SIGNAL//SYNTH
Ai

Lab-in-the-Loop at Compute Speed — Ian Quigley, Leash Bio

aired Oct 14, 2026 · 109.0m
Signal
84.3/ 100
High signal
confidence 0.99
Orig54.7
Actn100.0
Dens100.0
Dpth100.0
Clty61.7
Summary

My co-founder Andrew Blevins and I were at Recursion looking at a small molecule on a screen. And Mike was like, hey, looks like a Stuart Shribert molecule.

Why listen

It goes beyond the title with direct discussion of like, yeah, these, including: And then a medicinal chemist walked by us that's named Mike Cucarisi.

Key takeaways
  1. 01And if the models can identify the chemist by looking at the molecule alone, then maybe the models can infer chemist intent, which means maybe they don't learn anything at all
  2. 02And you can use that representation and then train a model to predict activity from Kemble
  3. 03When you train models on the exhaust, you might be in trouble
Best for
listeners looking for a practical AI episode debrief