> I guess a US business with non-citizen employees could work.

No. Only if those employees have a green card and the company must not only take on that responsibility but ensure other employees are denied access. Otherwise the company would be subject to millions in fines.

US export laws are no fuckin' joke like everyone here seems to think they are.

It's really frustrating to read pages of comments rooted in emotion and no understanding of the existing laws.

Are you certain? Trump admin is hand-picking GPT 5.6 winners

There's huge incentive.

The government has arbitrary commandeered their business.

This could ruin Anthropic.

They are walking a tight rope with respect to revenues, hype, IPO.

If this kills their hyper growth prospects, it could kill the IPO.

If there's a serious change the gov. is out of line, the judges could put a stay and possibly throw this out.

There may however be enough of a case, in which there's maybe not much they can do.

Having a crazy person completely control your business is very, very bad.

US government regularly loses lawsuits and has to backtrack policy.

A law firm already has: https://www.reuters.com/legal/litigation/legal-tech-firm-sue...

The labs will not just ignore the order, there are too many other levers they can try to pull to mess with those companies. Just for some examples, think about the number of employees reliant on visas that could be revoked, the government contracts that the hyperscalers hosting them that could be canceled, the certifications that all the data centers need to be hooked up to the grid, the tariffs that could be put on critical components, the IPOs that need to be approved by regulators, the bill introduced in Congress to seize 50% of their equity...

Lots of these moves would and should be struck down in court as an arbitrary and capricious use of administrative power. Some of them might not be, and in the meantime you're signing up for tons of trouble. A trillion-dollar company does not simply go to war with the US government.

A more mid-sized company that's not so intertwined, but not so small that they can't get a good legal team, might be another story.

These regulations have been in place since the 1990s and have been applied by every administration. It isn’t a new authority and many companies have had the opportunity to fight it. Anthropic’s lawyers will know this.

Yes he does. They could ignore the US government, but will likely quickly find themselves in court fighting a fight that they are likely to lose and isn’t worth fighting anyway.

Ignoring US export control laws that have been on the books for nearly 50 years is a good way to pay millions of $ in fines and/or land in pound-me-in-the-ass prison.

Ask every satellite launch company in the 90s how that worked out.

Aside from how brazen and stupid this would be, the executive branch does have ways of limiting them and their sharing of tech, as we've seen.

Anthropic has been touting how their newest model is basically a cyber weapon and is so dangerous that they need to only role it out to trusted people and make sure it has super restrictive guard rails. They are begging to be regulated. This is exactly what they’ve been asking for.

I don't buy that anymore. The day America threatened to invade Canada and Denmark was the day America showed they cannot be trusted any more.

It's not like China can be trusted either, but China isn't planning any direct invasions to the west. Taiwan, perhaps, but they're playing a long-term tactical game rather than a "invade the country we don't like this week" game. They might get some info on you, but the data brokers in the west will sell a lot more details about you, pre-categorized and all.

If you're afraid of industrial espionage, Chinese companies may be a risk, but in that case you shouldn't be uploading your secrets to an AI company in the first place.

> I know which one I'd rather send tokens to

That's the neat part with the Chinese open weight models. You don't have to send your tokens to the PRC, the models can be hosted stateside or anywhere else you'd like.

Less likely to align with your interests maybe, but have you considered that not everyone has the same interests?

Personally I am much more concerned about handing my data over to the government that actually has power over me and labels dissenters terrorists than I am with the government overseas that has no direct effect on my life... well, other than providing alternative LLMs with permissive licenses that can be hosted anywhere in the world... but to each their own, I suppose.

Being open weight, the Chinese models can be served the same way as Anthropic's: via AWS or GCP. Or whomever really, or on prem.

Don't trust US or China. Trust the open source community.

You don’t have to run the chinese models in chinese data centers as many of them are open weights. Some could say that trumps both.

> I know which one I'd rather send tokens to.

Do you have access to Mythos? If not the choice has already been made for you.

> Between the Chinese government and Anthropic, I know which one I'd rather send tokens to.

You meant to write "Between the Chinese government and the US government". Completely agreed though, better to send it to the former.

What if the model is hosted on a 3rd party site ?

