Unsupervised Learning Ep 72: Co-Founder of Chai Discovery Joshua Meier on 99% Faster Drug Discovery, BioTech’s AlphaGo Moment, Building Photoshop for Molecules
44 snips
Aug 13, 2025 
Joshua Meier, co-founder of Chai Discovery and ex-Chief AI Officer at Absci, shares insights on the groundbreaking role of AI in drug discovery. He discusses how the field has evolved, achieving unprecedented success rates. The conversation reveals how AI models differ from human chemists and the strategic choices behind open-sourcing foundational models. Meier also highlights Chai 2's innovative capabilities in testing biological hypotheses and the promising future of AI in revolutionizing therapeutic development and enhancing pharmaceutical innovation.
AI Snips
Chapters
Transcript
Episode notes
Three Waves Of AI Drug Discovery
- AI drug discovery evolved in three waves: early research, mid-stage modeling, and current foundation-model companies.
- Wave three borrows NLP/vision advances and achieves much higher lab success rates than prior generations.
Atomic-Level Microscopes Enable Design
- AI can move discovery from lab to computer by predicting atomic interactions and designing molecules.
- Accurate structure and interaction prediction acts like an atomic-level microscope that enables computational design.
Models Can Generalize Across Modalities
- Base models can be modality-agnostic and handle proteins, small molecules, DNA and RNA.
- Commercial focus may vary by manufacturing and partnerships, not fundamental model limitations.