The Funnel as Capital: A Semantic Economic Reading of the Application Process

Companion to ERS-001 v0.6 (The Application as Extraction Surface, DOI: 10.5281/zenodo.20330670). Where the predecessor paper inaugurated the measurement apparatus and quantified at $4,500–6,000 the unpaid labor extracted from one applicant in one cycle of the Anthropic Fellows screening, this paper reads the same case structurally: the screening process is not a selection apparatus that happens to externalize costs; it is a capital-accumulation apparatus that uses the form of selection to organize unpaid labor. v0.2 revisions incorporate cross-substrate review (unprimed ChatGPT). Six changes from v0.1: (1) the strict-equivalence rhetorical claim 'Applicant labor input = Institutional capital gained' replaced with 'Institutional capital formation is downstream of applicant labor input' — same structural force, fewer attack surfaces; (2) formal definition of calibration intelligence added at first use — the durable operational knowledge derived from aggregate applicant interaction with a screening apparatus; (3) recursive accumulation dynamic named explicitly — the rejected pool from cycle n trains the apparatus that structures cycle n+1; this is what makes the apparatus platform capital rather than ordinary unpaid labor; (4) explicit distinction added between 'selection is fictitious' (not the argument) and 'selection coexists with structural extraction' (the argument); (5) AARC computation reframed as order-of-magnitude modeling with explicit defense: 'whether the actual figure is $1.5M or $6M does not alter the structural relation'; (6) one line added to the closing cadence: 'The labor refines the factory.' Core argument. Selection is real. Selection is also the visible cover under which a much larger production process accumulates. The advertised product of a structured screening process is a ranked list of finalists; the actual product is calibration intelligence — funnel-shape data, assessment-instrument calibration, applicant-population behavioral signatures, process-refinement intelligence — which accrues to the screening operator as durable institutional capital. The labor input that produces this intelligence is supplied by the applicant pool, including (especially) rejected applicants, whose uncompensated labor recursively improves the institution's capacity to extract uncompensated labor from future applicants. The structural form is wage labor with the wage parameter set to its lower bound: zero. Relation-laundering through entity multiplication. The three-party structure (applicant → intermediary firm → principal) is named as a visibility device that ensures no two parties to the relation stand in a recognizable labor relation. The applicant applies (intransitive verb), the intermediary provides services (denominated to the principal, not to applicants), the principal sponsors a fellowship (charity-adjacent framing). The labor relation disappears from the institutional ledger. Aggregate Applicant Reliance Cost (AARC). Defined as Σᵢ Nᵢ · ARCᵢ across funnel stages. Worked order-of-magnitude example for one cycle of Anthropic Fellows places per-cycle aggregate labor extraction at approximately $3 million. Per-cycle aggregate is in seven figures; annual aggregate across the contract in eight figures; operator-wide aggregate across all clients in nine figures or more. None of this labor appears on any institutional ledger as labor. Generalization. The structure applies wherever a three-party screening apparatus induces preparation labor under opacity while capturing calibration intelligence as institutional capital. Cases: technical-hiring platforms (HackerRank-style), graduate admissions consultancies, venture accelerator portals, AI interview firms (HireVue/Karat), ghost-job posting ecosystems, talent-marketplace intermediaries. Document class: Extractive Reliance Study 002. Series: Semantic Labor Accounting / Cross-Domain Liquidation Studies. Companion to ERS-001 v0.6. v0.2 incorporating cross-substrate review, ~5,900 words.