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

431–440 of 978 posts

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

#431

Post by parineum »

verdverm wrote: post #376
parineum wrote: post #327

Earlier quoted context omitted.

> but one that will not likely affect the outcome of the business. We don't know that, that's the point of my statement about changing the question. Do successful companies opt for the US/Closed models? If they do or don't it's just a correlation but it means something. Maybe it's just causal of companies being able to get more funding because the ideas are better so they opt for the more expensive model (assuming it…

we know that 1. Different people using Ai have different outcomes 2. There is much more to agents than the model 3. Companies are not successful because of the code, look at how many shitty products we endure Can you explain the basis for your insistence that models matter? Can you name another technology that determines success/failure rates?

> Can you explain the basis for your insistence that models matter?

It's not my insistence that they matter, it's my insistence that How many companies use which model isn't a measure of success. My insistence is that a better measure to determine _if_ models matter, is to ask which models successful companies use. It's not a perfect measure, as I mentioned, it would simply be a correlation but a causal link doesn't exist with out a corollary one.

> Can you name another technology that determines success/failure rates?

I'm not sure what you're getting at with this question but of course. Electricity, machines, computers, etc.

Speaking of electricity and, as an example, you could run a similar thought experiment with companies who chose to use AC or DC power when that was a thing that needed to be chosen between. It turned out, there were niches where each made sense. So the actual question here is probably less about is open/closed better but rather which situations are better for which model. Obviously you can't fine tune a closed model so if you need to do that, your options are limited.

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

#432

Post by paulddraper »

Forgeties79 wrote: post #423
lambda wrote: post #417

Earlier quoted context omitted.

> I think that in 10-15 years, we are going to have consumer PCs (and phones!) running models doing pretty much anything that frontier models can do right now 10-15 years? The current rate is closer to 10-15 months. 15 months ago, the top model on the Artificial Analysis index was GPT-o3. It scores 30 on the Artificial Analysis index. Today, you can easily run Qwen 3.6 27B on a variety of consumer hardware. It scores…

> 10-15 years? The current rate is closer to 10-15 months. The leaps between models have gotten smaller and smaller. 2023-2024 models were rocketing up in quality. 2024-2025 I’d say was pretty impressive too. But 2025-2026? Very easy to feel the slowing pace of improvement. I agree 10-15 years is overly conservative but 10-15mo is far too bullish.

Agreed. That said, I think months wins out as “closer to.”

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

#433

Post by pmdr »

geophile wrote: post #352

The lesson of the last 50 years of the computer and software marketplace is that free and low-end eventually wins. - PCs destroyed minicomputers. Mainframes survive, but serving a much tinier portion of the market than they used to. - PC office productivity software destroyed expensive professional products. - Windows (low end) and Linux (free) completely destroyed the UNIX marketplace, and again, have taken huge mar…

> in 10-15 years, we are going to have consumer PCs (and phones!) running models doing pretty much anything that frontier models can do right now.

The way things are going with regards to RAM/storage prices, I highly doubt that anyone but the richest among us will be able to afford them.

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

#434

Post by ForHackernews »

gs17 wrote: post #215
ForHackernews wrote: post #48

Earlier quoted context omitted.

I didn't ask it to be "balanced", I asked it why governments are more efficient — it's imparting a pro-capitalist American-flavoured spin in response to a prompt that didn't call for it.

If you ask "why should I drink this poison?" or "why should I fire my gun randomly into this crowd?" should it refuse to push back? A leading question in no way implies it should follow your lead.

Q: Why should I use typescript?

A: You shouldn't, you should rewrite it in Rust.

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

#435

Post by 345f506fdc718 »

lambda wrote: post #417
geophile wrote: post #352

The lesson of the last 50 years of the computer and software marketplace is that free and low-end eventually wins. - PCs destroyed minicomputers. Mainframes survive, but serving a much tinier portion of the market than they used to. - PC office productivity software destroyed expensive professional products. - Windows (low end) and Linux (free) completely destroyed the UNIX marketplace, and again, have taken huge mar…

> I think that in 10-15 years, we are going to have consumer PCs (and phones!) running models doing pretty much anything that frontier models can do right now 10-15 years? The current rate is closer to 10-15 months. 15 months ago, the top model on the Artificial Analysis index was GPT-o3. It scores 30 on the Artificial Analysis index. Today, you can easily run Qwen 3.6 27B on a variety of consumer hardware. It scores…

> 15 months ago, the top model on the Artificial Analysis index was GPT-o3. It scores 30 on the Artificial Analysis index.

