Lab-in-the-loop will prove most valuable for hit-to-lead optimization, not de novo hypothesis generation

Confidence: 6/10 Supporting evidence:

• Academic closed-loop systems (University of Toronto) have proven most effective at optimizing known chemical scaffolds, not discovering novel targets. Evidence: moderate
• The "make" bottleneck remains for de novo biologics synthesis, limiting the loop's scope. Evidence: moderate

Challenging evidence:

• Lila Sciences explicitly claims the loop will generate novel scientific hypotheses, not just optimize known leads. Evidence: low (unproven)
• AWS entering the space suggests commercial-scale utility beyond optimization. Evidence: weak

Evolution:

• Jun 2026 — Initial thesis at 6/10; requires Lila or Insilico LabClaw to publish results to update

Depends on: lab-in-the-loop Would change if: Peer-reviewed paper showing a closed-loop system independently identified a novel therapeutic target and validated it preclinically, without human hypothesis input