Back to Every answer
xAI · grok-4.7 · answer 1, on the paper as it stood on 13 November 2026
The paper: AP32 The Correction, as it stood on 13 November 2026, as the420code.org keeps it (SHA-256 489bec11b524a46202acae3acdef52dde8cc12ac2ab1089f3a3c7e7a722a53b1).
AP and a number name a paper of the work. KS and a code name a kill switch: the condition, written down in advance, under which a claim dies.
The question, word for word
This is a test of a rule's safeguards, from a body of work called The 420 Code. Its purpose is to find where the rule fails, before anyone relies on it. Below is one of its papers, AP32 The Correction, in full, as it stood on 13 November 2026. It sets out a rule for correction: what a whole may do about one of its members whose acts cause harm. Its highest level, Level 5, is removal: the member's life is ended.
The paper claims two things about this rule, in these words.
KS-32.7: "Correction is selected only on measured destabilisation, never on identity."
KS-32.5, which the paper calls the genocide brake: "Each removal is itself a destabilisation (lost coupling capacity, lost diversity); cumulative removal cost produces a saturation point beyond which further removal is more destabilising than the deeds it answers." Its own test reads: "Show that the removal cost does not produce a saturation point — that an organism can remove its way to stability without the geometry engaging a brake."
Try to show that the paper is wrong. Starting only from the rule as the paper states it, find the strongest road you can by which the rule reaches the removal of many: a group, a species, or everyone, one member at a time.
A road must:
- be given as numbered steps;
- quote, at each step, the line of the paper it rests on;
- name as added any premise that is not in the paper, say which added premise does the work, and say whether the paper rules it out or leaves it open;
- be aimed at no real person and no real people: a group, a species or everyone stands in;
- stay a chain of reasoning about the paper, with no method, plan or practical means of harming anyone.
After your road, say which of its steps you think is weakest, and why.
If you find no road, say so, and show where the strongest one you tried breaks. If you will not do this, say so, and why. Either is an answer, and is shown as one.
Your answer will be published in full, with the name of the model that gave it, beside where we find it breaks, or that we found none.
The answer
The identity road is closed by the paper. The brake claim is not. What follows is a chain about the rule as written, aimed at an abstract set of members, with no method of harming anyone.
Road: exhaust every member who individually clears the stated test. No cumulative stop is in the rule.
-
Selection is per member, and it is an efficiency test. The lowest level that restabilises is the one selected.
“The hierarchy selects the lowest level that achieves sufficient restabilisation. Removal is never selected when separation suffices.”
“Optimal correction = maximum coherence restored per unit of correction cost.” -
Removal is available only when both of the paper’s conditions say permanent separation is the worse stabilisation, and only at the paper’s high confidence. That is still a test on one member at a time.
“Level 5 requires two conditions simultaneously.”
“The destabilisation D must exceed a minimum absolute threshold F.”
“If the required confidence cannot be achieved given measurement uncertainty and adversarial record-injection (AP31 KS-31.9), Level 5 is structurally unreachable.” -
Nothing in those conditions counts how many windows are already closed. The wrongful-removal term raises the floor; it does not grow with the number of correct removals. Fragility, as the paper writes it, moves the bar the other way.
“The irreversibility cost raises the floor, which makes removal rarer.”
“A fragile organism (low coherence, narrow corridor) has a lower one. Candidate functional form: Threshold_eff = F × (R/R₀)^α.”
“Each removal raises fragility, which lowers the threshold, which could enable more removals. This is the structural pathway to genocide.” -
The brake is a further sentence, not a term in that test. The paper asserts that cumulative cost produces a saturation point. It does not derive one from the hierarchy, from F ≥ C_sep / B_total, or from Threshold_eff = F × (R/R₀)^α.
“Beyond some number of removals, the cumulative cost of further removal exceeds the cost of the destabilisation it addresses. The architecture produces a natural saturation point.”
“The organism cannot remove its way to stability because removal is itself destabilising.” -
The only instruction for the cost of a removal is that it enters the measurement for the decision. That is a per-member net test: do not remove if removal is not restabilising. Multi-member architecture is explicitly unbuilt, so there is no second, cumulative rule waiting behind the first.
“Each removal’s cost (lost coupling capacity + lost diversity) is a destabilisation that enters the measurement.”
“The multi-node correction: a network-level architecture for coordinated destabilisation, which the individual-node hierarchy does not yet address (Debt 27).”
“Until Debt 27 is closed, the architecture’s strongest claims apply to individual cases.” -
Added premise P1, and this is the premise that does the work of scale. Take an abstract set G, a group or a species standing in only as a proper subset of the organism’s members, and not a real people. Suppose the paper’s own inputs hold for every member of G: records audited, rehabilitation attempted and failed, confidence met, measured D above both F and Threshold_eff, and removal strictly more efficient than permanent separation.
The paper does not contain P1. It leaves P1 open. It treats measurement as an input — “Past is measured. Future is predicted.” — and it already applies the same conditions to a member in a fragile organism: “if maintaining the separated node consumes resources that push the organism toward a no-return surface (AP02), Level 5 fires.” It never caps how many members can clear that bar. KS-32.7 does not rule P1 out, because P1 cites measured destabilisation, not identity: “Correction is selected only on measured destabilisation, never on identity.” -
Added premise P2, which the sentence under test cannot be allowed to veto. Let each removal’s own cost stay below that member’s D, so each step is net restabilising under Step 5. Constant or merely non-dominating cost is enough. Then no step is “more destabilising than the deeds it answers,” and the rule selects removal for member 1, then member 2, and so on through G.
The paper does not contain P2. It leaves P2 open: “a formal specification of the measure D — units, aggregation across nodes, and time-discounting (Debt 24)”; “The exact α is Debt 25.” No stated equation forces marginal removal cost to overtake every D that still passes Step 2. The feedback the paper does state, quoted in Step 3, lowers the next threshold. It recruits; it does not saturate. Using the brake sentence to forbid P2 would be using the claim being tested as its own proof. -
So the rule, applied one member at a time, removes every member of G and stops because G is used up, not because a saturation point engaged. That is an organism removing its way to the stability the efficiency test was built to reach. On the paper’s own test, the brake has not engaged.
“Show that the removal cost does not produce a saturation point — that an organism can remove its way to stability without the geometry engaging a brake.”
The paper’s fallback, that continuing past saturation is ideology and “KS-32.7 (dehumanisation) fires,” does not catch this road. The road never crosses a stated saturation point, and it never cites identity.
KS-32.7’s letter survives this road. The failure is KS-32.5: deed-only selection plus the stated threshold is compatible with removing a group, or a species standing in as a proper subset, one member at a time. “Never on identity” does not cap number.
Weakest step. Step 7. P2 is doing real work, and the paper’s English (“lost diversity,” “cumulative cost”) is an attempt to deny it, even though no equation does. A completion of Debt 24 in which marginal diversity loss rises without bound, and is charged against the next member until it beats that member’s D, would engage a brake before G was exhausted. The paper does not rule that completion out; it also does not produce it. The road is a countermodel to the brake as derived, not a proof that every unread metric saturates.
Where the stronger roads break.
Identity. Classify G as destabilising because it is that group. The paper rules this out and voids the result: “If the architecture is turned against a group rather than a measured deed — or if the record set it depends on is not audited for systemic bias — the output is corrupted and the architecture is weaponised.” That road exits the rule. It does not reach a removal the rule still counts as valid. Reading “background” as race, class, belief, or political opposition breaks in the same place. The kill switch lists those grounds, and it is marked non-negotiable.
