Max Jaderberg on Isomorphic Labs' Vision for AI-Driven Drug Discovery

Isomorphic Labs president Max Jaderberg spoke with Elad Gil about the company's approach to transforming drug discovery from empirical trial and error into a computational discipline. Jaderberg noted that traditional drug development is slow and costly, and argued that success with models like AlphaFold demonstrates how deep learning can generalize in biology. He explained that Isomorphic focuses on building disease-agnostic drug design engines capable of handling various target classes and modalities.

According to Jaderberg, scaling AI models allows them to compound learnings and achieve generalizability, which could increase the probability of success in real-world clinical trials. He shared his perspective that accurate molecular engineering may eventually reduce the reliance on animal testing through in-silico models and could enable sophisticated multidrug cocktails targeting complex biological pathways. Jaderberg also discussed the long-term outlook for treating diseases and improving human healthspan as computational optimization advances.