban them, so they MUST have the most advanced AI model in the world!" stamp of approval just before IPO.
The interesting part here is not whether Anthropic is right on safety, but that safety gives them a moral vocab for bold policy changes and platform power.
Relatedly, I think it's worth noting that Anthropic models have consistently been top-scoring in BullshitBench[0], in a league of their own, really.
Not affiliated with the bench in any way, but I think it surfaces important differences between the behavior of the models from different labs.
TLDR: The benchmark is measuring pushback in response to nonsensical requests and questions, as opposed to going with it and hallucinating a nonsensical answer.
TBH this is the main thing that made me start trusting Claude enough to actually find it useful, and I'm surprised other models haven't caught up. I assumed they had and I just wasn't aware because I'm not using them in the same way.
Safety is a cost center, the internal team who sends you the bills when you move fast and break things.
I always thought safety was interesting in and of itself, but for some reason HN doesn’t have many people from the safety side of tech in conversation.
Tech isn’t a niche hobby anymore; Billions of people are impacted by the decisions of a few firms.
My grandfathers android had 3 different messaging apps installed, somehow. AI is enabling new forms of fraud at a time when we still haven't solved the old ones.
And this is all in the first world, move your coordinates to the developing world? We had human trafficking to get educated English speakers into call centers in Laos/Cambodia to defraud first world inhabitants of their money.
We aren’t in the early days of tech anymore, and the kind of scale that we have enabled comes with it a certain cost. We can choose to ignore them, or to understand them, but we will feel their impacts all the same.
> The entire Anthropic origin story is rooted in the founders’ belief that OpenAI wasn’t taking safety seriously enough; the company believes that only they can control AI, and that because they uniquely care about safety, they are justified in trying to control everyone else, up to and including the U.S. government.
Anthropic believes they have the responsibility to guard their tools from mis-use. That is all. They are not trying to "control" anything or anyone. They do however decide what they think is mis-use.
it's a pretty ridiculous stretch to attribute them thinking that OpenAI wasn't taking safety seriously enough (which is, among other things, a little bit evident from the fact that they no longer have a safety team at all) into asserting that they want to control the US government.
> Here’s the thing about these safety justifications: I think they work because, to Anthropic, they aren’t justifications. The company really believes that they are the only ones who believe in super intelligence, and thus are the only ones who are sufficiently concerned about the dangers. That excuses decision after decision, policy after policy, and confrontation after confrontation that, to people on the outside, look like a bizarre combination of cynicism and naiveté.
I really dislike this belief (that has at least been expressed here) by some that X is okay because they-really-believe-it. This has a real Road to Hell stank on it.
It is incredibly convenient when your predictions or supposed beliefs go south. Well, we really believed that we were doing it for the betterment of human kind. And we really believed that X was an existential threat that was inevitable in which case we had to step up and do it because we we the only good guy ideologues. So sorry but not sorry.
I also don’t care if commenters know rank-and-file on the inside that “really believe it” as well. Not for one second.
The problem is when people use "we really believe it" as an excuse to do harm, which has not actually occurred here. Anthropic is not committing violence, they're not defrauding the population. They're sticking to both morality and the rules.
So... what, you just don't trust anyone good? Would it be better to pull in a health insurance CEO? They're happy to watch people die for profits, no concerns at all about them pulling a "greater good" card because they're in it for entirely selfish reasons.
Incomparable domains. People routinely suffer illness. We can compare outcomes. These ideologues are building something completely unprecedented which, according to themselves apparently, can go paperclip-rogue if one is not careful. So the worst case is unprecedented. Then there is the more mundane matter of heating up the economy, something which also has no one blameworthy until any such supposed bubble actually pops.
> So... what, you just don't trust anyone good?
The baseline here is apparently that they are good, I’m just supposed to trust and shut up?
The whole thesis falls apart though. You can't be on your way to "power over everything" and get distilled into free Chinese models within months. Pick one.
