Systems don't fail because of technology. They fail when they become hard to understand.
The Week I Chose Not to Build the Feature
Jun 25, 2026 3 min read
I identified a feature that would directly support a commercial step I want to take. The feature was buildable. Nothing was broken. I could probably have shipped it this week. I chose not to.
Depth of Thinking Now Has a Proportional Return
Jun 9, 2026 7 min read
For most of my career, thinking too carefully about a problem was a luxury I could not afford. Implementation cost was the ceiling. That constraint has shifted — and it changes what it means to build well.
When architecture decisions age well
Jun 3, 2026 4 min read
A conversation about llms.txt turned into unexpected validation.
The content architecture built for clarity and ownership
turned out to already be structured the way AI systems want to consume knowledge.
The Missing Layer Between Data and Prompts
May 31, 2026 1 min read
A small lesson from building with LLMs: when generated messages feel unnatural, the issue is often not the prompt or the model — it is how the underlying information is represented.
Why I deliberately chose a statically generated website
May 20, 2026 4 min read
A reflection on static generation as a foundation for content ownership, signal clarity, and more understandable systems in the AI-native era.
Building Shift Advisory’s website: from edge services to a deterministic serving node
May 11, 2026 3 min read
A small DNS limitation unexpectedly turned into a broader architectural exercise around ownership, operational complexity, routing boundaries, and deterministic infrastructure design.