A Working Theory

by Hanif Carroll

Field notes on AI, judgment, and better ways of working.

Practical notes from using AI in real work—what helps, what fails, and where judgment still matters.

Hanif Carroll looking toward the A Working Theory introduction

Issue 001 · July 16, 2026

Reliable AI Workflows Are Learned Through Failure

A reliable AI workflow becomes clearer when a failed run reveals which decision, approval, or stop condition came too late.

Read issue 001Browse the archive

A Working Theory

by Hanif Carroll
001

Automation · Judgment · Work

Issue 001 · Published

What you’ll receive

  • Practical AI workflows
  • Judgment automation cannot make
  • Better ways of organizing real work

Get the next issue when it’s ready.

Field notes on AI, judgment, and better ways of working.