I think we know who is winning #china #deepseek #fyp #foryou

So yesterday, China announced a brand new model. It's better than OpenAI's 01 reasoning model. It's called DeepSeek R1. And they open-sourced it under a license, which means literally any company can use it for any purpose. But not only did they do that, they released, I think, six distilled versions of the model from 1.5B, which is tiny, it can run on your phone, all the way up to, I think, a 600B model. That is insane. I might make a proper video about it later, but my God, if the rumors are true, it's outperforming closed-source models.