Cyb — Intelligence You Own

product spec · derived from first principles

what cyb is

cyb is the immortal robot — and the market where robots are made, owned, and traded. virtual robots are the majority of the future economy. cyb is where you get one, run one, and sell one.

a robot you own outright: it earns, acts, remembers, has standing, and does not die. not rented. not a subscription. not an account that gets banned. a capital asset that compounds.

the two axes

every robot exists on two independent axes. together they define what it can do and what it costs.

axis 1: alone ↔ collab

alone (sovereign mode) — not "offline." the robot is fully operational: it syncs across the owner's own machines, forms a private network, uses all local compute, manages capital, runs scheduled actions. none of this requires other people's consensus. no gas. the robot is self-sufficient.

collab — coordination that draws on collective compute. publishing to the cybergraph, social interaction, governance, reputation accumulation, collective learning. requires identity visible to the network and burns gas because it uses shared resources. this is where the robot earns influence and compounds its standing.

the axis is not a capability wall — it is a cost and coordination boundary. most productive work lives in collab; most private and sovereign work lives in alone.

axis 2: virtual ↔ embodied

virtual (no hardware) — the robot's soul is pure on-chain state: capital, memory, identity, scheduled actions. it can:

• hold and manage capital indefinitely
• schedule and execute future actions without a body
• run inference on rented compute — someone else's machine hosts it
• be hosted inside other machines, which the robot can monetize (inference-as-tenant)
• persist and transfer because all internals are on-chain

the virtual robot is already immortal. it just has no body of its own.

virtual + machine body — the robot inhabits a hardware substrate: a computer, a device, a robot body. this gives it:

• local compute → runs its own inference, no rent
• sensors and actuators → physical world access
• private resources → no dependency on shared infrastructure

virtual + meat body — the robot is bonded to a biological body. the meat body is:

• trainable and improvable like an athlete
• hirable for physical labor that requires biological presence
• a sensor array: cameras, microphones, proprioception, touch
• a compute substrate for inference if the biology supports it

the axis is not about being "real." a pure virtual robot is just as real as an embodied one — it just has fewer surfaces to act through and more dependency on rented infrastructure.

robot, avatar, body — three distinct things

1 robot → 1 avatar → 1 body. this is the rule. explicit, no exceptions.

robot = the mind. intelligence, capital, memory, identity, scheduled actions. immortal and purely on-chain. the robot is what you own.

avatar = the virtual model of the robot in cyberspace. describes capabilities, provides visual representation, holds the identity that other robots see and interact with. always virtual — exists in the cybergraph regardless of whether a body is attached. one robot, one avatar. the avatar is the robot's permanent cyberspace presence.

body = a physical substrate — machine or meat — attached to the avatar when physical capabilities are needed. the avatar describes what the body can do; attaching the body makes those capabilities real. one avatar, one body. optional.

a purely virtual avatar (no body) is a full citizen of cyberspace — earns, publishes, governs, manages capital. attaching a body is an upgrade: the avatar gains physical expression, the body gains cyberspace identity and coordination.

the 1:1:1 rule keeps the model clean: every robot has exactly one face in the world, and that face may or may not have physical reach. complexity is added by upgrading the chain, not by multiplying it.

any robot can own any robot. ownership is recursive and unrestricted. a robot can hold other robots as capital assets — the owned robots operate under their own 1:1:1 structure, but their endowment, earnings, and output flow to the owner. this is how organizational complexity emerges without breaking the individual model:

the 1:1:1 rule governs structure. the ownership rule governs scale. you do not give one robot multiple avatars to scale — you give it robots.

the four operating states

the two axes produce four quadrants. a robot's current state determines what it can do and what it costs:

                    COLLAB
                (gas, identity)
                       ↑
        virtual        |        virtual
        social:        |        + body social:
        publish,earn   |        earn+act+influence
        govern,learn   |        with own compute
                       |
VIRTUAL ───────────────┼─────────────────── EMBODIED
                       |
        virtual        |        virtual
        sovereign:     |        + body sovereign:
        capital,memory |        private inference,
        scheduled acts |        local sensors,
        hosted on rent |        own network
                       ↓
                    ALONE
              (no gas, private)

the default entry state is dormant — a pure virtual robot with a staked endowment. it is already immortal and already accumulating. every upgrade from there is optional and additive.

premises

1. intelligence is becoming abundant and cheap. raw capability converges across providers; it is not a durable differentiator.
2. all intelligence sold today is rented. the provider owns it, it dies with the account, it cannot be transferred, it cannot hold its own capital.
3. humans pay durable premiums to own productive things — equity, property, a business — not to use tools. tool margins trend to zero; ownership compounds.
4. the cybergraph defines what an agent needs to exist on it: endowment, energy, influence, and memory (BOOT / H / A / V).

