Live data from Hacker News

Startup founders urge U.S. government not to shut off Chinese open weight AI

politico.com

311–320 of 983 posts

Re: Startup founders urge U.S. government not to shut off Chinese open weight AI

#311
post #275

Earlier quoted context omitted.

> by insisting that the major instance of military intervention by a world power that's happened in the last year shouldn't be counted I’m not. I’m saying exempting that, we have an even field of conflict mongering. Counting Iran, the field is biased, but we have significant sampling-timing issues for something like “is there war.” China has violently annexed sovereign countries in the past decades. Its threats to do…

>"China has violently annexed sovereign countries in the past decades." An outright lie. The only possible candidate is Tibet but it's former sovereignty is very questionable and not clear cut

> it's former sovereignty is very questionable and not clear cut

Same as Taliban-run Afghanistan, same as Iraq, same as Ukraine, same as Taiwan. If you’re Tibetan, the line is academic—you faced, by modern definitions, genocide all the same.

Calling the territory you’re invading and/or annexing illegitimately independent, illegitimately governed is how these always go.

Re: Startup founders urge U.S. government not to shut off Chinese open weight AI

#312

“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.

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.

Re: Startup founders urge U.S. government not to shut off Chinese open weight AI

#314

Earlier quoted context omitted.

I think they're referring to occasional fist fights between the Indian and Chinese deployments at the top of the Himalayas where there's a dumb territorial dispute. There were a couple of times decades ago where the forces shot at each other and the governments had to scramble to calm things down and deescalate, so somehow they've reached this gentleman's agreement to have fistfights instead. https://en.wikipedia.org…

> We're talking about this kind of thing balancing out, say, Fallujah or shock and awe over Baghdad Nonsense. For that comparison you’d look to the invasion and annexation of Tibet. (And it’s not ancient history if it’s currently endorsed and un-apologized for.)

Ok, then we expand the US horizon back to the 50s as well. We've added Tibet, and the 1979 3 week Vietnam war to the Chinese column, plus them being involved in the Korean war and to a lesser degree the American Vietnam War. This comes with way, way more added to the American column. At least it's no longer a shutout.

Any fair look at both countries' records puts the US as doing 10-100x more military intervention, on a global rather than regional scale. It is actually a little bit "sheeple" to look at that scoreboard and start calling China aggressive. They're less aggressive than every other UNSC member without even going to a pound for pound basis.

Re: Startup founders urge U.S. government not to shut off Chinese open weight AI

#315

Proprietary model weights are IP, their outputs are not IP and I can't imagine a court decision that would rule otherwise because that would set an extremely far reaching and dangerous precedent, even for American businesses. I don't think distillation as 'stealing IP' has any legal legs. They can probably claim violation of ToS at best because the terms of use do prohibit use for training rival models.

> Proprietary model weights are IP

I don't think they are. Not copyright at least. There may be some "trade secret" stuff for them but they're not copyrightable.

Re: Startup founders urge U.S. government not to shut off Chinese open weight AI

#316

If there were ever a sign I need to get mirroring, this is it. What are the best Chinese models on HuggingFace today? Bucket by ideal RAM: Text generation, image generation, TTS, etc

qwen3.6-27B, gemma4-it-31B, poolside/Laguna-S-2.1 are all good text gen models generally in the Note, Laguna provided quants have some issues and they are reworking/updating them.

Re: Startup founders urge U.S. government not to shut off Chinese open weight AI

#318

Earlier quoted context omitted.

> We're talking about this kind of thing balancing out, say, Fallujah or shock and awe over Baghdad Nonsense. For that comparison you’d look to the invasion and annexation of Tibet. (And it’s not ancient history if it’s currently endorsed and un-apologized for.)

Ok, then we expand the US horizon back to the 50s as well. We've added Tibet, and the 1979 3 week Vietnam war to the Chinese column, plus them being involved in the Korean war and to a lesser degree the American Vietnam War. This comes with way, way more added to the American column. At least it's no longer a shutout. Any fair look at both countries' records puts the US as doing 10-100x more military intervention, on…

> Any fair look at both countries' records puts the US as doing 10-100x more military intervention

Sure. I never claimed otherwise. I’m saying it would be news to anyone in Asia-Pacific (or Eastern Europe) that America is or has been the destabilizing force. That was the original claim.

If the choice is 10x or 1x, go with the latter. But that is a false dichotomy. You can be anti-invasion and annexation broadly. There is no moral high ground in choosing Chinese models; no more than a person in China choosing Anthropic or OpenAI.

Re: Startup founders urge U.S. government not to shut off Chinese open weight AI

#319

Earlier quoted context omitted.

Specific use case is way too important for a list like that to be very helpful. I only know image gen well enough to use as an example, but style, resolution, inpainting, img2img or text2image, lora compatibility, etc. all lead to different “best” models. Even more complex for image2video, but there’s fewer models to choose from there at least.

Sanctions would cancel the whole business rationale for the Chinese government having/supporting open weight models. You can't commoditize your competitors with open-source if your product is blocked.

The open weights models would be even more effective by incentivizing people and companies to switch jurisdictions. It might even break SFs monopoly on AI, or at least weaken it, everybody isn’t there exclusively to funnel money to openAI and Anthropic.

It would also boost research in non US jurisdiction. Who is going to be wooed by “come to our lab where you’re only allowed to work with closed models!”

Re: Startup founders urge U.S. government not to shut off Chinese open weight AI

#320

Proprietary model weights are IP, their outputs are not IP and I can't imagine a court decision that would rule otherwise because that would set an extremely far reaching and dangerous precedent, even for American businesses. I don't think distillation as 'stealing IP' has any legal legs. They can probably claim violation of ToS at best because the terms of use do prohibit use for training rival models.

> Proprietary model weights are IP I don't think they are. Not copyright at least. There may be some "trade secret" stuff for them but they're not copyrightable.

Trade secrets are a type of IP.
Post reply on HN