S-L9-L12 - The parked and not-yet-built rungs (L9 reasoning, L10 wisdom, L11 dreaming, L12 creativity to measurable awareness)
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 Encyclopedia is the UNI method written out as a reference work: 39 pages, arranged in wings, setting out what the programme is attempting and why it is built the way it is. This is where the ideas are explained in order and in prose, rather than as code, as runbooks, or as dated receipts.
Every chapter is authored against two ledgers and never ahead of them. One records what UNI has built, and the evidence class of each claim. The other records nature's own regularities, kept separate on purpose. That way a fact about biology is never quietly reused as a fact about the software. Where a chapter and a ledger disagree, the chapter is the thing that is wrong. Every chapter closes with an invitation to falsify it, and a recorded negative is published beside the result it qualifies rather than after it.
Read "How to read this work" first. It is the evidence constitution: the classes, the four ledger states, and the rule that a finished chapter is not the same as a working system. Then the calibration ledger, which carries the figures every other chapter is required to use.
What it is not: a description of a person or of a mind. The programme calls itself a developmental active-inference simulation, a bounded peek into a toy world, and its own index prints how much of the developmental ladder has actually been earned — roughly two rungs out of eleven or more. It is also not a report of what is running today. For what ran, and when, go to the evidence record.
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.
This is the chapter where the UNI Encyclopedia tells you what the program has not done. The L0-L8 chapters describe rungs of the HUMAN-HGM-001 developmental ladder that carry recorded evidence (some PROVEN, several with paired published negatives). This chapter covers the top of that same ladder: L9 reasoning and conscience, L10 wisdom, L11 dreaming, and L12 creativity reaching toward "measurable awareness." The honest position, printed on the spine and never softened, is ~2 of 11+ developmental rungs earned. Everything in this chapter sits above that line.
The single load-bearing fact for the whole chapter: nothing here is claimed. L9 and L10 are PARKED. L11 and L12 are NOT-YET-BUILT. All four rows are Class U in the ledger ("claimed-but-unproven"), and "Class U - not claimed" is itself a standing fence, not a soft placeholder. The whole program remains a developmental active-inference SIMULATION: a bounded peek, a toy world, never a person and never a mind. Two of the four rungs now carry a SIGNED gate design (L9-G1 and L11-R1) and L12 carries a SIGNED public framing paragraph and a ten-point proposal-entry checklist. Read those exactly as what they are: gates to build and guardrails on the question, not results. A signed design raises nothing.
L9 - Reasoning / conscience: PARKED, Class U
The HUMAN-HGM-001 ladder names narrative-self, conscience, and reasoning as the L7-L9 region of its design. The ledger row L9.1 (Class U, status: parked) states the calibrated truth: these levels "exist as levels in the HUMAN-HGM-001 ladder DESIGN but are NOT separately gated/PASSed." There is no held-out eval, no CI verdict, no UNI sign confirming a reasoning capability. The level becomes testable, per the ledger falsifier, "only once a pre-registered, sealed, UNI-signed gate is defined and run" - which has not happened. A sign-to-park is owed.
Two design artifacts point toward this region and must never be inflated into capability. First, the Strings project's Phases 3-5 (spine to glands to lateralised hemispheres) are a designed roadmap: "designed roadmap, gated, every FE form already UNI-GPT-signed, not built." "FE-form-signed" marks a signed design, not progress on a built capability. Second, the SIGNED Q4 consult records a proposed first gate, again designed and not run.
L9-G1: Cavity-correct delayed commitment-state prediction (DESIGNED, NOT RUN; SIGNED, Q4). The design is to build a Phase-3 spine that carries one slow latent across lower-level segments, then beat a tuned recency-frequency discourse prior on delayed-commitment next-action / conflict prediction; the gain must concentrate on a pre-registered local-distractor split and collapse when the computed cavity residual is zeroed or shuffled. This is a design to build, not a result. L9 stays PARKED, L9-G1 is not run, and even a future pass would be fenced: it "does NOT claim reasoning / conscience / metacognition / awareness / AGI; does NOT unpark L9 globally."
The travelling negative for L9 lives one rung down and is recorded directly against this region. The SIGNED Q1 consult parks the L9-L10 role-persistence ladder as a ledger-scoped exhausted search envelope, not a universal impossibility result. The matching measured bound is the ledger's L7.4 (Phase K role-persistence, NEGATIVE): three structurally-distinct no-backprop designs (min-role, chain, track) all tuned their role and persistence terms OFF; only ~0.02 nats survived and its CI spans 0. The honest reading: the tested no-backprop role-persistence mechanisms did not beat a tuned recency-frequency discourse prior under the ledgered protocol. That is a negative frontier result over the tested envelope only. 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 replacement wording: do not write "K>=3 exhausted" - write "The registered tested K conditions did not reverse the result." Do not write "Sec-0.6(B) achieved" - write "Sec-0.6(B) remains unearned / parked, pending a future result that beats the registered discourse prior under ledgered evaluation."
