Remix.run Logo
holmesworcester 3 hours ago

I think we should see this as simply silly behavior by a government.

Export control is not an effective tool for controlling a consumer facing technology developers everywhere want to use (see:VPNs) so there was no good faith policy justification for imposing an export control.

This is an administration that seems to be keeping track of who its friends are and aren't, and likes to be the center of every story. They also seem to like extracting concessions and reciprocal favors. We saw some of this behavior in the last administration too. US voters deserve better.

libraryofbabel 3 hours ago | parent | next [-]

I am saying this probably is "silly behavior by a government" and it is a milestone that points towards what the future may look like. Why can't it be both?

It's easy to wave this aside as the current administration playing political games. But I don't think there is any reason to assume that the current era of open availability of models is going to continue indefinitely. Do you think that Chinese labs will continue to release open models forever, even why they get to the level that Mythos is at now, and beyond? And do you think that a competent US government would have no interest in regulating and restricting model access in 2 years time, assuming that model capabilities continue to improve? I think we bias towards thinking the status quo is the norm and will continue, but this news invites us to question that assumption and think about different ways the future could go.

gpm 3 hours ago | parent | next [-]

> Do you think that Chinese labs will continue to release open models forever

Yes.

I think the Chinese government either already has, or will soon, grasp that if they train the models that people use they dictate what people believe (at least around the margins where that's malleable), and they will happily throw resources at that.

And simultaneously that the only way they can actually get everyone to use their models is if it's possible for us to run them on our own hardware.

(This isn't exactly a utopian view of the future)

jychang 2 hours ago | parent | next [-]

This is going to age very poorly when the best Chinese labs ALREADY just started not open sourcing their models.

Qwen 3.7 is not open source; previous Qwen versions would have open source releases, but Qwen 3.7 plus does not. The second best Chinese model, Minimax M3, is testing the waters by taking longer and longer between “model release” and open sourcing it. This time, they spent 2 weeks after release before open sourcing it. There’s also a lot of rumors of GLM and Deepseek not open sourcing future models.

It’s pretty obvious that you cannot take Chinese models as open source for granted, they’ll be closed source soon.

ls612 2 hours ago | parent | next [-]

The main reason the Chinese labs are releasing models as open weights is because they don't have the compute necessary to provide all of the inference. For the US frontier models something like 80-90% of the lifetime compute required for the model is inference rather than training. China wants to shepherd as much of their limited compute as possible towards training to keep up in the race.

londons_explore an hour ago | parent [-]

With nearly everyone using inference accelerators, the pool of hardware is no longer shared between training and use.

zardinality an hour ago | parent | prev [-]

[dead]

nine_k 2 hours ago | parent | prev | next [-]

The US administration restricting the use of US-trained models is one of the best gifts it could make to the Chinese LLM producers, and to the PRC government.

dozerly 2 hours ago | parent [-]

This entire administration is a gift to everybody but the US. It’s either in service of Russia, China or whoever is willing to pay Trump the most.

rjzzleep an hour ago | parent [-]

Chinese have a nickname for Trump. 川建国. Trump the nation builder(meaning China). But Biden actually continued most of Trumps policies.

tw1984 3 hours ago | parent | prev [-]

> I think the Chinese government either already has, or will soon, grasp that if they train the models that people use they dictate what people believe (at least around the margins where that's malleable), and they will happily throw resources at that.

that doesn't require the model to be SOTA, it can be just a compact model capable of running on some inexpensive hardware. that is vastly different from SOTA models like Mythos which can potentially disrupt lots of things.

strangegecko 2 hours ago | parent [-]

Of course it requires SOTA, people will always choose better models over some compact thing that is obviously more limited. You can't control the truth with models nobody wants to use.

columnarx3 an hour ago | parent | next [-]

People choose SOTA right now because of the heavily subsidised model subscriptions. People aren't going to pay 20x the price for a model that's maybe 10% better.

ezst 39 minutes ago | parent [-]

And the fact that "better" is highly subjective and domain/task/vibe-specific

adrianN an hour ago | parent | prev | next [-]

Why do I want the model I use for coding to know Shakespeare or vice versa?

rjzzleep an hour ago | parent | prev [-]

Small models are the future.

deanishe 2 hours ago | parent | prev | next [-]

> Why can't it be both?

