Chai Discovery's Bitter Lesson: Drug Design Is Another Scaling Problem

Chai Discovery is industrializing drug discovery by transforming biology into an engineering discipline, shifting away from serendipitous trial-and-error toward a design-oriented paradigm. Co-founders Josh and Matt emphasize the "bitter lesson" of scaling compute, data, and model architectures to achieve high-success-rate molecular design. By increasing binding success rates from 0.1% to 15% with their Chai 2 model, the company enables rapid, iterative improvements that were previously impossible. Unlike firms developing internal drug pipelines, Chai provides infrastructure to major pharmaceutical partners, ensuring rigorous model validation through real-world application. This approach aims to unlock "undruggable" targets and facilitate the creation of highly specific, "last-in-class" therapies. By treating amino acid sequences as prompts for generative models, the team is building a computer-aided design suite that accelerates the path from initial hypothesis to testable molecule, ultimately reducing the time and cost required to bring breakthrough medicines to patients.