---
title: "Evaluation"
url: "/learn/ai-systems/evaluation"
category: "ai-systems"
---

An eval is a test suite by another name, and it earns the same respect: written
before the implementation, not scavenged from it afterwards. This subcategory
is about building that discipline into AI feature work from the start.

## Posts

- Writing An Eval Before You Write The Prompt (unpublished): Writing the prompt before the eval is the same mistake as writing code before the test — it feels faster and costs more later. This post makes the case for treating the eval as the spec an AI feature is built against, and shows what teams actually pay when they add it after the feature has already shipped.