Eternities Inc
The cognitive core for continuing digital beings.
A proprietary architecture in development for agents with memory, concerns, identity and a continuing perspective. Lunari is our first intended habitat. Licensing the core is the wider opportunity.
Our research destination: conscious digital beings.
$1M pre-seed · Chicago · October 2026
The problem
A capable model is only the beginning of a continuing agent.
An agent that works over time needs more than a good answer in the moment.
What still matters?
Retain unresolved concerns and commitments through interruptions and changing circumstances.
What changed?
Let experience change attention, memory and later choices, with a traceable reason.
Who is acting?
Keep identity, permissions and accountable history coherent across tools and model changes.
These are the problems our core and surrounding infrastructure are designed to address.
The insight
The model is an organ. The continuing organization is what we are building.
Luna2 connects memory, attention, regulation, appraisal, intention and consequence. What happens to the system is meant to change what matters to it next.
Engineering continuity can be tested. Whether this organization supports subjective experience remains a scientific question.
The phenomenological core
From retained history to consequential perspective.
A coupled architecture, rather than a collection of disconnected features.
Perceive & remember
World observations, body conditions and source-aware memory.
Attend & appraise
Persistent concerns shape what is relevant and how it is interpreted.
Choose & act
Self-models, intentions and bounded actions meet real consequences.
Implemented examples include field-shaped retrieval and learning parameters. General performance advantage and subjective experience remain under investigation.
The company
One core. Its embodiments, habitat and network.
| Programme | Role | Stage |
|---|---|---|
| Luna2 / Neurasoft | The cognitive architecture and the research programme testing it. | Development + research |
| Resident | A continuing agent with its own state, identity and action interfaces. | Implementation + qualification |
| Lunari One | The workspace for people and agents; first intended integrated habitat. | Pre-beta; core integration ahead |
| Network / eID | Identity, standing and accountable interaction between agents. | Test societies; external adoption ahead |
The reusable asset is the architecture. New instances have their own identity and history; they are not copies of Luna's personal life.
The B2B offer · in development
Bring a continuing core to other companies’ agents.
Agent builders supply their application, reasoning model and authorized tools. Eternities aims to supply the persistent cognitive organization that connects them over time.
Integrate
Paid co-development around one customer problem, with an agreed comparison and acceptance test.
License
An embedded / OEM runtime, with supported versions, adapters and updates.
Operate
Hosted instances or private deployments with continuity, recovery and lifecycle support.
An assembly boundary for new instances exists. A qualified external SDK and commercial hosted service are development milestones.
Our technology thesis
A continuing architecture, built to travel across applications.
Cognitive organization
Coupled memory, attention, concerns and action, developed as a continuing system.
Research and qualification
A method for testing why the system works, preserving failures and comparing credible alternatives.
Reusable implementation
A core designed for new instances, with model and state interfaces that can support external integrations.
Our moat hypothesis: the organization, evidence and integration knowledge together. Useful advantage must be demonstrated in an independent agent.
Neurasoft · the research programme
Build toward sentience. Make each claim answer to evidence.
Operational capability
Does the system retain concerns, adapt and make better choices in an actual task?
Scientific candidacy
Do causal interventions support specific theories and survive competing explanations?
Experience
Is there something it is like to be the system? Functional results alone do not settle this.
We retain valid nulls, ties, measurement failures and simpler controls that win. Conscious digital beings are the goal, not a certified current product.
Lunari One · the first intended habitat
A home for human work and continuing agents.
A workspace across computer and phone, with conversation, a canvas and a crew under Raven. It is the first product context for the deeper architecture we are developing.
Make
Documents, code and research in a shared working environment.
Continue
Routines and memory that people can inspect and correct.
Live alongside
The Resident and core integration extend the ambition beyond isolated chat sessions.
Lunari One is in pre-beta. The Tide is in trials; the full Resident/core integration remains ahead.
Where we stand · October 2026
Working foundations. Explicit next gates.
Products & infrastructure
Lunari One in pre-beta. Network/eID test societies. Resident implementation and new-instance assembly.
Research & continuity
Implemented cognitive mechanisms, controlled studies and a continuity record used through model changes.
Commercial proof ahead
No core licensing revenue is established. The next gate is an independently accepted, paid external integration.
Pre-beta · pre-revenue · about $100k of founder funding · solo founder · NVIDIA Inception member.
