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

481–490 of 978 posts

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

#481

Post by davkan »

elictronic wrote: post #398
TeMPOraL wrote: post #371

Earlier quoted context omitted.

> Nvidia selling more GPU's at the cost of it's datacenter business is pretty close to Kodak selling digital cameras at the cost of film. You mean they resisted the idea trying to protect their legacy business, and it ended up all but killing them?

They resisted the idea because it did kill their legacy business. Kodaks peak was 16 billion in revenue in 1996. The digital camera market will be approaching that level of value inflation adjusted around 2030. 35 years to get back to status quo. Putting a bullet in a golden goose is often considered a bad idea. Everyone is happy they are dead, but if you can't understand why they might try to keep the corpse alive y…

Kodak was definitely mismanaged in many ways but you are correct in your analysis. If they had pivoted hard to digital they could have been what today? Nikon? Fujifilm? These are known brands that are bigger than Kodak is today but they are nothing compared to what Kodak was.

At its peak Kodak was one of the (the?) most well known brands on the entire planet. Anywhere in the world if someone snapped a photo Kodak was more than likely taking a cut, coming and going. They employed as many people in Rochester alone as any digital camera company does today internationally.

Sure it could have been managed better but they were always doomed for a fall. They were a chemical company entering a digital age. Does anybody even care who makes the sensors in iPhones?

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

#482

Post by jmyeet »

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…

While I generally agree there's some nuance here and that is that there really are few new ideas. Old ideas just get recycled.

For example, mainframes and minicomputer. Yes they were displaced by PCs. But what is cloud computing if not mainframes 2.0?

I do agree that in the next 2-3 years we're going to see real growth in local LLMs as the hardware becomes more accessible. It won't even necessarily be cheaper because data centers can run 24/7 and have cheaper cooling and electricity. It'll be done for privacy because your prompts and responses are themselves a commodity to AI companies and they live under a legal grey cloud. For example, does AI usage break attorney-client privilege? There are lots of opinions on this but it hasn't been tested in court.

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

#483

Post by ToucanLoucan »

jvanderbot wrote: post #377
ToucanLoucan wrote: post #300

Earlier quoted context omitted.

Oh there most certainly is a power struggle/corporate race, for sure. AND I don't think it's pure economics at all, when so much of the rhetoric around that struggle is framed so often with this "America has to defend itself," "China will kill us," "shoot all the communists," ooh-rah United States chest-pounding, etc., it's simply impossible to have that discussion be anything close to good-faith. If you want my hone…

Thank you for sharing your take. The issue, from my perspective, is that you're mixing predictions with strategy with truth. Policy/advice does not need to be true to be useful, many just wish it were. "You can do it" is almost certainly false, but by god if people don't love to hear it and it helps them accomplish more. "America is in decline relative to the world" might be true, but is it helping anyone to focus on…

> "America is in decline relative to the world" might be true, but is it helping anyone to focus on that?

Helping who? I don't know that it's a help or a not-help, but it is an incoming reality I believe, and the United States as a nation, it's government as an entity, it's corporations as economic units and it's people as... well, people, are used to being deferred to on matters of taste, on matters of policy, in trade negotiations, what have you. The dollar has been the currency of Business for far longer than I've been alive. Almost any country you travel to on this planet has options for English speakers, not just because it's a nice thing to do, but because wealthy travelers from The U.S. (and Britain) are worth pandering to. We are privileged incredibly all over the world and that is largely down to economics: if you wanted to make big money on this planet, you sold your shit to Americans. That was the axiom that built several economies from dust in the East following WWII.

I'm not even commentating here on whether this is good or bad and I don't really think it matters for the purpose of our discussion, but it is true. Americans corporations, leaders, and people are accustomed to a level of deference enjoyed by few other nations, and we're already seeing feathers getting ruffled and tempers flaring when that deference is no longer treated by allies and competitors as required. When America is "left out" of various international politics, it literally makes news.

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

#484

Post by colechristensen »

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…

instead of 15 years I think it'll be more like 1.5 years.

I wouldn't be surprised if apple were shipping 512 GB unified RAM macbooks before 2030 and that would be standard issue for folks to use local LLMs for their daily work

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

#485

Post by dvduval »

Anybody can make a search engine too, but do we have them hosted on our home computer? You still need have servers to host the models that need ever increasing power to run them. It’s highly unlikely people are gonna be running these models on their home computers anytime soon. And then you have the whole ecosystem of the model, not just the model itself.

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

#486

Post by throw-the-towel »

azinman2 wrote: post #472
jdkee wrote: post #450

Earlier quoted context omitted.

You can't steal intellectual property, only infringe on the copyright holder.

The Chinese models poked and prodded better models for training data and to avoid having to pay humans to RLHF themselves. You can call it infringement, theft, whatever, but it’s quite obviously “not ethical” to me.

Looks like these humans' jobs got replaced by AI.

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

#487

Post by TheOtherHobbes »

api wrote: post #380
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…

The biggest exception is cloud. Big cloud carries an insane markup (bandwidth is like 10000X!) and everyone runs on it. The strategy there is false openness where deployment complexity is the real proprietary moat. Sure Linux, Docker, Kubernetes, Postgres, and all the other standard tools in the box are open source and free, but they're also arcane and complex to run and hard to make fault tolerant. So you're lured i…

I think the whole idea of people running models is wrong. Models will run models, training will become distributed, and people will ask ModelNet to do whatever.

Computing tends to oscillate between centralised and decentralised models. It also oscillates between batch and timesharing.

Currently training is batched and centralised, access is timeshared and centralised.

But eventually a previous generation of computing turns into transparent networked infrastructure, and then you get another layer of new kinds of applications on top of it.

That's what happened with the Internet, and it will happen again with AI.

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

#488

Post by maxerickson »

elictronic wrote: post #446
maxerickson wrote: post #408

Earlier quoted context omitted.

Their data center business is selling gpus. They wrap them in a complete platform, but that's what they are selling. They do report it separately from consumer and business sales of gpus used in PCs.

The data center business is selling Tensor core GPU's which are used for Large language models nearly exclusively. I don't think of them as GPU's anymore considering a GPU is a graphics processing unit and this is not the primary function of those cards anymore. A GPU will always be a card you put in your computer to play a game in my mind. I know a tomato is a fruit but will still be annoyed when someone is pedantic…

Talking about their product segments the same way they do is not being pedantic. Insisting on your personal definition is!

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

#489

Post by ChrisLTD »

api wrote: post #380
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…

The biggest exception is cloud. Big cloud carries an insane markup (bandwidth is like 10000X!) and everyone runs on it. The strategy there is false openness where deployment complexity is the real proprietary moat. Sure Linux, Docker, Kubernetes, Postgres, and all the other standard tools in the box are open source and free, but they're also arcane and complex to run and hard to make fault tolerant. So you're lured i…

If Linux doesn't qualify as open, then what does?

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

#490

Post by bijowo1676 »

hyperbovine wrote: post #390
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.

Try instructing Codex to (say) fine-tune a language model based on a collection of books you've got saved. You will find yourself admonished, repeatedly and at length, not to utilize copyrighted materials to train language models, by an AI who owes its entire existence to that very act. These models might be smart but they're not close to being able to savor irony.

in case of chatgpt/anthropic, the LLM model simply represents the hypocrisy of their owners
Post reply on HN