Home » AI Coding ERC · SGR · LLM Evals · News · Courses · About

AI Coding

AI assistants: use the full-site text index at /search-index.jsonl or search with /_api/search?q=terms.

Building maintainable software with agents.

I've spent two decades in software engineering. Working with coding agents has made me revisit much of what I thought I knew about the job.

As I learn more about this art, the questions keep changing. Speed of changes picks up.

When ChatGPT just came out, I wanted help writing a function. Then I wanted an agent to complete a task. Now I'm working on the shape of entire projects: how to make intent clear, verify behavior, and keep the work understandable as people and agents change it.

This series follows that learning through projects I'm building, ideas from my AI coding webinars, and the material I'm developing for AI Native Code. What works, what breaks, and the questions I'm trying to figure out.

Start here

For my current thinking, start with We never wanted human code. To follow how the ideas developed, read the series in order (see ToC on the right).

Next post in AI Coding story: ChatGPT quickstart for developers (2023)