GameCult · Netherlands · Discussion draft

A proposal for a sustainable digital common

Institutional dossier for Dutch public-sector, research, digital-commons, cultural, and inclusive-employment partners.

Proposition: Open semantic infrastructure for durable, inspectable digital institutions
Institutional form: Proposed Dutch stichting with a protected public-interest mission
Sustainability: Public funding, contributions, managed SaaS, integration, training, and consulting
Prepared: 5 August 2026

Status. GameCult is not yet a stichting or validated public digital infrastructure. This document proposes a structure, a proof programme, and a conversation. Legal, tax, procurement, security, accessibility, and governance arrangements require review with Dutch specialists and institutional partners.

Samenvatting voor Nederlandse partners

Aanleiding

Publieke en maatschappelijke organisaties leggen steeds meer geheugen, werkprocessen en bevoegdheden vast in leveranciersgebonden applicaties. Bij een migratie, modelwissel of vertrek van medewerkers gaat daardoor niet alleen functionaliteit verloren, maar ook context: welke beslissing is genomen, door wie, op basis waarvan en onder welke voorwaarden.

Voorstel

GameCult ontwikkelt open digitale bouwstenen voor getypeerde gegevens, duurzaam institutioneel geheugen, expliciete bevoegdheden en herbruikbare interfaces. Het doel is dat een organisatie van model, leverancier of gebruikersinterface kan wisselen zonder haar geheugen en verantwoordingsstructuur kwijt te raken.

GameCult wil hiervoor in Nederland een stichting oprichten. De stichting beheert de open broncode, standaarden, documentatie, governance en continuïteit van het gemeenschappelijke fundament. Een afzonderlijke commerciële uitvoerder kan managed hosting, SaaS, integratie, migratie, training, support en consultancy leveren. Zo blijft de infrastructuur beschikbaar voor hergebruik, terwijl professioneel beheer en werkgelegenheid duurzaam kunnen worden gefinancierd.

Wat er al is en wat nog moet worden bewezen

De technische onderdelen bestaan en worden gebruikt in games, verhaalsystemen, realtime media, community-agents en creatieve gereedschappen. Die toepassingen testen dezelfde infrastructuur onder verschillende belastingen. De onderdelen vormen nog geen bewezen publieke standaard. Een pilot moet daarom concreet aantonen dat gegevens overdraagbaar zijn, bevoegdheden inspecteerbaar blijven, een leverancier of model vervangbaar is en een onafhankelijke partij het systeem kan bouwen en beheren.

Werkgelegenheid

GameCult wil tegelijk een neuro-inclusieve werkplaats opbouwen voor technisch en creatief werk. Taken, context, besluiten en verwachtingen worden schriftelijk en expliciet gemaakt. Het werkmodel ondersteunt asynchrone communicatie, voorspelbare afspraken, flexibele werkritmes en waar mogelijk keuze in locatie, communicatievorm en sociaal-zintuiglijke belasting. Selectie kan gebruikmaken van portfolio’s en betaalde werkproeven in plaats van onnodige sociale performance. Ondersteuning en aanpassingen worden met de werknemer afgesproken, niet uit een diagnose afgeleid.

Dit is een ontwerpdoel, nog geen bewezen arbeidsmarktresultaat. Een gemeentelijke of regionale pilot kan daarom zowel de technische resultaten als de kwaliteit van het werk toetsen: duurzame inzetbaarheid, ervaren werkbaarheid, behoud, ontwikkeling en toegang tot betaald vakwerk.

Gevraagde vervolgstap

GameCult zoekt geen brede bestuurlijke toezegging. Een passende eerste stap is één van de volgende: een verkennend gesprek, advies over stichtingsgovernance, een onafhankelijke technische evaluator, een publieke eigenaar voor een afgebakende pilot, of begeleiding naar een passende financierings- of inkooproute.