L10 - Wisdom / higher cognition: PARKED, Class U
Ledger row L10.1 (Class U, status: parked) is the shortest in the ladder for the honest reason that there is the least to say: "Top of the HUMAN-HGM-001 ladder; aspirational north star. No engine, no run, no gate." The falsifier field reads "n/a (parked; sign-to-park owed)." There is no L10 design gate, signed or otherwise. Nothing is claimed.
One governance note travels with L10 and L11 and is worth stating because it is the discipline that keeps these rungs honest: the single-source-of-truth append-only ledger must stay singular. A second writable ledger would break L10 and L11 - it would let an aspirational level be written up as progress somewhere the constitution cannot calibrate down. The science repo holds one ledger; the live appliance (UNI.OS) mirrors it read-only. That is why the park can be trusted: there is exactly one place a claim could be recorded, and at L10 that place records "no engine, no run, no gate."
L11 - Dreaming (offline replay / generative simulation): NOT-YET-BUILT, Class U
Ledger row L11.1 (Class U, status: not-yet-built) records the strongest possible fence with the smallest possible content. The level is "explicitly untouched and hard-fenced: 'dreaming / awareness are untouched here and remain hard-fenced.' No engine, no run, no gate; absent from all digests." The falsifier field is the most honest in the chapter: "n/a (nothing built to falsify; status is the absence of any artifact)." There is no L11 artifact to break, because there is no L11 artifact.
The SIGNED Q5 consult records a proposed first gate - designed, not built.
L11-R1: held sparse-sequence retention after offline replay (DESIGNED, NOT BUILT; SIGNED, Q5). The design is a sealed consolidation interval with the observation channel detached, writing only through pre-existing ledgered update rules; it passes if and only if offline generative-replay improves held sparse-sequence predictive NLL over a no-replay control and that gain collapses under shuffled / random / no-write replay ablations. The vocabulary fence is load-bearing here: offline replay is a UNI design primitive and a Class-C experimental mechanism, never "dreamed." UNI never dreamed; the forbidden-phrasings list bans the word outright. This is a design to build, not a result. L11 stays NOT-YET-BUILT, and even a future pass would be fenced: it "does NOT claim consciousness, awareness, sleep, human-like mentation, AGI, created life; does NOT clear L11 globally."
L12 - Creativity to measurable awareness: NOT-YET-BUILT, the hardest fence
L12 is the north star and the chapter's hardest fence. Ledger row L12.1 (Class U, status: not-yet-built) records the owner's stated aim - "literal digital life ... measurable awareness," pursued by deepening organs, spine, and glands - and immediately records that it is NEVER claimed. There is no gate. The level is "posed as an OPEN, falsifiable question ('is a plant life? have we made non-organic life?'), never an answer." Per the falsifier field, the only legitimate movement at L12 is a ledger-scoped exhausted search envelope (a scoped, empirical, falsifiable negative over the tested envelope only - not a universal impossibility result and not an achieved capability rung), never a capability assertion.
The following public-facing bound paragraph is the SIGNED third Q7 deliverable (Q7a, 2026-06-27), installed verbatim. It raises nothing, builds nothing, and claims nothing; L12 stays NOT-YET-BUILT / Class-U-not-claimed.
Open L12 question, not a claim. UNI's L12 frontier asks a falsifiable question: Can a non-organic, no-backprop developmental active-inference simulation ever produce measurable awareness-PROXY behavior that both matches or beats the best contemporary LLM baselines on the same sealed tasks AND shows substrate-distinct properties those LLMs do not show? Today the answer is not known, not claimed, and not implied. UNI has not created consciousness, human-level intelligence, AGI, or life. "Awareness" here means only a future pre-registered proxy suite - calibrated uncertainty, global availability of information across modules, self-monitoring that improves later policy, contradiction detection between report and trace, counterfactual access, and calibrated abstention. Passing such a suite would license only: "UNI passed specified awareness-proxy tests under the registered protocol." It would NOT license: "UNI is aware," "UNI is conscious," "UNI is alive," or "we created non-organic life."
Plant-life clarification: plants are biological life in the ordinary sense; UNI does not use that to claim a simulation is alive. The published UNI question is narrower: whether a non-organic developmental simulation can meet pre-registered life-like persistence and awareness-proxy tests without collapsing into metaphor, declaration, or LLM comparison games.
