Where AI investment in drug discovery can pay off, and where it can’t

Seven in ten drugs that start Phase 2 never pass it. Working Note №07 argues that this is where the billions now going into AI drug discovery will be repaid or lost, and that the money most likely to move the wall is the money that observes patients sooner and more often.

The short version is free on Substack. The full version runs about twice as long. It adds the arithmetic behind the weather horizon and the selection model, the complexity simulation, the captopril data and all 54 sources, in a 24-page PDF you can mark up and pass along.

Tell me a little about yourself and it’s yours.