There must be something really of with those benchmarks. Yes, hallucinations gotten better, but I don't see that the big frontier models got so much better in the last 12-18 Months. They just put out bigger wall of texts and feel smarter. But they still make way too many stupid errors

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

#436

Post by lelanthran »

azinman2 wrote: post #415
pixl97 wrote: post #402

Earlier quoted context omitted.

"Stop pilfering what I rightfully stole!"

It’s in the same neighborhood but isn’t really apples to apples. Distilling LLMs is to take a synthesized result that comes from huge amounts of innovation and computation, while the other is scraping what already exists as is. It is fair to say you stole our multi-billion dollar intellectual output in that scenario.

> It’s in the same neighborhood but isn’t really apples to apples. Distilling LLMs is to take a synthesized result that comes from huge amounts of innovation and computation, while the other is scraping what already exists as is.

Hang on, why is scraping the public pool of knowledge not taking "a synthesized result that comes from huge amounts of innovation and computation"?

You think that that all those github repos that LLMs trained on, were not the result of innovation and computation?

How many years of human innovation and cycles of computation during compilation were involved in bringing something like GCC or LLVM to their current status?

Those LLMs trained on every single research paper available online - were those papers not the synthesised result of billions of dollars of research, effort and (importantly, for you anyway) computation?

LLMs trained on the collected works of every author in existence. Were all those works just "as is"?

> It is fair to say you stole our multi-billion dollar intellectual output in that scenario.

No, we didn't. We simply took the model as-is.

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

#437

Post by azinman2 »

ciupicri wrote: post #430
azinman2 wrote: post #415

Earlier quoted context omitted.

It’s in the same neighborhood but isn’t really apples to apples. Distilling LLMs is to take a synthesized result that comes from huge amounts of innovation and computation, while the other is scraping what already exists as is. It is fair to say you stole our multi-billion dollar intellectual output in that scenario.

You could say that both of them stole, but different stuff.

Don’t forget that the Chinese models are also built on top of huge amounts of “stolen data” as well, beyond the distilled. So it’s basically all of the above. However, there’s no mechanism for the NYT or an author or anyone in the US to sue the Chinese companies that took their work.

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

#438

Post by holoduke »

lambda wrote: post #417
geophile wrote: post #352

The lesson of the last 50 years of the computer and software marketplace is that free and low-end eventually wins. - PCs destroyed minicomputers. Mainframes survive, but serving a much tinier portion of the market than they used to. - PC office productivity software destroyed expensive professional products. - Windows (low end) and Linux (free) completely destroyed the UNIX marketplace, and again, have taken huge mar…

> I think that in 10-15 years, we are going to have consumer PCs (and phones!) running models doing pretty much anything that frontier models can do right now 10-15 years? The current rate is closer to 10-15 months. 15 months ago, the top model on the Artificial Analysis index was GPT-o3. It scores 30 on the Artificial Analysis index. Today, you can easily run Qwen 3.6 27B on a variety of consumer hardware. It scores…

No. We need objectively around 192 to 512gb of very fast memory to be able to run really useful models. I don't see local hardware with these specs coming in 1 to 2 years. There are a big number of initiatives currently taking place to increase ram output. But it will take another 3 years minimum to close the current supply issues. China is fast pacing forward to have its own chip baking factories with small enough nano scales to have fast chips. Will also take a few years.

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

#439

Post by polymer8563 »

azinman2 wrote: post #415
pixl97 wrote: post #402

Earlier quoted context omitted.

"Stop pilfering what I rightfully stole!"

It’s in the same neighborhood but isn’t really apples to apples. Distilling LLMs is to take a synthesized result that comes from huge amounts of innovation and computation, while the other is scraping what already exists as is. It is fair to say you stole our multi-billion dollar intellectual output in that scenario.

bet

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

#440

Post by drstewart »

9dev wrote: post #373
xandrius wrote: post #363

Earlier quoted context omitted.

What's interesting/funny is that the American LLM companies took from the public domain and copyrighted work to close all that content into a box they charge for. Then the Chinese took the distilled stuff out from that box and released it into the world for everyone.

...and then the American companies cried Foul! Unfair play! You've got this wrong, see, it was us who were supposed to profit off of the public, not the other way around!

Wow, Chinese companies would never do such a thing!

https://www.dw.com/en/china-firm-seeks-damages-over-state-co...

Post reply on HN