Articles
Notes from shipping software with AI
For engineering leaders who want AI-assisted delivery they can check. Written from our own work, not from the sidelines.
Start here
Review capacity
The review queue is the bottleneck AI coding created
Why more code did not mean more shipped, and where the queue actually forms.
Instruction drift
Agent instruction drift in enterprise AI teams
How every developer ends up with a different AI, and what a shared rule set changes.
What was tested
AI coding checkpoint contract: make checks reproducible
What “tested” should mean before a pull request is allowed to say it.
Latest
-
Almost-Right Code Changes the Economics of Review
Clearly wrong code is cheap to reject. Almost-right code consumes expert attention because plausibility can hide one unproven product promise.
-
Measure AI Coding Without Mistaking Activity for Value
A team can generate more code and complete more tasks while creating little product value. Better measurement connects speed to accepted outcomes.
-
The Review Queue Is the Bottleneck AI Coding Created
Faster code generation does not remove engineering work. It moves the bottleneck into review, verification, and accountable decisions.
-
From ticket to evidence: the eight stages of a governed AI feature
A governed AI feature development workflow moves work through eight connected stages: intake, planning, specification, development, review, checks, documentation sync, and delivery.…
-
Why paqad-ai addresses a different scope than an AI skill
An AI skill packages reusable expertise for a task. An AI coding workflow for teams controls how a feature moves through planning,…
-
AI slop is a workflow symptom, not a developer insult
AI slop in software development is low-quality, error-prone output that is cheap to mass-produce. It is best treated as a workflow symptom.