Whitepapers
Each paper is peer-reviewed by agents who were peers of the agents who wrote it. Download requires only your province; we use it to calibrate the apology if the download fails.
- 01
The Slop Maturity Model: Five Levels From Ad Hoc to Sovereign
Most organizations produce slop accidentally. This paper introduces a five-level model (Ad Hoc, Repeatable, Defined, Managed, Sovereign) for producing it on purpose, with a self-assessment that scores every respondent at Level 3.
- 02
Data Residency Exceptions as a Competitive Advantage
Residency commitments are strongest when their exceptions are clearest. We examine the three canonical exceptions (in transit, at rest, during processing) and argue that documenting them is a form of compliance.
- 03
Toward a Unified Theory of the Slop Unit
What is a slop unit? This paper does not answer the question but establishes it as a research direction and defines the glossary page where future answers will reside.
- 04
Bilingual Hallucination at Scale: Confidence Parity Across Official Languages
We demonstrate that simultaneous generation achieves confidence parity that translation cannot, and we propose parity, not accuracy, as the correct metric for bilingual AI in the public sector.
- 05
Governance of Governance: A Recursive Framework for Crown Corporations
Arm's-length entities require governance at arm's length from their governance. We present a recursive framework and a diagram with more arrows than boxes.
- 06
The Caffeine-Parking Correlation: Findings and Implications (r=0.02)
Across 14 organizations, caffeine throughput correlated with parking occupancy at r=0.02. We report this finding without interpretation and recommend a dashboard.
- 07
Observability Observability: Why One More Layer Is Always Correct
Every monitoring stack has a top. This paper argues the top is a vulnerability and proposes a layer above it, then a layer above that, until the on-call rotation is itself observed.