After 30 years of coding, I learned AI works best when you treat it like a junior developer. Here are 5 skills that changed how I work with AI teammates.
The story
After 30 years of coding and 8 years working with AI, I’ve learned that getting the most out of AI teammates isn’t about magic prompts – it’s about treating them like junior developers.
Here are 5 skills that changed how I work:
1. Prompt like you’re writing a ticket
Vague instructions get vague results. I stopped asking AI to “write a sorting function” and started being specific: “Write a Python function that sorts a list of integers using quicksort, handles empty lists, and includes docstrings.”
2. Break everything into chunks
AI chokes on complexity. I never ask it to build an entire system. Instead, I delegate atomic functions and isolated modules – UI separate from backend, integrations handled one at a time.
3. Match the tool to the job
Copilot crushes boilerplate. Claude handles long-context reasoning. ChatGPT nails architectural advice. I’ve stopped using one AI for everything and started picking the right one for each task.
4. Review everything
I treat AI output like code from an intern – useful, but never blindly trusted. I run tests, check for logical errors, and enforce style guides through CI/CD. Version control is my safety net.
5. Set guardrails
Clear standards keep AI on track. I feed unit tests, type definitions, and coding standards into my prompts. Pre-commit hooks catch style drift. Static analysis finds hidden bugs.
AI doesn’t replace expertise. It amplifies it – if you know how to lead.
Recent Comments