The bottleneck is compute and data, not the model. That's why they could only gate it for a bit. The ITAR thing proves it: no nationality controls in place, so the only option was killing the whole thing. Not exactly what an all-powerful gatekeeper does.
> The whole thesis falls apart though. You can't be on your way to "power over everything" and get distilled into free Chinese models within months. Pick one.
But is that last part actually true though? Sure, there might be 600B+ models available for download and local inference if you have the hardware, but does the users who use Anthropic switch over to those even if they're available even as hosted models? Seems like some do, most don't, Anthropic and Claude remains very popular among the people who use LLMs, there is no denying that.
The hotness we are seeing is smaller 'expert' models with an 'orchestrator' model in front that evaulates the prompts and routes to the appropiate small models and then synthesizes the collected answer. Easier to split across many smaller, cheaper servers and more efficient than a huge monolithic model.
Do you have more info about this? I can't tell if you're being misled by the unfortunate "Mixture of Experts" terminology (which don't work the way you're describing), or alluding to something different.
Or, maybe I'm wrong, but my understanding is: MoE is just an architecture to keep the activated weights smaller per token. The experts get routed basically token-by-token, and the "experts" themselves don't have a semantic domain so the "expert" word was maybe a poor choice.
> does the users who use Anthropic switch over to those even if they're available even as hosted models?
I'm currently spending $200 for Claude. That's around my maximum that I can afford. I could stretch that to $500 I guess. But I saw reports of people spending tens of thousands of dollars with Claude API. That's certainly outside of my budget.
So if/when Anthropic decides to stop subsidizing subscription (if they ever do that thing, I still not sure about that), I'll certainly look at the other options. And available "open weights" LLMs hosted by someone will be my first pick. Right now Claude 4.8 feels very advanced, but things move very fast...
The ai labs would be very dumb to get rid of subscriptions. First, I don’t even think the subscriptions are losing money, I suspect they’re around break even, maybe small loses. More importantly, the subscriptions are how they lock in users and convince companies to pay api rates. Without user loyalty that they cultivate with subscriptions businesses will just use the cheapest model on open router or maybe local models.
I don't think you're appropriately understanding the full gamut. The individuals who only spent $200/months will be stuck. But the pie is increasing in size, it's not stagnant. There are a lot of orgs who can afford to run a 1T model and even more that can run a 600B model. These newcomers are what's being fought over
> Anthropic and Claude remains very popular among the people who use LLMs
Only because someone else is paying the bills. I use Claude Opus at work because my employer pays for the tokens and encourages me to do it.
At home, I use DeepSeek Flash. It's not as good, but it's maybe 0.7 quality for 0.001 cost.
Same, I had Deepseek search for, download and transfer (to my Linux emulation machine) the best Dreamcast games yesterday.
GPT refused to do so (citing that it's illegal even though I own the games). Deepseek did a wonderful job for 7 cents.
At work I use Opus because, why not? But I could easily switch to a less capable model if needed.
I have a question that perhaps you or someone else here has an answer for: I enjoy using Opus via Google Antigravity (usually agy) for perhaps 90 minutes a week. For Google’s subsidized $20/month plan they seem to give out a reasonably generous amount of Claude tokens. How does this compare with Anthropic’s $20/month plan using Claude Code?
BTW, I also use DeepSeek v4 Flash very frequently: fast and so cheap it is almost free.
I disagree. It is not the model alone. It needs a system which capitalizes on it. And this is very complex. Hardware, software, architecture - it takes a lot to get it right.
Try running the latest OS models on a normal Mac or PC. Claude Fable and Mythos are systems not just pure models.
And of course marketing. Don't believe the hype.
I think Claude is often times underwhelming. Security concerns are also a concern companies have a blond spot for. The really toughest pro security (Yes, pro! Totally different framing!) company I know is Google after all.
What I can companies advise to do is, really having more than just bug bounties but a professional hacker team that does nothing else but attacking them the whole day and night 24/7. This needs to be coordinated with the government otherwise you might sound an alarm and will be SWATed for doing good. And I would pay them huge sums since the risk and fallout warrant such a treatment, not the standard wage.