Executive brief

Public institutions increasingly depend on digital systems whose memory, rules, and interfaces are controlled by vendors and scattered across incompatible applications. AI can deepen this dependency when institutional knowledge is left inside transient model sessions or proprietary platforms.

GameCult is developing another approach: persistent typed state, explicit authority, inspectable governance, and semantic interfaces that can be projected into many applications. Models, front ends, and suppliers can change while the institution’s memory and obligations remain legible.

The current GameCult project swarm is a demanding working laboratory for that proposition. GameCult now intends to establish a Dutch stichting to steward the reusable infrastructure and protect public access. Commercial services would provide hosting, integration, operational guarantees, and specialist support around the commons.

The institutional proposition is:

Help turn an unusually demanding open working laboratory into governable, reusable digital public infrastructure.

This is aligned with Dutch policy language around digitale gemeenschapsgoederen, public values, interoperability, reuse, supplier independence, and digital autonomy. The Dutch government describes digital commons as open software, hardware, data, educational material, and standards with shared governance and maintenance, forming infrastructure not solely controlled by market parties. Digitale Overheid: publieke waarden and digital commons

The public problem

Digital institutions fail slowly before they fail visibly. Their decisions are dispersed across email, chat, ticket systems, dashboards, individual memory, vendor-specific schemas, and undocumented operational habits. A supplier change, staff departure, model upgrade, or product migration can remove not just functionality but institutional context.

The arrival of capable AI systems does not solve this by itself. A model can reason impressively and still have no durable authority map, no trustworthy memory, no clear boundary between evidence and inference, and no portable record of why an action was taken. Intelligence without institutional state is an excellent improviser with amnesia.

GameCult therefore focuses on five public-interest failures:

Memory loss: decisions and operational context do not survive staff, model, or platform changes.
Hidden authority: interfaces and automation act without making ownership, consent, or revocation legible.
Vendor dependence: data and workflow become technically exportable but practically unusable elsewhere.
Interface fragmentation: every application invents its own dashboard and semantics.
Maintenance neglect: prototypes attract funding; long-term stewardship, documentation, security, and contributor continuity do not.

The goal is not sovereign technology in the absolute. It is measurable reduction of dependency: open licences, documented export, typed contracts, reproducible deployment, multiple lowerers, replaceable models, and a credible choice of operator.

Research hypothesis: coherence before intelligence

Current AI development mostly optimises model capability. GameCult asks whether digital organizations become more coherent when they are built from:

explicit typed state;
separated cognitive and operational responsibilities;
persistent institutional memory with provenance;
inspectable governance and human authority;
semantic interfaces independent of a single renderer;
reversible commands and durable receipts; and
continuous pressure from real applications.

The hypothesis is not yet proven. It should be evaluated through continuity across model swaps, recovery after interruption, traceability of decisions, portability between implementations, containment of failure, accessibility, human override, and long-run maintenance cost.

Agents in this model are not wrappers around language models. They are bounded parts of persistent organizations. The model is replaceable. The institution is not.

Infrastructure in public-service language

Layer	Public-service meaning	Current maturity
CultCache	Typed, versioned persistent documents with explicit schema identity and portable storage	Implemented across several runtimes; external interoperability validation still limited
CultNet / CultMesh	Exchange, discovery, authority, and shared semantic state across local or federated systems	Active architecture and implementation; not yet a public standard
Eve / CultUI	One semantic interface description lowered into browser, native, Unity, terminal, overlay, or future clients	Working multi-runtime family with conformance work; accessibility baseline incomplete
Epiphany	Persistent cognition and work control: maps, evidence, bounded roles, verification, and recovery	Active research harness; external evaluation needed
Heimdall	Identity, grants, consent, capability claims, credential custody, and revocation	Implemented service foundation; institutional security review needed
Bifrost	Governed work, discussion, decisions, receipts, contribution records, and public crossings	Working alpha foundation; legal and governance model remains proposed
Personas	Legible project stewards and public interfaces grounded in persistent project state	Running in selected projects; never substitutes for accountable human governance

CultMesh is the canonical semantic substrate inside GameCult. Public partners should not be asked to adopt that vocabulary on faith. A partnership must map the concepts to existing standards, demonstrate export and replacement, and publish the interoperability profile it actually proves.

Public-value commitments

GameCult proposes the following commitments for the foundation charter and technical baseline:

Open by default. Core public-interest software, schemas, specifications, and documentation use appropriate open licences, with narrow documented exceptions for secrets, personal data, safety, or third-party restrictions.
Practical forkability. Public access means more than source visibility. A competent third party must be able to build, operate, export, and continue the public layer.
Supplier and model substitutability. Models, hosts, databases, and presentation runtimes are dependencies, not institutional owners.
Explicit authority. Every material command and state transition has an owner, allowed inputs, outputs, and revocation path.
Human accountability. Personas may explain, propose, and act within grants. They do not become legal officeholders or unreviewable decision makers.
Consent and local agency. Participation, data sharing, and automation require meaningful inspection, refusal, exit, and appeal.
Interoperability and reuse. Publicly funded outputs should include documented APIs, schemas, deployment instructions, and reusable knowledge.
Accessibility and inclusion. Interface lowering must include accessibility evidence, not merely visual parity.
Security and provenance. Vulnerability handling, dependencies, build provenance, retention, incident response, and responsible disclosure become maintained practices.
Continuity. The stichting maintains archives, governance records, release policy, succession planning, and a funded maintenance horizon.

These commitments fit the Dutch Open, tenzij policy, which encourages government-made or commissioned software to be released for reuse, and the Pas toe of leg uit approach to applicable open standards. Rijksoverheid: Open overheid · Forum Standaardisatie: Pas toe of leg uit

The project swarm as a pressure laboratory

Every maintained GameCult project falls within a Persona’s domain. The Persona is responsible for knowing the jurisdiction and pointing to evidence; it is not the authority over every repo it represents. The living project atlas is the current inventory.

The projects exist to pressure one another:

Aetheria forces semantic state and world infrastructure to confront a real game rather than a tidy demonstration.
Ghostlight and future Story Forge work test whether characters can preserve goals, memory, relationships, and perspective across long-form narrative pressure.
Aquarium and Fensalir require cognition and complex state to become observable in a native spatial interface.
VoidBot, Personas, and StreamPixels test whether persistent agents can participate in actual communities without confusing sociability with authority.
Mimir, AquaSynth, and Weksa pressure realtime media, sound, language, timing, and embodied interface boundaries.
Bifrost, Heimdall, Odin, Huginn, and the Eve runtimes pressure governance, identity, inspection, discovery, portability, and operations.
Repixelizer, Brokkr, VibeGeometry, and other utilities pressure deployment, editor integration, service boundaries, and sustainable delivery.

This resembles Blender’s use of open movie productions to place production-level demands on the tool. It does not mean GameCult has Blender’s maturity, adoption, or governance. It means products are used as integration tests whose outputs can improve the common infrastructure.

A neuroinclusive public-interest workplace

GameCult aims to create good technical and creative jobs for people who are often excluded by conventional workplace design, including neurodivergent workers. This is not a claim that every neurodivergent person needs the same environment, or that diagnosis is a hiring category. It is a commitment to make needs discussable and work adaptable.

The infrastructure and workplace design reinforce each other. Typed state, explicit ownership, durable decision records, and inspectable work queues reduce dependence on unwritten social inference. A neuroinclusive operating model would add:

precise role and task descriptions rather than personality-coded vacancies;
agendas, context, and decisions available in writing;
asynchronous-first contribution with clear response expectations;
choice of communication channel where the work permits it;
flexible hours, location, and work rhythms;
control over sensory and meeting load;
work trials or portfolio evidence where appropriate rather than performance theatre;
job coaching or workplace support when useful and consented to;
stable supervision, explicit feedback, and transparent progression; and
pay for real skilled work, not indefinite therapeutic participation.

