A Field Guide
Go to Working with AI→

Coding with AI

Jesal Mehta

Building things that remain understandable, editable and maintainable after the conversation is gone

AI-assisted coding is not only about generating working code. The larger goal is to build systems that:

CORE PRINCIPLE
Good AI-assisted development is not just about getting AI to write working code. It is about building something you can still understand when the conversation is gone.
COMPANION PRINCIPLE
The chat should help build the project; the project should not depend on the chat to explain itself.
Start With Intent
  1. 01 Describe the Outcome Before the Implementation →
  2. 02 Separate the Problem From the Solution →
  3. 03 State the Constraints →
Make The Project Legible
  1. 04 Make the Repository Explain Itself →
  2. 05 Record Intent, Decisions and Current Reality →
  3. 06 Keep Documentation in Sync With Code →
Structure For Human Editing
  1. 07 Separate What Changes for Different Reasons →
  2. 08 Keep Routine Content Human-Readable →
  3. 09 Expose the Things You Expect to Change →
Build Tools For Yourself
  1. 10 Use AI to Build Exploratory Tools →
  2. 11 Separate Product Code From Maintenance Tools →
Work Safely With AI-Generated Code
  1. 12 Make Small Changes and Define What Must Stay Stable →
  2. 13 Compare Implementation With Intent →
  3. 14 Test Behaviour and Use Version Control →
Build Toward Independence
  1. 15 Reduce AI Dependence as the Project Matures →
Summary
  1. 16 It All Comes Down to Five Things →

Good AI-assisted development is not just about getting AI to write working code. It is about building something you can still understand when the conversation is gone.

Jesal Mehta