← Index of published work

A multi-part build: from why a few good rules beat a mountain of data, to a playable lending world that judges your AI.

1. Part · SEP 1, 2026

   Hello (Hammer) Worlds

   Frontier world models for people who do not merely produce numbers. They have to defend them months later.

2. Part · SEP 24, 2026

   The Answer Key and the Instrument

   Hammer's harness experiments expose uneven AI judgement. The work is choosing the right actuarial method, using evidence, and completing a defensible analysis.