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

621–630 of 1001 posts

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

#621

Post by qwertox »

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…

They're standing on the shoulders of giants, not only in terms of re-using expensive computing power almost for free by using the outputs of expensive models. It's a bit of a tradition in that country, also in manufacturing.

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

#622

Post by tensor »

smitelli wrote: post #404
tensor wrote: post #389

Earlier quoted context omitted.

That's not correct. First of all, training off of data generated by another AI is generally a bad idea because you'll end up with a strictly less accurate model (usually). But secondly, and more to your point, even if you were to use training data from another model, YOU STILL NEED TO DO ALL THE TRAINING. Using data from another model won't save you any training time.

> training off of data generated by another AI is generally a bad idea Ah. So if I understand this... once the internet becomes completely overrun with AI-generated articles of no particular substance or importance, we should not bulk-scrape that internet again to train the subsequent generation of models. I look forward to that day.

Yes, this is a real and serious concern that AI researchers have.

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

#623

Post by zelphirkalt »

bilekas wrote: post #95

> “It’s also extremely hard to rally a big talented research team to charge a new hill in the fog together,” he added. “This is the key to driving progress forward.” Well I think DeepSeek releasing it open source and on an MIT license will rally the big talent. The open sourcing of a new technology has always driven progress in the past. The last paragraph too is where OpenAi seems to be focusing their efforts.. > we…

Actually the "our IP" argument is ridiculous. What they are doing is stealing data from all over the web, without people's consent for that data to be used in training ML models. If anything, then "Open"AI should be sued and forced to publish their whole product. The people should demand knowing exactly what is going on with their data.

Also still an unresolved issue is how they will ever comply with a deletion request, should any model output personal data of someone. They are heavily in a gray area, with regards to what should be allowed. If anything, they should really shut up now.

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

#624

Post by zelphirkalt »

buyucu wrote: post #454
bilekas wrote: post #95

> “It’s also extremely hard to rally a big talented research team to charge a new hill in the fog together,” he added. “This is the key to driving progress forward.” Well I think DeepSeek releasing it open source and on an MIT license will rally the big talent. The open sourcing of a new technology has always driven progress in the past. The last paragraph too is where OpenAi seems to be focusing their efforts.. > we…

I'm willing to bet ''ban DeepSeek'' voices will start soon. Why compete, when you can just ban?

Actually asking for banning DeepSeek would be the ultimate admit of defeat by ClosedAI.

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

#625

Post by HarHarVeryFunny »

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…

DeepSeek-R0 (based on DeepSeek-V3 base model) was only trained with RL, no SFT, so this isn't at all like the "distillation" (i.e SFT on synthetic data generated by R1) that they also demonstrated by fine tuning Qwen and LLaMa.

Now, DeepSeek may (or may not) have used some O1 generated data for the R0 RL training, but if so that's just a cost saving vs having to source some reasoning data some other way, and in no way reduces the legitimacy of what they accomplished (which is not something any of the AI CEOs are saying).

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

#626

Post by znpy »

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 really got me thinking that open ai should have no ip claim at all, since all their outputs and stuff are basically a ripoff of the entire human knowledge and IPs of various kinds.

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

#627

Post by nullc »

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…

There is a big difference between being able to train on the reasoning vs just the answers, which they can't against o1 because it's hidden. There is also a huge difference between being able to train on the probabilities (distillation) vs not, which again they can and did do with the llama models and can't directly with OpenAI because the conceal the probability output.

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

#628

Post by SubiculumCode »

If you have a set of weights A, can you derive another set of weights B that function (near) identically as A AND a) not appear to be the same weights as A when inspected superficially b) appear uncorrelated when inspecting the weight matrices?

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

#629

Post by gosub100 »

cumulative00x wrote: post #614

There is a saying in Turkish that roughly goes like this, it takes a thief to catch a thief. I am not a big fan of China's tech, too, however, it amuses me to watch how big tech charlatans have been crying over Deepseek shock.

It's true irony to see thieves getting stolen from.
Post reply on HN