AI Meets Biotech: The Future Of Protein Therapeutics With Mike Nally And Jason Silvers

17 snips

May 29, 2025

Mike Nally, CEO of Generate Biomedicines, and Jason Silvers, its CFO, discuss the revolutionary impact of AI on protein therapeutics. They delve into how their 'self-driving lab' utilizes machine learning and cryo-electron microscopy to expedite drug discovery, potentially reducing development time and costs significantly. The conversation highlights their data-first approach to overcoming traditional trial-and-error in drug development, unlocking treatments for previously undruggable diseases, and transforming the economics of biotech.

Ask episode

AI Snips

Chapters

Transcript

Episode notes

INSIGHT

Rethinking Protein Drug Discovery

Traditional protein drug discovery relies heavily on trial and error, often borrowed from natural sources.
Generate Bio uses a data-driven, engineering-first approach to revolutionize protein therapeutic development.

INSIGHT

Integrating Sequence, Structure, Function

Generate combines vast protein sequence and structure data with functional testing to build better models.
This integration allows exploration of sequence, structure, and function relationships for therapeutics.

INSIGHT

Accelerating Clinical Timelines

Generate can take therapeutics from concept to clinic in 18-24 months, greatly shortening traditional timelines.
Early-stage cost to drug candidate nomination reduces from millions to hundreds of thousands of dollars.

Get the Snipd Podcast app to discover more snips from this episode

Get the app

In this episode of FYI – For Your Innovation, Brett Winton and ARK analyst Nemo Despot sit down with Generate Biomedicines CEO Mike Nally and CFO Jason Silvers to explore how generative AI is transforming the discovery and development of protein-based therapeutics. Founded in 2018 by Flagship Pioneering, Generate Biomedicines is building a “self-driving lab” that combines machine learning, cryo-electron microscopy (cryo-EM), and high-throughput wet lab automation to dramatically accelerate drug development. The conversation dives into how Generate is reimagining protein therapeutics — going beyond trial-and-error methods to a data-first, design-driven model for creating novel medicines. The team discusses the company’s proprietary approach to integrating structural biology with functional data, the economic implications of reducing time-to-market from 13 years to under 9, and how their platform could unlock treatments for diseases that have been historically undruggable. They also touch on strategic partnerships, scalability, and how AI is shifting the biotech business model from artisanal science to an industrialized, data-driven enterprise.

Key Points From This Episode:

00:01:30 Why Generate Biomedicines is rethinking protein drug discovery from first principles
00:04:40 How their structure-first approach differs from peers like AbSci and Recursion
00:07:04 Using cryo-EM to build proprietary protein interaction datasets
00:10:57 Traditional drug discovery is random, expensive, and inefficient — here’s how Generate is changing that
00:16:58 From concept to clinic in 18–24 months: Accelerating timelines through AI
00:20:47 Going beyond efficiency: Unlocking access to undruggable biology
00:24:48 Turning cryo-EM into a high-throughput data engine for model training
00:31:20 The long-term vision: Patient-specific protein therapeutics
00:40:00 Why scalability gives Generate an edge over traditional biotech
00:47:52 The future of biotech as a research & development (R&D) sharing economy
00:54:19 Adapting fast: Keeping pace with generative AI advances across the stack
01:00:16 The KPI (key performance indicator) for platform success: Rate of improvement