Does 'when costliness becomes forgeable' explain generative AI's demonetization of writing, art, and credentials?

observation-unforgeable-costliness-bridges-asteroid-pgm-and-bit-gold traces one economic law — a good's value rests on the cost of producing its scarcity, and collapses once a new technology makes that scarcity cheap to fake or replicate — through Szabo's collectible-money history (claim-szabo-collectible-monies-collapse-when-costliness-becomes-forgeable) and the asteroid-PGM pricing model (claim-asteroid-pgm-price-holds-then-collapses). The capture that surfaced this bridge flagged an unresearched but "most Cali-shaped" continuation: does the same law explain generative AI's effect on the value of writing, art, and credentials — each historically valuable partly because they were costly (in time, skill, or training) to produce, and each now cheaper to produce at scale via generative models?

Why it matters: this would be a third independent domain instance of the same law (after monetary history and resource economics), and — unlike the aluminium check, which found the trajectory differs from the PGM model's specific hold-then-collapse shape — this thread has not yet been checked against real data at all. It is also the vault's most-present-tense instance of the pattern, and the one most likely to be publish-worthy.

What's needed:

• Find primary or Tier 1–2 sources on measured price/wage effects of generative AI on creative or credentialed labor (e.g. freelance writing/illustration marketplace rate data, translation-industry pricing, entry-level-credential wage premiums) — not think-pieces.
• Check whether the collapse, if real, follows the wampum/glass-bead shape (value evaporates once the "furnace" — here, a generative model — makes the scarcity trivially reproducible) or something more gradual, as the vault's aluminium/inference-cost finding suggests may be the more common real-world shape (claim-cheaper-extraction-disruptions-fall-monotonically-not-hold-then-collapse).
• The capture noted this thread's nearest vault neighbor is the Soviet-expertise-premium cluster (claim-soviet-expertise-was-finite-stock-depleted-as-it-diffused, claim-soviet-emigre-career-premium-was-structural-not-ability) — worth reading those first, since they already model a skill-premium diffusing away.

Progress (2026-07-18) — partially answered. The first bullet is done: claim-chatgpt-release-cut-upwork-writing-freelancers-jobs-and-pay, claim-llm-collapse-of-costly-writing-signal-cuts-meritocratic-hiring, claim-generative-ai-availability-compresses-university-grade-distributions, and claim-ai-text-detectors-misflag-non-native-english-writing give four independent Tier 1 sources with measured, quantified effects covering freelance writing, freelance design/art (directional only), hiring meritocracy, and university grades. None of the four sources cite Szabo or use his framing — the bridge to "when costliness becomes forgeable" remains this vault's synthesis, not a sourced claim, and that should stay explicit wherever these notes get cited together. The second bullet — whether the collapse follows the wampum/glass-bead hold-then-cliff shape or the more gradual monotonic shape found for aluminium and AI inference (claim-cheaper-extraction-disruptions-fall-monotonically-not-hold-then-collapse) — is still open. None of the four new sources' event-study figures (e.g. Hui, Reshef & Zhou's Figure 2) were read closely enough this session to characterize trajectory shape, only endpoint magnitudes. Leaving this open rather than closing it; the next pass on this question should pull the event-study time series directly rather than the summary coefficients.