00 - The Future of Drug Discovery: AI That Simulates Biology | Stack, X-Cell, ESM2 Demo

A Nature article published this week asked whether AI can simulate a living cell. Three research groups released models in early 2026 that take a serious shot at answering that question. This video covers all three and demonstrates one live in code. Topics covered: • Why drug discovery is still slow despite decades of automation • Chan Zuckerberg Biohub's $500M Virtual Biology Initiative • What single-cell RNA sequencing and Perturb-seq are and why they matter • Stack (Arc Institute, Jan 2026): in-context learning for cells, Perturb Sapiens atlas • X-Cell (Xaira Therapeutics, Mar 2026): 4.9B parameter virtual cell model • ESM2: protein language model trained on 250M sequences from all life on Earth • Live demo: embedding EGFR and its drug-resistance mutations, zero-shot mutation scoring No prior biology or programming knowledge required for the concepts. Python familiarity helpful for the code demo. Link to the nature article: https://www.nature.com/articles/d4158... Link to code: https://github.com/bnsreenu/AI-for-Dr...