Google's VERID grounds AI agents in wet labs, finds nanomolar cancer protein binders
CurieuxExplorer · x · 2026-11-21
Google Research unveiled VERID (Verified Infrastructure for Discovery Laboratories), a framework that connects AI agents to physical lab robotics to close the gap between high-level reasoning and the error-prone reality of wet labs.
- Core idea: a semantic digital twin of the lab workcell — a live, stateful 3D volumetric entity tree tracking everything from robotic arm geometry to well volumes and reagent identities.
- Motivation: frontier models still reason unreliably about 3D geometry and physical state, and wet-lab throughput lags far behind agentic hypothesis generation.
- Results: closed-loop molecular design cycles where agents learn how to learn discovered single-digit nanomolar protein binders for cancer targets.
- Three accompanying papers: VERID, agentic aptamer sandbox, and epitope excision.
More from Embodied
- FOSDEM 2027 Robotics & Simulation devroom opens call for talks, deadline Nov 23 — facontidavide · 2026-11-21
- SuperMimic took 11 months of code and experiments; sim-to-real may move past PPO — ChongZzZhang · 2026-11-21
- Autonomous cars found draining batteries overnight by constantly reparking themselves — mtizard · 2026-11-21
- Huawei's clip-on Ruiying Z10 brings 10x optical zoom, journalists ditch cameras — CodeByPoonam · 2026-11-21
- Refining the morphogenesis-to-locomotion pipeline for automated robot design — zzznah · 2026-11-21
- Uncertainty-Aware Robotic World Model Talk Draws Crowd at CoRL 2026 Workshop — breadli428 · 2026-11-21