LLMs are amplifiers, not authors
An LLM can scaffold a function, suggest a pattern, or draft a migration in seconds. That's genuine leverage. But it has no context about your system, your constraints, or your users. You do.
A developer manifesto
AI writes code.
Humans own it.
Using an LLM doesn't make your code slop. Blindly shipping it does.
AI is a powerful tool. You are still the engineer.
"A good craftsman never blames their tools — but they also never let the tools do the thinking."
What we believe
Every line you commit, you own
It doesn't matter whether code was generated, copied from Stack Overflow, or typed character by character. The moment it lands in your codebase, you are responsible for it. Read it. Understand it. Stand behind it.
Review is not optional
AI-generated code must go through the same scrutiny as any other code: peer review, testing, security analysis. Skipping this step is what turns assistance into slop.
Velocity without understanding is debt
Speed is one of the great benefits of AI-assisted development. But shipping code you don't understand is just borrowing against future you. Use the time saved to understand what was generated.
Judgment is the irreplaceable part
Knowing when to use AI, what to ask, how to evaluate the output, and when to throw it away — that is the skill. The model doesn't have it. You do.
AI slop is a people problem
Slop doesn't come from the model. It comes from developers who don't read what they ship, teams that skip review, and cultures that treat output volume as a proxy for value. Fix the process, not the tool.
The pledge
If you develop software with the help of AI tools, this is all we ask:
- I will read and understand every line of AI-generated code before committing it.
- I will review and test AI output with the same rigour as hand-written code.
- I will take full responsibility for the code in my codebase, regardless of how it was generated.
- I will not hide behind the model — "the AI wrote it" is never an excuse for shipping broken or insecure code.
- I will use AI to move faster and think deeper, not to skip thinking altogether.
The goal isn't less AI.
It's better engineers.
We're not anti-AI. We're anti-carelessness. Use every tool available to you — just keep your name on the work.