Lab-in-the-Loop at Compute Speed — Ian Quigley, Leash Bio
TL;DR
How Leash Bio tamed DELs, the assay that traumatizes everyone, and used the data to catch cheating models. And why that same data is perfect for AI agents learning to think like chemists. How Leash Bio tamed DELs, the assay that traumatizes everyone, and used the data to catch cheating models. And why that same data is perfect for AI agents learning to think like chemists.
Nauti's Take
Small biotech and AI teams should test the data layer first: Are the DEL measurements reproducible, well labeled, and independent enough to expose model gaming? Teams building drug-discovery agents can evaluate this data as both a training and verification layer, then validate it against real assays before allowing autonomous chemistry workflows.
Sources