Writing
Notes on AI and the built environment.
Longer pieces published here, and shorter posts I share on LinkedIn. Each one starts from a question a leader in engineering, construction or infrastructure might ask.
-
The coordination ceiling
AI is getting better at single tasks. Infrastructure changes when it gets better at teamwork. What Google DeepMind’s Organizing Intelligence paper, by Alex Imas, Andrew Koh and Shane Legg, means for water, transport and energy, with interactive charts.
Read the explainer → -
What does the evidence say about AI in your work?
Why I rebuilt the 38thMove website around a database of published examples, then 301 of them, and the patterns that stand out.
Read on LinkedIn → -
Engineering’s design loops, and the productivity catch
Engineering is goal-driven and iterative: design, analyse, check, revise. AI that can work through those loops could matter far more than faster single tasks, but saved hours do not automatically mean more output.
Read on LinkedIn → -
A hundred examples, and still no reliable productivity number
What more than a hundred published examples do show: a pattern of claims around progress reporting, asset inspection, energy and technical information, much of it still supplier-reported.
Read on LinkedIn → -
AI and the business model of built-environment consulting
If AI means fewer hours for the same work, who gets the saving? The firm, the client, the software companies or a new competitor. It does not automatically belong to the consulting firm.
Read on LinkedIn → -
Why I started collecting the evidence
People want a number to hang an AI decision on, and decision-grade public evidence is surprisingly hard to find. The start of the evidence map, with its first 69 cases.
Read on LinkedIn → -
Follow for new posts
I post on LinkedIn about what the evidence shows and what it means for firms and asset owners.
Follow me on LinkedIn →
Contact
Want to talk about any of this?
A short note about what interests you is enough. I will reply personally.