How I Stopped Fighting My AI Code Assistant and Started Building Better Software
https://medium.com/@riccardo.tartaglia/how-i-stopped-fighting-my-ai-code-assistant-and-started-building-better-software-51490b2fe42e
I'm a software developer with 14 years of experience. I know how to code, and I've been doing it for a long time.
Then something changed: Here's my transition into AI in my day-to-day work.
This resonates — code assistants aren’t just about speed of output, they’re about how they fit into developer workflows without adding cognitive debt. In most real projects I’ve worked on, the real challenge is not just generating code but making sure the output integrates predictably with existing patterns, passes tests reliably, and doesn’t introduce regressions as the codebase evolves.
In practice, one of the first metrics I watch when adopting AI-assisted workflows is test suite stability under incremental generated changes — if tests start failing more often or require constant overrides, that’s a signal the integration isn’t helping.
Curious — in your experience, when you shifted from “fighting the assistant” to trusting it, what signal or behavior told you that it was actually improving your workflow (e.g., fewer fixes per generated change, fewer manual rewrites, faster loop time with tests)?
I'm curious to hear how AI has impacted your workflow and productivity. Has it freed up your time to focus on more strategic tasks?
AI has significantly improved my workflow.
Overall, it allows me to focus on design decisions and code quality instead of routine work.