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

Deepseek: The quiet giant leading China’s AI race

chinatalk.media

Post reply on HN
Print view View original

101–110 of 476 posts

Re: Deepseek: The quiet giant leading China’s AI race

#101

Post by janice1999 »

andy_ppp wrote: post #77
qwertox wrote: post #36

Earlier quoted context omitted.

It remains to be seen how stable a totalitarian government can be. China has the benefit of having full control over its people and therefore gets to decide what is important and what not, and currently people are ok with handing that control over to the government. But it's also a very fragile state, which can only be retained through full repression.

Do the governments elected recently in the West look stable to you?

They do. Ireland just had an election I voted in and is forming a new government with multiple parties. The UK, after years of TV worthy Tory party drama, had its most transformative election in over a decade. I see active and engaged multi-party democracies with peaceful transitions of power and long established and respected laws for calling elections, no confidence votes (e.g. France, yes it happens) and so on.

Re: Deepseek: The quiet giant leading China’s AI race

#103

Post by cced »

carom wrote: post #44
djaouen wrote: post #43

Earlier quoted context omitted.

Protip: Some of us were not involved in the desecration caused by the East India Tea Company. Just because we look British means we should suffer like them, too?

They are referring to the fact that the US ignored European IP in its early days and relating that to what China is doing to the US now.

What’s also funny is that the promoters of the “China is stealing all of our IP in exchange for their labor” folks never mention why corporations don’t just pull out?

Are these IP thefts or technology transfers? If corporations are having their IP stolen, why don’t they just leave?

These narratives never explain or mention this. Idk why people still latch onto them, they are completely uninteresting “China is stealing all our IP and there’s nothing we can do about it except for continuing to allow our IP to be stolen” is an IQ test and trope.

Does “theft of IP” outweigh, or not, “access to very cheap labor (read: jobs)” ?

We need to stop simping for corporations and start thinking critically about these things.

Re: Deepseek: The quiet giant leading China’s AI race

#104

Post by tossandthrow »

joyeuse6701 wrote: post #98
tossandthrow wrote: post #19

Earlier quoted context omitted.

> ... why deepseek can never outperform ... This read more like a "western supremacists" post. 1. Only until China produces more compute than the west. 2. You don't have to ask ChatGPT / Claude many questions before realizing the grave censorship these are under - DeepSeek has access the roughly the same corpus of data as their western counter parts. 3. It is naive to think they only develop open source or will not s…

Xi has knee-capped anything a threat to his power, this Xi-ceiling as I call it, will prevent true cutting edge dominance compared to the West. Sure, there’s censorship in the West, but it’s not nearly as scary or effective as the East’s. Genius does not regularly spring under the sword of Damocles.

I am unconvinced that it is more technically complex to censor a historical event from an llm than it is to remove instructions on how to create explosives.

Re: Deepseek: The quiet giant leading China’s AI race

#105

Post by echelon »

jdietrich wrote: post #32
nsoonhui wrote: post #13

I find that the gushing around deepseek is fascinating to watch. To me there are a few structural and fundamental reasons why deepseek can never outperform other models by a wide margin. On par maybe--as we reach the diminishing returns with our investment in the models, but not win by a wide margin. 1. The US trade war with china which will place deepseek compute availability at disadvantages, eventually, if we ever…

I don't think it's necessarily about DeepSeek, but about the wider competitive picture. There are two tacit assumptions being made about LLMs - that having a SOTA model is a substantial competitive advantage, and that the demand for compute will continue to grow rapidly. DeepSeek's phenomenal success in reducing training and inference cost points to the possibility of a very different future. If it's the case that SO…

>> 3. Most importantly, deepseek is open source, which means that the other models are free to copy whatever secret source it has, eg: Whatever architecture that purportedly use less compute can easily be copied.

> I don't think it's necessarily about DeepSeek, but about the wider competitive picture. There are two tacit assumptions being made about LLMs - that having a SOTA model is a substantial competitive advantage

Everything is a game of ecosystems.

Windows lost to Linux on servers because it was cheap and easy to deploy Linux. Thousands of engineers and companies could build in the Linux playground for free and do whatever they wanted, whereas Windows servers were restrictive and static and costly.

Dall-E lost to Stable Diffusion and Flux because the latter were open source. You could fine tune them on your own data, run them on your own machine, build your own extensions, build your own business. ComfyUI, IPAdapter, ControlNet, Civitai... It's a flourishing ecosystem and Dall-E is none of that.

It'll happen with LLMs (Llama, Qwen, DeepSeek), video models (Hunyuan, LTX), and quite possibly the whole space.