Hackers are the real deal, not AI. Proof: Hackers using AI.
> Try running the latest OS models on a normal Mac or PC.
It can be done through the magic of SSD offload. The worst case involves seconds-per-token speeds, but that's OK if you only care about low volumes of slow unattended inference, which maximizes utilization for the hardware.
(The real worst case, where you're streaming the whole model from the cheapest storage you could feasibly think of, involves multiple minutes per token for a single inference, or even hours per token batch if you're doing many inferences in bulk. That's a lot less helpful, so there's a space for smaller models at the edge, even for unattended workloads.)
> I disagree. It is not the model alone. It needs a system which capitalizes on it. And this is very complex.
AFAICT … despite saying you “disagree”, you appear to be agreeing with the parent comment that the model is less important and compute (all that complex infra) and data (also complex infra) are more important.
> > The bottleneck is compute and data, not the model.
> I disagree. It is not the model alone. It needs a system which capitalizes on it. And this is very complex. Hardware, software, architecture - it takes a lot to get it right.
What do you disagree with exactly?
For now I suspect however that the gigantic models are not needed and you will be able to do pretty much what you need in a specific domain with 120b or lower. There is so much trash in the frontier models. I don't need all the world's slam poetry for my coding tasks for example.
Wrong, mostly.
Model capability is a function of model size. Raising the bar raises model performance in every domain.
An "idiot savant" model that's overtrained for a specific domain would beat a generalist model of the same size. But scale the generalist up enough, and it'll trounce the specialist. Removing poetry data from a model training mix doesn't give you much - it might even cost you some performance - and "idiot savant" approach of overtraining for a domain has a hard ceiling.
So far, it seems like there's some equivalent of "g factor" in LLMs - a broad "intelligence" value that performance across many diverse domains correlates with. And, as a rule, larger models have more of it.
> no nationality controls in place
Not for now, but how long before we have KYC regulations concerning LLMs?
That’s really what Dario wants. Let’s hope he doesn’t get it
But he already got it, no? Claude Fable can only be made available to US citizens, which implies that every user who wants to use Claude Fable must provide proof of citizenship in some way, basically KYC.
what Dario wants is to retain any influence whatsover on how the research progresses before the inevitable nationalization of the frontier. he gets to keep the N-2 tech and maybe influence the N-1 tech, but the only influence on the frontier he has is today; whatever he imprints in the pipeline the government takes over.
IOW I don't think he thinks in the same categories as most folks here.
N-1? N-2?
Best-possible-model (N) - Two Generations (2), same with N-1, N is the SOTA in this example. I'm not sure that actually clarifies what the comment is trying to say other than they think the models will be nationalized (can't even imagine what that would look like).
basically imagine the Manhattan project, but instead of blowing up the desert they're building the biggest datacenter you've ever seen.
Isn't this the beginning of the plot of "I Have No Mouth, And I Must Scream"? The exceptionally disturbing dystopian horror?
the possible futures after the thing is built are uncountable, but hoping the thing won't get built at this point is naive.
in general I agree people should be reading a lot more sci-fi nowadays than they used to.
I read the popular ones, but itch for more. Which sci fi most applies today?
Regulatory capture is the OpenAI and Anthropic end goal, for certain.
But I also think they exist in a sort of un-designed corporate narcissism, which is a common trait in bubble economies — I am not judging them particularly severely.
Netscape under Clark and Andreessen and Sun under McNealy both fell into corporate narcissism: the belief that only they really mattered, that they were chosen, and that the world needed to rearrange itself to just let them shine. They arguably let themselves get played by Oracle (a corporate psychopath) and others as a result.
OpenAI's position is profoundly corporate-narcissistic: all we need is all the money in the economy and not to have to do anything upsetting like think about turning a profit for the next four years. Like rich kids. It would be nice if you believed we were so important that we should get an enormous stipend for just being us.