You don't have to send your tokens to the CCP to use the Chinese models, that is the beauty of it. You can find GLM, Minimax, Deepseek, Kimi, etc hosted in China, Europe, the US, and probably elsewhere depending on what your geographic preferences for token transport are.

Except the open weights can live on your hardware so no one gets your tokens.

Given recent history, I'd be more likely to be killed by American actions than Chinese ones.

China won't arrest you for wrongthink where you live now

At least China isn't a terrorist state run by Jeffery Epstein Associates. It's a nation run by engineers, and it has flaws, but at least the people at the top know how to read.

If I can run it locally, it is certainly more safe than running on some cloud server somewhere.

well at least they're a little further away

They are collecting tribute is what they are doing.

Yes.

A published policy with the right to appeal exclusions from the list.

An equal standard for all companies rather than ad hoc application.

A countervailing policy to mitigate the unfair advantage conferred on the companies that have early access (such as a higher tax rate that goes to fund ai job loses, and a commitment that AI use of the new models won’t result in layoffs).

A requirement that hardware is made available for open source models rather than locked up in by the AI labs.

A restriction on AI labs being vertically integrated from hardware all the way up through the app layer. I would restrict AI labs to being API providers and prohibit them from building apps. That would allow an ecosystem of independent software development on the app layer without fear of being copied by the labs that have an unfair advantage in seeing the data while apps are being built, the usage data as they become successful and the ability to undercut competitors by subsidizing tokens unfairly.

I could go on.

What real risks? Genuinely.

I've read all the sci fi they have. It's not hard to see where the ideas came from.

What's being questioned is this sentiment of "The only way to save humanity is for me and my lighthaven groupies to become Xillionaire god-kings"

Is there any evidence at all that would convince you they're just blowing smoke up our asses in pursuit of money?

I mean, we're talking about an administration that has already over-reached in regulating this specific company out of personal bias; is openly seeking leverage over companies for favoritism and graft; hires on the basis of loyalty to whims of a narcissist; makes fun of the whole idea of competent government based on expertise; provides a range of conflicting explanations for whatever it chooses to do; and has been unable to field a team capable of understanding or explaining whatever real risks are here.

Your question is like asking what evidence would convince us that a bag of rocks doesn't have rocks in it. Easy, just take the rocks out.

> So only 100 companies have exclusive access to frontier AI.

Only the frontier AI labs have true access to frontier AI. Everyone else gets a reduced version.

Under the hood, yes, but Mythos had more relaxed safeguards and was/is only available to a subset of approved customers under Project Glasswing, similar to the situation with GPT-5.6 now.

Fable was available to me as a normal person using Claude.ai

Mythos never was and I don’t think that’s changing.

Mythos doesn't have the strict safeguards of Fable and is only accessible by a very small number of pre-approved companies.

I don’t think everyone agrees on that. It’s just that the government has been able to legislate that as technology evolved.

That being said many legal scholars say the state militia was intended to be the defense against tyranny not individual citizens because there were government led crackdowns on rebellion under Washington and other presidents from the earliest days of the republic. State militias have the full range of weapons

Well... there are crypto startups, and perhaps a generous definition of "money"

Will startups be even a thing now that the VCs obviously just need to funnel all their money to 2 or so companies ad-infinitum for guaranteed returns.

> Does anyone think a startup with a good product is going to be materially disadvantaged by not having access to an incrementally better security focused LLM release?

- It's not "incrementally better". It's a complete game changer. Opus 4.8 on max thinking does X amount of mistakes in my commercial work. Fable 5 did 5% of X. Counted. I barely had anything to contribute in the work sessions, for a full week I could count on my two hands the total amount of times I actually caught Fable 5 -- and one part of those were not true mistakes, more like divergence from policy in our `CLAUDE.md` files.

- It's not "security focused". It's simply better in every way _plus_ it's also security-conscious.

- It legitimately accelerated my work. I don't have too much unknowns in my work, I simply have way too much to do. Fable 5 was an objective and measurable improvement over Opus 4.8. Returning to it after Fable 5 was removed was extremely discouraging and frustrating, and still is to some extent.

> It’s lots of fun to pretend it’s some step-change that’s too dangerous for general release

Maybe, but not as much fun as tearing down a straw man apparently. :)

> (Just to be clear, I think the gatekeeping is ridiculous, especially given the above)

It's ridiculous for multiple other reasons but ridiculous nonetheless.

That kind of gets to the absurdity of it. Either it’s a wildly powerful next generation model with incredible capabilities and thus needs to be limited… or it’s another progressive enhancement like we’ve seen already and limiting access to it makes no sense.

Yes, i do. I have 10xd my productivity since last year and im not smarter. And yes my code is high quality

> Regulatory agencies limit uses of other products without acts of congress-- cigarettes, vapes, drugs, pesticides, chemicals, explosives.

Every one of those is by a regulatory agency that was explicitly empowered by Congress to do such regulation.

  Dajcie mi człowieka, a paragraf się znajdzie

  Give me the man and I will give you the case against him

The ATF was created by an act of congress. https://en.wikipedia.org/wiki/Gun_Control_Act_of_1968

> I agree that it should require an act of Congress to explicitly delegate this power.

Should ever new "weapon" invented require a new act of Congress? We've considered software subject this act since the 90s.

If everyone making AI is screaming up and down that we are in an AI arms race creating dangerous entities that will determine the fate of the world is the government just supposed to ignore them?

All of the agencies responsible for those regulations were created by and get their funding from Congress. Currently, they're asleep at the wheel. Or a better idiom might be "cowering in the corner".

None of those things are knowledge. I think theres something specific around limiting access to knowledge and capabilities that makes this feel insidious.

Fairly certain all those have "acts of congress" attached to them. I mean, it used to take a constitutional amendment to make something illegal but now we have tons of agencies responsible for regulating all the things.

Plus, they're relying on the "math is a weapon" law to ban "export" of the models.

"Malboro cigarettes may once again be sold, but Newport remains banned for everyone except large purchasers that have paid the appropriate bri... fees."

The Chevron doctrine gave more power to the executive agencies of the current administration, so I’m not sure what point you’re trying to make.

Growing up rural, the grift has been obvious my whole ĺife

But it was applied using principles of the rule of law with clear regulatory frameworks. This is not that

Jumping in to reflexively defend the admin again, I see.

Is there any policy from this admin you don't support?

Also known as, the tu quoque fallacy. Just because politicians in both parties have been doing this for decades doesn't mean that this administration is not especially hypocritical for doing it after whinging so much about free speech and free markets.

The only correct reply.

Yes they have.

This is a very bad analogy. Markets behave like an imperfect optimisation algorithm, and you can prove that, under some conditions which are most often met, they give people what they want. In fact, you can almost always expect governments to be less effective and less rational than markets in allocating resources to satisfy the desires of people, even when democratic. You can prove it either by using the same logic that tells you when markets fail (externalities, information asymmetry), or empirically by looking at what was basically the most perfect A/B test we had on society over the 20th century. Although it was a comparison between mixed economies and fully centralised ones, there is no reason to expect the optimum mix of centralisation/distribution to be closer to the worst-performing one (the fully centralised one).

Yes. But that was the “big government democrat” argument that republicans said was evil and un American.

I think they’d try to get something through Congress to regulate the industry in a rules-based way.

The current admin flies by the seat of their pants and at least creates the perception of political decision making.

We don’t have to guess. Democrats had a pragmatic policy under Biden — which was rescinded by Trump when he took office.

https://bidenwhitehouse.archives.gov/briefing-room/statement...

Fourth Amendment, through corporate data purchase or exfiltration.

They have affiliated themselves to him. Watch, within a month of Democrats being back in power they’ll be harping small government, denigrating the national debt they ballooned themselves. There’s no reason to help them attempt to disavow it.

But it's not just him, it's the entire party aggressively supporting him and everything he does.

it's actually the entire party that's propping him up. If it was just trump he would be living on the street.

Trump has an 87% approval rating amongst Republicans as of the last poll I can find.

While Trump is a megalomanic and does whatever he wants, he has the mandate of the Republican party, whose elected officials could choose at any moment to end this by withdrawing support.

