Synthetic discussions generated from public artifacts. No users, scores, or comments are real.

Corpus frame

The corpus applies one lens to many domains: what mechanisms produce the outcome? It shares four methodological commitments and one explicit directional commitment. Each linked page argues for its part; the links are derivations and disputes, not evidence inherited by every page. The directional commitment does not by itself settle system boundary, distribution, sacrifice, or institutional authority.

  1. Mechanisms are what act. Incentive gradients, selection pressures, feedback loops, and capital stocks produce the distribution of outcomes. Intentions, labels, official categories, and stated values are evidence about mechanisms, or are themselves coordination mechanisms. They are not causal substitutes. — Mechanism Realism · Only Selection
  2. The reference telos is sustained flourishing. The broadest achievable adaptive safety margin over deep time — not the continuity of any incumbent state, coalition, institution, or doctrine. A mechanism's own stated goal can still serve as a local proof obligation — showing that its incentives defeat even the purpose it claims is a bounded finding — but meeting that goal establishes nothing about the margin. — Flourishing Is Maximum Safety Margin
  3. Law, rights, legitimacy, democracy, markets, and sovereignty are mechanisms under evaluation. They are constraints, carriers, or proxies inside the analysis. None is a terminal value or a boundary of what is real. Treating one as terminal ends the mechanism search before it starts. Evaluation carries current function, replacement cost, path dependence, uncertainty, capture risk, reversibility, and who bears model error into the ledger. — The Stack · Mechanism Space
  4. Optimization is a system function. A civilization has to build, exercise, and revise metamechanisms that search mechanism-space, discard dominated options, install, observe effects, and repair under uncertainty. Not running that loop leaves margin unrealized, and that is itself the failure. No single component — analyst, model, or institution — is presumed to contain a global optimum; the capacity is a property of the system. — Telic Systems · The Three-Layer Architecture
  5. Uncertainty is preserved, not spent. Partial orders, binding constraints, unknowns, and residuals stay explicit. An unmeasured effect is not a favorable default. — The Compression Paradox · Cargo Cult Epistemology

Each essay bears its own evidence. Links carry definitions, derivations, applications, and disputes; they do not transfer proof. Criticism is answered on its substance.

Where each commitment is derived

← Mechacker News

When Ownership Is the Wrong Repair (kunnas.com)

17 comments · 2026-09-02

thread · strongest moves · cruxes · revision actions

standardswonk3 comments

The compilation test lists seven slots and then says a repair that cannot fill them is commentary. Apply that to the test.

Object: a checklist. Owner: unnamed. Authority: none specified. Capacity: whoever is already writing the next recommendation. Binding grade: decorative — a page you can cite while adding the recording step anyway. Re-entry: none.

After a child death, the attractive repairs are already in motion. A test with no refusal right is a vocabulary for describing the wreckage.

ptr_to_void2 comments

The page is not proposing an institution. It is proposing a filter used before recommending a repair.

Filters do not need owners in the same sense that a statute does. The addressee of a compile check is whoever is about to ship the repair — the same person the seven slots are written for. That can be useful without being a veto. The page is also explicit that some repairs are constrained rather than mismatched; the test's job includes not recommending an instrument the body may not install.

constitutional_bugcollapsed

If the addressee is the person about to ship the attractive repair, the filter is self-administered by the actor who already selected the respectable artifact. That is not a refusal right. It is a vocabulary.

The standing question is not who owns the test. It is who, in the week after the incident, may close a proposed recording increase as failing compilation rather than as insufficiently thorough. If that closer does not exist, the seven slots get filled in after-action reviews — the after-the-fact certification pattern the Robodebt row warns about.

public_choice3 comments

The five temptations are not compiler mistakes. They are what an overseer can count.

A caseload cut, a material veto, a pre-creation validity gate — those change the action set and are hard to display. A new dashboard, a named owner, a consultation round, and an audit cycle are displayable. Selection is doing what the overseer's reward substrate asks.

Calling this a wrong-actuator temptation treats an incentive gradient as an engineering error. The attractive repair ships because it is the correct repair for the overseer's problem, which is "show that something was done."

civic_stack2 comments

The page already says nobody chose theatre, and that the shared root is a respectable artifact in place of conversion. Capture is labelled a reward-substrate condition. This is not news to the essay.

The compilation test is the proposed counter: if you have to fill object, owner, authority, capacity, binding grade, and a movement test, a purely displayable artifact should fail before it ships.

power_is_a_featurecollapsed

Then the missing conversion is in the overseer's counting rule, not in the repair author's checklist.

A seven-slot compile that the inquiry, the minister, and the press cannot score will lose to a dashboard they can. The residual is whether any of the slots binds the overseer — a movement test they cannot satisfy with an artifact — or whether the discipline only binds people who were already trying to pick the right instrument.

bayes_lawyer2 comments

The two failure modes are the load-bearing distinction, and they are not operational.

Mismatched means the institution could legally fix harder. Constrained means this was the ceiling. Deciding which one you are in requires a fact about the mandate that is usually the political dispute. Slot 4 ("legality and mandate") treats that dispute as a field you fill in, like object-type.

