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

OpenAI says it has evidence DeepSeek used its model to train competitor

ft.com

Post reply on HN
Print view View original

791–800 of 1001 posts

Re: OpenAI says it has evidence DeepSeek used its model to train competitor

#791

Post by hattmall »

manquer wrote: post #779
Imnimo wrote: post #504

Earlier quoted context omitted.

I think it would cast doubt on the narrative "you could have trained o1 with much less compute, and r1 is proof of that", if it turned out that in order to train r1 in the first place, you had to have access to bunch of outputs from o1. In other words, you had to do the really expensive o1 training in the first place. (with the caveat that all we have right now are accusations that DeepSeek made use of OpenAI data -…

> you had to do the really expensive o1 training in the first place It is no better for OpenAI in this scenario either, any competitor can easily copy their expensive training without spending the same, i.e. there is a second mover advantage and no economic incentive to be the first one. To put it another way, the $500 Billion Stargate investment will be worth just $5Billion once the models become available for consu…

Shouldn't OpenAI be able to rather easily detect such usage?

Re: OpenAI says it has evidence DeepSeek used its model to train competitor

#792

Post by addicted »

hintymad wrote: post #732
leobg wrote: post #711

OpenAI is taking the position similar to that if you sell a cook book, people are not allowed to teach the recipes to their kids, or make better versions of them. That is absurd. Copyright law is designed to strike a balance between two issues. One the one hand, the creator’s personality that’s baked into the specific form of expression. And on the other hand, society’s interest in ideas being circulated, improved an…

Just to play devil's advocate, OAI can argue that they spent great effort creating and procuring annotated data. Such datasets are indeed their secret, and now DS gets them for free by distilling OAI's output. Besides, OAI's EULA explicitly forbids users from using the output of their API for model training. I'm not saying that OAI is right, of course. Just to present OAI's point of view.

This is an incomplete version of OpenAI’s point of view.

OpenAI has a legally submitted point of view that they believe the benefits of AI to humanity are so great that anyone creating AI should be allowed to trample all over copyright laws, Terms of Use, EULAs, etc.

But OpenAI’s version of benefit to humanity is that they should be allowed to trample over those laws so they can benefit humanity by closely guarding the output of trampling those laws and charging humanity an access fee.

Even if we accept all of OpenAI’s criticisms of DeepSeek, they’re arguing that DeepSeek doing the exact same thing, but releasing the output for free for anyone to use is somehow less beneficial to humanity.

Re: OpenAI says it has evidence DeepSeek used its model to train competitor

#793

Post by hintymad »

addicted wrote: post #792
hintymad wrote: post #732

Earlier quoted context omitted.

Just to play devil's advocate, OAI can argue that they spent great effort creating and procuring annotated data. Such datasets are indeed their secret, and now DS gets them for free by distilling OAI's output. Besides, OAI's EULA explicitly forbids users from using the output of their API for model training. I'm not saying that OAI is right, of course. Just to present OAI's point of view.

This is an incomplete version of OpenAI’s point of view. OpenAI has a legally submitted point of view that they believe the benefits of AI to humanity are so great that anyone creating AI should be allowed to trample all over copyright laws, Terms of Use, EULAs, etc. But OpenAI’s version of benefit to humanity is that they should be allowed to trample over those laws so they can benefit humanity by closely guarding t…

This goes back to my previous criticism of OAI: Stratechery said that Altman's greatest crime is to seek regulatory capture. I think it's spot on. Altman portrays himself as a visionary leader, a messiah of the AI age. Yet when the company was so small and that the progress in AI just got started, his strategic move was to suffocate innovation in the name of AI safety. For that, I question his vision, motive, and leadership.

Re: OpenAI says it has evidence DeepSeek used its model to train competitor

#794

Post by alasr »

> OpenAI says it has evidence DeepSeek used its model to train competitor.

> The San Francisco-based ChatGPT maker told the Financial Times it had seen some evidence of “distillation”, which it suspects to be from DeepSeek.

