Drug Development

On the same scale: a measurement only travels when it is tied to a shared standard.

A number you can act on: when a method's stated confidence is worth trusting.

The AI everyone discusses in drug development is the AI regulators have barely reviewed

The constraint is not model quality. It is a missing track record, and that is an application-layer problem

Qualified: Why Virtual Control Groups Are a Decision Problem, Not a Statistics Problem

The Infrastructure Nobody Wants to Build

Why Metadata, Not SEND, Is the Rate-Limiting Step for Virtual Control Groups

The EU animal use data does not say what most people think it says

NAMs in 2026: the regulator went AI-first. The data did not.

Agent Standards Are Not a Tech Detail, They Are a GxP Decision

AI Pain Scores in Drug R&D: Useful, Until You Trust Them

Novartis just published what its liver-chip got wrong. That is the win.