Raising Health Building AI Foundation Models for Molecular Design
17 snips
Jan 8, 2026 Jeremy Wohlwend and Gabriele Corso, co-founders of Boltz, delve into the transformative power of AI in molecular biology. They discuss Boltz’s mission to create open-source models that revolutionize biomolecular interaction and drug discovery. The conversation highlights breakthroughs post-AlphaFold, the rapid adoption of Boltz models by the pharma sector, and the launch of Boltz Lab—designed to streamline protein and small-molecule design in scientific workflows. They envision a future where AI accelerates iteration and enhances the drug discovery process.
AI Snips
Chapters
Transcript
Episode notes
AlphaFold As The Inflection Point
- AlphaFold marked an inflection point showing AI could solve core biology problems at surprising quality.
- That success made researchers ask what other biomolecular tasks AI could now tackle.
Model Evolution From Structure To Design
- Boltz1 matched AlphaFold 3 levels and Boltz2 added binding strength prediction for small molecules.
- BoltzGen then generalized to design proteins that bind arbitrary biomolecular targets with experimental validation.
Open Source To Automate Pilots
- Open source the models to accelerate adoption, external validation, and community contributions.
- Make engineering and ease-of-use central so non-experts can run and build on the models.