@kentcdodds: You know how when you tell an agent to remove something, it adds a test assertion to make sure that thing is no longer …
Summary
Kent C. Dodds describes how AI agents often add unnecessary test assertions when removing code, and shares that he had an agent create an oxlint plugin to prevent this behavior.
View Cached Full Text
Cached at: 09/06/26, 12:51 PM
You know how when you tell an agent to remove something, it adds a test assertion to make sure that thing is no longer there?
That’s super annoying and it’s fine for verification but stupid to keep.
So I had the agent make an oxlint plugin to stop it. https://t.co/7wBNxCxrdx
Similar Articles
@kentcdodds: When you were writing software in your codebase, you felt the pain of friction and would grease the wheels as a result.…
Kent C. Dodds discusses the need to proactively manage friction when using AI agents in software development, as agents may work around issues without reporting them.
@kentcdodds: Stop wasting your time! Have you thought about... letting your agent hit the merge button? Close the agent loop and the…
Encourages developers to close the agent loop by allowing AI agents to automatically merge code, saving time.
@kentcdodds: If your agents are making slop, don't give up or fight the agent with cleanup. Work on the primitives the agents have t…
Kent C. Dodds advises improving the primitives used by AI agents to reduce low-quality outputs instead of relying on cleanup, illustrating the point with a linked thread.
@mattpocockuk: "No tautological tests"
A tweet criticizes the AI model Opus for adding excessive unit tests for simple constant strings, highlighting issues with tautological testing in AI-assisted development.
@gwenshap: One quirk of AI generated code is excessive guard rails. Recently, I wanted to test a new API with a local stack. I ask…
A developer shares an experience where OpenAI's Codex added an excessive guard rail by inserting a runtime extension existence check into an API, which a human engineer would never do.