Researchers Push Back on "DeepSeek Is Just Distillation" Claims

Steve Hsu and others rebutted claims that DeepSeek merely distills OpenAI models, arguing its efficiency stems mainly from published architectural innovations including MoE, efficient training, and R1 reinforcement learning.

2026-09-12 ~ 2026-09-13 · 2 related posts

• Steve Hsu: DeepSeek's efficiency comes from architecture, not distillation — jedisct1 · 2026-09-12
• DeepSeek's MoE, efficient training and R1 were original work — distillation wasn't the whole story — TheMoonMidas · 2026-09-13