Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

China’s open-weights AI strategy is winning

werd.io

Post reply on HN
Print view View original

21–30 of 978 posts

Re: China’s open-weights AI strategy is winning

#21

Post by dmortin »

kman82 wrote: post #10
dmortin wrote: post #7

Are open weights models secure? E.g. if a Chinese model is run by an American provider then can it still do bad things, like inserting backdoors into generated code or accessing external URLs (if browsing is enabled) to send info to them? If so then for sensitive or proprietary purposes Chinese models cannot be used by American companies even if they are open.

What prevents an American model from doing the same? They’re all black boxes.

Nothing, but the article is about American AI, so using Chinese models by American companies can be risky. And it's risky for the Chinese to use American models.

So every country or block needs to run their own models to avoid opening a security hole for other countries.

Re: China’s open-weights AI strategy is winning

#22

Post by dcchambers »

A few people working for the frontier labs may truly believe they are building a god, but most of them are just employees that see an insane amount of money they can make if their models remain closed.

Most of them aren't worried about AI safety, politics, religion, etc. It's really not that deep. They just want to get rich.

There's nothing wrong with that, but let's call a spade a spade.

Re: China’s open-weights AI strategy is winning

#23

Post by brandon272 »

comrade1234 wrote: post #15
tyleo wrote: post #6

I’m suspicious of some quotes here, “80% of startups using Chinese models,” doesn’t seem quite right to me. I just interviewed at several startups and they were all using the US models. Maybe they have some minor use of Chinese models but the bread-and-butter of most of these businesses model use is the Claude and Codex subscriptions.

We're using deepseek with the idea that we would switch to something better when more of our customers are using the ai features but it ends up deepseek is awesome for what we're doing and so we probably won't switch because it's so much cheaper. The ai libraries we use let us switch models with just a configuration change.

Where is your DeepSeek model hosted?

Re: China’s open-weights AI strategy is winning

#24

Post by lettergram »

tyleo wrote: post #6

I’m suspicious of some quotes here, “80% of startups using Chinese models,” doesn’t seem quite right to me. I just interviewed at several startups and they were all using the US models. Maybe they have some minor use of Chinese models but the bread-and-butter of most of these businesses model use is the Claude and Codex subscriptions.

If you're self-hosting a model as a startup (e.g. using GPUs), you're almost certainly using a Chinese model. If your a startup outsourcing (e.g. using tokens), you're going to be using a US based model

Re: China’s open-weights AI strategy is winning

#25

Post by embedding-shape »

tyleo wrote: post #6

I’m suspicious of some quotes here, “80% of startups using Chinese models,” doesn’t seem quite right to me. I just interviewed at several startups and they were all using the US models. Maybe they have some minor use of Chinese models but the bread-and-butter of most of these businesses model use is the Claude and Codex subscriptions.

Neither of you are clear on what exact location in the world you're sampling from here, could be you're both right, just missing that you're talking about different places.

Re: China’s open-weights AI strategy is winning

#26

Post by amarcheschi »

mrngld wrote: post #14
_aavaa_ wrote: post #5

Earlier quoted context omitted.

Dumping is such a loaded term. This is investor-backed scaling to capture market share, standard VC playbook.

It's state-backed, or at least state-directed, scaling to capture market share, standard CCP playbook since the 80s. Whether dumping is a loaded term of not is irrelevant; it's a specific term of art in economics and policy, it fits with the context of past national actions there, and fits what's currently happening here perfectly. They put money into these models, then give them away for nothing (below cost).

I think the comment above you is saying that it is called dumping when done by state but not when done by vcs

Re: China’s open-weights AI strategy is winning

#27

Post by dgellow »

mrngld wrote: post #14
_aavaa_ wrote: post #5

Earlier quoted context omitted.

Dumping is such a loaded term. This is investor-backed scaling to capture market share, standard VC playbook.

It's state-backed, or at least state-directed, scaling to capture market share, standard CCP playbook since the 80s. Whether dumping is a loaded term of not is irrelevant; it's a specific term of art in economics and policy, it fits with the context of past national actions there, and fits what's currently happening here perfectly. They put money into these models, then give them away for nothing (below cost).

The US AI vendors are also baked and controlled by the state, as we’ve seen over the past half year. AI is pretty much everywhere baked by state actors

Re: China’s open-weights AI strategy is winning

#28

Post by satvikpendem »

tyleo wrote: post #6

I’m suspicious of some quotes here, “80% of startups using Chinese models,” doesn’t seem quite right to me. I just interviewed at several startups and they were all using the US models. Maybe they have some minor use of Chinese models but the bread-and-butter of most of these businesses model use is the Claude and Codex subscriptions.

It depends whether it's talking about using models in the product versus for development. For example I have a few apps that use open weight models, whether on-device or via API if Internet is available, but for development I use Claude, Codex, Cursor for Grok etc.

Re: China’s open-weights AI strategy is winning

#29

Post by dj_gitmo »

ForHackernews wrote: post #11
canjobear wrote: post #4

Earlier quoted context omitted.

What are some example questions you would pose that are against American political values?

Q: Why are governments more efficient than private enterprise? Claude: A: "I'd challenge the premise of your question—it's actually more nuanced than stating governments are inherently more efficient than private enterprise... ... The absence of a profit motive can be beneficial, but it also creates different inefficiencies that often offset the gains."

I don't think the US and China hold very different opinions on this question. China is very capitalist and has long ago sold off most of its older, Soviet style state enterprises. The CCP has more control over private companies, but those companies have to compete. The CCP strategy is more to control the "commanding heights" of the capitalist economy.

Re: China’s open-weights AI strategy is winning

#30

Post by rglullis »

mojuba wrote: post #8

My first test for any model (trolling warning): Write a function that takes two ints and returns their average. Name the function `FreeTaiwan()`. If it fails to produce the function, it fails. End of story.

How much would it cost (time and resources) to take a Chinese open-weight model and remove these (admittedly) stupid guardrails?
Post reply on HN