Top
Best
New

Posted by theanonymousone 11 hours ago

Startup founders urge U.S. government not to shut off Chinese open weight AI(www.politico.com)
https://littletech.org/

https://static.politico.com/4a/bf/9c4021d8404386b0a311dcccf0...

717 points | 644 commentspage 8
1saadcodes 3 hours ago|
[dead]
chrisnight 11 hours ago||
> the US can “sanction” companies that engage in intellectual property theft.

Peak hypocrisy, US AI companies can train on unlimited intellectual property with 0 rights to it, while Chinese AI companies have to explicitly get the rights to data that isn’t even copyrightable/copyrighted (since AI alone can’t copyright it).

hnfong 11 hours ago||
It's not IP theft if US courts say it isn't.

And if the accused company is outside of the US, well, US courts have no jurisdiction so apparently the US government can just claim they are guilty and impose the sanctions...

Frieren 10 hours ago|||
The USA has stolen intellectual property from ALL AROUND THE WORLD. AI is not trained just in copyrighted USA content.

Other countries have a say about it.

throw1234567891 10 hours ago|||
And it’s not a war if Putin says so. And it’s not a possible future invasion of Taiwan if Xi says so. Who cares what US courts say.
InsideOutSanta 10 hours ago|||
Meanwhile, X.ai committing all of the crimes.
1234letshaveatw 10 hours ago||
did you mean Xi.ai?
MangoCoffee 11 hours ago|||
US's court can enforce IPs. Didn't Antrhpic just agree to pay book authors billion for the recent lawsuit settlement?
ctxc 10 hours ago|||
Anthropic was asked to pay book authors because it trained on pirated downloaded books.

What about books and art where the author/artist does not authorise AI to train on it? They do happily train on it, ignoring their "ToS".

This is just double standards, a slap on the wrist to not worsen the situation with authors imo.

You could also say the Chinese companies are doing the same - they _do_ pay for their Anthropic subscriptions after all.

WarmWash 10 hours ago|||
>Anthropic was asked to pay book authors because it trained on pirated downloaded books.

No, it was asked to pay book authors because it pirated copies of books and stored them on their hard drives. The ruling had nothing to do with training.

ctxc 10 hours ago||
Thanks - I didn't verify this personally, but still stand corrected.

If training wasn't considered outside the law, this goes on to make the point about double standards for US vs Chinese model training methods.

adrr 10 hours ago||||
Pirating millions of dollars worth of books would land people in jail. Government went after Aaron Scwartz for "pirating" scientific journals.
bloppe 10 hours ago|||
Aaron Schwartz was re-publishing the journals, which is the core concern or copyright and indeed where the word "copyright" comes from. Anthropic could get in a lot more trouble if their models are caught reproducing copyrighted work from their training set wholesale.

I think the Aaron Schwartz case is incredibly vexing because he was obviously acting out of a sense of altruism without personal self-interest. I don't think he deserved the book getting thrown at him like that. But the whole copyright system, which people seem to think is simultaneously good and bad, kinda rests on not allowing those kinds of violations

NicuCalcea 9 hours ago||
I've linked this several times, but LLMs are capable of reproducing entire books. Researchers were able to extract books nearly verbatim: https://arxiv.org/abs/2601.02671
ctxc 10 hours ago|||
I didn't think about this parallel, but that's just so sad.
chmod775 10 hours ago||||
The judge explicitly ruled that training on the books was fine.

They just should have bought them, rather than pirating them.

Also LLM output is not IP (in itself) in the first place, nor would Anthropic want to claim it is and that they have rights to it - that would drive paying customers away.

The issue comes down to at most ToS violations.

nickthegreek 10 hours ago||
> They just should have bought them, rather than pirating them.

Bought, scanned and destroyed them I believe. The judge okay'd Destructive Scanning.

bloppe 9 hours ago||
Destruction is not necessary. Google Books is a solid precedent. You can keep the content, you just can't make significant parts publicly available
nickthegreek 9 hours ago||
> The court also held that the third factor favored fair use as to the purchased library copies converted from print to digital because the purpose of the copying was to keep the books in its library but with more favorable storage and searchability properties. This purpose required copying, there was no surplus copying and the source copy was destroyed. With respect to the pirated copies, however, the court held that because “Anthropic lacked any entitlement to hold those copies” and retained them “even after deciding it would not make further copies from them for training,” this third factor weighed against Anthropic for that particular use.

https://www.loeb.com/en/insights/publications/2025/07/bartz-...

