Two different jobs that both say 'AI': using a coding agent to build and maintain a repo, and building AI into the product itself. How to drive Claude efficiently, what agents and MCPs actually are (with Figma/Linear/GitHub examples and the gh-vs-MCP call), and—when you want a feature that talks to a model—how to structure an AI service and wire it into the frontend and backend templates from the last few articles. The condensed version of everything I've learned shipping both.

AI is a part of our life now, we need to learn what they are and how they work, the objective of this post is to get into that.