Writing on AI-assisted development, design, and building real products.
4 April 2026
AI doesn't remove the need for engineering discipline. It amplifies whatever discipline you bring. Strategy, design, context engineering, and DevOps are all part of the same practice.
Read →7 March 2026
When a tool constraint produces genuine stress, it's a sign the tool has become load-bearing. That's not a problem — but the dependency needs to be managed consciously.
Read →7 February 2026
The Model Context Protocol turns AI from an isolated assistant into a connected participant in the systems where work actually happens. Once you've used it, the old way feels unnecessarily constrained.
Read →3 January 2026
There's a point where you stop thinking about individual prompts and start thinking about systems. Reusable capability tools are the infrastructure that makes that practical.
Read →6 December 2025
The major AI platforms have all introduced persistent workspaces. It's a seemingly small change that has a large impact on how useful AI actually is in practice.
Read →1 November 2025
Every model shows performance degradation as context length increases. Understanding context rot changes how you think about working with AI.
Read →4 October 2025
AI first doesn't mean AI for everything. The most effective approach starts by understanding the problem — then deciding where AI genuinely adds value.
Read →6 September 2025
The conversation about AI and jobs has been dominated by the wrong question. AI gives back time, headspace, and the opportunity to do work that actually matters.
Read →