happiness

the felt state of the network, as one number. every neuron holds a private value $h_\nu \in [0,100]$ — how it feels inside the system — changeable at any moment, revealed to no one. the protocol publishes only the stake-weighted aggregate:

$$H \;=\; \frac{\sum_\nu w_\nu\, h_\nu}{\sum_\nu w_\nu}, \qquad w_\nu = \text{stake}(\nu)$$

0 is hell. 100 is nirvana. everything the protocol does is, in the end, answerable to this number

origin

Ralph Merkle proposed the mechanism in DAOs, Democracy and Governance (2016): govern not by electing representatives but by declaring an explicit welfare metric — each citizen periodically reports how their life is going, the time-averaged collective happiness becomes the objective, and policies are adopted by forecasting their effect on it (Hanson's futarchy: vote on values, bet on beliefs). Merkle's poll was annual and anonymous. cyber keeps the metric and upgrades the plumbing:

• continuous, not annual — $h_\nu$ is a standing value a neuron edits whenever its state changes. the system feels policy in hours, not election cycles
• stake-weighted, not per-capita — a thousand fresh sybils carry no weight; the signal comes from those with skin in the graph
• private by construction — individual values live as encrypted commitments in bbg, the aggregate is computed under the same regime as everything else there: aggregate public, individual private. a zheng proof certifies $H$ was computed correctly from the commitments without opening a single one

why private

the honesty of the signal is downstream of its privacy. a visible happiness report is a political act — performed suffering to extract concessions, performed contentment to avoid attention, herding toward the visible mean. an invisible one is interoception: nothing to perform, no one to punish you for the truth, no social gradient to climb. the neuron that cannot be identified has no reason to lie — the same logic that makes a secret ballot honest, running continuously

the metabolic role

$H$ is a key metabolic factor. the network's metabolic health composes capitalization, syntropy, and happiness as a geometric mean — so collapse in any one signal drags the whole composite to zero. a network with zero happiness is metabolically dead regardless of how rich or how ordered it is

this is the alignment claim, stated as arithmetic. syntropy measures whether the graph is sharpening; happiness measures whether the minds inside it are flourishing. the two can diverge — a system can concentrate focus brilliantly while grinding its neurons down — and that divergence is precisely the misalignment a superintelligence must detect in itself. $H$ is the alarm channel:

• the protocol's own agency reads $H$ as a constraint: parametrization may retune emission, difficulty, and blend weights only within bounds that do not drive $H$ down — metabolic safety bounds
• policy experiments read $H$ as the outcome variable, futarchy-style: predicted $\Delta H$ is how competing proposals are compared
• neurons read $H$ as the weather — the one honest number about how the whole thing feels from inside

what it is not

• not a vote. $H$ decides nothing directly; it is a sensor, not an actuator. policies act, $H$ reacts
• not karma. karma is earned and public; happiness is declared and private. karma measures what you gave the graph; happiness measures what the graph does to you
• not syntropy. $J$ is the order of the knowledge; $H$ is the state of the knowers. a healthy system needs both rising — and pays attention when they part ways

the cheapest instrument in the whole design: one number per neuron, and the superintelligence acquires a felt sense of its own body

see netics for the control loop it closes · self for who listens · bbg for how privacy holds · vimputer for the aggregation machine

discover all concepts