Stoneman4-HI1107×1388 324 KB

Elton Stoneman

Getting an AI tool to produce code is one thing. Deciding whether that code belongs in your project, keeping changes manageable, and trusting what you ship are harder questions. That’s the part of Elton’s book that caught my attention.

Using Claude Code and Claude Cowork, you work through a growing collaborative workspace application, something like a hybrid of Slack and Figma. The project gives you a concrete setting for writing specs, directing parallel development work, and checking the results against your architectural decisions.

The approach stays rooted in familiar engineering practices: behavior-driven development, test-driven development, and documentation. It also tackles the less glamorous details that matter once you move beyond experimenting, including token budgets and maintaining a healthy skepticism toward AI-generated output.

You may know Elton from Learn Docker in a Month of Lunches or Learn Kubernetes in a Month of Lunches. Here, he brings that practical teaching approach to working with AI agents across the software development lifecycle.

Claude at Work is in Manning’s Early Access Program, so you can read the available chapters while the book is still being written and contribute feedback along the way.

• Full details: Claude at Work - Elton Stoneman

Don’t forget you can get 45% off with your Devtalk discount! Just use the coupon code “devtalk.com” at checkout