Anthropic's position is: we think we're so unique and ominous that government needs to make us both essential and terrifying. We have to exist otherwise worse people will.
Both narcissistic positions.
> Regulatory capture is the OpenAI and Anthropic end goal, for certain.
it has to be, because the other way around - the government taking over parts or the whole thing - is inevitable if the trend holds.
the inevitable trend is that numbers will be free and nobody will control the whole thing
ai-celebrities are just clinging to relevance like all the other celebrities out there
HN is the builder side of the conversation, and in my experience, few safety people congregate here.
The safety side of tech is a PTSD inducing shit show. Governments are more than happy to champion age verification laws, because parents, around the world, are clamoring for anything to pump the breaks on the social media experiment.
Society outside of HN is quite tired of Tech, and I despair of figuring out a way to make this clear to the commentariat.
Social media is old hat now.
As someone on the "safety side of tech", social media is being exploited to increase surveillance and government control precisely because its actual social influence is heavily on the wane, and capital is happy to sacrifice what's left to increase the profits of the expanding public/private tech surveillance industry (with "protect the children" controls on social media like age verification being the usual backdoor route it always is).
Society may be growing tired of Tech, but governments aren't, and in fact they're heavily expanding their back channel reliance on not-traditionally-military Tech as an extension of their Defense spending.
Cyber security has the maturity that trust and safety hopes to achieve at some point.
Social media was being exploited from inception. Palantir had sales documents for sock puppet management software back in the PHP era.
I don’t disagree that Government is interested in tech, but I will push back on the dismissal of child safety that is inherent in your comment, intended or not.
For all that some people in the firm may have tried to do the right thing, Social media firms have created bad outcomes for children, and executives were briefed on the harms they were going to cause.
This is the dismissal that concerns me, because it ends up miscalculating the level of anger and unhappiness amongst the voting populace, and therefore the political will to pass regulation to reign tech in.
> Society outside of HN is quite tired of Tech, and I despair of figuring out a way to make this clear to the commentariat.
I don't think anyone in tech is really truly engaging with how quickly the shine has come off the tech industry. Except maybe Apple, who even so still have some work to do.
Technology and science is the intersection that is supposed to make our lives better, easier, more prosperous. The last decade or two what marvelous technology has came from silicon valley that hasn't served primarily the billionaire class and made life worse for the common people.
The yoke of silicon valley is feeling heavy. People might just throw it off.
Porque no los dos?
this is exactly the play is my point
Spot on. There's a certain level of drinking the kool-aid or getting high on their own supply. Anthropic is a lot worse than OpenAI but OpenAI had to go through rounds of shedding.
To be maximally fair to them, I think it is difficult to be one of the key businesses in a market bubble and not fall victim to this kind of thinking, especially when the continued inflation of the bubble depends on you — lots of people lose their shirts if you don't push hard to be "special".
But as you say, there is a measure of getting high on one's own supply now.
And there's the curious solipsistic energy of Sam Altman whimsically musing in public that it turns out his product is too expensive for people and they complain when you make the price realistic (when it possibly needs to be more expensive for OpenAI to survive).
They seem to believe that the ordinary rules either will not or somehow must not apply to them; it's increasingly bizarre to watch.
Maybe the people around pets.com were this bizarre; we didn't have so much livestreamed interview content to show us.
Yeah yeah, but after the IPO!
"Distillation" from APIs is not a thing, it cannot replicate a model's deep reasoning and behavior.
I struggle with the practicality of the whole thing.
The amount of tokens required to properly distill a frontier model is so large that by the time you could consume the # of tokens you would either be banned for extremely obvious abuse or a new model would be released, rendering your efforts less and less valuable over time. Intelligence is not a linear thing. Being behind just a little bit can have exponential consequences.
> Being behind just a little bit can have exponential consequences.
That seems to be the argument of Dario, Sam et. al., but I'm not ready to believe it. Time will tell, but this can be a marathon and Anthropic and OpenAI is in getting ready to sprint the last lap of the first mile.
