Blog
-
Stay Out of the Way: The VP Engineering Job Description Nobody Posts
Standard VP Engineering JDs describe a different job than the one actually done. 129 additive verbs. Five semantically-subtractive ones. Here is the job that nobody posts.
-
The AI Force Multiplier: What DORA Actually Found
Three years of DORA research on AI in software delivery, summarized. AI is a force multiplier, and force is indifferent to the result. Here is the data, the seven capabilities that decide where the force points, and what the 2024 paradox actually meant.
-
The Andon Cord in Software Teams
The lean manufacturing tool everyone wants to install. Why the cord is the cheap part, and what a decade of research actually says about making it work in a 6-10 person agile team.
-
Should I Kill My Change Approval Board?
A decade of research across two programs and 45,000+ respondents agrees: change approval boards hurt speed and stability simultaneously. Here is the data, the regulatory objection answered, and what to do instead.
-
Theory of Constraints for Engineering Teams
Goldratt's 1984 manufacturing novel keeps getting translated badly into software. Here is the actual mechanism, and why most engineering teams have a constraint they have not named and an improvement budget pointed everywhere except it.
-
WIP Limits Are Not Suggestions
Most teams have a WIP limit. Almost none enforce it. Why the number above the column is the cheap part, and what hard enforcement actually surfaces about your system.
-
One Piece Flow in Software Delivery
The manufacturing principle everyone cites but few apply correctly. What one piece flow actually means for software teams, and why the unit of flow is a well-sized story, not a commit.
-
The Slop Code Taxonomy
5 root causes, 9 failure categories, 100+ signals: a field guide to what goes wrong when AI writes your code and nobody reviews it properly.
-
The State of the State of DevOps
Fourteen reports. Two research programs. A decade of data on what makes software teams perform. Here is what survived, what didn't, and the uncomfortable conclusion the research keeps circling but can't quite prove.
-
DORA Metrics: What the Research Actually Says
DORA is the most cited framework in software delivery. But 2024 found AI adoption made pipelines worse while developers felt more productive. Here is what the research actually supports.
-
The Iron Laws of Agentic Coding
Sixteen constraints for building with AI agents: the kind that make speed possible by making catastrophe unlikely.