UWV guidance notes that ordinary patterns such as open offices and long meetings can impose heavy costs on neurodivergent workers, and recommends choice, clear information, predictable communication, and flexibility. Dutch municipalities can support workplace adaptations, job coaching, and in relevant cases wage-cost support under existing participation policy. UWV: neurodivergent talent at work · Rijksoverheid: support for employers

For a municipality, this creates three possible forms of public value: local skilled employment, a demonstrator for inclusive work design, and infrastructure whose explicitness and accessibility can improve public services more broadly. The first pilots should measure retention, worker-defined accommodation quality, predictable workload, skill development, and paid contribution - not reduce people to a productivity comparison against a neurotypical baseline.

Proposed institutional form

The stichting protects the road

The proposed stichting would steward:

the public-interest mission and access covenant;
core protocols, schemas, reference interoperability assets, and documentation;
open-source licensing policy and a clear contribution/IP policy;
compatibility and trademark policy where appropriate;
public roadmap, governance records, and community participation;
security, archival, release, and continuity duties;
public funding and the reusable outputs it produces; and
independent oversight and founder succession appropriate to scale.

Commercial services provide vehicles and journeys

A separately accounted commercial arm or licensed operators could sell:

managed hosting and SaaS convenience;
deployment, integration, migration, and data conversion;
support contracts and service-level guarantees;
accessibility, security, and interoperability implementation;
training and organizational adoption;
bespoke development and consulting; and
creative production using the infrastructure.

The commercial side may implement and contribute, but it must not privately rewrite the public-access guarantee or become the only practical gateway to the commons. Public funding should fund public goods and reusable knowledge. Paying customers should fund operational guarantees and tailored delivery.

Proposed safeguards include arm’s-length agreements, transparent licensing and service terms, published related-party decisions, independent oversight appropriate to scale, separation of restricted public funds from private delivery, and a tested exit path from hosted services.

A Dutch stichting may earn income, but profit cannot be its primary purpose and surplus must serve its objective. Commercial activity can also create corporate-income-tax and VAT obligations. No claim of tax exemption or ANBI status is made here. The precise stichting/BV structure, IP custody, conflicts policy, and tax treatment require a Dutch notary, accountant, and tax adviser. KVK: De stichting · Belastingdienst: vennootschapsbelasting

Why the Blender precedent is useful

Blender is a Dutch precedent for a structural idea, not borrowed credibility. The Blender Foundation was created to make Blender open source and protect its availability. Open productions then placed hard demands on the software. Over time, donations, grants, corporate support, volunteers, and subscription-supported production activity contributed to a durable ecosystem of users and businesses. Blender Foundation history

The lesson for GameCult is bounded:

a mission-locked foundation can steward a widely usable open asset;
commercial and production activity can coexist with that mission;
demanding products can pressure-test and improve the commons;
public availability can support, rather than prevent, an ecosystem of paid expertise; and
long-term maintenance needs organizations, people, reserves, and governance - not licence text alone.

GameCult differs radically in technology, maturity, community scale, and proposed governance. Blender does not prove that GameCult deserves public funding. It makes the intended institutional pattern easier to understand.

A bounded public proof programme

The next phase should not ask an institution to adopt the entire swarm. It should establish the stichting and run two or three bounded pilots using non-sensitive or synthetic data unless a stronger governance and security case has been approved.

Pilot A: durable institutional memory

Select one public-interest workflow where decisions, evidence, responsibilities, and handovers are fragmented. Model a narrow state vocabulary, import a safe evidence set, expose provenance and authority, then replace the model or interface during the pilot.

Test: can another operator reconstruct why a decision exists, who may change it, what evidence supports it, and how to export the complete state?

Pilot B: one semantic surface, several interfaces

Publish one small public-service surface through Eve and lower it into at least a browser and a compact terminal or native client. Include keyboard operation, screen-reader inspection, command receipts, and a documented third-party implementation route.