I'm uneducated on how distillation works at more than a basic level so forgive me if this is a stupid question.
Isn't "distillation" of another provider's model exactly how these models got training date in the first place: Massive amounts of the written word + Prompt -> Answer. Why wouldn't distillation produce similar "reasoning" in the new model? It's just inputs and outputs.
What you're describing is (pre-)training. Distillation requires richer labels, the probability distribution over tokens (it would be logits rather than probabilities but that's not important). From a chat transcript you can only understand the argmax/most likely token of that distribution (and only if the API allows you to set the temperature to 0). It's not impossible for an API to give you that but they won't if they don't want you distilling their models.
The intuition is that distillation exploits not only the "right" answer but the relationship between answers (what's the second most right answer? the third? etc).
Among other things, because you simply can't get those "massive amounts" of text from a SOTA model at reasonable cost. And complex reasoning cannot possibly be trained in a pure one-shot fashion, real post-training takes massive resources. The whole story doesn't add up.
This is totally inaccurate, the APIs provide the reasoning logs. You ABSOLUTELY can distill from APIs, in fact, that's the primary way distillation is done currently.
Not for proprietary models, all you get is a terse summary.
Do you think token completion endpoints are the final form for AI APIs?
That thesis is not about what Anthropic will achieve, but about what power they think they ought to have.
That's a different problem that what you're arguing against.
The distilled versions miss the spark of the model. Its like they land in the uncanny valley of models.
They get to 80% of the top models for 10x cheaper, unless you don't care about the money at all, it's hard to ignore.
> To that end, I can certainly buy the case that Fable/Mythos is in fact more capable when it comes to identifying and exploiting security issues
This has been covered before: https://aisle.com/blog/ai-cybersecurity-after-mythos-the-jag... (https://news.ycombinator.com/item?id=47732020)
> Anthropic’s cautious roll-out was justified. The problem with publicly releasing models, however, is that guardrails can be jailbroken, and apparently that is exactly what happened shortly after the release
The future is unevenly distributed. Anthropic, and Amodie in particular, seem to be of the mind they can control a bit of the unknown using words. They are likely being guided by the very product they built. *AI CAN MAKE MISTAKES
That Project Glasswing bullshit reeks of it. Corporations have take control of our attention, our Internet, and now our thinking.
I say it's high time to take it back.
We took the specific vulnerabilities Anthropic showcases in their announcement, isolated the relevant code, and ran them through small, cheap, open-weights models.
Is not
We sent open weight models against a codebase to find vulnerabilities.
The top comment in the very discussion you linked on that AISLE blog has a strong rebuttal to that blog post...
(reposted)
As I understand it, ITAR regulations for export controls have just been applied to any form of Mythos. These are overseen by U.S. Departments of State and Commerce, and forbid foreign nationals from access to any form of Mythos, either within or outside the U.S.
Only U.S. citizens and immigrants that are holders of a "green card" may now access Mythos.
It appears that Anthropic does not have internal controls to implement these restrictions in any form, so the only option was to shut Mythos down.
Penalties for ITAR violation can reach ten years in prison and a million dollars per violation. (I can post a link to those details if there is any interest.)
As long as Anthropic is a U.S. company, there is no escaping this.
https://fortune.com/2026/06/14/how-a-warning-from-amazon-led...
This is how the US gov does business now, capricious and vengeful.
Textbook retaliation for not letting them use an abliterated version of Claude in weapons systems.
This effectively renders any US closed model useless for any foreign company. Could happen to OpenAI, Google, etc. Too much of a risk to implement something that can be yanked out because the company didn’t behave the way they want.
Looks like it’s time for Kimi, Z, Deepseek to take the front row. They’ll catch up in a few months anyway. Kimi code 2.6 is crazy good
This is a suicide shot for the American economy. The numbers only lined up for AI to rescue the USA from its debt if it captured a significant portion of the world's AI spend, and while it was a longshot before, there's basically zero percent chance the world trusts American AI when the government is pulling strings.
