Skip to content
Free since 2014. Built for the engineers who keep things running.

cat /articles/tests-as-data-ensuring-ai-generated-code-does-what-we-want.md

Git

Tests as data: ensuring AI-generated code does what we want

Matteo Vaccari argues that keeping acceptance tests as readable data, separate from the test runner, gives teams a practical way to specify what AI-generated code should do. The approach makes expectations easier to review and harder for coding agents to alter unnoticed.

freek.dev15 Nov 2026 · 1 min read · summary

article --summary

The summary

The article looks at how acceptance tests written as plain, readable data can serve as a specification for software produced by AI coding agents. Vaccari's central idea is to store these expectations apart from the code that runs them, so they are not buried inside test scripts. This separation has two practical benefits. People can review the expected behaviour without reading test code, and the expectations are easier to protect from changes a coding agent might make without anyone noticing.

Editorial note: This is a summary of an article first published by freek.dev. The full piece is on their site.

Reading now

Topic
Git
Source
freek.dev
Published
15 Nov 2026
Format
Curated summary

Have a take on this?

Write it up and submit it. It goes on the list once an editor approves it.

Submit an article

articles --latest

Latest articles

New from the blogs we follow and the community.

All articles

newsletter --weekly

LaraChat Weekly, minus the noise.

The best reads, new jobs and upcoming events, in one email a week.