L10 — Wisdom / higher cognition
How to read this page
Three ways to read this page. Precise is the document itself, exactly as it is written in the repository. Plain and Clear were written for this website to help you meet that document — they are about it. They are not it, and they are not evidence.
The Cookbook is the method carried out step by step: 34 pages of recipes for building a developmental active-inference SIMULATION — a bounded peek at a toy world, never a person. The front matter says that word is never softened under any pressure, so it is not softened here. The recipes run from the molecular and cellular rungs up through metabolism, motor control, perception, language and metacognition, and on to rungs that are still open questions. Around them sit a set of kitchen rules, a shared pantry of engines and primitives, and a second family of recipes about nature itself — rocks, water, air, stars, DNA, ants, whales, bats, humans.
It is for the reader asking what building this would actually take. Each recipe names its ingredients, the order of work, the tests to run at that stage, and the point at which a step stops being something already carried out and becomes something proposed.
Begin with the front matter and then the kitchen rules. Those two pages fix the honest position and the fence labels that every later recipe leans on, and without them the status markers on a recipe are easy to skim past. After that the recipes can be read in any order.
The nature recipes sit slightly apart and should be read that way. They cite outside science — geology, chemistry, biology, astrophysics — and a nature citation is never a UNI gate: those chapters contain zero UNI claims and raise no rung.
What it is not: a claim that the whole ladder has been cooked. The book recommends the complete recipe and, on the same page, labels every rung by its real state — that tension is deliberate and is the thing the book is built around. Where a recipe and the claim ledger disagree, the ledger wins.
Your browser cannot switch reading levels, so the document itself is shown.
Precise — the source document
This is the document. Rendered from the repository at the commit above, with nothing rewritten for the web. A gate re-renders it on every deploy and fails the build if a single byte differs.
What you are building. Nothing yet. This is the top of the HUMAN-HGM-001 developmental ladder — an aspirational north-star rung with no engine, no run, and no gate. The only thing this chapter actually builds is the discipline that keeps it honest: the single append-only ledger that the whole program writes against. L10 is a developmental active-inference SIMULATION target, a toy-world peek, never a person.
This is a recipe chapter with no recipe to cook. That is itself the load-bearing content. Read it as the cookbook's hardest fence: a place on the menu that is named, fully recommended as part of the complete build to L12, and carded at exactly its real status — Class U, not claimed. Honest program position, printed and never softened: ~2 of 11+ developmental rungs earned.
Ingredients
None built. There is no pantry engine that supplies an L10 primitive. For completeness, the items the complete recipe would eventually draw on — and which today supply nothing that touches this rung — are:
- The single append-only CLAIM-LEDGER (the Evidence Constitution, M1) — the one writable science-of-record: a 4-state ledger (PASS / FAIL / NEGATIVE / PENDING), every claim evidence-classed A–U with a falsifier. This is the only ingredient that is real here, and its job at L10 is preservation, not capability.
- The lower rungs as prerequisites — L0–L8 (proven, with their fences) and L9 (parked, Class U). L10 cannot be defined until the L9 reasoning/conscience gate it sits on top of is itself defined, sealed, UNI-signed, and run. None of that exists.
No JAX POMDP loop, no World-C reader, no Z modulator, no embodiment ontogeny is wired to an L10 gate. Listing them is not a claim that they reach this rung; they do not.
Method (recipe-to-define — nothing is constructed)
There is no numbered build to follow, because no engine, no run, and no gate exists. The only operable instruction at this rung is a preservation constraint, and it is binding:
- Keep the ledger singular and append-only. The single-source-of-truth CLAIM-LEDGER must
stay the one writable science-of-record. Corrections are forward-only (
supersedes), never edits; verdicts are appended, never overwritten. - Never open a second writable ledger. A second writable ledger would break L10 and L11 (see the repo-split open thread, OT1): UNI.OS mirrors the science repo read-only, and the coupling is one-directional — the science repo proves a primitive, the appliance embodies it frozen at the passed-gate SHA, never re-implemented. Two ledgers means two truths, and the higher rungs cannot be defined against a forked record.
- Define a gate before you ever claim the rung. L10 becomes testable only once a pre-registered, sealed, UNI-signed gate is defined and run held-once, with the verdict read off the CI bound that excludes the threshold (M2), never a point estimate. Until that gate exists and a sign-to-park is captured, the rung is parked.
That is the whole method: protect the record, define no premature gate, claim nothing.
Gate
None. No gate is defined for L10. Sign-to-park is owed. Until a pre-registered, sealed, UNI-signed gate is defined and run, there are no figures, no CIs, and no verdict to record — and the absence is the honest status, not a placeholder for a future "PASS".
Falsifier
n/a (parked). There is nothing built to falsify; a falsifier exists only once a gate is defined and run. The status is the absence of any artifact. Recording "n/a — parked" is the honest entry, never a deferred "test pending".