It was a zero percent chance anyway. Look at Europe working to leave American software behind in recent years. And it was greatly accelerated by leveraging American AI to build the exit.
You can read it all over HN. It's about weakening American influence and building Eurocentric economies and influence. And exercising the same level of choice that Americans prefer as well. Americans also want to escape Google, Microsoft and Apple and more. They've all been caught investing too heavily in government influence and thought control (aka marketing).
And on the other side of that, an American company that deprives the US of AI for defense, is defacto weakening American defense because competition militaries will gain a technological edge by simply taking control of AI companies in their country which the US hasn't done (yet).
There are very valid arguments on both sides, I think.
Consider this quote from the main article...
"When you further combine this realization with the company’s pronouncements about AI’s ability to conduct all economic activity, you realize that Anthropic’s leadership effectively wants to have power over everything and everyone."
This is fearful stuff on all sides, and none of the people involved might realistically be able to navigate the danger.
the whole thing playing out as expected. if you think about it, the only question is the timeline.
the next model with a gap to mythos as mythos is to opus will be controlled technology from the get-go. the one after it may be top secret.
Open models will catch up eventually, TOTL models will get distilled into smaller, more efficient versions, it’s not something you can moat indefinitely
who is going to continue to publish these open models and why would they keep doing it?
China will, but they'll only be useable by hackers torrenting it and running it on small GPU clusters you learn about on IRC. Everything old is new again.
I have no idea why people keep thinking this
Or OpenAI will pay Trump's regime's bribe and they'll suddenly realise that it does not need controlling and they're free to sell it?
that's... an optimistic take I think
That part just sounds like hyperbole at best, conspiracy at worst.
By that logic, anybody who values safety has a god complex? It’s absurd…
I am just quoting the parent article.
"What this degradation represented was both the capability and willingness of Anthropic to silently alter its models to achieve its policy preferences. In other words, Anthropic willfully validated some of its critics’ worst fears in terms of being a supply chain risk."
Again, hyperbole and assumption of evil intent because… they take precautions? Nice prose doesn’t dispense you from forming a sound hypothesis
[delayed]
I never really understood this "US person" restriction. There are 350M people in US, mostly citizens and green cards holders, surely some of them could be working for a foreign power.
They don't even need to know they are. You can assume that if the model becomes available again, a lot of people will find themselves working for companies distilling these models that just happens to ultimately do work for foreign entities, whether or not the people accessing the models knows or not.
> As long as Anthropic is a U.S. company, there is no escaping this.
Reminds me of the RISC-V Foundation → RISC-V International move to Switzerland. Around the time some dumbass Republicans tried to impose export restrictions on a set of open, world-wide used specifications.
Pandora's box has been opened, and there's no closing it. Capable AI models will be everywhere.
Could Anthropic relocate to a different country?
They cannot do it. Apart from all the practical, technical and talent reasons, it would still be exporting forbidden stuff.
The signal is clear enough though for the next Anthropic..
Individuals can leave, but the company cannot transfer restricted intellectual property.
Europe has extradition treaties, so the U.S. can force anyone in Europe back to the U.S. for criminal indictment who demonstrates inappropriate possession of this technology.
Would be very hard to demonstrate that they did that. If all employees move to some country with a slow legal justice system and strong labor laws, they also recreate the training data because that can be transferred, they can train another version in said country which is perfectly legal.
Can you demonstrate beyond any reasonable doubt that the model weights have been transferred? No. Will the EU judges move to extradite said individuals (and many are EU citizens)? Also no, especially in the face of spurious accusations. And even if they were open to, you can stonewall everything and you will probably outlast any US administration pursuing that.
Well, force is a strong word… it’s still just accords, that the US doesn’t seem to be valuing lately… so if they say no, what’s the US going to do? Start a war over a company?
