Systems, taste,
and the long arcs
underneath both.

Origin

The question came before the medium.

My first sustained interest was psychology: how thought is formed, represented, and translated into observable behaviour. That led naturally toward cognitive science, symbolic systems, and computing: different ways of asking how internal models become decisions, actions, and shared reality.

Software gave those questions a practical medium. A data model is a representation. An interface is an argument about what matters. A product encodes assumptions about people, incentives, memory, and choice. Engineering became the discipline through which these abstractions had to survive contact with use.

So most of what looks like product judgment is judgment about shape: where a boundary belongs, what an interface promises, which abstractions earn their keep, and what a system teaches the people who use it.

Practice

For now, I build products.

For now, I work through these questions by designing and building software products. I have worked in greenfield and legacy systems, founded and joined early teams, rebuilt customer-facing platforms, led delivery, and helped integrate products after acquisition.

The recurring work is not tied to one layer. It moves from business model to product architecture, from data and interfaces to infrastructure and operations. The aim is dependable software that remains understandable as the product, team, and organization change.

The ideas here were earned in production: software used at scale, natural-language analytics, real-time systems, migrations, and products that had to keep working while their foundations changed. The record is here.

Working posture

How I work with people and systems.

I am usually the person asking questions until the business and the engineering describe the same system. I look for the hidden assumptions, the real boundaries, and the decisions that will become expensive later.

I lead when the work needs shape, follow when someone else has the clearer view, and try to leave teams with better mental models rather than dependence on me. Reliability matters to me in both senses: the software should keep working, and people should be able to rely on the person responsible for it.

The convictions underneath that, continuity over novelty, explicit boundaries, restraint as an engineering discipline, are collected here.

Writing

Two bodies of writing, one underlying inquiry.

My technical and product-engineering writing lives here: software architecture, product systems, Staff Engineering, AI-assisted development, reliability, and the practical judgment required to build and operate software.

In Principle holds the wider inquiry: learning, culture, cognition, memory, language, continuity, and the technologies that shape how people understand and act. It began from the belief that education survives beyond the facts we can still recite: that what remains is a way of seeing.

Education survives when what has been learnt has been forgotten.

The boundary between the two is porous, but useful. One stays close to the engineering practice. The other follows the human questions beneath it.

Both have been shaped by people who treated clarity, representation, language, and composition as serious responsibilities. Listening belongs to the same training: sequence, timing, restraint, and attention to what is missing; ko3 is the product expression of that practice.

Longer arc

The questions remain larger than the medium.

The longer arc remains close to cognition, symbolic systems, computing, and the ways representations shape human behaviour. I may eventually return to those questions through more formal study or adjacent research. For now, I work through them by building products, studying systems, and writing.

ReferentGround is one current expression of that inquiry. As AI agents write more software, the problem is no longer only whether the code runs. It is how a system represents intent, preserves evidence, and remains understandable to the people responsible for it.

In Principle follows the same concern at a different scale: what people, cultures, and institutions inherit; what remains after explicit instruction is forgotten; and how tools quietly shape the thought they appear only to carry.