Is the government going to fund all further development? Hard to imagine investors continuing to throw billions at products they aren't allowed to sell.

locknitpicker 2 hours ago | parent | prev [-]

> I am saying this probably is "silly behavior by a government" and it is a milestone that points towards what the future may look like. Why can't it be both?

Here is why it's unlikely this is anything other than "silly behavior by a government":

- some benchmarks show GPT-5.5, Gemini 3.1, and even Claude Opus outperforming Claude Fable, and yet it's Fable which is restricted.

- some benchmarks still show the likes of Kimi 2.5 outperforming any Claude model, and DeepSeek is getting equivalent scores (a few tenths of a percent difference)

> Do you think that Chinese labs will continue to release open models forever (...)

That's immaterial to the discussion. Even if China forced Chinese labs to restrict access to all models, the truth of the matter is that Trump's administration to restrict access to US-based models does not prevent others from having access to models that are as capable or even better.

So what's exactly the point of this?

rileyphone 2 hours ago | parent | next [-]

All that says is some benchmarks aren’t worth the tokens it takes to evaluate them. Mythos is clearly capable of finding zero days other models can’t, and Fable is close enough to be lumped with it.

mullingitover 33 minutes ago | parent [-]

> Mythos is clearly capable of finding zero days other models can’t

I'm unconvinced that this is anything more than proof of work and marginal improvement that other models will catch up with, perhaps as early as to next week. Lots of other current-gen models will find vulns that can be chained together if you're willing to burn enough tokens on the task, and Fable is an absolute token incinerator.

solumunus 2 hours ago | parent | prev | next [-]

You’re completely overrating these benchmarks and it’s landing you at a nonsense opinion. Just actually use the models and you will see that the gap is significant.

kolinko an hour ago | parent | prev [-]

Did you use the models yourself?

geuis 3 hours ago | parent | prev | next [-]

I still remember when Netscape had outdated ssl for a few years because more advanced cryptography was classified by the US gov as armaments or something. Basically used export restrictions to prevent better security technology from being adopted into commercial products.

jasonfarnon 3 minutes ago | parent | next [-]

I'd forgotten all the government attempts at controlling crypto like PGP in the early internet days. It is one straightforward way to look at what's happening here without resorting to speculation about this administration's motives.

londons_explore an hour ago | parent | prev [-]

Which was a clear as day message that "We have ways to decrypt this, but can't yet decrypt that, so please use the one we can snoop on".

Yet somehow we're always forgetting that lesson and surprised when government is found snooping.

Aeolun 3 hours ago | parent | prev | next [-]

> This is an administration that seems to be keeping track of who its friends are and aren't

This seems to be an administration that is dead set on antagonizing every single part of the world. Either enemy or ally.

locknitpicker 2 hours ago | parent [-]

> This seems to be an administration that is dead set on antagonizing every single part of the world. Either enemy or ally.

Let's not forget the Trump administration threatened two separate NATO allies with invasion and annexation, and then had the gall to complain they were not helping them attack Iran.

themgt 2 hours ago | parent | prev | next [-]

I think we should see this as simply silly behavior by a government. ... We saw some of this behavior in the last administration too.

So it's silly behavior, as typified by the last decade of American governance? Is there "serious" American leadership we should be expecting to see soon, e.g. 2029 AOC elected on a platform of unlimited 10GW datacenters and universal basic Mythos 8 models?

It may seem subjectively silly to you, but e.g. getting executed for refusing to point at a deer and call it a horse is pretty silly stuff as well, at least for those not living in the Qin Dynasty.

US voters deserve better.

