Doge Chubbuck

AI Summary

→ WHAT IT COVERS Liam Vettas and Doge Chubbuck discuss building Periodic Labs, which creates AI systems that conduct real physics experiments to discover new materials like superconductors. → KEY QUESTIONS ANSWERED - How do AI models learn from physical experiments versus simulations? - Why can't current language models effectively do scientific discovery? - What makes superconductivity research ideal for AI physicist development? → KEY TOPICS DISCUSSED - Experiment-in-the-loop training: Periodic Labs uses physical lab experiments as reward functions for reinforcement learning, replacing math graders with real-world physics verification systems. - AI physicist applications: The technology targets advanced manufacturing, semiconductors, space, and defense industries where materials research requires iterative physical experimentation and high R&D budgets. → NOTABLE MOMENT Vettas and Chubbuck first met eight years ago at Google Brain while attempting to flip a massive tire that required two people to move successfully. 💼 SPONSORS None detected 🏷️ AI Research, Materials Science, Superconductivity, Reinforcement Learning