I read that deepseek was trained on western llm output. So it is expected to hav... | Hacker News

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

		raindear 5 months ago \| parent \| context \| favorite \| on: DeepSeek releases Janus Pro, a text-to-image gener... I read that deepseek was trained on western llm output. So it is expected to have the same biases.

benterix 5 months ago [–]

Did the creators actually say so? I'd rather expect them to train on pirated books just like OpenAI and Meta.

Consider applying for YC's Fall 2025 batch! Applications are open till Aug 4
Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact