by Amit Shekhar · 25 August 2026
AI Is Only as Good as Our Definition of Done
One simple idea, AI is only as good as our definition of done, which means the clear line we draw between a task that is finished and a task that is not. We will also see what a definition of done really is, why AI feels magical on the tasks that a machine can check, why it feels unreliable on the tasks that only a human can judge, how the loop that AI runs behind the scenes creates this exact gap, and how we can write a definition of done that turns a vague task into a checkable one.
Before you read, guessWhat requirement ensures a function behaves predictably for specific inputs?
Ten seconds, a guess, then read — a wrong guess still makes the answer stick.
What this article covers
- What is a definition of done
- Tasks where the definition of done is exact
- Tasks where the definition of done is fuzzy
- Why AI is so strong exactly where we can measure
- How to write a better definition of done
The article lives on outcomeschool.com. Read it there, then come back: the tutor in the margin has read it and will answer questions, and the questions below check what stayed.