Faster Science, Better Drugs

In this episode, Erik Torenberg speaks with Patrick Hsu from the Arc Institute and a16z's Jorge Conde about the ambitious goal of making science as fast as software. Their focus is on developing 'virtual cells' that can simulate human biology and guide experimental processes. The guests discuss the slow pace of scientific research—tracing it back to multifaceted issues such as funding, incentive structures, and the complex nature of multilevel biological systems. By increasing collaboration across disciplines, they believe they can enhance the efficiency of scientific discovery.

The conversation dives into the reasons why scientific research is slow, largely due to entrenched incentive structures that discourage collaboration. Hsu and Conde argue that traditional academic settings are often spread out, making effective communication and collaboration difficult. They suggest that by bringing various disciplines under one roof, as they are doing at the Arc Institute, they can increase the 'collision frequency' of ideas and enhance innovation.

A significant portion of their discussion centers around the concept of virtual cells and how AI models can revolutionize drug discovery. Hsu explains that current models, while useful, still face challenges in fully simulating biological processes due to the complexity and our incomplete understanding of biology. They express optimism about achieving an 'AlphaFold level' moment in cell biology, where virtual models can accurately predict cell behavior and inform drug design.

As they discuss the intersection of science and business, the guests stress the importance of translating scientific breakthroughs into commercially viable products. They highlight the high failure rates of drugs in clinical trials and discuss the role of AI in potentially improving these odds by informing better choices in drug design. Both Hsu and Conde acknowledge that lowering capital intensity and compressing timelines in drug development are necessary for the biotech industry to thrive.

The episode concludes with reflections on the future of AI in biology and drugs. Hsu emphasizes the need for groundbreaking technologies that will allow scientists to tackle complex problems more effectively. They foresee a future where the integration of AI into various aspects of drug discovery could set the stage for significant scientific advancements and public health benefits.