I’ve gone through the hype cycle of AI coding tools — from early disappointment to genuine productivity gains — and the journey has taught me when and how these tools actually earn their keep.
The Early Days: GitHub Copilot in VS Code
When I first tried GitHub Copilot inside VS Code, I wanted it to be magic. Instead, it was mostly noise. The suggestions were often irrelevant, and the time I spent correcting them outweighed any gains. It felt easier to just write the code myself.
Cursor and Zed: Incremental Improvement
Later, I tried Cursor and Zed. Cursor, at that stage, felt like a slightly unpolished VS Code clone, but I could see potential. Zed had far more polish — fast, clean, and pleasant to use — but the underlying AI capability was still the same idea: edit text in files.
The Breakthrough: Claude Code and OpenAI Codex
Everything changed when I started using Claude Code and OpenAI Codex. These tools could operate like a real assistant — moving files, running tests, executing Git commands — not just inserting text. Suddenly, AI wasn't just 'autocomplete on steroids'; it could actually work alongside me in my dev environment.
Over time, these capabilities made their way into Cursor and Zed, and the tool usage experience improved dramatically.
My Rule of Thumb: AI for the Gruntwork
The biggest win for me has been offloading anything that doesn’t feel bespoke:
- Boilerplate generation — DTOs, interface stubs, config files.
- Unit test scaffolding — AI gets you to a good starting point quickly.
- Build pipeline tweaks — YAML edits, CI/CD job wiring.
- Terraform & infrastructure scaffolding — repetitive, rules-based work AI excels at.
The creative or business logic stays firmly in human hands. That’s the part where context, nuance, and intent matter most.
Real Examples
Two examples from my own work:
- The Pride avatar post — AI handled the image transformation boilerplate while I focused on the creative concept.
- The GitHub avatar sync post — AI automated tedious API calls and formatting while I refined the logic.
Closing Thoughts
AI coding tools are no longer a gimmick in my workflow — but they work best when I treat them as a junior developer who’s brilliant at repetitive work. I delegate the gruntwork, keep the important decisions myself, and review everything before it ships.
Pro Tip: Use AI to create space for deep, creative thinking. If you’re spending less time wiring boilerplate, you have more bandwidth for architecture, problem-solving, and innovation.