Deepseek R1 was trained at least partially on the output of other LLMs. So, it m... | Hacker News

Hacker Newsnew | past | comments | ask | show | jobs | submit

		IncreasePosts 7 months ago \| parent \| context \| favorite \| on: OpenAI delays launch of open-weight model Deepseek R1 was trained at least partially on the output of other LLMs. So, it might have been much more expensive if they needed to do it themselves from scratch.

nomel 7 months ago [–]

Lawsuit, since it was against OpenAI TOS: https://hls.harvard.edu/today/deepseek-chatgpt-and-the-globa...

xcrunner529 7 months ago | [–]

Pretty hilarious considering how OpenAI got its data for training.

nomel 7 months ago | | [–]

No. Deepseek has access to the same data (see copyright law in China). The data is effectively free. That's not where the cost is.

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact