3 ms·Secrets of DeepSeek AI model revealed in landmark paper11 points by bookofjoe 1y agobookofjoe 1y agohttps://archive.ph/LGL8c https://archive.ph/LGL8crurban 1y agoThis is about the R1 reinforcement learning paper from January 2025, not something new.