News for hackers. Stuff that matters.

A comment on DeepSeek may have used Google's Gemini to train its latest model. Parent comment

Re:DeepSeek may have used Google's Gemini to train its lates... by zackangelo on Wednesday June 04, 2025 @10:54AM (#44181350)
There might be a plateau coming but I’m not sure that will be the reason.

It seems counterintuitive but there is some research suggesting that using synthetic data might actually be productive.

[ Parent | on HN ]
Re:DeepSeek may have used Google's Gemini to train its lates... by jsheard on Wednesday June 04, 2025 @10:59AM (#44181406)
I think there's probably a distinction to be made between deliberate, careful use of synthetic data, as opposed to blindly scraping 1PB of LLM generated SEO spam and force-feeding it into a new model. Maybe the former is useful, but the latter... probably not.
[ Parent | on HN ]