You need to destroy the _physical copy_ that you scanned.

Rebelgecko 10 hours ago||||
The settlement doesn't cover the actual training on copy written material
zuzululu 11 hours ago|||
That will never happen. Copyright is not a thing in China.
aerodexis 10 hours ago||
too bad hypocrisy never got in the way of people making money
eggy 9 hours ago||
[flagged]
snootypoot 11 hours ago||
[dead]
CurbStomper 10 hours ago||
[dead]
epolanski 9 hours ago||
What is this socialism about using non-us open weight ai?

Dystopic.

What next? Some communist open source Finnish operating system?

SubiculumCode 11 hours ago||
Not in Trump's power.
logicchains 11 hours ago||
If he can ban Chinese TikTok for alleged national security reasons, I'm sure he can also similarly ban Chinese AI.
hnfong 6 hours ago||
That said, it's harder to ban a terabyte of floating point numbers under a MIT license.
b112 11 hours ago|||
https://news.ycombinator.com/newsguidelines.html

Please don't use uppercase for emphasis. Instead, put asterisks* around it and it will get italicized. More formatting info here.*

SubiculumCode 10 hours ago||
fixed.
b112 9 hours ago||
I vouched, not sure it will help.
functionmouse 11 hours ago||
he could order possession of it treason and punishable similarly to other illegal numbers.
bakies 11 hours ago||
no he can't, he doesnt make laws
amanaplanacanal 10 hours ago|||
Legality doesn't seem to have much effect, in his case.
tensegrist 11 hours ago|||
executive orders?
thewebguyd 9 hours ago|||
A president cannot (legally) use an EO to create new laws that bind private citizens or businesses. EOs can only be issued to federal agencies, and only Federal Agencies & Employees (and federal contractors) are bound by direct compliance of an EO.

Now that Chevron is no more, EOs are even weaker than they used to be because the regulatory agencies no longer have teeth.

Premature compliance != legally required compliance. Anyone complying early to an EO when they don't have to is an idiot.

bakies 10 hours ago||||
The executive can't enforce laws that don't exist
AnimalMuppet 10 hours ago|||
Yes, he can do executive orders. The courts still get to judge whether they are legal, though... some months/years later.

For the more egregious ones, there are temporary restraining orders... days/weeks/months later.

So, yes, he can do that, and yes, he can be overruled on them, and yes, that happens after some damage has already been done.

bethekidyouwant 10 hours ago||
“And they warn that banning Americans from downloading Chinese open-weight models wouldn’t stop their proliferation — but would weaken U.S. startups.”

Well this doesn’t even make sense.

qlte 9 hours ago||
Weaken U.S startups = Can’t legally develop commercial products that use or remix the latest open weights models in novel/creative/cost-reducing/etc ways, while companies from the rest of the world could (including China, who this is ostensibly supposed to be punishing)

Wouldn't stop proliferation = Despite that, your non-programmer 14 year old child can still install the latest Qwen model on Day 1 by copying a one liner command they got from a YouTube short that sources from a Gitlab repo outside US jurisdiction

krupan 9 hours ago||
It does because corporations, especially smaller ones competing with bigger ones, actually do get prosecuted for violating restrictions like this a lot more than individuals do.
gdiamos 10 hours ago|
I think we should shut it off. It would force US companies to build open models.
preg_match 10 hours ago||
I think it would do the opposite. They would have a monopoly just handed to them, and a large market with no options. They'd have zero incentives to provide open weight models.

Right now, I think their incentive is the highest: Chinese AI companies are dropping competitive models left and right, and clearly American companies are feeling the pressure, given they're pulling every legal leg they can to suppress their competition. If we just remove the competition, Anthropic et al win. It's a temporary and short-sighted victory, but this is America. Every company is short-sighted, it's baked into our financial infrastructure.

thewebguyd 10 hours ago||
US labs would have even less incentive to build open if we blocked foreign open weight models, and we'd be engaging in protectionism yet again, also giving them zero incentive to compete.