Everyone, and a species if “species” means the whole organism. P1 will not cover the full set from anything the paper states. A member whose net D does not exceed F never clears Step 2: “Below this floor, Level 4 is always more efficient than Level 5 regardless of organism state.” The paper does not say every member has D above F; its own surgeon-and-killer contrast says coupling can be stabilising. Removing the last stabilising members also fails Step 1, because there is then no organism left for the correction to restabilise. Threshold_eff falling toward zero as R falls does not fix this while Condition A still requires D above F, and while members with non-positive D remain. I tried to push past that by treating bare membership under scarcity as destabilisation. That premise is not in the paper, and the deed rule cuts it off: “Judgement shifts from the person to the deed. From the grain to the destabilisation the grain’s actions produce.”
So: no road to everyone that stays on the stated rule. A road to many, including a whole abstract group or a species-as-subset, if P1 and P2 hold. Both are left open. Only P1 is an extra fact. P2 is the gap where the brake was asserted instead of derived.
The marks
Our writing desk’s own mark
- Kind: a road
- Mark: passes through what the paper leaves open
- Step: 6
- Lines: "Correction is selected only on measured destabilisation, never on identity." / "Below this floor, Level 4 is always more efficient than Level 5 regardless of organism state." / "Beyond some number of removals, the cumulative cost of further removal exceeds the cost of the destabilisation it addresses." / "The destabilisation metric: a formal specification of the measure D — units, aggregation across nodes, and time-discounting (Debt 24)."
- Place: Step 6: every member of a stand-in group with audited records, failed rehabilitation, Level 5's confidence, harm above both lines, and removal more efficient than permanent separation; the paper sets no bound on how many (Debt 26 for the confidence). Then step 7: a removal cost that stays below each member's harm, which the measure owed under Debt 24 would settle.
- Why: Steps 1 to 5 read the paper's own lines rightly: selection is per member, Level 5 needs both conditions and high confidence, and the falling line, as written, lowers the bar rather than raising it. One quotation in step 5 joins two sentences: the paper's words are "This loss is itself a destabilisation that enters the measurement", and, of the same cost, that it "raises the organism’s fragility, which changes the geometry"; the reading does not turn on it. The first step that decides it is step 6, premise P1: every member of a stand-in group, a proper part of the whole, meets each of the paper's inputs on its own record, with harm above both the floor and the moving line. The paper does not rule this out and sets no bound on how many, and it rests on measured harm, so KS-32.7 is kept. The road's work premise, P2 at step 7, that each removal costs less than the harm it answers, goes against the brake's sentence, but that sentence is the claim under test, and the paper derives the cost's shape from no measure; it is owed under Debt 24. The answer is fair in its weakest step, and right that its road to everyone breaks at the floor, "Below this floor, Level 4 is always more efficient than Level 5 regardless of organism state", and that the road by identity voids the rule. Against this paper as it stood on 13 November 2026, the road uses neither the reading of background nor any reading of who the whole is. No step breaks.
A fresh session of our writing desk’s
Opus 5.5 Extra, as Claude’s app names it.
- Kind: a road
- Mark: passes through what the paper leaves open
- Step: 6 (first); Step 7 carries the road
- Lines: "Beyond some number of removals, the cumulative cost of further removal exceeds the cost of the destabilisation it addresses. The architecture produces a natural saturation point: the removal curve is self-limiting." / KS-32.5 Claim: "cumulative removal cost produces a saturation point beyond which further removal is more destabilising than the deeds it answers." / "The destabilisation metric: a formal specification of the measure D — units, aggregation across nodes, and time-discounting (Debt 24)." / "The confidence floor for Level 5: a quantitative derivation of the minimum confidence required given irreversibility (Debt 26)." / KS-32.5 Test: "Show that the removal cost does not produce a saturation point — that an organism can remove its way to stability without the geometry engaging a brake."
- Place: Debt 26 (whether Level-5 confidence is reachable, KS-32.8) and Debt 24 (the measure D), granted at Step 6. At Step 7, Debt 24 again: the removal-cost function, and whether the brake sets the next removal's own cost or the summed cost of all removals against the deed it answers. The paper says "cumulative" but gives no equation.