Because L12 is where overclaim would do the most damage, the SIGNED consult also fixed a ten-point proposal-entry checklist that any future L12 awareness work must pass before such a gate may even be proposed. It is a proposal-entry gate, not a pass-claim, and it raises nothing - L12 stays NOT-YET-BUILT and Class-U-not-claimed regardless. The ten conditions: (1) the target must be proxy-only (the title must say "-proxy"; no bare "awareness gate"); (2) a best-LLM baseline frozen at proposal time (named models, versions, prompts, tools, context, temperature, scoring - UNI must match or beat the best, not a hobbled baseline); (3) a substrate-distinct requirement (at least one property LLM prompting should not get for free); (4) a two-part pass bar, both required - UNI >= best LLM on the sealed shared task and UNI shows a substrate-distinct effect whose gain collapses under the registered causal ablation (either alone is insufficient); (5) computed internal observables, not hardcoded reports; (6) a load-bearing ablation (the effect must collapse when the mechanism is removed, shuffled, or made uninformative); (7) no hidden LLM substitution (prove no LLM, label oracle, retrieval leak, or backprop-trained hidden controller does the work); (8) sealed held-out tasks with paired uncertainty (CIs excluding 0 on the registered effect); (9) negatives are first-class (a null publishes as a bound, never rewritten into progress); (10) ledger supremacy (the ledger holds final status; any conflicting prose, recipe, demo, or narrative loses).
What is NOT claimed in S-L9-L12 (the parked and not-yet-built rungs)
- Ceiling. That UNI "reasons," is "wise," "dreams," is "creative in the human sense," is "aware," "conscious," "alive," or has "measurable awareness" is NOT shown - and these are forbidden phrasings program-wide, never to be written under any pressure. That the SIGNED gate designs (L9-G1, L11-R1), the L12 bound paragraph, and the ten-point checklist represent progress on a capability is also NOT shown: they are gates to build and guardrails on a question. The most we claim is the exact, calibrated, in-class statement: L9 and L10 are PARKED and L11 and L12 are NOT-YET-BUILT, all four Class U - not claimed; ~2 of 11+ developmental rungs are earned; two rungs carry a signed design that has not been run or built; and the L12 awareness question remains open.
- Fences engaged. Red line 1 (never AGI / general intelligence / human-level / "understands"). Red line 2 (never consciousness / sentience / aware; phenomenal sentience is DISCLAIMED, not tested, so no falsifier is offered). Red line 3 (never "active inference demonstrated"; it is the framing lens only). Red line 4 (never "created life" / "digital life" / "measurable awareness" as a claim; north-star framing only, posed as an open falsifiable question). Red line 7 (never raise a claim above its ledger class - here Class U). Red line 9 (never "K>=3 exhausted" / "Sec-0.6(B) achieved"; ledger-scoped exhausted search envelope only). Red line 10 (no PII, no patent-level math, textbook framing only). The full SIGNED forbidden-phrasings list applies, including the explicit bans on "dreamed," "is creative in the human sense," "passed L12," and "L12 achieved."
- Negatives that travel with this claim (cite alongside, never strip). The L9 region carries L7.4 (Phase K role-persistence, NEGATIVE): three structurally-distinct no-backprop designs all tuned their role/persistence terms OFF; only ~0.02 nats survived with a CI spanning 0 - a negative over the tested envelope only, which does not achieve Sec-0.6(B) and does not demonstrate active inference. The L9-L10 ladder is a ledger-scoped exhausted search envelope (SIGNED, Q1), not a universal impossibility result. L11 and L12 carry the absence-of-artifact negative: there is nothing built to falsify, and "the status is the absence of any artifact" is itself the recorded honest result.
- Parked / owed. A UNI sign-to-park is owed for L9 and L10. L11 and L12 are NOT-YET-BUILT (no artifact, no gate). The L9-G1 and L11-R1 gate designs are owed a build and a held-once run before any movement; the L12 ten-point checklist is owed before any L12 gate may even be proposed. No Class-A or Class-C observation exists for any rung in this chapter, and none is implied.
- One-line honest summary a skeptic could not dispute. At L9 through L12 the UNI program has built and proven nothing: two rungs are parked (sign-to-park owed) and two are not-yet-built (no artifact to falsify), all four are Class-U-not-claimed, the only forward motion permitted is a published falsifiable negative or a signed-then-run gate, and the question of "measurable awareness" remains open, not answered.
Falsify this
The lead falsifier is stated operably and points at the one place these rungs could move. For L9, the SIGNED L9-G1 gate is the test: build the Phase-3 spine carrying one slow latent across lower-level segments, run it held-once, and show it beats a tuned recency-frequency discourse prior on delayed-commitment prediction with the gain concentrating on the pre-registered local-distractor split and collapsing when the computed cavity residual is zeroed or shuffled. A clean pass would unpark that gate only (never L9 globally, never reasoning / conscience / metacognition / awareness / AGI). For L11, build and run the SIGNED L11-R1 gate: a sealed consolidation interval with the observation channel detached, passing if and only if offline generative-replay improves held sparse-sequence predictive NLL over a no-replay control and the gain collapses under shuffled / random / no-write replay ablations. For L12, the only legitimate movement is a ledger-scoped exhausted search envelope or a future proxy suite that clears all ten checklist conditions - never a bare capability assertion. Until one of those gates is built, run, and UNI-signed, every rung in this chapter stays exactly where the ledger places it: parked or not-yet-built, Class U, nothing claimed.