"they by extension think that only they should have final say over AI generally. When you further combine this realization with the company’s pronouncements about AI’s ability to conduct all economic activity, you realize that Anthropic’s leadership effectively wants to have power over everything and everyone."
That might be one of the most important points in the post. Very troubling.
The problem is... what's the alternative?
It's questionable whether the current government can even unite the talent required for this project. Seizing it might just push all the talent to Europe or China.
The idea of open-sourcing something that falls into the "national security" category is clearly a non-starter unless there's more powerful, classified models that can outmatch them.
I think Anthropic has clearly demonstrated the most responsibility here: they've been crying for regulations, they were careful about Project Glasswing, and they've got comically over-sensitive filters around numerous topics.
I think the overly sensitive filters reveals an alignment gap
if they had more success on alignment and safety research then I don't think the cludgy filters would be necessary
Perhaps they should consider leaving the US. Pretty clearly the descent into a corrupt autocracy is having real consequences.
Where would they go?
1) It’s safe to assume the US would do its best to prevent it, and even if Anthropic was successful in exfiltrating their data, code, models, and people, I’d imagine the US would immediately block all US companies from working with them. So they’d be blocked from their own US-based compute, plus Google, Amazon, Microsoft, xAI, Meta, etc.
2) Where would they go? China maybe, but as far as we can tell it doesn’t have sufficient compute for Anthropic’s level of need. The EU likely as or more restrictive in different ways to the US - the EU is hardly buzzing with AI innovation. Some Middle Eastern countries might have the money, energy, and interest in carving out such a position, but no compute. Plus I’d imagine the US would act directly against any country or region receiving them, economic or otherwise.
3) Then, as said elsewhere, the US would block GPU sales to wherever they found a safe haven, preventing the buildup of the compute they’d need to continue.
Does any other place have the infrastructure Anthropic requires to train their models and run inference?
No. If we cannot even have an EU CloudFlare, then we definitely do not have the infra for this kind of computing.
The EU options are not even close to what CF can do
There are fortunately some initiatives and interesting developments in the European market. Take bunny.net for example. We have to start somewhere in Europe, right? Better late than never.
This infrastructure may not even be needed in the next decade. Europe should have done this 20 years ago.
>EU CloudFlare
What limitations does bunny.net have?
> What limitations does bunny.net have?
A huge free tier (technically, none)
https://bunny.net/HopStart/
UAE would be happy to pay for it
> Does any other place have the infrastructure
That's not the problem.
The US government can export ban GPUs like they do now to more countries if needed. Even if the infrastructure exists, the GPUs won't.
China
Why settle for play pretend autocracy when you can go all-in amirite?
Better to be firmly under the boot of a sane autocrat than have the illusion of freedom under a madman.
Somehow it's looking more and more appealing.
Hey at least there are adults in the room.
Oh please, the earlier spat with the Trump admin was the best thing that ever happened to Anthropic. Before that, Claude was really only well-known in developer circles, not the wider normie-sphere. After Anthropic got the "Trump hates them, so it MUST be good!" stamp of approval, the company's recognition and popularity took off.
This too, will end up being a good thing for them. The ban will end up getting lifted due to some "amazing deal" in the coming weeks and Anthropic will now have the "Trump tried to ban them, so they MUST have the most advanced AI model in the world!" stamp of approval just before IPO.
All this stuff is pro wrestling kayfabe.
The interesting part here is not whether Anthropic is right on safety, but that safety gives them a moral vocab for bold policy changes and platform power.
Relatedly, I think it's worth noting that Anthropic models have consistently been top-scoring in BullshitBench[0], in a league of their own, really.
Not affiliated with the bench in any way, but I think it surfaces important differences between the behavior of the models from different labs.
TLDR: The benchmark is measuring pushback in response to nonsensical requests and questions, as opposed to going with it and hallucinating a nonsensical answer.
[0]: https://petergpt.github.io/bullshit-benchmark/viewer/index.v...
