I explore new opportunitites before they become obvious — and I build in public while I do it. Experiments, essays and conversations from inside the work, not after it.
I build small working experiments to test what's actually possible. You get to see it running, not just argued about.
A living style guide and component library for this site — spacing tokens, type scale, color roles, and layout primitives, all documented in one place. Still catching up with the actual pages as new patterns show up faster than I can formalize them.
View experiment →
A sandbox for comparing prompt variations against the same eval set side by side.
View experiment →
Visualizing where a request actually spends its time across a multi-agent pipeline.
View experiment →
Testing CRDT-based sync as a replacement for a central source of truth.
View experiment →Most takes on emerging technology are either too optimistic or too cautious. I argue a position instead, and invite you to push back.
Cybernetics is not a theory of machines, but of viable systems under change. This reflective piece explores how that perspective reshaped my approach to AI-driven product design.
Read essay →
Resistance to AI is rarely about mindset or education. This essay argues that adoption emerges naturally when systems invite participation in their understanding.
Read essay →
The real value of agent systems is not smarter output, but clearer accountability. This piece reframes agents as a way to scope responsibility rather than centralize intelligence.
Read essay →
Workflows assume predictability; AI introduces variation. This essay explains why assessment must precede flow when systems are asked to interpret uncertain inputs.
Read essay →I sit down with people worth listening to. You get the conversation, unfiltered.
A conversation about where clean architecture meets messy reality.
View conversation →
A conversation about where clean architecture meets messy reality, and why most process failures are systems failures in disguise.
View conversation →
What changes when the thing you're building can be wrong, and how trust gets designed rather than bolted on.
View conversation →
Borrowing from control theory to think about product loops, and why most dashboards measure the wrong thing.
View conversation →I build the mental models that help me make sense of what I'm seeing. You get the vantage point without doing the climbing.
All mental models →
A cybernetic blueprint for how an organization stays functional as it grows.
View model →
How feedback loops and structure shape behavior more than any single decision does.
View model →
The study of feedback, control, and communication in systems — machines, organisms, and organizations alike.
View model →
Four team types and three interaction modes for organizing around fast flow instead of org-chart inertia.
View model →
How empowered product teams are organized to solve problems, not just build and ship features off a roadmap.
View model →