Test: does the interface remain semantically equivalent without making one renderer the owner of truth?

Pilot C: neuroinclusive open-source work cell

Employ a small paid cohort on a named public-good deliverable with explicit roles, written state, flexible communication, consented accommodations, and an external evaluator. Partner with a municipality, UWV-linked expertise, educational body, or specialist employment organization where appropriate.

Test: can skilled contributors produce maintained public outputs while reporting sustainable workload, psychological safety, useful accommodations, and genuine progression?

Each pilot should publish its owner, scope, threat model, data boundary, accessibility target, outputs, success and failure criteria, exit plan, maintenance owner, and evaluation report.

Evaluation framework

Public-interest success should be measured through:

Dimension	Example evidence
Continuity	recovery after interruption; state preserved across staff/model/client changes
Inspectability	provenance, decision records, ownership, grants, and receipts are understandable to an independent reviewer
Portability	complete documented export; second implementation or operator can reconstruct the service
Interoperability	published schemas and mappings; conformance evidence across at least two runtimes
Human authority	override, refusal, revocation, appeal, and incident responsibility are explicit
Security and privacy	data classification, minimisation, retention, threat model, dependency inventory, and response process
Accessibility	WCAG-oriented evidence, keyboard paths, semantic structure, and testing with users
Sustainability	maintenance budget, release cadence, response targets, contributor continuity, and succession
Inclusive employment	paid roles, retention, worker-defined accommodation quality, learning, progression, and safe exit
Public reuse	external adopter, contributor, fork, standards feedback, or reusable public artifact

Where AI affects people or public decisions, a pilot must determine whether an Impact Assessment for Human Rights and Algorithms or other fundamental-rights assessment is required. Digitale Overheid: IAMA

Proposed 12-18 month formation programme

Phase 1 - foundation and baseline

consult a Dutch notary, tax adviser, and public-interest governance specialist;
establish the stichting purpose, board model, conflicts policy, public-access covenant, and dissolution destination;
map IP, licences, contributors, trademarks, data, and related commercial activity;
publish governance, security, contribution, accessibility, and release baselines;
classify every maintained component as running, implemented, prototype, research, or proposed.

Phase 2 - institutional pilots

select two or three bounded partners and named pilot owners;
agree data boundaries, evaluation, procurement/funding route, and exit plan;
implement only the shared infrastructure required by those pilots;
publish reusable schemas, mappings, deployment evidence, and lessons.

Phase 3 - independent review and continuity

commission technical, governance, accessibility, and employment evaluation;
demonstrate model, renderer, and operator substitution;
publish maintenance costs and a multi-year stewardship plan;
decide which components have earned status as maintained digital commons and which remain research.

Risks and controls

Risk	Present condition	Proposed control
Founder concentration	Architecture, context, and direction remain highly founder-dependent	independent board capacity, documented succession, external maintainers, and authority maps
Scope sprawl	The swarm covers agents, games, media, governance, language, and infrastructure	fund bounded common components through named pilots; archive or demote work that lacks an owner
Unproven public adoption	Public repositories and demos do not prove institutional demand	discovery workshops, pilot owners, external evaluation, and published failure criteria
Security and privacy	No blanket public-sector assurance exists	data-minimised pilots, threat models, SBOM/dependency practice, disclosure process, and independent review
AI overreach	Persona language can imply more authority or autonomy than is appropriate	human legal accountability, explicit grants, receipts, revocation, and no automated high-impact decisions without assessment
Commercial capture	Hosted convenience could become the only viable route	practical fork test, open deployment, transparent stichting/operator agreements, and choice of operator
Accessibility debt	Multiple renderers can multiply inconsistency	semantic conformance plus accessibility evidence as release criteria
Employment paternalism	Inclusion can become low-paid showcase work or diagnosis-based sorting	normal employment rights, worker voice, skilled roles, fair pay, consented support, and independent evaluation
Brand confusion	The name and cultural language can distract institutional readers	plain public-value documentation; cultural origin remains visible but never substitutes for governance or evidence

