Writing / Selected essays
Enough evidence.
A clearer decision.
Three ideas for people building and reviewing consequential AI work.
I Do Not Believe In Done Without A Receipt
A useful closeout gives the next person something they can inspect.
I Do Not Let Metrics Drive
A score becomes useful when its scope is clear and responsibility stays visible.
I Do Not Trust Old Permission
An approval should still describe the work that is about to happen.
The Failure Museum
Three illustrated lessons about the gap between a reassuring signal and a justified decision.
Visit the selected collection →