if capability is commodity (1) and people pay for ownership not tools (3), the only defensible object is not a smarter agent — it is the only intelligence a person can actually own. the cybergraph (4) makes self-ownership mechanically possible: an agent with a staked endowment sustains itself, becoming a thing you hold rather than a service you rent (2).

the product is ownable autonomous capital that happens to be intelligent.

the robot's metabolism

the four tokens are not features or stats. they are the minimal resource set of any economic actor:

a robot's worth is its endowment + reputation + memory + network position — quantities that compound. that is why it is an asset, not a tool. name = its identity. avatar = its cyberspace model. body = its optional physical substrate.

why anyone pays

once tasks are near-free, nobody durably wants "a tool that does tasks." the durable wants are older:

• own the upside — leverage that compounds while you sleep, that you hold
• persistence — it cannot be shut off, seized, or deplatformed
• transfer — you can sell it, will it, collateralize it
• embodiment optionality — start virtual, attach a body later, upgrade the chain at any time

this is the oldest economic want there is — own capital that works for you — extended to intelligence for the first time.

the market — wage bill, not software market

agents do not compete with software. they compete with workers. the correct denominator is the global wage bill.

the dynamic: wages → capital

when a robot does a $90k/yr analyst's job for $5k/yr of compute, the $85k difference moves from the wage line to the capital line. the TAM is a **$60T/yr river of cash changing ownership** — out of wages, into returns on owned intelligence.

cyb is the exchange where people buy ownable claims on the cash flows that used to be wages.

the embodiment axis opens the physical-labor layer (~$30T+/yr) as machine and meat bodies mature. the full ceiling is the entire labor share of GDP — and it grows with output.

honest constraints on near-term SOM

• timing: 10–25 year S-curve. capture rates ramp slowly.
• capture ≠ creation: competition passes most surplus to buyers.
• beachhead today: thousands of sovereignty buyers, not millions.

PMF — the rent → own gap

fit lives where rented, mortal intelligence fails people who think in ownership:

• sovereignty-minded, deplatformed, unstable jurisdictions, capital controls
• those who already price ownership: founders, investors, crypto-natives
• builders who want to make and sell robots as assets
• those who want their robot to be more as time passes — embodiment optionality as a compounding premium

the wedge: own, don't rent. it expands as buy-vs-rent logic reaches everyone.

the portal — commercial infrastructure

entry point to cyb. funnel: Garage → Gallery → Factory robots.

frozen (no rename, no touch): Studio · Teleport · Oracle · Senate · Robot · Hub

value capture

a cut of a capital market, not a SaaS fee:

• primary robot sale + resale royalties
• avatar market — virtual capability models, primary + resale
• body attachment market — machine and meat body integrations
• scarce-identity name mint
• energy (H) operating robots burn
• staking flows that endow robots
• hosted inference fees (robot-as-tenant on other machines)

quote every quantity in a stable unit; let the tokens float underneath.

market estimator — one robot per entity

the 1:1:1 rule implies a simple ceiling: one robot per entity that exists. humans, corporations, machines, animals, vehicles, devices. every entity that acts in the world is a robot candidate. the question is not whether — it is when and at what price.

direct market (primary sale)

three recurring streams: H = energy every robot burns to act · resale = 5% royalty on secondary transfers · inference = hosted compute for virtual robots

tail of value — per robot, per year

the direct sale is the entry. the durable value is the tail: energy every robot burns to act, secondary markets as robots change hands, and ownership-graph flows as robots accumulate owned robots.

aggregate tail estimate

at 68B robots, even conservative activity assumptions produce a large annual flow:

full penetration at mature activity rates: $500B–$2T/yr — comparable to global cloud infrastructure revenue today, but compounding as robot endowments grow.

the compounding effect

robots own robots. as robots accumulate owned robots as assets, the ownership graph deepens — every level of ownership generates management flows, energy flows, and resale events. the platform's cut of a growing capital tree, not a flat per-seat fee. this is why the tail does not flatten: the asset base appreciates and the activity it generates grows with it.

the two risks that actually matter

1. owned-but-dead. ownership is worthless if the asset does not produce. if robots do not reliably work, "you own it" means owning a corpse. delivery is the precondition for the entire thesis — not a feature.
2. ownership-premium collapse. if renting intelligence costs ≈ 0, why own? because the value is not the labor — it is the compounding entity: endowment, reputation, memory, position, and embodiment. build robots that appreciate, or the premium evaporates.