UNI Encyclopedia & Cookbook

Two books and a build, in one repository.

• encyclopedia/ — a science-honest reference work about the UNI program: what it has actually built, at what evidence class, with every negative published and every falsifier live.
• cookbook/ — the buildable recipes: the L0–L12 developmental ladder, and now the NATURA recipes (build DNA, a sperm, an ant colony, a dinosaur, a whale, a bat, a rock, water, air, a star, a human — and the open question beyond).
• gpt/ — a 20-file pack that loads this corpus into the UNI Active Inference Guide custom GPT, built reproducibly from the two books by tools/build_gpt_pack.py.

The honest position, printed here and never softened: ~2 of 11+ developmental rungs earned. The UNI program is a developmental active-inference simulation — a bounded peek, a toy world, never a person. "Full human" and "beyond human" are permanent open questions, never targets.

The two sovereign ledgers (the rule that organizes everything)

This repository carries two evidence vocabularies. They are sovereign, and merging them is the cardinal error.

A nature citation is never a UNI gate. Reading Kleiber's law raises no UNI rung. Citing published biology makes no UNI claim proven. This is the same discipline the repo already applies to UNI.OS ("substrate evidence is never general-AIF evidence"). Cross-reference the two ledgers by explicit link only — never by merge.

Where any prose disagrees with its ledger, the ledger wins and the prose is wrong.

Map

Nature as the authority — stated precisely

The doctrine is defensible in exactly one form, and the corpus holds it there:

Nature is the authority because it is the only system that has already run the experiment — a very long parallel search under real physical constraints, in which the failures were deleted. Convergent evolution is evidence of a constraint-optimum.

And it is only honest with its counterweight, which travels with it everywhere:

Not every trait is an adaptation (Gould & Lewontin 1979). Drift, phylogenetic inertia, developmental constraint and frozen accidents are real — the vertebrate retina is wired backwards. "Nature does it this way" is a hypothesis generator, never a proof. The bio-inspired design still has to beat a tuned conventional baseline on a pre-registered metric, or it is recorded NEGATIVE.

That second paragraph is what separates this from cargo-cult biomimicry.

Rebuilding the GPT pack

python tools/build_gpt_pack.py

The build merges the corpus into exactly 20 knowledge files (the OpenAI hard cap), validates the JSON, gates the Instructions file against the 8,000-character limit, writes gpt/MANIFEST.json, and zips to dist/. It fails loudly rather than truncating silently.

Never hand-edit a gpt/knowledge/K* file: it is a build artifact, and your edit is overwritten on the next build with no record of why. Edit the chapter in encyclopedia/ or cookbook/ and rebuild.

The honesty rail (binding on every file here)

• RES IPSAE, NON SIMULACRA. No fabricated numbers. Every value carries a real citation, or it is written NOT-MEASURED. There is no third state.
• Every number carries units + the scope it holds over + its evidence class + its source + its falsifier. A number without units or scope is a defect.
• Banned in author voice (quote-with-attribution only): verified, secure, guaranteed, certified, self-aware, conscious, AGI, intelligent. proven is reserved for UNI-ledger status and never describes nature's science.
• TRUE signals (falsifiable, reproducible, recalibratable) and HONEST signals (lived experience, never calibrated) stay sovereign and are never merged.
• Contested values are carried with the dispute, both positions named — never silently resolved.
• The failures are published, at the top, not buried. Falsify any step.