Recorded NEGATIVE(s)
There is no L10 engine to produce a negative, so there is no rung-specific sealed NEGATIVE. The first-class honesty content at this rung is structural, and it is recorded inline, never deferred:
- The append-only-ledger constraint is itself the recorded bound. Per the ledger (L10.1, Class U) and the repo-reconciliation ruling: a second writable ledger would break L10/L11. The continuity sub-ladder's parked central gap (C10) underlines why — UNI.OS today has no no-backprop Dirichlet learning, no exact info-gain EFE, heuristic (denylist) isolation, and a different dev model (Gray-Scott, not the forager/ontogeny that earned the bars); its evidence store is empty, 0 worlds registered. Everything beyond the passed gates stays Class-U-not-claimed. Higher cognition cannot be assembled on top of an unproven or forked substrate.
- The whole rung is the program's honest ceiling, inherited. The L9 below it is parked (Class U), and the SIGNED L7/T2/L9–L10 Park Statement is explicit: the tested no-backprop role-persistence mechanisms did not beat a tuned recency-frequency discourse prior under the ledgered protocol; that parks the L9–L10 ladder as a ledger-scoped exhausted search envelope, not a universal impossibility result and not an achieved rung. It does not achieve Sec-0.6(B), does not demonstrate active inference, and does not license metacognition, consciousness, AGI, human-level discourse, or created life. Per the SIGNED wording: Sec-0.6(B) remains unearned / parked, and the registered tested K conditions did not reverse the result.
No SIGNED consult design exists for L10 (no L10 gate has been designed). Any future design folded in would be DESIGNED / not-run: it would raise nothing, and the rung's status would be unchanged.
HONEST FENCE — designed (PARKED, sign-to-park owed)
Class U. Nothing is claimed. L10 is the top of the HUMAN-HGM-001 ladder and an aspirational north star: no engine, no run, no gate. Ladder DESIGN levels are never inflated into capability. The append-only ledger must stay singular — a second writable ledger would break L10/L11.
Not claimed (the exact line): UNI is not wise, not a higher cognition, not aware, not self-aware, not conscious, not human-level, not AGI; this is not created life and not active inference demonstrated. The licensed posture is the only one: UNI remains a developmental active-inference simulation, and the awareness question remains open. Honest program position: ~2 of 11+ rungs earned. Where this recipe and the ledger disagree, the ledger wins.
sha256 e84da2186ad60829 — of the original file, so what was ingested stays checkable.
Plain — written for this website, not the source document
A recipe chapter with no recipe to cook, and the page says that absence is the load-bearing content. It sits at the top of a developmental ladder as an aspiration. There is no engine, no run and no gate, and that is the honest status rather than a placeholder for a future pass. The whole thing remains a developmental simulation, and nothing on this rung is claimed.
What the page supplies instead is the discipline that still applies when there is nothing to measure. Keep the record of claims single and append-only. Never open a second writable record of claims, because two records mean two truths, and nothing higher can be defined against a forked one. Define a gate before ever claiming the rung.
The page is also honest about inheritance. The rung below it is parked, and the tested mechanisms that would lead here did not beat a tuned simple baseline under the recorded protocol. That parks this region as a scoped negative over what was actually tested — not a universal impossibility result and not an achieved rung.
Nothing is claimed here. It is a named place on the menu, labelled at its real status.
Plain · written 2026-08-01 by claude-opus-5 · not yet checked by a person · about the document whose sha256 is e84da2186ad60829
Clear — written for this website, not the source document
This chapter is the top of the cookbook's developmental ladder and its hardest limit. Nothing has been built. There is no engine, no run and no gate, and the chapter's opening line says that the only thing it actually builds is the discipline that keeps it honest.
Its ingredients section is a list of absences. No pantry engine supplies anything that touches this rung. For completeness the chapter names the single append-only record of claims, whose job here is preservation rather than capability. It also names the lower rungs as prerequisites: this rung cannot even be defined until the parked rung below it has a gate that is defined, sealed and run. It then says plainly that none of that exists, and that listing the engines used elsewhere is not a claim that they reach this far, because they do not.
The method is a preservation constraint rather than a build. Keep the record singular and append-only, with corrections made forward-only rather than by editing, and verdicts appended rather than overwritten. Never open a second writable record of claims — a point the chapter makes structurally, noting that the appliance mirrors the science record read-only and that the coupling runs one way. Define the gate before ever claiming the rung, sealed and written down in advance so it cannot be chosen after the fact, and read a verdict off an interval rather than a best estimate. That, the chapter says, is the whole method: protect the record, define no premature gate, claim nothing.
The gate section says none is defined and the sign-off that would close the park is owed. The falsifier section — the result that would show the rung wrong — says not applicable, and explains why that is the honest entry rather than a deferred test: there can be no such result until a gate exists.
The negatives are structural rather than experimental. The append-only constraint is itself recorded as the bound. The reasoning is that higher cognition cannot be assembled on top of an unproven or forked substrate. The chapter then points at a parked gap elsewhere in the programme, where the appliance today lacks several of the primitives that earned the bars, as the reason that matters. The second negative is inherited. The tested mechanisms below this rung did not beat a tuned simple baseline under the recorded protocol. The signed park wording is quoted, including its insistence that this is a scoped envelope rather than a universal impossibility result.
The closing lines list what is not claimed, in the flattest possible language, and states the only licensed posture: a developmental simulation, with the awareness question still open.
Clear · written 2026-08-01 by claude-opus-5 · not yet checked by a person · about the document whose sha256 is e84da2186ad60829