The third episode of TechBio Talks features host Chris Gibson in conversation with leading AI investor Nathan Benaich of Air Street Capital. Learn about the evolving landscape of AI in drug discovery, the biggest trends identified in Nathan’s new State of AI Report, and what AI investors are looking for, now and in the future.

In the second episode of TechBio Talks, host Chris Gibson talks to Anne Carpenter, Institute Scientist and Imaging Platform Senior Director at the Broad Institute of MIT and Harvard, a pioneer in image-based profiling, who played a key role in Recursion’s origin story.

Anne talks about her early discovery that phenomics-based profiling could be as powerful as mRNA profiling; how she decides what data to make by determining the problems she can solve well; how these tools can be deployed to produce safer chemicals; the critical role of open science in advancing new technologies; and why the race to the virtual cell “is more like a mosh pit.”

⏱️ Timestamps:

00:00: Welcome to TechBio Talks

02:33: The dinner conversation that helped lay the foundation for Recursion

03:06: Discovering that cell morphology could be as powerful as mRNA profiling

05:32: Behind the open-source tool, CellProfiler

06:21: The importance of relevant biology

07:21: In drug discovery, framing the problem is the real bottleneck

08:55: Which types of data are most important for drug discovery

10:52: Why we need open data

12:30: What it’s like to see others utilize and advance her technology

14:02: How the virtual cell race is “more like a mosh pit”

15:27: Evolving the definition of the virtual cell

17:38: From systemization to major industry shift

18:51: Anne’s “big bold bet”

20:48: The proposed cost to test existing medicines against rare diseases

22:29: Parting thoughts: “If you are interested in data sciences and you're tired of mRNA profiles – check out image-based data”

In the first episode of TechBio Talks, host Najat Khan talks to Regina Barzilay -- Distinguished Professor at Massachusetts Institute of Technology and AI faculty lead at MIT Jameel Clinic, MacArthur “Genius Grant” recipient and member of the National Academies of Medicine and Engineering.

Regina and Najat discuss everything from Regina’s work on early cancer detection models to what sparked the collaboration behind Boltz-2, the first open-source AI model for highly accurate protein binding affinity predictions from MIT alongside Recursion – and how it’s already leading to new discoveries.

⏱️ TIMESTAMPS:

(00:00): Introducing MIT's Regina Barzilay

(01:32): Early Detection of Breast Cancer

(04:55): What Do You Do In the Clinic When You Discover High-Risk Patients?

(06:00): Bringing the Algorithm to the Patient06:19: Ensuring Your Results Are Relevant

(07:50): Slow Adoption of AI Tools in Clinical Protocols

(08:43): Understanding the Mechanism of Action

(10:49): The Origins of Boltz-2

(14:06): Noise and Intelligent Data Curation

(14:57): How Recursion Is Integrating Boltz-2

(16:17): Applying Vision to Open Source Tools

(17:24): The Next Frontier

(18:38): New AI Tools for Metastatic Patients

(21:41): Clinical Trials Underway at MD Anderson

(25:08): Be Critical and Rely on Logic and Evidence