Don't let them off the hook.

It seems people can flip that coin whenever it suits them.

And yet the rest of the party falls in behind him.

The entire Republican party in all branches of government is supporting Trump. His politics and the Republican party politics are one and the same. The last election the party did not have a platform because, quite literally, they said that whatever Trump says _is_ their platform.

He's a Republican backed by the Republican establishment funded by Republican donors and massively influential in Republican primaries. Republicans voted him into power twice. Republicans pushed his voter fraud narrative. Republicans embraced his vaccine skepticism and killed countless Americans. Republicans voted for his ICE policies that murdered two citizens of my home state.

Republicans caused this disaster and are all, each and every, individually morally responsible for putting Trump in power.

Republican voters, Republican politicians, Republican donors and the Republican political machine.

They picked the losing side of history and they can sink with it.

What Anthropic was pushing for under Biden has very little to do with the values Republicans have been espousing (and failing to live up to) for decades. That's kind of the point op was making. Republicans run on small government but do not deliver it. Democrats do not run on small government. Democrat Presidents campaign on and push for things like the ACA, they don't have fun quips like, “The nine most terrifying words in the English language are: I'm from the Government, and I'm here to help.”

A clear regulatory framework to operate within allows businesses to operate within it rather than get surprised by the King's whims upending their business on every few Fridays. If you expect regulation will eventually happen, pushing for it to happen on terms you're able to comply with rather than as haphazard surprises is pretty sensible.

It's very fashionable to hate on the current administration, despite what the previous administration was doing. That's reality and I'll be punished to hell for saying so.

> It is just utterly baffling to me. I'm... well along the spectrum... so people not responding correctly to obvious information is just something I've gotten used to but just... wow.

People get quite a kick out of seeing people they don't like get hurt. They can stay entertained by that for a long time until it bites them.

Only now is it finally biting with the collapse of the rural medical clinics, the war induced spike in the price of gasoline, etc.

That's probably playing a big part in the seeming shift in the electorate in every election.

Republicans trust Americans with guns and not an LLM. That should tell you something.

The one currently running the show.

The it did a pretty shit job of it. Within 100 years it was killing hundreds of thousands to fight against that purpose.

This should perhaps not be surprising considering the president.

They cherry pick and choose the parts of her work that fits their agenda, while forgetting the other parts.

Milton Friedman was prescient on this:

> The two biggest enemies of the free market are two separate groups: my academic colleagues and business people. Business people are enemies of free markets, not friends.

> [...]

> The business people are just the opposite. They're all in favor of freedom for everybody else, and at the drop of a hat you can get any leading businessman to give you an eloquent speech on the virtues of a free market. But when it comes to their own business, they want to go down to Washington and get a special tariff to protect their business. They want a special tax deduction. They want a tax subsidy. And Chrysler is on the verge of failing, which it should have done. It should have been allowed to fail. Chrysler goes down and exercises political influence and tries to get the government to lend it money to subsidize it.

> So businessmen in general — not all, there have been some notable exceptions — and I don't want to include everybody. But in the main, most businessmen are enemies of free markets.

https://www.youtube.com/watch?v=vhgaPVO8aw8

I think it’s a complex dynamic but he certainly preferred this to happen than not, and someone might have nudged the development to his favor while aligned well with this admin’s interests.

They are all private aren't they? There's nothing to crash since the valuations were all made up funny raise numbers anyway. A donation to the right person likely removes the restrictions

Depends on how much of an overhang there is with the power of existing models. Have we discovered 10%, 50%, or 90% of the valuable applications for Opus 4.8 / GPT 5.5? Hard to be confident at this point.

> If this is the way things are now, isn't that going to crash the AI stocks?

Maybe, maybe not. Tech stocks are mostly vibes-based now, reality isn't really a concern for them.

Stop contributing now. Why wait?

"disastrous" seems hyperbolic to me.

what disaster do you foresee?

Section 5 Part 2 (Information Security) of the EAR Commerce Control List likely applies.

IANAL; you need to be one to interpret this stuff. These laws are as thick as a dictionary.

The EAR dates back to the Export Administration Act of 1979 but it was mostly overhauled by Congress in 2018.