Beyond Simulation: How AlphaFold 3 and AI are Forging a New Paradigm in Scientific Discovery
July 29, 2025 11:00 AM - 12:00 PM CDT
Online
Embedded Files
About the Session
Join us for a deep dive into Google DeepMind's revolutionary AlphaFold 3, the AI model that is redefining the boundaries of molecular biology and demonstrating how AI methods can provide scientific capabilities and discovery advantages over simulation methods in some cases. Felix and Josh will explore how its novel diffusion-based architecture can predict the structure of complex interactions involving proteins, DNA, RNA, and ligands with unprecedented accuracy and speed. This session will discuss the model's groundbreaking improvements in accuracy and performance, along with some example use cases and some model limitations. Finally, we will discuss the profound implications for the future of research, positioning AI not merely as a replacement for simulation, but as a powerful new partner that augments and accelerates traditional methods, charting a new course for scientific discovery.
Speakers
Josh Abramson
Research Engineer, Google DeepMind
Josh joined GDM 8 years ago. He was a key author of AlphaFold 3 and has been collaborating with Isomorphic in ongoing work since then. Earlier projects at GDM involved agent memory, human-agent interaction and reinforcement learning, where he provided technical leadership as well as contributing to the research. Before joining GDM, Josh did a PhD in Statistics at UC Berkeley and then learnt to code while working in quantitative finance for 4 years.
Felix Schürmann, PhD
Senior HPC Technologist, Google
Google Sites
Report abuse