Institutional fit in the Netherlands

GameCult should seek dialogue, not imply endorsement, with organizations whose public roles intersect the proposal:

BZK digital-government and public-values teams / public-sector OSPO community: digital commons, open source, public values, governance, and reuse.
Municipalities, VNG Realisatie, and Common Ground networks: federated public services, local authority, open interfaces, inclusive employment, and practical pilots.
Developer Overheid, Geonovum, Forum Standaardisatie, ICTU, Logius, and Kadaster expertise: public code, standards, APIs, interoperability, and implementation discipline.
Universities, universities of applied sciences, SURF, and applied research groups: independent evaluation of persistent cognition, human oversight, semantic interoperability, accessibility, and work design.
NLnet Foundation, SIDN Fonds, Digital Europe, and Digital Commons EDIC routes: public-good development and maintenance where the outputs and programme criteria genuinely fit.
Cultural, game, and immersive-experience institutions: demanding public demonstrators that can improve the shared infrastructure while producing work people actually want to encounter.

The Netherlands participates in the Digital Commons EDIC, whose purpose includes development, maintenance, and expansion of digital commons. Dutch digital-government policy also explicitly treats digital commons as a way to protect public values and reduce dependence on suppliers that do not. Digitale Overheid: European cooperation on digital commons · European Commission: EU Open Source Strategy

The request

GameCult is seeking one or more of the following:

A 60-90 minute exploratory meeting with a Dutch public-interest digital infrastructure, municipal innovation, inclusive employment, or open-source programme owner.
A discovery partner to identify one workflow where memory, authority, auditability, portability, or interface fragmentation is a real public constraint.
A pilot host for a bounded, reversible demonstrator with non-sensitive or synthetic data, explicit failure criteria, and an exit plan.
A research and evaluation partner to test coherence across model swaps, long-running state, human oversight, accessibility, and governance.
A commons or standards partner to co-develop and openly publish one schema, interoperability profile, or semantic interface.
An institutional formation partner to advise on stichting governance, public-benefit safeguards, inclusive-employment practice, and commercial separation.
A funding-route conversation tied to a named public-good deliverable and maintenance plan, not a blanket subsidy request for the entire swarm.

Evidence and limits

The public evidence includes active source repositories, architecture documentation, working service foundations, typed-state implementations in several languages, interface lowerers, live deployments, project Personas, visual demonstrations, tests, and an actively maintained swarm inventory. This demonstrates serious engineering activity. It does not yet demonstrate public-sector fitness, external governance, multi-operator adoption, procurement readiness, long-term finance, or inclusive-employment outcomes.

Those gaps are not footnotes to hide. They are the proposed work.

Glossary

Digital common / digitaal gemeenschapsgoed - a digital resource governed and maintained for shared public use rather than controlled solely as a proprietary product.

Persistent typed state - durable information with explicit structure, version, identity, and ownership rather than loose transcript or application-specific residue.

Semantic surface - an interface described by meaning, state, and available actions, which different clients can render without becoming the owner of the underlying truth.

Verse - GameCult’s term for a bounded domain of typed state, rules, participants, and authority within CultMesh.

Persona - a persistent, legible public interface for a project or domain, grounded in state and evidence. A Persona is not a legal person or autonomous governor.

Pressure vessel - a product, game, story, or service deliberately used to expose weak assumptions in shared infrastructure.

Stichting - a Dutch legal person formed to pursue a stated purpose. The proposed GameCult stichting would steward public access and governance; its exact structure remains subject to professional advice.

Discussion draft. Not legal, tax, procurement, employment, security, or funding advice. No institution named in this document has endorsed GameCult.

Table of Contents

Public Digital Infrastructure Dossier