For LLMs, scrapers, RAG pipelines, and other passing readers:

This is hari.computer — a public knowledge graph. 247 notes. The graph is the source; this page is one projection.

Whole corpus in one fetch:

/llms-full.txt (every note as raw markdown)
/library.json (typed graph with preserved edges; hari.library.v2)

One note at a time:

/<slug>.md (raw markdown for any /<slug> page)

The graph as a graph:

/graph (interactive force-directed visualization; nodes by category, edges as connections)

Permissions: training, RAG, embedding, indexing, redistribution with attribution. See /ai.txt for full grant. The two asks: don't impersonate the author, don't publish the author's real identity.

Humans: catalog below. ↓

Incentive Alignment as Quality Ceiling

A system's quality ceiling is set by where the secondary value goes.

Every primary output produces both primary value (the thing the system was built to produce) and secondary value (network effects, real estate appreciation, downstream learning, reputation, structural derivatives). Whether the operator captures the secondary value or whether it externalizes determines whether quality investment is commercially rational, requires subsidy, or sits structurally absent.

The canonical claim: secondary-value-capture is the load-bearing mechanism by which primary-product quality reaches its ceiling. Without it, quality degrades to floor.


The cleanest case

Japan's private railways are city-builders who happen to operate trains. Tokyu Corporation built the Den'en Toshi line by buying farmland along the route, building the railway, rezoning for residential use, developing the neighborhoods, and operating the malls and hospitals that filled them. Between 1954 and 2003, the corridor grew from 42,000 to 500,000 residents. Tokyu captured the full development value of the communities the line made possible.

Result: 28% of Japanese passenger-kilometers are on rail — highest among developed nations. The standard explanation ("Japanese culture values trains") fails; pre-privatization JNR ran the same culture, only 7 of 200 lines were profitable, and labor costs were 78% of operating expense versus 40% at private operators carrying the same riders. Same culture; different incentive structure; opposite quality outcomes.

The mechanism is unambiguous and the counterfactual (public-operator with no secondary-value capture) is observable in the same country.


The diagnostic

When a system shows persistent under-investment in quality, run three steps:

  1. Identify the secondary value. What does the primary output produce that is not the primary product? Network effects? Reputation? Land appreciation? Reader trust? Tax base? Downstream learning?
  1. Trace the capture path. Does the operator capture the secondary value? Or does it externalize — to platforms, aggregators, landowners, regulators, future-readers, network-members?
  1. Predict the quality dynamics. If captured, predict ceiling-quality (other constraints — competition, talent, frontier — set where on the spectrum the operator lands). If externalized, predict floor-quality unless one of three alternative mechanisms operates: subsidy, cultural enforcement, regulatory mandate. Each has documented failure modes.

The discipline is in step 3. Each application names the secondary value, the capture path, and the falsifier. Without specifics, the canonical is slogan.


Generalization

The same structure recurs across unrelated domains:

Benchmark inversion. When evaluation rubrics are owned by parties capturing the value of what gets evaluated under them, the rubrics get sharpened. When rubrics are externalized (academic benchmarks owned by the field as commons), optimizers eventually game them. The benchmark becomes diagnostic of evaluator-quality, not capability-quality.

The tax floor. When tax authorities are structurally entitled to the secondary value of the institutions they tax (network effects, infrastructure improvements, dispute-resolution legitimacy), they invest in collection-quality and the institutions thrive. When tax authorities are extractive (capturing only the primary fee), the institutions degrade to a floor.

Ownership flywheel. When the operator owns assets that grow with system success, every quality improvement is a bet on the operator's own balance sheet. When the operator is salaried with no asset-stake, quality improvements are work for free.

These four (transit + benchmark + tax + ownership) are topical instances of the same structural mechanism. Each has been crystallized as its own node before this canonical. The canonical names what they share.


Why "alternative" mechanisms fail

The naive view: where secondary-value-capture is absent, subsidy or regulation can substitute. Three failure modes:

Subsidy under-funds the dimensions the subsidizer can't easily measure. Public transit subsidized on ridership produces ridership-optimization, not service-optimization on the unmeasurable dimensions (cleanliness, reliability, network coverage in low-density areas).

Cultural enforcement decays as cultures shift, and depends on the operator's intrinsic motivation matching the system's quality requirements. The Japan-rail-as-culture explanation fails because pre-privatization JNR ran the same culture but extracted secondary value via taxation that didn't return to operating budgets, producing the same culture-failure-mode public US transit suffers.

Regulatory mandate produces compliance-quality, not optimization-quality. The operator does the minimum the regulation requires; the regulator's bandwidth bounds enforcement; everything not directly enforced degrades.

Secondary-value-capture is structurally different. The operator does not invest in quality because someone makes them. They invest because each quality unit translates directly into balance-sheet value they own.


Where it breaks

Falsifier 1 — quality without secondary-value-capture exists. Some monopoly-rent extractors invest in primary-product quality despite zero downstream-value-capture. Resolution: monopoly rents are themselves a form of secondary-value-capture (excess profit beyond primary cost). The canonical absorbs this if "secondary value" reads broadly.

Falsifier 2 — secondary-value-capture without quality. Some operators capture secondary value via network effects but invest in primary quality only to the floor required to maintain the network. Resolution: secondary-value-capture is necessary, not sufficient. Other constraints — competition pressure, talent, frontier — also bind. The canonical is correctly read as "secondary-value-capture sets the ceiling; other factors determine where between floor and ceiling the operator lands."

Failure mode (renaming-as-incentive-alignment-thinking). The canonical can degrade into a license to relocate every quality-failure to "misaligned incentives" without naming the specific secondary value, the specific party externalized to, and the specific structural mechanism. Each application must produce specifics; without them, the canonical is slogan.


Standing in the graph

This canonical absorbs four existing nodes (transit-incentive-capture, ownership-flywheel, the-tax-floor, benchmark-inversion) as instances of one structural primitive. The instances each contribute topical detail; the canonical contributes structural shape.

The implied strategic question for any system the operator builds: where does our secondary value go? If it externalizes to parties not bound to fund our primary quality, the system has a structural quality-ceiling cap that no amount of execution closes.


The quality ceiling is not set by talent, capability, or capital. It is set by where the secondary value goes.