Me, and the AI thingy.
First-hand write-ups on AI in brownfield systems. Costs and all.
-
Which of Your Words Is Which
Pointing an agent at your codebase makes ambiguity worse, not better. What grounds it is a small, pure-business lexicon and a phase order, not more context.
The context you author -
The Bigger the Window, the Quieter the Rot
A bigger context window raised the ceiling on what an agent can do. It never bought immunity from context rot. The fight is still the fewest and best tokens.
The context you author -
Prompting Does Not Survive a Real Codebase
A coding agent fails on a real codebase for structural reasons, not because you phrased it badly. This series builds the machine that makes it reliable, one mechanism at a time.
Building a harness for a coding agent -
Harness Engineering: The Machine Around the Model
The harness is everything that isn't the model. Harness engineering is the discipline of the machine that turns a raw model into an agent, and it sets the ceiling on what that model can reliably do.
-
Context Engineering: Deciding What the Model Sees
Context engineering is the discipline of deciding which tokens the model sees on every step of a task, not the craft of writing one good prompt.
-
What a Context Window Actually Is (and What It Costs)
A context window is the finite set of tokens a model reads on every step. Here is how it is built, why a longer one costs more than linearly, and why capacity is not reliability.