1h 19mMistral Is SO Back: Inside Le Chonk, the West’s Biggest Open Model (Pierre Stock, VP Science, Mistral)
Pierre Stock is the VP Science at Mistral. We cover why open weights make AI access harder to cut off, how Mistral Large 4’s 1 trillion-parameter mixture-of-experts architecture activates 50 billion parameters per token, and why reinforcement learning remains unsaturated across coding, cyber, finance, and chip-design tasks.
Oct 22, 2026