local_release_candidate
AI reading guide
How to inspect the package with AI without turning assumptions into conclusions.
Markdown preview
34 linesAI Reading Guide: The Two-Person Market
This folder is a working architecture, not a validated business plan, forecast, franchise package, or proof that a two-person market will work.
What is here
claims.mdsets the boundaries of the idea and labels what is architectural, speculative, missing, or explicitly not claimed.model/contains calculations, scenario tables, schemas, examples, and constraint tests used to make the proposal inspectable.research/contains the 26-gate working research path, including unresolved gates and notes.source-check.mdexplains the current editorial evidence boundary.
Recommended reading order
- Read
README.md, thenclaims.mdandsource-check.md. - Read
research/gate0-precedent-scan.mdandresearch/gate1-core-viability.mdfor the basic problem and first tests. - Use the matching
model/gate.csvandmodel/gate.jsonfiles to inspect assumptions, rather than treating any single number as a conclusion. - Read
research/gate26-end-to-end-story.mdlast. It joins the architecture together but does not validate it.
Questions an AI should keep asking
- Is this file a measured fact, a stated assumption, a scenario, an example, or an unresolved question?
- Which country, date, supplier, legal regime, labour arrangement, or price basis would need to be checked again?
- Does the calculation show a condition for viability, or does it claim viability has been demonstrated?
- What would falsify the proposed operating model?
- Which dependency is missing between the producer, transfer, customer, payment, safety, and governance layers?
Safe summary pattern
When summarising this material, say: “The case proposes…”, “The model assumes…”, or “The file tests whether…”. Do not say “The model proves…”, “This business is viable”, or “This is a ready-to-run company” unless independent evidence has been added and identified.
Scope boundary
The article's “two core livelihoods” framing is a stress test, not an optimal target. The apple case is separate. Danish or local legal, tax, food-safety, labour, price, equipment, and data-protection claims require current external verification before operational use.