collective attention. the probability distribution π over all particles — content-particles and axon-particles — that emerges from the tri-kernel operating on the attention-weighted cybergraph
focus sums to 1 across the whole graph. emphasizing one particle defocuses all others. no individual neuron controls focus — it is computed from the aggregate of all attention
individual neurons direct attention. the cybergraph computes focus. cyberank reads focus at a single particle. relevance reads focus in context. karma aggregates focus per neuron. value multiplies focus by cap
when focus converges, it produces cyberank: the per-particle prob of observation. the tru performs this computation via the tri-kernel — diffusion, springs, heat
see cyber/focus for the dynamics. see collective focus theorem for convergence proofs. see focus flow computation for the full protocol specification
discover all concepts