Researchers push back on autonomous labs: automating the easiest step isn't the hard problem

arjunrajlab · x · 2026-11-16

Researcher xinchenw argues that while 24/7 autonomous labs for scaling biological data generation are trendy, automating well-established in vitro assays with existing hardware solves the easiest, least painful step rather than what matters for drug discovery — like training self-driving cars on decades of closed-course data and hoping they navigate Manhattan. The most valuable data sits at the intersection of novel assay technologies and complex samples (animal models, organoids, primary tissue, not immortalized cell lines). arjunrajlab agrees: automation handles known unknowns well but not unknown unknowns, a role academia should lean into.

Original post →

More from AGI Musings

AGI Musings channel →