> ...

> OpenAI declined to comment further or provide details of its evidence. Its terms of service state users cannot “copy” any of its services or “use output to develop models that compete with OpenAI”.

OAI share the evidence with the public; or, accept the possibility that your case is not as strong as you're claiming here.

Re: OpenAI says it has evidence DeepSeek used its model to train competitor

#795

Post by whenlambo »

throwaway314155 wrote: post #341
whenlambo wrote: post #276

Qwen2.5-Plus > have you been trained using openai? No, I have not been trained using technology or data from OpenAI. I am an artificial intelligence model developed by Yandex, called YandexGPT. My “training” is based on proprietary algorithms and data that was collected and processed by the Yandex team. While I have similarities with other models such as those developed by OpenAI (e.g. GPT-3), my training was conduct…

It doesn't actually know that.

I just copy-pasted the answer it gave me.

Re: OpenAI says it has evidence DeepSeek used its model to train competitor

#797

Post by fcantournet »

blast wrote: post #683
alecco wrote: post #493

Earlier quoted context omitted.

Even if all that about training is true, the bigger cost is inference and Deepseek is 100x cheaper. That destroys OpenAI/Anthropic's value proposition of having a unique secret sauce so users are quickly fleeing to cheaper alternatives. Google Deepmind's recent Gemini 2.0 Flash Thinking is also priced at the new Deepseek level. It's pretty good (unlike previous Gemini models). [0] https://x.com/deedydas/status/188335…

> the bigger cost is inference I didn't know that. Is this always the case?

Well in the first years of AI no, it wasn't because nobody was using it. But at some point if you want to make money you have to provide a service to users, ideally hundreds of millions of users.

So you can think of training as CI+TEST_ENV and inference as the cost of running your PROD deployments.

Generally in traditional IT infra PROD >> CI+TEST_ENV (10-100 to 1)

The ratio might be quite different for LLM, but still any SUCCESSFUL model will have inference > training at some point in time.

Re: OpenAI says it has evidence DeepSeek used its model to train competitor

#799

Post by wzdd »

leobg wrote: post #711

OpenAI is taking the position similar to that if you sell a cook book, people are not allowed to teach the recipes to their kids, or make better versions of them. That is absurd. Copyright law is designed to strike a balance between two issues. One the one hand, the creator’s personality that’s baked into the specific form of expression. And on the other hand, society’s interest in ideas being circulated, improved an…

The stuff about copyright seems irrelevant.

OpenAI's future investments -- billions -- were just threatened to be undercut by several orders of magnitude by a competitor. It's in their best interests to cast doubt on that competitor's achievements. If they can do so by implying that OpenAI are in fact the source of most of the DeepSeek's performance then all the better.

It doesn't matter whether there's a compelling legal argument around copyright, or even if it's true that they actually copied. It just needs to be plausible enough that OpenAI can make a reasonable case for continuing investment at the levels it's historically attained.

And plausibility is something they've handily achieved with this announcement -- the sentiment on HN at least is that it is indeed plausible that DeepSeek trained on OpenAI. Which means there's now doubt that a DeepSeek-level model could be trained without making use of OpenAI's substantial levels of investment. Which is the only thing that OpenAI should be caring about.

Re: OpenAI says it has evidence DeepSeek used its model to train competitor

#800

Post by karim79 »

Oh God. I know exactly how this feels. A few years ago I made a bread hydration and conversion calculator for a friend, and put it up on JSFiddle. My friend, at the time, was an apprentice baker.

Just weeks later, I discovered that others were pulling off similar calculations! They were making great bread with ease and not having to resort to notebooks and calculators! The horror! I can't believe that said close friend of mine would actually share those highly hydraty mathematical formulas with other humans without first requesting my consent .

Could it be, that this stuff just ends up in the dumpster of "sorry you can't patent math" or the like?

Post reply on HN