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

691–700 of 1001 posts

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

#691

Post by anon373839 »

miki123211 wrote: post #677
Imnimo wrote: post #320

I think there's two different things going on here: "DeepSeek trained on our outputs and that's not fair because those outputs are ours, and you shouldn't take other peoples' data!" This is obviously extremely silly, because that's exactly how OpenAI got all of its training data in the first place - by scraping other peoples' data off the internet. "DeepSeek trained on our outputs, and so their claims of replicating…

> This is obviously extremely silly, because that's exactly how OpenAI got all of its training data IANAL, but It is worth noting here that DeepSeek has explicitly consented to a license that doesn't allow them to do this . That is a condition of using the Chat GPT and the OpenAI API. Even if the courts affirm that there's a fair use defence for AI training, DeepSeek may still be in the wrong here, not because of cop…

> DeepSeek has explicitly consented to a license that doesn't allow them to do this.

You actually don’t know this. Even if it were true that they used OpenAI outputs (and I’m very doubtful) it’s not necessary to sign an agreement with OpenAI to get API outputs. You simply acquire them from an intermediary, so that you have no contractual relationship with OpenAI to begin with.

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

#692

Post by s17n »

KennyBlanken wrote: post #605
s17n wrote: post #600

Earlier quoted context omitted.

> This is obviously extremely silly, because that's exactly how OpenAI got all of its training data in the first place - by scraping other peoples' data off the internet. OpenAI has also invested heavily in human annotation and RLHF. If all DeepSeek wanted was a proxy for scraped training data, they'd probably just scrape it themselves. Using existing RLHF'd models as replacement for expensive humans in the training…

"We spent a lot of labor processing everything we stole" is...not how that works. That's like the mafia complaining that they worked so hard to steal those barrels of beer that someone made off with in the middle of the night and really that's not fair and won't someone do something about it?

Oh, I don't really care about IP theft and agree that it's funny that openai is complaining. But I don't think its true that deepseek is just doing this because they are too lazy to scrape the internet themselves - its all about the human labor that they would otherwise have to pay for.

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

#696

Post by dang »

glitchc wrote: post #638
dang wrote: post #388

Earlier quoted context omitted.

Would you please not do this here? We're trying for an opposite sort of conversation. https://news.ycombinator.com/newsguidelines.html

Sorry dang. I'll do better.

Appreciated!

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

#697

Post by nicce »

twobitshifter wrote: post #401
thiago_fm wrote: post #310

Earlier quoted context omitted.

The most interesting part is that China has been ahead of the US in AI for many years, just not in LLMs. You need to visit mainland China and see how AI applications are everywhere, from transport to goods shipping. I'm not surprised at all. I hope this in the end makes the US kill its strict IP laws, which is the problem. If the US doesn't, China will always have a huge edge on it, no matter how much NVidia hardware…

You don’t even need to visit china, just read the latest research papers and look at the authors. China has more researchers in AI than the West and that’s a proven way to build an advantage.

It is also funny in a different way. Many people don't realise that they live in some sort of bubble. Many people in "The West" think that they are still the center of the world in everything, while this might not be so correct anymore.

In the U.S. there is 350 million people and EU has 520 million people (excluding Russia and Turkey).

China alone has 1.4 billion people.

Since there is a language barrier and China isolates themselves pretty well from the internet, we forget that there is a huge society with high focus on science. And most of our tech products are coming from there.

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

#698

Post by dmitrygr »

miki123211 wrote: post #677
Imnimo wrote: post #320

I think there's two different things going on here: "DeepSeek trained on our outputs and that's not fair because those outputs are ours, and you shouldn't take other peoples' data!" This is obviously extremely silly, because that's exactly how OpenAI got all of its training data in the first place - by scraping other peoples' data off the internet. "DeepSeek trained on our outputs, and so their claims of replicating…

> This is obviously extremely silly, because that's exactly how OpenAI got all of its training data IANAL, but It is worth noting here that DeepSeek has explicitly consented to a license that doesn't allow them to do this . That is a condition of using the Chat GPT and the OpenAI API. Even if the courts affirm that there's a fair use defence for AI training, DeepSeek may still be in the wrong here, not because of cop…

> DeepSeek has explicitly consented to a license that doesn't allow them to do this.

By existing in USA, OpenAI consented to comply with copyright law, and how did that go?

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

#699

Post by almostdeadguy »

Hope Sam Altman is getting his money's worth out of that Trump campaign contribution. Glorious days to be living under the term of a new Boris Yeltsin. Pawning and strip-mining the federal apparatus to the most loyal friends and highest bidders.

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

#700

Post by lolinder »

dartos wrote: post #685
miki123211 wrote: post #677

Earlier quoted context omitted.

> This is obviously extremely silly, because that's exactly how OpenAI got all of its training data IANAL, but It is worth noting here that DeepSeek has explicitly consented to a license that doesn't allow them to do this . That is a condition of using the Chat GPT and the OpenAI API. Even if the courts affirm that there's a fair use defence for AI training, DeepSeek may still be in the wrong here, not because of cop…

TOS are not contracts.

Citation? My understanding was that they are provided that someone has to affirmatively accept them in order to use your site. So Terms of Service stuck at the bottom in the footer likely would not count as a contract because there's no consent, but Terms of Service included in a check box on a login form likely would count.

But IANAL, so if you have a citation that says otherwise I'd be happy to see it!

Post reply on HN