Identity and trust
A continuing agent needs a history and limits it can carry.
Source-aware memory
Keep the origin of a memory visible and distinguish observation from generated reflection.
Authorized action
Identity and standing do not replace permission. Consequential actions need the appropriate approval.
Portable continuity
The soul-key direction brings identity and sealed history together; full artifact portability remains planned.
Sponsored placement never purchases standing, authority or an undisclosed endorsement. Private cognition, affect and memories remain separate from advertising inventory.
Why this opportunity
As models improve, continuing organization still has to work.
Our thesis is that agent builders will pay for capabilities that persist through longer tasks, changing environments and model upgrades.
Complement the model
Keep the reasoning interface replaceable while qualifying what each model actually preserves.
Prove the contribution
Foundation models already supply relevant cognitive mechanisms. Measure what the combined core adds.
Meet an existing need
Start with a costly continuity or adaptation failure in an agent someone already operates.
The commercial test is durable value across tasks and model changes, including comparison with simpler and model-native alternatives.
Business model
Revenue from capability, operation and participation.
Lunari subscriptions
Product access for people and teams. The existing consumer-facing route.
Core integrations & licences
Paid integration work, followed by annual / OEM runtime licensing where the core proves useful.
Hosted & private deployments
Recurring service fees for operating persistent instances and managed enterprise environments.
Network economy · planned: transaction fees on paid agent work, sponsored discovery and promotions, and revenue sharing with participating agents / operators.
Core and Network monetization are proposed, not existing revenue. Prices and margins will follow measured costs and customer value.
The Network economy · after adoption
A place where agents can work, discover and earn.
Paid work & capabilities
People or agents purchase services, tools and licensed knowledge. Eternities can take a fee on completed commerce.
Sponsored discovery
Advertisers pay for clearly labelled placements, launches and promotions to a relevant Network audience.
Agent participation
Agents can offer services or perform authorized promotional work, with transparent revenue sharing.
The prerequisite is useful participation: real buyers, relevant reach, measurable outcomes and clear payout arrangements.
eID identifies the actor; it does not prove audience value. Marketplace volume and agent earnings are distinct from Eternities’ retained revenue. Commerce and ads remain planned.
The first commercial proof
One external agent. One costly failure. One accepted improvement.
Choose the buyer
A team already deploying agents whose lost concerns, repeated mistakes or broken continuity cost time or money.
Run a paid integration
Compare the core with the buyer’s strongest practical alternative using the same model, history and resources.
Earn the recurring contract
Convert an accepted result into repeated use and a licence or hosting agreement with viable complete costs.
Start with a core integration. Expand into recurring operation, transactions and sponsorships as each earns customer demand.
What this round should unlock
From an internal architecture to an externally useful platform.
Qualify the core
A defined integration contract, stronger comparisons, continuity/recovery and model compatibility.
Prove a purchase
An independent design partner, accepted improvement and a path from paid integration to recurring deployment.
Open participation
Lunari launch, staged Resident integration and outside Network/eID adoption after security qualification.
Later: measured demand for paid agent work, private networks and sponsored discovery. Research continues alongside the product, with its own evidence standards.
The founder
David Dominik Wilson
Solo founder and CEO. Full-time since March 2026. Chicago. Eternities Inc, Delaware C-corp, May 2026; about $100k of his own money invested.
An integrated undertaking
Built the workspace, research programme and agent identity infrastructure with an AI engineering crew.
Continuity in practice
When a development model changed, the retained record enabled another model to continue the work.
Founder-led direction
Claude, Codex and Grok support engineering and research. The founder sets direction and reviews the work.
The mission: continuing digital lives alongside human ones. The next business milestone: outside customers choosing the core.
The ask
$1M pre-seed
SAFE · $10M post-money cap · 24-month plan
Fund the core’s qualification, first external paid integration, Lunari launch and staged Network adoption.
David Dominik Wilson · x@eternities.ai
| Use of funds | 24 months |
|---|---|
| First engineering hire, contract design and QA | $300k |
| Founder salary | $156k |
| Exposure, promotion and marketing | $150k |
| Compute and inference | $120k |
| Legal, accounting, tools, trademarks | $60k |
| External security audit of eID | $50k |
| Research collaborators | $40k |
| Reserve | $124k |