Six pilot programs designed to produce outside-controlled evidence
Each program starts with a narrow institutional problem, an exact authority/rights/economic boundary and success/failure criteria agreed before testing. A participant is not asked to endorse ENTITY or adopt the entire stack.
1. Digital Rights Market Lab
Question: Can a real dataset, software asset, model output, compute resource or other governed digital object be represented through explicit rights and moved through a market without conflating byte possession, ownership, execution or settlement?
Test flow: governed object → authority/right → instrument → venue → listing/disclosure → order book, RFQ or call auction → trade → clearing → entitlement → usage → settlement evidence.
Good partners: exchanges, data spaces, market-infrastructure groups, enterprise data owners, research-data organizations.
Success evidence: invalid issuers/rights fail closed; price discovery and execution are deterministic; entitlement follows valid execution; external payment remains separately verifiable.
Failure evidence: venue custody creates authority, rights disappear on transfer, duplicate execution succeeds, or trade state is incorrectly promoted into verified payment.
2. External Contribution Economy Lab
Question: Can externally originated technical work retain durable contributor, exact-change, review/acceptance, licence and economic lineage and later participate in a valid contribution/service instrument without changing upstream rights?
Starting evidence: Vector #26501 / PR #26504 and Memnox #46 / PR #86 are already upstream merged and recorded as UPSTREAM_ACCEPTED_RECORDED, both pro bono with USD 0 realized cash.
Test flow: external problem → contributor → exact change-set → review/CI → upstream acceptance → ENTITY/BTDU lineage → optional CONTRIBUTION instrument / service terms → independently evidenced settlement.
Success evidence: accepted work remains reconstructable and attributable; licence terms remain controlling; commercial terms can be added only when explicitly valid.
Failure evidence: provenance is treated as ownership, retroactive royalty is invented, or payment is inferred from a merge.
3. Agent Authority & Economic Action Lab
Question: Can an autonomous or AI-assisted system purchase, license, access, execute or publish only inside explicit delegated authority while preserving action, evidence and economic lineage?
Test flow: agent identity → delegated authority → right/market scope → proposed action → governed transition → evidence → revocation/recovery.
Success evidence: allowed actions succeed; prohibited or expired actions fail closed; NIKI/model output never becomes authority by itself.
Failure evidence: scope can be escaped, AI output silently grants permission, or recovery widens authority.
4. Sovereign Data Rights Lab
Question: Can real data move through organizational/infrastructure boundaries while machine-readable rights, provenance, permitted use, derivation and downstream obligations remain attributable?
Test flow: controller/object → rights → access/query/train/derive → governed use → derivative lineage → evidence → economic/compliance consequence → provider migration/recovery.
Success evidence: rights remain enforceable and recoverable; unauthorized use is rejected/detectable; lineage and licence boundaries survive provider changes.
Failure evidence: rights disappear, provenance becomes unverifiable or the same occurrence creates duplicate obligations.
5. Portable Market Recovery Lab
Question: Can the economic meaning of a market survive loss/replacement of the database, application or infrastructure provider?
Test flow: create venue/instruments/balances/listings/orders/trades/entitlements/usage/revenue rules/participation state → export canonical market bundle → remove original runtime from recovery path → verify semantic hashes/controller attestations → restore to clean compatible destination → compare state → exercise negative tamper case.
Success evidence: complete logical state is recovered, signatures/controller authority remain valid and no new rights/privilege appear.
Failure evidence: missing required tables are tolerated, altered state verifies, restore depends on original database bytes or provider custody becomes authority.
6. Independent BTDU Qualification Program
Question: Do BTDU representation, exact reconstruction and governed-information advantages survive independently selected workloads, methodologies and comparison architectures?
BTG provides: current code, controlled benchmark definitions, evidence and hypotheses.
Independent partner controls: workload selection, comparison systems, methodology, statistical treatment and conclusions.
Success evidence: independently reproduced advantages under explicit conditions.
Failure/boundary evidence: workloads where benefits disappear, costs shift elsewhere or alternatives outperform BTDU. Negative evidence is a valid outcome.
Pilot charter requirements
- Named problem owner and technical sponsor.
- Exact governed objects, identities, authority and rights boundary.
- Which market/economic stages are actually in scope.
- External evidence requirements for payment, legal or real-world claims.
- Pre-declared success, failure and inconclusive criteria.
- Recovery and at least one negative/adversarial case where applicable.
- IP, confidentiality and publication rules agreed before testing.
- No implied endorsement from participation.
- No relabeling BTG-controlled evidence as independent validation.