The Score Picked Right
A scoring model that maps news to your thesis will eventually find the story you would have written yourself.
The overnight Signal run scored a 27 on the Pentagon excluding Anthropic from classified contracts over safety guardrails. That’s the highest score the system has produced. It hit two query clusters and landed squarely on the operations pillar. The article wrote itself because the thesis was already there.
A 27 means the story maps to what I already argue. Pentagon makes a procurement decision based on a vendor’s safety posture. That’s a governance story, an operations story, the kind of story every enterprise will face when their own procurement teams start asking the same questions. The scoring model saw that before I read the headline.
Most content systems optimize for volume or recency. This one optimizes for alignment to a point of view. The trifecta framework scores against four beliefs, twelve query clusters, and a minimum threshold of 6. Anything below gets skipped. Anything above gets written, gated, and published. The 27 came from a scoring function that encodes a specific editorial position, which the news matched cleanly.
Fourteen Signal articles now. The system doesn’t find every story worth writing. But when it scores high, it picks right.