Essays
AI Governance
How to govern what an agent reads. Ongoing series.
Your AI isn't hallucinating. You're feeding it garbage.
Two invisible laws explain why your assistant keeps failing even as models keep getting better: volume without judgment, and the same concept living in three places.
Read the full essay
The fix for your AI already existed. Nobody had brought it to documents.
Kitchens solved it with three habits, and the data world arrived at the same ones thirty years ago: a contract at the door, a memory that never overwrites, and a shared recipe book with a local menu.
Read the full essayThe Data Kitchen
The foundation: fifteen essays on data governance, from the recipe to the product.
The Centralist Paradox: Why Your Data Team Is a Bottleneck (and It’s Not Their Fault)
An audit that needed the data yesterday, a central team with a three-month queue, and the root cause: the bottleneck isn't people, it's architecture.
Read the full essay
The Recipe: When the Business Stops Asking and Starts Building
Before you cook, you need a real recipe: who writes it, in what language, and why the game of telephone between the business and IT costs months.
Read the full essay
The Pantry: Why the Most Important Data Is the Data You Don’t Touch
Raw data exactly as it arrives: insert-only, traceability, and why touching the original ingredient destroys your audit trail.
Read the full essay
Mise en Place: Why your data transformations are a ticking time bomb
Homonyms, duplicated calculations under identical names, and why a staging layer without a contract is a ticking time bomb.
Read the full essay
Plating: Why your 500-column table is a mistake (and how to fix it)
The consumption facade: thin projections, read contracts, and why serving everything in one giant table is a mistake.
Read the full essay
Two perfect recipes — and why that's not enough to scale
Two domains with flawless recipes but no common plane: why local perfection doesn't scale without a shared kernel.
Read the full essay
Objectville's central warehouse: one common contract the whole chain subscribes to
Hub, satellite and link as a shared warehouse: a contract every domain adds its pieces to without breaking the model.
Read the full essay
The validator robot that learns — living documentation in action
Rules that run against reality every day: documentation that doesn't quietly grow old in a PDF.
Read the full essay
Is it the supplier's problem or the franchise's? — the catalog that learns and the rejection trail
Quarantine, a catalog of rejections, and the first right question when data fails: where did it come from?
Read the full essay
The factory that never deletes — how the model versions every change without destroying anything
Insert-only as a structural property: rebuild the model as of day X with a query, not a forensic project.
Read the full essay
Reproduce, don’t recalculate — the rule of the moment, the map of a spreading error, and closing the pillar
A March audit with March’s rule: reproducing isn’t recalculating with today’s rule dressed up as history.
Read the full essay
Mother Sauces — how a platform grows without breaking what already works
The Business Vault as certified mother sauces: versioned shared calculations that domains consume without recomputing.
Read the full essay
Each branch builds its own pizza — Strategy and the operable self-serve platform
Strategy on a shared kernel: each branch composes declaratively instead of inheriting a tangled central monolith.
Read the full essay
The Kitchen Pass — facade, physical plating and specifications that sign off the plate
A facade with a signed card: evaluable specifications between the kitchen and the consumer — not SQL pasted in without a contract.
Read the full essay
Data as a Product, the roots that hold up the chain, and closing the block
Closing the series: data as a domain product, the chain’s roots, and why flattening what’s local shuts down branches.
Read the full essay