Deserve's got nothing to do with it.

hardbass 23 minutes ago | parent [-]

How does a data center harm me? I have seen how incredibly stupid the average (dem or republican both)'s reasons are. If not outright lizard brain radiation beams, more than once I have seen claims of it producing "toxic waste" which is absolutely absurd.

sersi 7 minutes ago | parent [-]

Cheaply designed datacenter that don't use a close loop for watercooling and use too much water are a problem

sreekanth850 an hour ago | parent | prev | next [-]

Every F35 is exported with a killswitch. and you think this is a silly decision? its not silly, its gatekeeping, iam sure this will get much strict in future, where even developing a frontier model can get sanctions from US. IMO Every country need AI sovereignty and its right time to form a group or consortium of nations to fund and build an equally capable frontier model that is accessible to all others. AI should not be confined to certain nations, the way nuclear capabilities are restricted.

thazework 3 hours ago | parent | prev | next [-]

GEO blocking is not the same as blocking based on nationality. I'd like(?) to think someone in this decision chain realized "restricting to US nationals" meant effectively restricting it to all and chose this route knowing Antrophic would need to just pull the model (so engaging in censorship without calling it that, possibly less susceptible to court challenges).

LordDragonfang an hour ago | parent [-]

I'm less confident in that. To me the way the announcement reads as malicious compliance -- this administration is extremely petty in its dealings, and it's not outside the realm of possibility they asked for and would have accepted an essentially symbolic ban, something that anyone with technical knowledge and a VPN could bypass.

Anthropic would have been able to talk to someone and explain how it wasn't possible to ban just "foreign nationals", and would have pointed out how nonsensical such a request was. The fact that the post does not mention any such discussion, and leaves the nonsensical request as the only stated reason, makes this feel like a power move by Dario, simply complying in the most dramatic and rage-inducing way and announcing it in a way to direct that rage at the USG. (Which is, IMO, a savvy move)

ilaksh 2 hours ago | parent | prev | next [-]

It's also possible that they literally are too dumb to realize they asked for something infeasible. For example, the same main character who apparently gave up a career as an extra in made-for-TV WWII German movies to become a very high ranking government official.

nl 2 hours ago | parent | prev | next [-]

VPNs won't work when they do document (passport) verification.

bluegatty 2 hours ago | parent [-]

Your company won't allow you to use export restricted technology or risk going out of business instantly.

blueblisters an hour ago | parent | prev | next [-]

> We saw some of this behavior in the last administration too. US voters deserve better.

With due respect, this take is very deluded. US voters have very little to lose if the tech is not available to the rest of the world. US politicians and elite, regardless of political inclination, understand the enormous strategic potential of this technology and will ITAR the shit out of frontier models and/or use them as leverage for extracting concessions out of other countries.

The main losers are Big AI labs, their investors, foreign employees and rest of the world.

Fwiw, China and other countries would’ve done the exact same thing. It’s perhaps the game theoretic optimal approach when your comparative advantage is so vast (capital, compute, talent, embedded knowledge) and keeps growing especially if RSI is real (making it nearly impossible for anyone else to catch up)

yoyohello13 2 hours ago | parent | prev | next [-]

We need to stop making light of these things. Governments don’t do ‘silly’ things. When you wield that kind of power over people’s lives, everything you do is deadly serious.

pants2 3 hours ago | parent | prev | next [-]

I don't see your point why export control is a silly tool. There's a difference between a VPN which I can prop up on my home server or a $5 VPS, vs a Mythos-scale closed source model running on millions of dollars of hardware

holmesworcester 3 hours ago | parent [-]

I mean, if the stated intent of an export control is to allow domestic use but prevent export, achieving the stated intent is impossible, because every developer in the world wants the latest models and will get a VPN.

vr46 3 hours ago | parent | next [-]

A VPN won't work in this instance without a US credit card. So it's completely possible.

esseph 2 hours ago | parent [-]

I'm so glad none of those US credit cards have never been stolen.

Can you imagine the disaster???

jack_pp 3 hours ago | parent | prev [-]

Harder to hide payment info than ip origin

ElProlactin 3 hours ago | parent | next [-]

It is trivially easy for nation states, non-nation bad actors, etc. to use US payments. I'd guess that most of the financial scams targeting Americans rely on US-based mules and their American bank accounts.

Also, foreign nationals legally residing in the US can have access to US-based payments. There's no way when accepting a credit card payment from a US card issuer to ask whether the card holder is a natural born citizen versus Green Card holder, etc.

LastTrain 2 hours ago | parent | prev [-]

For peons sure. For anyone who is an actual security threat it would be easy. That is why this is either a) stupid or b) yet another lever to make it easier for this administration to incarcerate people.

bluegatty 2 hours ago | parent | prev | next [-]

"effective tool for controlling a consumer facing technology developers everywhere want to use (see:VPNs) "

No - it's extremely effective.

