AI Coding Tools

AI coding tools are the most immediate, measurable AI productivity gain available to a software engineer. But the gap between developers who get a genuine multiplier and those who get a frustrating, bug-introducing toy is entirely about how the tools are used.

This section is practical and tool-agnostic: the categories and skills, not a review of this month’s products.

The mental model

An AI coding tool is a fast, widely-read, eager junior pair programmer with no memory of your codebase and no judgment about your standards. Used as a collaborator you direct and review, it’s a real multiplier. Used as an oracle you trust blindly, it ships subtle bugs.

In this section

The Landscape The categories of AI coding tools — autocomplete, chat, agentic IDEs, CLI agents — and how they work.

Working Effectively Context, prompting, verification, and where AI assistance helps versus where it hurts.

Agentic Coding Workflows Delegating multi-step work to coding agents — planning, reviewing, and the pitfalls to avoid.

What you’ll be able to do

Choose the right kind of tool for a task, drive it with good context, verify its output rigorously, and delegate larger work to coding agents without losing control of your codebase.

Prerequisites

None specific — though LLM Engineering and AI Agents explain why these tools behave as they do.