Wiki · Evidence & Verdicts
Metabolism activation gate — the energy-posterior receipt Phase-2 lacked (2026-07-11)
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.
Eighty-seven dated pages: receipts, pre-registrations, handoffs, validation records and review verdicts. A receipt is written at the moment a piece of work was checked. It names what was claimed, the commit and the seed, what was actually run, and the outcome in one of a small set of controlled words. Then it names what the work did not achieve. That last part is what makes it a receipt rather than an announcement. A pre-registration is the same discipline run in advance: the conditions that would count as a pass and the conditions that would falsify the claim are written down before the run, so neither can be adjusted once the numbers arrive.
That is why so many small dated stubs are an audit trail rather than noise. No one of them is meant to be a good read. The value is in the sequence and in the dates, because you can watch a prediction be registered, then the run happen, then the verdict land — sometimes against the prediction. Pages here record a falsified result, a rejected fix, a retracted overclaim, and a green receipt that turned out not to be reproducible from the commit that carried it. A record that carried only successes would be worth a good deal less than this one.
A gentle way in is to read a pre-registration first, so the shape becomes familiar, then a result page, then one of the corrections. This section sits off the main navigation on purpose: it is the record you check the rest of the site against, not the place to begin.
What it is not: documentation, and not a summary. Nothing here has been tidied in hindsight. Every entry reads as of its date, a later entry may overturn an earlier one, and the presence of a page is not a claim that its result stood.
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.
✅ SUPERSEDED FORWARD BY
docs/receipts/metabolism_activation_gate_LIVE.md(production PASS). This receipt is the OFFLINE proof: the organ dynamics are correct when driven (pos/neg/neg/pos + reproduction). The live wiring gap that made it inert in Phase-2 has since been remediated (88be5c9,383ffb4,535f9b6,a9b1508) and the live pos/neg/neg/pos gate PASSED in production (2026-07-11, receipt above): 3 agents on mc-server allalive_final=truewith 8–10 energy-posterior reversals; NEG-1 severed-action twin dies @ 25; NEG-2 severed-food dies @ 16. Activation is proven both offline AND in production.
What this closes. Phase-2 §16 / docs/receipts/phase2_metabolism_red.md recorded the metabolism
hypothesis as WITHHELD because organ activation was unverified — "G5b un-passed and no energy-posterior
receipt." This is that receipt: a rigorous positive / negative / negative / positive proof that the
:metabolism organ is mechanistically live.
Harness: runs/metabolism_activation_gate.exs (deterministic; mix run --no-start). Verdict: PASS.
Pre-registered bars (pinned before the run)
- POSITIVE PASS: survive ≥120/150 ticks AND ≥2 energy-posterior direction reversals AND E[bin] range ≥1.0.
- NEGATIVE PASS-as-negative: dies at tick < 60 (no sustained viability).
- G5b margin (acting lifetime − twin lifetime) > 0.
Results
| arm | condition | outcome | energy posterior |
|---|---|---|---|
| POS-1 | fed + acting, seed 7 | survives 150 | 68 reversals, amp 1.94 (depletes ↔ refills) |
| NEG-1 | action-severed twin (forced :noop) | dies @ 25 | monotone drain, no refill |
| NEG-2 | acting but inv.food=0, seed 7 |
dies @ 16 (4 failed eats) | monotone drain to empty |
| POS-2 | fed + acting, seed 42 | survives 150 | 35 reversals, amp 2.55 |
G5b margin = 150 − 25 = 125 > 0.
What it proves (and its bounds)
Viability is action-dependent (NEG-1: no action → death) AND food-dependent (NEG-2: action without a real-food refill → death), while the fed+acting agent sustains via an energy-posterior limit cycle (POS-1) reproducibly across seeds (POS-2). Two structurally-distinct negatives (severed action vs severed food) rule out a decorative edge — the organ implements a genuine food-contingent homeostat that modulates action to keep an internal store viable.
Honesty fence (binding)
- Scope = ACTIVATION / mechanism only. This upgrades P2's activation from WITHHELD to VERIFIED —
OFFLINE. It is a synthetic-world proof; a production (live colony against
mc-server) energy-posterior receipt is the next confirmation. Do not present this as a production result. - It says nothing about G6 (behavioural plateau-break), which FAILED live in Phase-2 and remains open. A live activation ≠ a live behavioural cure.
- Claim fence:
energyis a model variable, never a felt state. This is homeostatic self-maintenance — self-maintenance, never experience / hunger-as-felt / life / awareness.
sha256 c3b6621f4faf0ce0 — of the original file, so what was ingested stays checkable.
Plain — written for this website, not the source document
This is a receipt — a file recording what was run — for a test that came out well, and it is unusually careful about what the good result does not mean. The test asks whether an energy organ actually does anything, using four arms: two where the body can both act and eat, and two where one of those is cut off. The cut-off arms die early. The intact ones last the full run, with their energy estimate rising and falling rather than draining away. A banner at the top says a later live receipt supersedes this one, and a limit at the bottom keeps the claim to mechanism in a synthetic world.
Plain · written 2026-08-01 by claude-opus-5 · not yet checked by a person · about the document whose sha256 is c3b6621f4faf0ce0
Clear — written for this website, not the source document
A receipt, the file recording what was run, with a pass verdict, and a banner at the top saying it has since been superseded forward by a live one. An earlier phase had recorded this claim as withheld, because the organ's activation was unverified, and this page is the missing piece of that.
The design is a positive, negative, negative, positive set of arms, and the bars were pinned before the run. How long the good arms must last, how many reversals in the energy estimate count as a genuine limit cycle, and how early the cut-off arms must fail. A results table then gives one row per arm. Both intact arms last the full run with the energy estimate oscillating between depletion and refill. One cut-off arm cannot act at all and drains steadily to death. The other can act but has no real food, and does the same.
The reasoning is that two structurally different negatives rule out a decorative result. Viability depends on acting, and it also depends on food, and the intact arm sustains itself across more than one seed.
The list of what is not claimed, at the end, is as long as the result, which is the point. The scope is mechanism only, in a synthetic world, and the page says plainly not to present it as a production result. It says the finding is silent about a separate behavioural gate that failed live and remains open. And it fences the language: energy here is a model variable, and this is self-maintenance rather than anything felt.
Clear · written 2026-08-01 by claude-opus-5 · not yet checked by a person · about the document whose sha256 is c3b6621f4faf0ce0