The Flour Mill Problem: Why 'Good Enough' AI Changes Everything
On August 20, a model called Ox Alpha appeared on OpenRouter with no company name, no press release, and no logo. Just a stealth label and a price tag of zero. For rou...
AI systems · governance · application
Efficient AI systems people can use, govern and trust.
I build efficient AI systems for complex organisations, then write about the engineering, governance, and practical decisions that make them useful in production.
Selected writing
Writing for the decisions around AI
From the journal / 01
Trust by Default Is a Cost Problem
One night YC quietly gave its AI agent full access to the production database. The agent became 10x more useful. That experiment, told in the Lightcone episode Inside ...
The Dilemma of B2B Software in the Agent-First Age
I read a post by Paul Graham on X a while back. It was mostly about something else, but somewhere in the middle he made a one-sentence statement about the future of so...
Selected work
Projects you can explore
Working systems and experiments
LLM Intelligence vs Cost
Compare the price of intelligence across the models people actually use.
Open dashboard →
Sumo Elo Explorer
A personal data project that makes strength across sumo eras explorable.
Open project →The common thread
The work follows three practical questions
Explanation, not navigation
The writing and projects share one concern: making AI useful in the settings where people must also understand, govern, and live with its consequences.
Build useful systems
How models, data, tools, and workflows become reliable products rather than impressive demonstrations.
Make them governable
How organisations set guardrails, make decisions, and keep AI accountable as it enters real work.
Put them to work
Where AI creates value in complex organisations, and where human judgement still matters most.
Browse the site
Choose a route
Start with what you need
The latest notes
Recent writing
From the journal / 02