Writing & experiments
Field notes from the build.
Observations about software architecture, AI collaboration, product systems, and the practical choices that turn an idea into durable software.
The Demo Is Not the Product
What changes when an impressive AI prototype has to become a reliable, governable software system that people can actually depend on.
Software Architecture for Founders
The early technical decisions that preserve options, reduce expensive rework, and help a product grow without premature complexity.
Designing for Real-Life Variability
Lessons from building Tasked With, where a useful plan has to make room for missed days, changing priorities, and family life.
Local-First AI: Rethinking Memory, Privacy, and User Control
Why personal intelligence should have inspectable memory, explicit permissions, and an owner who keeps the keys.
When a Rewrite Is the Wrong Answer
A practical way to decide whether a troubled system needs replacement, rehabilitation, or a careful migration through smaller seams.
What I Learned After Building an AI Engineering Team
A candid follow-up on agents, context, review loops, parallel work, and the parts of software engineering that still resist automation.
Build, Buy, or Integrate?
A framework for deciding where a product team should invest its engineering time, accept a dependency, or keep an escape route.
Building an AI-Powered Engineering Team with CrewAI
How role-specific agents, well-scoped tasks, and deliberate orchestration can simulate a functional engineering team while keeping control and privacy close.
Want to compare notes?
I’m always interested in thoughtful software problems and new product ideas.
Send me a note ↗