Cost of Delay in an AI World: A Map for When the Ground Moves
Efficiency and cost of delay are not rivals. In the AI era, they are tools for different terrain. A Wardley Map shows when one should lead and the other should follow.
Latest Post
Efficiency and cost of delay are not rivals. In the AI era, they are tools for different terrain. A Wardley Map shows when one should lead and the other should follow.
Read the postEfficiency and cost of delay are not rivals. In the AI era, they are tools for different terrain. A Wardley Map shows when one should lead and the other should follow.
A prediction that the near-future repository contains only intent, scenarios, data shapes and a UX sketch. The running application becomes disposable, regenerated on demand, and personal to a single user.
How do teams and organisations build, recognise, and leverage prompt engineering skill at scale? On hiring, pairing, team topology, and the talent pipeline problem when AI is doing the work junior engineers used to learn from.
Domain expertise is a force multiplier when working with AI. A look at the Dreyfus skill model applied to prompt engineering, and why what you know about the system you are building matters more than any prompting technique.
My daughter and I spent a few hours with GitHub Copilot planning, building, and shipping a real browser game from scratch. Plan it, build it, play it, fix it — the simplest development cycle, and it works.
Small, contract-bound, independently operable units change what AI agents can reliably deliver. A look at the architectural hypothesis worth testing — and an invitation to try to break it.
Give an AI agent one reference point and it produces something generic. Give it three or more — tests, contracts, documentation, standards — and it can triangulate. The geometry of context, and how to use it.
An introduction to Deneb — why it exists, what it stands for, and what honest AI-assisted writing looks like in practice. The internet has enough AI hype. This is something different.