Direct answer
A source report is evidence, not automatic truth. ModelBreeder.com uses maturity labels to decide how aggressively a claim can be promoted into public teaching, architecture, tooling, and .uai active memory.
| Label | Meaning | Site handling |
|---|---|---|
| Established practice | Supported by common engineering practice or mature tooling | Can become direct implementation guidance. |
| Emerging practice | Used in current tools or papers but still shifting | Add caveats, compatibility rules, and tests. |
| Conceptual synthesis | Coherent framework built from multiple ideas | Use as theory, not as product claim. |
| Speculative analogy | Useful metaphor or future scenario | Keep clearly labeled and bounded. |
| Implementation note | Package or site-specific record | Use for current project operations. |
| Rejected directive | Source instruction that conflicts with the site contract | Preserve in /docs, record rejection, do not implement. |
Promotion process
PROCEDURE promote_source_claim(claim)
label <- CLASSIFY_MATURITY(claim)
SWITCH label
CASE "Established practice": ADD_TO_GUIDE(claim)
CASE "Emerging practice": ADD_WITH_TESTS_AND_LIMITS(claim)
CASE "Conceptual synthesis": ADD_TO_THEORY(claim)
CASE "Speculative analogy": ADD_TO_SPECULATION_BOUNDARY(claim)
CASE "Implementation note": ADD_TO_OPERATIONS_OR_UAI(claim)
CASE "Rejected directive": RECORD_REJECTION(claim)
END SWITCH
END PROCEDURECurrent application
The Rust runtime files are source-code evidence. The Four Fs and local tiny-model reports are conceptual and emerging practice. The instruction to remove safeguards is a rejected directive because it conflicts with governed release, no-op, reversibility, and mutualist persistence.
Source reports used for this guide
These reports are preserved verbatim in the site archive. The guide above is an editorial synthesis and may narrow, qualify, or reorganize claims from the source material.