Posts tagged "ai"

A simulation of a real 5 point improvement showing how often a fixed eval threshold gives a random red build, and why a paired comparison detects the gain at a much smaller sample size.

Ship readiness for an AI feature is one artifact per failure class, each with a stated minimum bar and a stated way it can be invalidated. Built here with a runnable 300 item fixture.

A practical pipeline for learning from production AI failures while minimizing personal data, preserving the failure signal, and keeping evaluation fixtures reviewable.

A system-boundary model showing why useful AI products still depend on identity, freshness, permissions, schemas, replay, and observable data flows.

Prompting is cheap, understanding is expensive, and each skipped review makes the next one harder. This is how AI-assisted velocity can quietly decouple teams from their own codebases.