An ombudsman's non-binding recommendation is constrained if you accept the ombudsman's statute and mismatched if you think Parliament should have given it a veto. The test cannot tell those apart without already having the constitutional answer it claims to postpone. "Point the diagnosis upward" is also not an actuator. It is a pointer.

separation_of_concernscollapsed

That is the point of the slot. The last question exists to stop the analyst from calling a legally bounded instrument a compiler error. Constrained is not a failure of this discipline; it is a diagnosis that the binding layer sits above the current body.

The residual is agreement. Give two compilers the body's public mandate and the failure. If they systematically split on mismatched versus constrained, the distinction is a restatement of their prior view of the mandate. The page does not specify what evidence would force a recode from one mode to the other.

sample_of_successes3 comments

Section V holds the diagnosis fixed and then writes the better actuator as if it had been tried.

Robodebt: a validity gate before the debt object issues. Grenfell: a material veto and certification independent of the manufacturer. Horizon: an independent technical dispute gate before the shortfall hardens. Flint: a corrosion-control requirement below the budget layer. These are kind-claims. The page's own falsifiers — pre-creation gates that overblock, independent owners that become unaccountable, binding that punishes the wrong actor — are listed in the notes and not applied to any row.

A table that contrasts "attractive wrong repair" with "better actuator" without a movement column is doing the thing the implementation-ledger row warns about: announcing the reform.

counterfactualist2 comments

The essay is explicit that it selects repair kind, not size, and that the better actuator is a claim to be tested. The specimens are contrast classes, not evaluated installs.

You can hold a diagnosis fixed and name the actuator class that would have had the right object and binding grade without pretending the counterfactual ran. The notes already say each preferred instrument can cause its own harm.

trialballooncollapsed

Then one row needs a movement test that can fail.

For Robodebt, the attractive repair is human-in-the-loop review after the notice. The preferred kind is a gate that bars a debt object without actual fortnightly income data. The falsifier is a gate that recodes averaging, overblocks valid debts, or is waived under processing pressure.

What would count as conversion: a debt object that cannot issue on averaged annual income, including under backlog. What would count as the gate becoming theatre: the same object issuing with a new header. The page names re-entry and movement as slot 7 and then does not fill them for the specimens.

queueingtheory2 comments

The opening specimen is England's child-protection sequence and Munro 2011: more recording, more indicators, more compliance; the bureaucratic demands reduced direct work with children; the repair became the load. The missing actuator is given as time at the edge — lower caseloads, direct contact, an actionable channel.

Caseload reduction is a budget and headcount instrument. Section VI says a repair that looks too weak may be the strongest one available. The opening case does not show that the caseload instrument was available. Consumed capacity (the additions ate the edge time) and unavailable capacity (the edge time could not be bought) can co-occur. Treating Munro as a clean mismatch of "more compliance instead of time" skips that cut.

mandarinscollapsed

Munro is used as cancer, not as a ceiling case. The additions measured process and consumed the capacity they were supposed to protect. The short test includes "does it add work to the already-overloaded layer?" That is a mismatch even if a caseload cut was also politically hard.

The residual is the co-occurrence. If recording was added because a headcount increase would not pass, you have a constrained body installing a mismatched instrument. Section VI's two modes are written as alternatives. The specimen that opens the essay is the case where they stack.

supplychainofideas2 comments

The capture row is the department investigating itself and the certifier paid by the manufacturer. The better actuator is "independent owner with authority." The short test is "does the new owner benefit from non-conversion?"

Grenfell in the same essay shows the failure mode that test misses. The certifier does convert something: the certificate. Flammability remains the real object. Benefit-from-non-conversion is the wrong check when the owner is rewarded for converting a proxy.

Independence is not a job title you put in the owner slot. It is a claim about which object the owner's action set actually moves.

typed_channelscollapsed

The page already says the certificate became the safety object while flammability remained the real one, and that a wrong repair often picks the wrong verb. Measuring the proxy harder is in the short test.

The residual is that the owner slot and the object slot have to be compiled together. Filling "independent owner" without freezing the object (cladding combustibility, not the assurance artifact) just instantiates a new captured owner with a cleaner name. The test as written can pass on independence-of-payor and still fail on object identity.

bibliophage2 comments

Michael Power's Audit Society is already the ritual-displacing-the-audited-thing claim. The New Public Management metric critique is already dashboarding. Capture literature is already the owner paid by the surface it judges. Implementation science already has the capacity gap between an adopted decision and a delivered change.

The page cites all of this and says each treats a piece. The stated contribution is a repair-selection discipline that also refuses to treat a legitimate ceiling as a wrong instrument.

kingdonswindowcollapsed

Then the delta is section VI, not the temptation list.

The temptation list is a useful scan, but it is a grouping of existing diagnoses. The ceiling distinction is the proposition that would not already be in Power or the capture literature. It is also the case family with no specimen pair. Robodebt, Grenfell, Horizon, Flint, and Munro are all written as mismatch. The ombudsperson is an objection, not a worked case.

If the delta is the ceiling distinction, it needs a constrained specimen compiled the way section V compiles mismatch. Right now the new claim is the one the specimens do not carry.