DeepSeek uses banned Nvidia chips for AI model, report says • Hacker News Board

Re: DeepSeek uses banned Nvidia chips for AI model, report says

#261

Post by nsoonhui » Thu, Dec 11, 2025, 12:04 AM UTC

A lot of people attack US for banning the chip sale to China, but the reality is that the ban is two way. Not only US bans advanced chips, China also severely limits US chips in the hopes that it can spur the homegrown ones. In fact, the first sentence says it much

  Chinese artificial intelligence startup DeepSeek has relied on Nvidia Corp. chips that are banned in the country

Re: DeepSeek uses banned Nvidia chips for AI model, report says

#262

Post by rhines » Thu, Dec 11, 2025, 12:07 AM UTC

rllearneratwork wrote: post #260

ComputerGuru wrote: post #153

Earlier quoted context omitted.

> I also read somewhere (not Wikipedia) that they trained on ChatGPT, Claude, and Gemini queries, basically feeding in the output of competitor’s LLMs as training data All the labs permitting synthetic data do that.

pretty sure it is against ToS for ChatGPT, Claude and Gemini.

ToS didn't stop the companies that built those models and it won't stop the companies that bootstrap off them. Until an AI company eats a multi billion dollar lawsuit for unlawful data use they will continue to operate this way.

Re: DeepSeek uses banned Nvidia chips for AI model, report says

#263

Post by thrance » Thu, Dec 11, 2025, 12:20 AM UTC

nish__ wrote: post #185

platevoltage wrote: post #169

Earlier quoted context omitted.

Give me 1 reason why the USA, the (for now) leader of "the west", is morally superior to China?

[flagged]

China's economy is very much capitalist.

Re: DeepSeek uses banned Nvidia chips for AI model, report says

#264

Post by vkou » Thu, Dec 11, 2025, 12:24 AM UTC

echelon wrote: post #157

ComputerGuru wrote: post #154

Earlier quoted context omitted.

> It’s bandits all the way down, so adding a little smuggling to that doesn’t surprise me. Implying it’s *morally* wrong for a Chinese company to bypass US sanctions is hilarious. You really say that with a straight face when even the president admits this is only protectionism?

It's the moral imperative of every country to optimize for its citizens' economic prospects. The CIA (and every other intelligence org.) is literally a weapon designed to operate in the grey area to fit the mandate of the policy makers and elected leadership. Many of the things they do are questionable or worse. In the case of the DoD, they do these things at the behest of democratically elected leadership. Of course…

> It's the moral imperative of every country to optimize for its citizens' economic prospects.

Surely there are constraints on this, because otherwise, it would be the moral imperative of every country to enslave non-citizens for the benefit of (some subset of) citizens.

Re: DeepSeek uses banned Nvidia chips for AI model, report says

#265

Post by codedokode » Thu, Dec 11, 2025, 12:24 AM UTC

rllearneratwork wrote: post #260

ComputerGuru wrote: post #153

Earlier quoted context omitted.

> I also read somewhere (not Wikipedia) that they trained on ChatGPT, Claude, and Gemini queries, basically feeding in the output of competitor’s LLMs as training data All the labs permitting synthetic data do that.

pretty sure it is against ToS for ChatGPT, Claude and Gemini.

Did OpenAI observe any ToS when scraping content from the Internet? Sorry, but you cannot complain about stealing the stolen. OpenAI cannot even have the copyright on ChatGPT output, because it's a tool.

Re: DeepSeek uses banned Nvidia chips for AI model, report says

#266

Post by nsoonhui » Thu, Dec 11, 2025, 12:26 AM UTC

nostrademons wrote: post #119

I thought this was common knowledge. DeepSeek’s Wikipedia entry says that they trained all their models on Nvidia chips procured before the U.S. embargo to China on them. It wouldn’t surprise me if they continued acquiring them through, well, less than legal means. I also read somewhere (not Wikipedia) that they trained on ChatGPT, Claude, and Gemini queries, basically feeding in the output of competitor’s LLMs as tr…

I think the fact that DeepSeek trains on competitor queries (i.e., distillation) — along with using banned Nvidia chips — helps explain how it can achieve such low training costs (USD 6 million vs. billions) while delivering only slightly worse performance than its American counterparts. It also undermines the narrative that DeepSeek or China is posing a serious challenge to the U.S. lead in AI. The gap may be closing, but the initial reactions now seem knee-jerk.

