What we learned from OAI and Molecule.one's "Near-Autonomous AI" Discovery of OAI-M1-03

Join Molecule.one 's CTO / CSO Stan Jastrzębski in conversation with independent experts: Stanford's Le Cong and Shane Lewin (Clockwork Bio, formerly AI VP at GSK), on the lessons from OAI-M1-03: OpenAI and Molecule.one's ground-breaking, near-autonomous AI discovery that boosted the yield of Chan–Lam coupling, a chemical reaction important to producing certain drugs. In addition to the blogpost & paper, OpenAI published an abridged chain-of-thought, which provides remarkable insights about how the discovery came to light, from the AI's point of view. Stan unpacked the chain-of-thought in an essay recently, where he singled out the key attributes of success: Cross-pollination of ideas Open-ended prompting Engineered serendipity Non-linear ideation LLM limitations in chemical reasoning Join is to hear Le, Shane & Stan dive deep into their perspectives on where AI-led discovery in life sciences goes from here.