Writing
Essays on solution architecture, AI-assisted software development, and the practice of designing systems in regulated environments. Published when the thinking is complete enough to be useful — not on a schedule.
A long-form piece on AI-assisted software development is being finalised for publication. It draws on the work that produced AryaWealth and LADE, and argues that the current category of AI coding tools has mis-allocated its labour in a way that better frontier models alone will not correct.
Expected publication: late April 2026.
Essays published here will be long-form rather than short — the shortest piece is likely to be two or three thousand words, and the longer ones substantially more. The working assumption is that readers who are willing to spend the time are the audience worth writing for, and that a smaller number of considered pieces is more useful than a frequent stream of shorter posts.
Subject matter is drawn from the practice itself: architectural patterns that survive contact with real systems, the economics and mechanics of AI-assisted development, and observations from working across regulated domains. Writing that is genuinely specific to a client engagement will not appear here; the pieces that do appear are about the general questions those engagements surface.