Do you realize the difference between a 'few people using VPNs + fake IDs at 2-person companies ... vs companies all companies globally not allowed to use tech?

If 'Bank of Montreal' were caught using export controlled technology it could be devastating - so they're not going to be using it along with any little mom and pop shop.

We don't know what the Administration is doing other than 'This is Extremely Heavy Handed' and will have devastating consequences if it goes on.

bigyabai 3 hours ago | parent | prev | next [-]

I think this is also overly naive. We live in a world of hardware attestation and passkeys, the baseline requirements to use new models can increase to cryptocurrency-levels of KYC. If this becomes the new norm (which it easily could), then the best models will impose increasingly restrictive requirements.

holmesworcester 3 hours ago | parent [-]

The statement said that even foreign nationals within the US would be barred. That seems intentionally unworkable to me, and makes me think that the intent was to be more restrictive/disruptive than even an export control. It is hard to tell what the internal discussions are, but given the last run-in between the administration and Anthropic, and given the administration's politicization of nearly everything, I think it's likely that this is not necessarily a long term across-the-board policy plan.

I agree that it's really hard to tell from the outside, but if I had to guess I think we still have more to worry about on the side of "Wall Street races to superintelligence" than on the side of "KYC for AI". I could be wrong though.

sandcat_ 2 hours ago | parent | prev | next [-]

Agreed. This is no different from the US government attempting to control SSH, or restrict the sale of the Apple Power Mac.

10 years from now we’ll look back and laugh at how silly it all was.

https://en.wikipedia.org/wiki/Crypto_Wars

jcutrell 3 hours ago | parent | prev | next [-]

Silly or not, precedent matters and labeling it silly is rhetorical. The impact is going to be critically important.

hbarka an hour ago | parent | prev | next [-]

> We saw some of this behavior in the last administration too.

Can we stop with this bothsides-ing. The level of co-opting by this administration is unprecedented. There’s the strong-arming to get Intel equity stake, Nvidia/AMD revenue share, U.S. Steel golden share, Lithium Americas equity stake, Big Law pro bono pledges, TikTok forcible acquisition, Paramount-CBS-Skydance favor, it’s just unbelievable the stark use of power.

Schmerika 35 minutes ago | parent [-]

No, we can not stop with the both sidesing.

We are here in many ways as a direct result of the last admin, particularly the way they threatened tech companies. This moved tech companies to feel emboldened to go all-in on Trump. Don't think I'm justifying that - it's just what happened, in basically the tech bros own words.

The Dems then proceeded to lose to Trump, despite being extremely well funded themselves. They accomplished this through a spectacular series of "own goals": arming genocide, vetoing ceasefires, forcing deeply unpopular candidates, allowing a certain attempted insurrectionist rapist run out the clock on justice [0], awful elitist messaging on the economy, keeping the Epstein files under wraps, etc.

The red side is worse than the blue side, so the blue side demand immunity from criticism. The red side sets everything on fire, on purpose. The blue side prevents progressives from real change. The cycle rachets and repeats. This has been going on for decades, at the cost of millions of lives and trillions of dollars - but people who point it out get accused of saying both sides are the same.

0 - "That Biden was a placeholder president – a stop gap to streamline an aspiring American autocracy into an entrenched one – was obvious by mid-2021. The first, rather large clue was the lack of urgency toward sedition." - https://sarahkendzior.substack.com/p/behold-a-pale-horse-rac...

hbarka 26 minutes ago | parent [-]

> This has been going on for decades

With statements like that, if it’s been going on that long then it’s either our culture and normal way of life or you’re on some QAnon cuckoo rabbit hole.

defrost 16 minutes ago | parent [-]

> then it’s either our culture and normal way of life

That's the one - from an outside PoV the two US parties are two sides of the same coin, barely a perineum twixt them, both ceding the votes of many people to the cash of a very few.

It's baked into the US zeitgeist that it's better to sit back and watch "the government" go tits up and then wade in with guns hoping for a better outcome than it is to properly manage communal resources and common ground.

willsmith72 an hour ago | parent | prev [-]

silly or corrupt?