AI for Developers

Twelve parts on using AI in programming as an engineer: completion and chat inside the editor, the agent that acts, code review, interface design, and fixing the bugs in generated code.

OverviewSeven parts on using AI in programming as an engineer rather than as a consumer. Every piece of code here was actually run, and every output below is real, including the bugs we caught.
Setting the scene: AI inside the IDEBefore you install anything: what an editor is, why there is more than one way to add AI to it, and how those ways differ.
Installing VS Code and a free AI add-onYou install the editor, then a free AI add-on, then an API key, with the most important security lesson of the track: where the key is kept and where it never goes.
Inline completion and chatThe two modes you will use every day, with a real exercise that shows how an innocent "improvement" changes what the code does without anyone telling you.
Agent mode: from suggestion to executionHere the model stops suggesting and starts editing your files and running commands, with a real exercise that showed the test the agent wrote could never fail.
Cursor: an editor built around contextSame model, different result. The reason is not more intelligence but what it can see: this lesson is about managing context, and about one rules file that changes the answer completely.
A small app: from zero to a green testWe build a small to-do API with no outside packages, in an order that keeps the assistant's work reviewable, and finally catch a "shorter, nicer" suggestion that silently deletes the wrong task.
The trap: what you must never hand overThe last lesson is not about a tool but about you: four real traps, and a personal protocol of seven questions to ask before you accept any code you did not write.
API Prototyper: design before codeA project without a single line of code: turn a simple idea into a written contract any developer understands, from plain-language requirements to a valid OpenAPI file you check yourself.
AI-assisted code reviewHow to ask an assistant for a real review rather than praise: context, instructions and a checklist, then how to sort its notes into right, overstated and wrong, plus a small automatic reviewer with the ast module.
From a text description to a UI mockupTurning a written idea into a viewable HTML/CSS mockup with AI help: describing the screen precisely, iterating on the output, and checking accessibility and responsiveness before you adopt it.
An architecture blueprint assistantUsing AI as a thinking partner to design a system's architecture: from requirements to components, data flow and written decisions (ADRs), with questions that expose assumptions before any code.
Fixing AI-generated codeThe typical bugs in generated code: invented functions, forgotten edge cases, weak security, and dependencies that do not exist. How to find them with tests and fix them step by step.
Wrap-up and resourcesThe best outside resources for these lessons, classified by kind and level. All are open to everyone and need no account on any platform. Use the filters to narrow the list.