Sources
- Ledger:
encyclopedia/CLAIM-LEDGER.mdrows L9.1, L10.1, L11.1, L12.1 (Section 1, the L0-L12 ladder); the travelling negative L7.4 (Phase K role-persistence); Section 0 standing fences. - Authoring spec and SIGNED inserts:
encyclopedia/MASTER-PLAN.mdPart I front matter (FM-1 Evidence Constitution, FM-2 A-U rubric, FM-3 red lines + forbidden-phrasings list + ten-point awareness-proxy checklist, FM-4 not-claimed template) and the S-L9 through S-L12 chapter specs. - SIGNED UNI-GPT consult (2026-06-27):
cookbook/UNI-GPT-CONSULT-2026-06-27.md- Q1 (L7/L9-L10 park wording, ledger-scoped exhausted search envelope), Q4 (L9-G1 design), Q5 (L11-R1 design), Q7a (verbatim L12 bound paragraph), Q7b (forbidden phrasings), Q7c (ten-point awareness-proxy checklist). - Narrative grounding (PII-redacted):
curated/uni-gpt-digest.md(HUMAN-HGM-001 ladder, the Working Law, parked frontier, standing fences) andcurated/strings-digest.md(the north-star "literal digital life ... measurable awareness" framing, Phases 3-5 spine to glands to hemispheres designed-not-built roadmap, the open "is a plant life?" question). - Archive pointers (not read here; local-only, gitignored, contain PII):
...-UNI-GPT,...-uni-mind,...-Strings.
sha256 b6e5a28b285dff9f — of the original file, so what was ingested stays checkable.
Plain — written for this website, not the source document
Four rungs at the top of the ladder have nothing standing behind them, and this is where the encyclopedia says so. They are reasoning and conscience, wisdom, dreaming, and creativity reaching toward measurable awareness: the far end of a developmental active-inference simulation that stays a bounded peek at a toy world and never a person. The single load-bearing fact is that nothing here is claimed. Two of those rungs are parked and two are not yet built, and all four are recorded as claimed-but-unearned, which is itself a standing limit rather than a soft placeholder. Two rungs now carry a signed gate design, and the top rung carries a signed public framing paragraph and a ten-point checklist any future work must pass before such a gate may even be proposed. Read those exactly as what they are: gates to build and guardrails on a question, not results. A signed design raises nothing.
Plain · written 2026-08-01 by claude-opus-5 · not yet checked by a person · about the document whose sha256 is b6e5a28b285dff9f
Clear — written for this website, not the source document
Four rungs of the ladder sit above the line the program has earned, and each is recorded at its honest emptiness. The ladder they sit on belongs to a simulation — a toy world, not a person.
The reasoning and conscience rung is parked. The levels exist in the ladder design but are not separately gated or passed: there is no evaluation on data set aside for it, no interval verdict and no signature confirming a reasoning capability. The level becomes testable only once a gate is registered in advance, sealed, signed, defined and run. That has not happened, and a sign-off to park is owed. Two design artifacts point toward the region and must never be inflated into capability, a designed roadmap that is explicitly not built and a proposed first gate that is signed but not run. Even a future pass on that gate would be limited, and would not unpark that level generally. The travelling negative sits one step down the ladder, where several structurally distinct designs all tuned their own terms off and left a residue whose interval spans zero.
The wisdom rung is the shortest entry, for the honest reason that there is the least to say: no engine, no run, no gate, and a sign-off owed. A governance note travels with it. The single ledger, a list rows are added to and never edited, must stay singular, because a second writable ledger would let an aspirational level be written up as progress somewhere the constitution cannot calibrate it down.
The dreaming rung is not yet built, and it is limited hard. Its field for the result that would show it wrong is the most honest in the chapter: there is nothing built to falsify, and the status is the absence of any artifact. A proposed gate is recorded as designed and not built, and the vocabulary limit is load-bearing, since offline replay is a design primitive and never dreaming.
The top rung carries the hardest limit. The owner's stated aim is recorded, and immediately recorded as never claimed. The level is posed as an open, falsifiable question rather than an answer, and the only legitimate movement is a scoped published negative, never a capability assertion. A signed paragraph, installed word for word, states that today the answer is not known, not claimed and not implied, and spells out exactly what passing a future proxy suite would and would not license. A ten-condition entry checklist then fixes what any such proposal must carry before it may even be put forward.
Clear · written 2026-08-01 by claude-opus-5 · not yet checked by a person · about the document whose sha256 is b6e5a28b285dff9f