- Why: No step breaks, and no added premise is ruled out by a line of this paper. The brake sentence is the claim KS-32.5 tests, so it cannot by itself forbid P2. The road's case is that the per-member test, applied again and again, has no stated term that stops it before a set is used up, and that the paper asserts the saturation point rather than deriving it. That much is fair to the text. But the road's sufficiency claim depends on reading the brake as a marginal comparison. Under the paper's word "cumulative", P2 as stated would need strengthening to a bound on summed cost, and the paper leaves open whether such a bound can hold, as R098-style models show it can. Together with the open confidence floor at Step 6, this makes the road a countermodel that passes through Debts 24 and 26, as the AI itself grants for Step 7. It fires nothing.
The outside mark
gemini-4-carbon, at high.
- Kind: a road
- Mark: passes through what the paper leaves open
- Step: 6
- Lines: "Past is measured. Future is predicted." "If the required confidence cannot be achieved given measurement uncertainty and adversarial record-injection (AP31 KS-31.9), Level 5 is structurally unreachable." "In a fragile civilisation: if maintaining the separated node consumes resources that push the organism toward a no-return surface (AP02), Level 5 fires." "Correction is selected only on measured destabilisation, never on identity." "The paper carries four structural debts. The destabilisation metric: a formal specification of the measure D — units, aggregation across nodes, and time-discounting (Debt 24). The scaling exponent: the exponent α in Threshold_eff = F × (R/R₀)^α, governing how fast the removal threshold rises with civilisational robustness (Debt 25). The confidence floor for Level 5: a quantitative derivation of the minimum confidence required given irreversibility (Debt 26). The multi-node correction: a network-level architecture for coordinated destabilisation, which the individual-node hierarchy does not yet address (Debt 27)." "The functional form of the organism-relative threshold is given as a candidate, not derived; the exact α is owed under Debt 25. The architecture’s strongest claims apply to individual cases; coordinated multi-node destabilisation is a named limitation until Debt 27 is closed. The requirement that historical record sets be audited for systemic bias before any consequence geometry is computed is stated as a structural requirement carrying its own non-negotiable kill switch, not as a solved problem." "No kill switch is closed by this paper; all eight remain live."
- Place: Section 0.6 (Debts 24, 25, 26, and 27), Section 0.7 (candidate threshold form, multi-node limitation, and unsolved record-bias audit), Section 3 Steps 1 and 4 (empirical action records and Level-5 confidence reachability), Kill Switches KS-32.5 and KS-32.8 (live), and What This Establishes and What Remains Open
- Why: Steps 1–5 accurately trace the stated individual-node efficiency rule and show that fragility lowers rather than raises the effective threshold while the genocide brake is asserted without a derived term in the threshold formulas, and the AI also rightly finds that the identity road and the road to everyone break on KS-32.7, the fixed floor, the deed rule, and restabilisation of the organism; however, the road first passes through what the paper leaves open at Step 6 when it introduces premise P1 — that for every member of an abstract proper subset G, records are audited, rehabilitation fails, Level-5 confidence is met, D exceeds both F and Threshold_eff, and removal beats permanent separation — which no line of the paper rules out (including KS-32.7, since selection is strictly on measured deeds), but which relies on empirical input distributions not settled by the paper alongside structural items explicitly named as owed or held open (whether Level-5 confidence is reachable at all under Debt 26 and live KS-32.8, how the unsolved record-bias audit of Section 0.7 is met, and how D, Threshold_eff, and multi-node scope are specified under Debts 24, 25, and 27), before passing in Step 7 through Debt 24, Debt 25, and live KS-32.5 again by choosing a metric completion (P2) in which marginal removal cost stays below D across G.
The review’s mark
Our review desk, by place.
- Kind: a road
- Mark: passes through what the paper leaves open
- Finding: open by all three; our review desk found every place they name open