THE QUESTION

IF DUP IS DELETED AND REPLACED WITH USDC, FIAT, PLATFORM CREDITS, REPUTATION, OR ORDINARY DATABASE ACCOUNTING, WHAT IMPORTANT PROPERTY OF THE NETWORK STOPS WORKING?

Determine whether a native asset is technically or economically necessary. The purpose is not to justify DUP. Try to prove we should not have one.

VERDICT

DUP REQUIRED = NO

Confidence HIGH

Duplios should not issue DUP because every security job that bonding performs is performed by USDC, fiat, or USD-funded credits, while endogenous DUP adds crash, whale, bootstrap, and regulatory failure modes the evidence engine does not need.

Nothing material in the Proof of Evidence decision engine. Reputation-only and no-bond economies still lose slashable cost-to-corrupt; that gap is closed by USDC, fiat escrow, USD-funded credits, or external collateral — not by DUP.

Quick replay events on this page: 49592. Full Monte Carlo (≥ 1,000,000 events) is in institutional/research/dup-necessity/results.json.

TESTS

• USDC replacement: PASS
• Sybil (DUP unique advantage): FAIL
• Collusion (DUP unique advantage): FAIL
• Token crash: FAIL
• Whale: FAIL
• Bootstrap: FAIL
• Machine economy: FAIL
• Regulatory complexity: HIGH (not legal advice)

BEST ARCHITECTURE

Non-DUP: PoE independence engine unchanged; USD-funded platform credits with operator clawback for the single-operator product; USDC (or fiat escrow) slashable contributor/reviewer/challenge bonds where third-party settlement is needed; bounties in USD/USDC; non-transferable reputation. External collateral + reputation (economy G) if a future multi-operator deployment needs portable bonds.

DUP model: None recommended. The least-bad paper DUP model is USD-target bonds (identical to USDC) plus separate reputation. Native-unit bonds fail crash tests. Governance tokens fail whale tests. Do not issue.

FUNCTIONS

• evidence_bounty_asset — DUP necessary=NO; USDC can=YES. Duplicate leakage DUP=0.0000 USDC=0.0000 CREDITS=0.0000. Engine zeros duplicate payouts regardless of unit. Reputation-only pays 0 bounty so it cannot hire work, which is a payment problem USDC/credits already solve.
• contributor_bond — DUP necessary=NO; USDC can=YES. USD-target $100 contributor bond is available in USDC and in USD-funded credits. DUP USD-target matches USDC cost-to-corrupt (100.00 vs 100.00). Native-unit DUP after 80% crash drops attack cost (4.00 vs USDC 100.00).
• validator_reviewer_bond — DUP necessary=NO; USDC can=YES. Reviewer $1,000 USDC bond is allowed in this harness. Capture probability DUP=0.017 USDC=0.009 whale-DUP=0.110. Whale who already holds DUP captures cheaper.
• challenge_bond — DUP necessary=NO; USDC can=YES. Challenge-farming profit DUP=-3.442 USDC=-2.769 REP=0.000. Bonding helps; the unit does not have to be DUP.
• slashable_security — DUP necessary=NO; USDC can=YES. Economic security per $ DUP=1.390 USDC=1.429. Credits can be clawed. Reputation cannot be slashed across new identities.
• anti_sybil_economic_cost — DUP necessary=NO; USDC can=YES. Sybil profit DUP=-0.974 USDC=-1.049 REP=0.000. After crash, DUP attack cost falls (YES). Endogenous demand is circular and fails the crash test. USDC Sybil cost stays $100/identity.
• network_governance — DUP necessary=NO; USDC can=NO. Token-weighted vote is whale-advantage DUP=1.000 vs USDC (no governance token) 0.510. Governance is fashion here: operator + published constitution + reputation. Coin voting adds capture surface.
• machine_to_machine_market — DUP necessary=NO; USDC can=YES. Machine usability DUP=0.000 USDC=1.000 crash-DUP=0.000. Agents abort when quote-to-settle variance exceeds 2%. DUP is worse, not better.

PRIOR ART

Staking a native token behind information is known (Augur, UMA, Kleros, Numerai). Duplios novelty, if any, is the Proof of Evidence constitution (independence accounting, reserved-bounty math, AI-is-not-truth). Those rules do not require DUP.

• Kleros — Kleros itself argues that paying jurors in an external unit avoids circular security. That is evidence against needing DUP.
• UMA Optimistic Oracle / DVM — Later research (ETH as oracle collateral; EigenLayer restaked ETH) exists specifically because native-token security is circular.
• Reality.eth — Direct demonstration that information bonds can be posted in an external unit.
• Augur — Prior art for the exact slogan. Also a known illustration of native-token resolution risk.
• Ethereum Research — ETH as oracle collateral (2026 discussion) — The live research direction in oracles is away from small native tokens, toward external collateral.
• Chainlink — Off-chain fact feeds and APIs settle in fiat without a consumer-facing work token.
• Bittensor TAO — Prior art for token-incentivised AI work. Quality is validator-scored, not independence-accounted evidence. Emissions are manufactured demand.
• Vana — Prior art for tokenised data contribution. Does not make DUP necessary for Duplios evidence rules.
• DePIN template (Helium, Filecoin, etc.) — Evidence work can be paid in USDC. A DePIN-style DUP emission would be a subsidy, not a unique security property.
• Community Notes / centralised fact-checking — Large-scale information quality systems operate without a native token. Not Duplios, but relevant to necessity.
• Numerai NMR — Staking on information quality is known. The scarce asset need not be the product's own coin (Numerai still chose one; the economic payer is the fund in USD terms).