Field notes
Practitioner essays on building production software with AI agents.
Operator-grade reports on the systems, decisions, and economics of shipping software with AI — what worked, what failed, and what changed.
Articles
-
The AI Soft Ban Assessment
Executive Deck ↗Exec summary ↗Listen ↗A browser-only scoring assessment for whether your company has quietly soft-banned AI in the SDLC while still calling itself AI friendly.
-
It’s Okay to Waste Tons of Money with Bad Consulting Partners, but Tokens Are Too Much Money?
Executive Deck ↗Exec summary ↗Listen ↗Before finance audits token spend, audit the expenses that somehow escaped fiscal responsibility: offshore teams, staff augmentation, systems integrators, consulting partners, Scrum layers, agile coaches, planning ceremonies, and delivery management. How much did they actually cost, and what success rate did they produce?
-
You Trust the Lowest Bidder. But Not the Best Frontier Model?
Executive Deck ↗Exec summary ↗Listen ↗You trust the lowest bidder's pull request. Why not the best frontier model inside a system that actually checks the work?
-
Find the Ceiling
Executive Deck ↗Exec summary ↗Listen ↗You are sitting in the Q3 budget review. The VP of agile transformation just got taken apart over the agile coach line. Cloud migration is six quarters late. The chief product officer is presenting another reorg and another offsite. You are up next, and for the first time in three years of these meetings you have real numbers. The winning numbers. The kind that beat expectations. The teams that doubled their month-over-month spend on Gen AI for building software shipped considerably more measurable value than the teams that did not. You are about to ask the room for more money. You are going to call it finding the ceiling.
-
Token Economics Is the Wrong Spreadsheet
Executive Deck ↗Exec summary ↗Listen ↗Three times a week now an engineering leader I respect tells me their CFO wants them to justify the ROI on AI tokens. The question itself is the part I am puzzled by. A token is a raw material, the way steel is a raw material at a foundry. Nobody audits the flour invoice at the bakery. The question is not a token question. It is a measurement question, and it is being asked on the wrong floor of the building.
-
I Want You Software Developers to Be Unhappy (Keep Reading, It’s Not What You Think It Is)
Executive Deck ↗Exec summary ↗Listen ↗The AI Engineer is a new job family, and it is the future of this profession. The "software developer" as the role has been defined for fifteen years is going through what farmers went through with the tractor. If you are still doing the old job and you are not unhappy, you have not been paying attention.
-
You Have a Sub-Five Miler. Your Relay Team Still Loses.
Executive Deck ↗Exec summary ↗Listen ↗You have ten engineers who could ship the next quarter alone. You have forty more who, through no fault of their own, cannot. Where do you put the AI investment, across the floor or into the ceiling? On the math, the answer most CTOs give is wrong, and the org chart you currently run is the residue of avoiding it for a decade.
-
For Five Days His Team Was Accidentally Allowed to Be as Good as They Actually Are
Executive Deck ↗Exec summary ↗Listen ↗My neighbor Bill manages a software team. His company accidentally gave them unlimited AI tokens for a week. They shipped more in seven days than they had in the prior six months. Then finance noticed the bill, and the door closed. Then Bill found a workaround.
-
Why Are You Deprioritizing the Most Important Training Your Org Will Ever Get?
Executive Deck ↗Exec summary ↗Listen ↗You told the board AI would compress delivery cycles by 30%. Then you scheduled the training for 3 PM Friday. This is the private conversation your peers will not have with you.
-
Without Writing Out the Standard, Your AI SDLC Will Struggle — Introducing the AI Software Engineer, a Silly Name for a Serious Problem
Executive Deck ↗Exec summary ↗Listen ↗You rolled out the training, bought the licenses, ran the pilots, and your engineering org is producing almost exactly what it was two years ago. The problem was never the training. The problem is that you never set a new standard for what it means to work at your company, and you are the only person who can. Introducing the AI Software Engineer — a silly name for a serious problem.
Essential or Ornamental
Three companies. Three choices. One satisfactory ending.
One does nothing. One maps the waste. One bets everything on twelve people in a warehouse.
Read free online →