Agile Drug Discovery: The Loop That Replaces Years
Chai Discovery's Neil Patil and Matt McPartlon explain how AI shifts drug discovery from slow waterfall to rapid, iterative loops. Build for the one-year product.
🔬Biology Is Turning Into Software — Matt McPartlon and Neil Patil, Chai Discovery
This episode features Matt McPartlon and Neil Patil from Chai Discovery, a protein design startup, discussing their AI models (Chai 1, 2, and 3) for protein and antibody design. They elaborate on their unique partnership-driven business model with major pharma companies, the technical challenges of scaling AI for biology, and their vision for transforming drug discovery into a precision engineering discipline. The conversation highlights the shift from traditional methods to AI-driven design, addressing critical bottlenecks in compute, data, and validation.
Key takeaways
- Traditional drug discovery follows a "waterfall model" where each stage, from target to optimization, takes months or years and costs a fortune to even try new things, according to Chai Discovery's Neil Patil. Read more →
- Neil Patil of Chai Discovery, a protein design startup, identifies "talent obscurity" as the biggest bottleneck for AI in biology, noting top ML talent often gravitates to LLMs or traditional software. Read more →
- Chai Discovery acts as a "neutral software factory" for medicines, focusing solely on providing AI models and a product platform rather than developing its own drugs. Read more →
- Epitope prediction is "ridiculously hard" for novel targets: Historically, finding the exact spot on a protein an antibody needs to block (the epitope) has been a massive challenge, often relying on trial and error. Read more →
- Protein design startups like Chai Discovery face brutal validation loops, where it takes months to know if an AI model's prediction was correct, severely slowing iteration. Read more →
5 articles from this episode
Talent Obscurity" Bottleneck: AI Bio Needs Your Generalists
Chai Discovery's leaders say AI for biology suffers from "talent obscurity." Stop chasing bio PhDs for ML roles. Recruit smart generalists.
Chai Discovery's Single-Tenancy Tactic for Pharma Trust
Chai Discovery partnered with Eli Lilly, Pfizer, and Novartis by deploying AI models as a 'neutral software factory.' Learn their secret: aggressive data segmentation and single-tenancy.
Precision AI: Why Brute Force Antibody Design is Dead
Chai Discovery's Matt McPartlon and Neil Patil explain how AI precisely targets antibody binding sites, moving drug design beyond brute force.
Your AI Models Need Durable Execution, Not Just More GPUs
Chai Discovery's Neil Patil explains how 'durable execution' saves complex AI projects from 'retry hell' when compute is flaky and expensive.
More Latent Space episodes
- The Future of Claude Code: Mods, Mutable Software, & Multiplayer Agents — Thariq Shihipar, AnthropicOct 4 · 4 articles
- Which GPU Clouds Are Actually Good? | ClusterMAX 3.0Oct 4 · 6 articles
- Recursive Language Models — Alex Zhang, MIT PhDOct 4 · 6 articles
- OpenAI’s New Agent Stack: Computer Use, Decisions API, UltraFast, Dots—Ari Weinstein & Nikunj HandaOct 4 · 6 articles