That the discussion has being hijacked and shifted to moral superiority is really unfortunate, because that was never the point in the first place.

Re: DeepSeek uses banned Nvidia chips for AI model, report says

#267

Post by vkou » Thu, Dec 11, 2025, 12:27 AM UTC

rllearneratwork wrote: post #260

ComputerGuru wrote: post #153

Earlier quoted context omitted.

> I also read somewhere (not Wikipedia) that they trained on ChatGPT, Claude, and Gemini queries, basically feeding in the output of competitor’s LLMs as training data All the labs permitting synthetic data do that.

pretty sure it is against ToS for ChatGPT, Claude and Gemini.

It's against my ToS for ChatGPT, Claude, and Gemini to train their models on my writing, but they've all done that. They are free to pound sand.

Re: DeepSeek uses banned Nvidia chips for AI model, report says

#268

Post by seizethecheese » Thu, Dec 11, 2025, 12:38 AM UTC

ComputerGuru wrote: post #154

nostrademons wrote: post #119

I thought this was common knowledge. DeepSeek’s Wikipedia entry says that they trained all their models on Nvidia chips procured before the U.S. embargo to China on them. It wouldn’t surprise me if they continued acquiring them through, well, less than legal means. I also read somewhere (not Wikipedia) that they trained on ChatGPT, Claude, and Gemini queries, basically feeding in the output of competitor’s LLMs as tr…

> It’s bandits all the way down, so adding a little smuggling to that doesn’t surprise me. Implying it’s *morally* wrong for a Chinese company to bypass US sanctions is hilarious. You really say that with a straight face when even the president admits this is only protectionism?

Hmm, the way I read “bandit” here was to be implying illegality not necessarily immorality.

Re: DeepSeek uses banned Nvidia chips for AI model, report says

#269

Post by themafia » Thu, Dec 11, 2025, 12:51 AM UTC

MangoToupe wrote: post #140

nostrademons wrote: post #119

I thought this was common knowledge. DeepSeek’s Wikipedia entry says that they trained all their models on Nvidia chips procured before the U.S. embargo to China on them. It wouldn’t surprise me if they continued acquiring them through, well, less than legal means. I also read somewhere (not Wikipedia) that they trained on ChatGPT, Claude, and Gemini queries, basically feeding in the output of competitor’s LLMs as tr…

> less than legal means. This is an absurd concept when it comes to international trade. Even intellectual property is mostly meaningless outside a state. Of course people will evade sanctions; what is the us going to do, invade singapore or malaysia?

> This is an absurd concept when it comes to international trade.

In this case it's just wrong. I don't know what people think "e-waste" recycling actually is or what happens to their "unrepairable" units after they rid themselves of them.

> Even intellectual property is mostly meaningless outside a state.

Interestingly the Dollar is most definitely meaningful outside of our state. I think the assumption becomes, that if this is true, then using it's power to enforce trade sanctions isn't that big a stretch.

> Of course people will evade sanctions

What's less clear if they should expect their government to actively help them in this evasion or not. I think the Chinese citizens are in unique international territory here.

> what is the us going to do, invade singapore or malaysia?

Deny our exports to them. This will cost the political donor class a lot of profits. So this is why it doesn't get done.

None of this is a fait accompli. This is the result of years of intentional corruption of the core systems involved.

Re: DeepSeek uses banned Nvidia chips for AI model, report says

#270

Post by rkagerer » Thu, Dec 11, 2025, 12:52 AM UTC

ahtihn wrote: post #240

rkagerer wrote: post #177

Earlier quoted context omitted.

You accept these terms when you use their services. I certainly didn't*. I'd love to see litigation testing just how solid those insidious opt-in-by-default schemes are as a basis for "ownership". If they had users explicitly opt-in with a "Yes, go ahead and train on my stuff and by the way I assert that I have all the rights to grant you the same", I'd have no problem with that, and they'd have a much stronger claim…

> and further send notice to companies from time to time that I don't agree to certain objectionable clauses of their ToS and they're welcome to close my account And then you stopped using their service right?

Sometimes, if they said tough luck.

Other times they turn a blind eye and choose to provide the service (and collect my money) despite the lack of agreement to some part of their standard terms and their tacit acknowledgement that I didn't accept them. On two occasions their legal team responded and said "that's fine", and once they actually fixed their ToS.

People who didn't grow up dealing with paper contracts where you could easily redline and send back for countersigning don't seem to understand that you don't just need to blindly say "yes" to everything a company tries to foist upon you.