GuidesBy professionSoftware developers and engineers
AI for developers who still own what they ship
For Software developers and engineers.
Use AI to draft code, explain unfamiliar codebases and speed up routine work, while you keep the design decisions, the testing and the responsibility for what actually runs in production.
When to use this
You have a feature to build, a bug to trace through code you did not write, and tests you know you should write but have not gotten to.
- Drafts a first version of a function, test or script from a clear description
- Explains an unfamiliar chunk of code or error message in plain terms
- Suggests a few ways to approach a problem so you can compare trade offs
Workflow
Describe the problem, the constraints and the existing patterns in your codebase
Ask for a draft or a few approaches rather than accepting the first answer as final
Read the generated code line by line, not just the parts you do not understand
Write or run tests against it rather than trusting that it looks correct
Check for security issues, especially around input handling and permissions
Prompts to try
Use it, adapt it, and make it your own.
Here is a function and the bug I am seeing: [paste code and error]. Explain what is likely going wrong and suggest a fix. Do not assume behaviour of other parts of the codebase I have not shown you, ask if you need more context.
Human lens
- You decide the architecture and the trade offs that fit this specific system
- You read and test every line before it is merged, especially the parts that look right
- You are accountable for security, performance and correctness in production
What to avoid
- Generated code can look correct while containing subtle logic or security errors
- It cannot see your whole system, so it may miss constraints that matter
- Do not paste secrets, credentials or proprietary code into a tool your employer has not approved
Tools that fit: ChatGPT, Claude, Gemini, Grok, Kimi · Reviewed 2026-07-19