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

901–910 of 1001 posts

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

#901

Post by fanfanfly »

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…

The data that OpenAI has certainly is better than what Deepseek has in your second argument. And OpenAI always has access to this kind of data, right?

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

#902

Post by patcon »

riantogo wrote: post #461
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…

Why would it cast any doubt? If you can use o1 output to build a better R1. Then use R1 output to build a better X1... then a better X2.. XN, that just shows a method to create better systems for a fraction of the cost from where we stand. If it was that obvious OpenAI should have themselves done. But the disruptors did it. It hindsight it might sound obvious, but that is true for all innovations. It is all good stuf…

Are we it rediscovering the evolutionary benefit of progeny (from an information theoretic lens)?

And is this related to the lottery ticket hypothesis?

https://arxiv.org/pdf/1803.03635.pdf

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

#903

Post by krystofee »

I dont know if point of this is just to derail public attention to narative “hey, chinese stole our model, thats not fair, we need computee”, when the deepseek has clearly done some exceptional technical breakthrough on R1 and v3 models. Which even if you stole data from OpenAi is its thing.

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

#904

Post by OsrsNeedsf2P »

harry8 wrote: post #773
ok123456 wrote: post #106

OpenAI's models were trained on ebooks from a private ebook torrent tracker leeched en-mass during a free leech event by people who hated private torrent trackers and wanted to destroy their "economy." The books were all in epub format, converted, cleaned to plain text, and hosted on a public data hoarder site.

Have you got some support for this claim? There's a lot of wild claims about, so while this is plausible it would be great if there were some evidence backing it.

He could be confusing it with Llama: https://www.wired.com/story/new-documents-unredacted-meta-co...

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

#905

Post by thih9 »

I don't mind and I believe that a company with "open" in its name shouldn't mind either.

I hope this is actually true and OpenAI loses its close to monopoly status. Having a for profit entity safeguarding a popular resource like this sounds miserable for everyone else.

At the moment AI looks like typical VC scheme: build something off someone else's work, sell it at cost at first, shove it down everyone's throats and when it's too late, hike the prices. I don't like that.

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

#908

Post by jamil7 »

mrkpdl wrote: post #676

The cat is out of the bag. This is the landscape now, r1 was made in a post-o1 world. Now other models can distill r1 and so on. I don’t buy the argument that distilling from o1 undermines deep seek’s claims around expense at all. Just as open AI used the tools ‘available to them’ to train their models (eg everyone else’ data), r1 is using today’s tools. Does open AI really have a moral or ethical high ground here?

Agree 100%, this was also bound to happen eventually, OpenAI could have just remained more "open" from the beginning and embraced the inevitable commoditization of these models. What did delaying this buy them?

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

#910

Post by isaacremuant »

grogenaut wrote: post #867
nelblu wrote: post #131

Hahaha I can't stop laughing... i dont know the validity of the claim, but immediately i thought of the British Museum complaining about theft.

there's an exhibit in the BM about how they're proud to be allowing the Egyptian government to take back some of the artifacts the British have been safeguarding for the world while Egypt was going through essentially "troubles". right next to it is an older exhibit about how the original curator took cuneiform rolls and made them into necklace beads for his wife and rings? for himself. either someone at the BM has a…

The safeguarding propaganda is a a typical go-to of the remnants of the British empire to keep their stolen goods.

They do it even with the Chile Moais when they never where in any danger.

It's all lies.

Post reply on HN