The a16z Show Building an AI Physicist: ChatGPT Co-Creator’s Next Venture
342 snips
Sep 30, 2025 
Liam Fedus, co-creator of ChatGPT and now co-founder of Periodic Labs, teams up with Ekin Dogus Cubuk, former head of materials science at Google DeepMind, to discuss their innovative approach to automating physical science discovery. They dive into the importance of integrating LLMs with experimental data to tackle complex problems like superconductivity and magnetism. Their lab focuses on real-world experimentation, emphasizing the necessity of iteration and learning from failures. With a unique hiring philosophy and a commitment to bridging ML and experimental cultures, they aim to revolutionize how we engage with hard sciences.
AI Snips
Chapters
Transcript
Episode notes
Scaling Laws Need Domain Contact
- Scaling laws still hold, but out-of-distribution scientific tasks may improve too slowly without domain-specific data.
- To make progress on physics you must put the target distribution (experiments/simulations) into the training loop.
Superconductivity As A North Star
- High-temperature superconductivity serves as a North Star because it bundles many subgoals: autonomous synthesis, characterization, and simulation.
- Superconductivity is technically robust and carries deep scientific payoff even before commercial products.
Ship Co-Pilots For Industry
- Do build commercial co-pilots for engineers in advanced industries as intermediate products on the path to scientific discovery.
- Focus on narrow, high-impact problems in spaces like semiconductors, space, and defense to gain adoption and funding.