One company can only do so much, and there is no real moat. You can't beat the rest of society once they overcome the activation energy.

And any third place player will be compelled to open source their model to get users. Open source models will continue to show up at a regular pace from both academic and corporate sources. Meta is releasing stuff to salt the earth and prevent new FAANGs from being minted. Commoditizing their complement.

Re: Deepseek: The quiet giant leading China’s AI race

#106

Post by GordonS »

joyeuse6701 wrote: post #98
tossandthrow wrote: post #19

Earlier quoted context omitted.

> ... why deepseek can never outperform ... This read more like a "western supremacists" post. 1. Only until China produces more compute than the west. 2. You don't have to ask ChatGPT / Claude many questions before realizing the grave censorship these are under - DeepSeek has access the roughly the same corpus of data as their western counter parts. 3. It is naive to think they only develop open source or will not s…

Xi has knee-capped anything a threat to his power, this Xi-ceiling as I call it, will prevent true cutting edge dominance compared to the West. Sure, there’s censorship in the West, but it’s not nearly as scary or effective as the East’s. Genius does not regularly spring under the sword of Damocles.

[flagged]

Re: Deepseek: The quiet giant leading China’s AI race

#107

Post by csomar »

nsoonhui wrote: post #13

I find that the gushing around deepseek is fascinating to watch. To me there are a few structural and fundamental reasons why deepseek can never outperform other models by a wide margin. On par maybe--as we reach the diminishing returns with our investment in the models, but not win by a wide margin. 1. The US trade war with china which will place deepseek compute availability at disadvantages, eventually, if we ever…

You are comparing apple to oranges. Claude is better, sure, and I'd probably use it over deepseek but deepseek is an open model. For me, this makes deepseek quite superior (not from a benchmark/output perspective) to all the other closed models.

Re: Deepseek: The quiet giant leading China’s AI race

#108

Post by jejeyyy77 »

nsoonhui wrote: post #13

I find that the gushing around deepseek is fascinating to watch. To me there are a few structural and fundamental reasons why deepseek can never outperform other models by a wide margin. On par maybe--as we reach the diminishing returns with our investment in the models, but not win by a wide margin. 1. The US trade war with china which will place deepseek compute availability at disadvantages, eventually, if we ever…

eh, none of your points support your argument.

Re: Deepseek: The quiet giant leading China’s AI race

#109

Post by GordonS »

Scea91 wrote: post #93
tossandthrow wrote: post #85

Earlier quoted context omitted.

Both in the EU and the US there is plenty of regulation that mandates these types of censorship - and with reason. In the US there is 18 U.S.C. § 842(p). In the EU there is the entire AI Act. But I am sure you can yourself chat your way through to figure out what legislation companies like OpenAI and Anthropic are under.

Is any of these equivalent in nature to, for example, censoring information about Tiananmen square events?

It's possible that China censors info about Tiananmen square because so much of what was published came from Western news orgs - and the West has form for using the "news" to attack other nations. Another example might be the supposed "genocide" of the Uyghur people - the MSM pushed the genocide narrative hard, while radicalising, funding and arming Uyghur Islamic extremists, so they could control the narrative. And of course, it largely worked.

Re: Deepseek: The quiet giant leading China’s AI race

#110

Post by antirez »

nsoonhui wrote: post #13

I find that the gushing around deepseek is fascinating to watch. To me there are a few structural and fundamental reasons why deepseek can never outperform other models by a wide margin. On par maybe--as we reach the diminishing returns with our investment in the models, but not win by a wide margin. 1. The US trade war with china which will place deepseek compute availability at disadvantages, eventually, if we ever…

1. The Chinese internal market is huge, and in case they develop models that are better than western models, not using them will be a disadvantage for us, not them. Also I can see many European countries (including my country, Italy) to buy Chinese AI regardless of US regulations.

2. Western has its own issues with data limits and extreme alignment that makes models dumber. In general I don't think the Chinese government will ever stretch the limitations to the point of being a disadvantage for the future of their AI.

3. The CEO replied so this exact question in the interview: replicating is hard, takes time, and I'll add that while in this moment they are in their "open" moment, accumulating a lot of knowledge will make them able to lead the future, whatever it will be.

Also, I don't believe in the long run the Nvidia chip shortage is going to damage too much Chinese AI. Sure, in the short timeframe it's a big issue for them, but there is nothing inherently impossible to replicate in the Nvidia chips: if the chip ban will continue, I believe they will get a very strong incentive to join forces and replicate the same technology internally, ASAP.

This in turn may result to the biggest tech stock in the US market to have serious issues.

Post reply on HN