TBH this is the main thing that made me start trusting Claude enough to actually find it useful, and I'm surprised other models haven't caught up. I assumed they had and I just wasn't aware because I'm not using them in the same way.
> I found my interactions with Fable to be extremely impressive; it made other models, including GPT 5.5 and Opus 4.8, feel small and dumb.
> Anthropic models have consistently been top-scoring in BullshitBench[0]
eyeroll I find that Anthropic models feel big and dumber.
https://www.endorlabs.com/research/ai-code-security-benchmar... puts Fable 5th, which seems about right to me.
I'm interested in code utility and correctness, even if the majority of AI use is not focused on that.
> has perfect alignment between talent and mission and business.
Do they have it or do they just sell it?
Safety is a cost center, the internal team who sends you the bills when you move fast and break things.
I always thought safety was interesting in and of itself, but for some reason HN doesn’t have many people from the safety side of tech in conversation.
Tech isn’t a niche hobby anymore; Billions of people are impacted by the decisions of a few firms.
My grandfathers android had 3 different messaging apps installed, somehow. AI is enabling new forms of fraud at a time when we still haven't solved the old ones.
And this is all in the first world, move your coordinates to the developing world? We had human trafficking to get educated English speakers into call centers in Laos/Cambodia to defraud first world inhabitants of their money.
We aren’t in the early days of tech anymore, and the kind of scale that we have enabled comes with it a certain cost. We can choose to ignore them, or to understand them, but we will feel their impacts all the same.
> The entire Anthropic origin story is rooted in the founders’ belief that OpenAI wasn’t taking safety seriously enough; the company believes that only they can control AI, and that because they uniquely care about safety, they are justified in trying to control everyone else, up to and including the U.S. government.
Anthropic believes they have the responsibility to guard their tools from mis-use. That is all. They are not trying to "control" anything or anyone. They do however decide what they think is mis-use.
it's a pretty ridiculous stretch to attribute them thinking that OpenAI wasn't taking safety seriously enough (which is, among other things, a little bit evident from the fact that they no longer have a safety team at all) into asserting that they want to control the US government.
> Here’s the thing about these safety justifications: I think they work because, to Anthropic, they aren’t justifications. The company really believes that they are the only ones who believe in super intelligence, and thus are the only ones who are sufficiently concerned about the dangers. That excuses decision after decision, policy after policy, and confrontation after confrontation that, to people on the outside, look like a bizarre combination of cynicism and naiveté.
I really dislike this belief (that has at least been expressed here) by some that X is okay because they-really-believe-it. This has a real Road to Hell stank on it.
It is incredibly convenient when your predictions or supposed beliefs go south. Well, we really believed that we were doing it for the betterment of human kind. And we really believed that X was an existential threat that was inevitable in which case we had to step up and do it because we we the only good guy ideologues. So sorry but not sorry.
I also don’t care if commenters know rank-and-file on the inside that “really believe it” as well. Not for one second.
The problem is when people use "we really believe it" as an excuse to do harm, which has not actually occurred here. Anthropic is not committing violence, they're not defrauding the population. They're sticking to both morality and the rules.
So... what, you just don't trust anyone good? Would it be better to pull in a health insurance CEO? They're happy to watch people die for profits, no concerns at all about them pulling a "greater good" card because they're in it for entirely selfish reasons.
Incomparable domains. People routinely suffer illness. We can compare outcomes. These ideologues are building something completely unprecedented which, according to themselves apparently, can go paperclip-rogue if one is not careful. So the worst case is unprecedented. Then there is the more mundane matter of heating up the economy, something which also has no one blameworthy until any such supposed bubble actually pops.
> So... what, you just don't trust anyone good?
The baseline here is apparently that they are good, I’m just supposed to trust and shut up?
This is the part where the USA and allied countries can gain a headstart from using such an overpowered model.
This only just shows how strong Mythos/Fable will be, once released to the public.
I'm guessing about 0.5 year till public.
> USA and allied countries
Doesn't this *exclude* allies countries?
They are probably thinking of the "Board of Peace"