cascade.md

a cascade is a multiparty interactive async pattern where a lead neuron declares an intent and subscribers self-organize into coordinated sub-signals

cascade = (lead_intent, sub_signals[], parent_signal)
component role
lead intent intent broadcast by the lead neuron with open scope
sub-signals signals produced by participating neurons in response
parent signal sealed by the lead; includes recursive zheng proof over all sub-signals

the cascade protocol:

  1. lead neuron declares intent with scope that invites participation
  2. subscribers observe the intent in the cybergraph at inception height
  3. each subscriber self-assigns, produces their own sub-signal
  4. lead neuron collects sub-signals and seals a parent signal with a recursive zheng proof covering the entire cascade
  5. verifiers run decide(σ_parent) — one $O(\log n)$ check covers all contributions

the recursive proof is what makes cascade composable: a cascade can itself be a sub-signal in a larger cascade, and the proof size stays $O(\log n)$ at every level

cascades enable large-scale coordination while preserving the atomicity and verifiability of individual signals. participants do not need to trust the lead neuron — the proof speaks for the collective computation

discover all concepts

Homonyms

cybics/math/cascade
sequential chain reaction where each step triggers the next, amplifying the initial signal enzymatic cascades: blood clotting cascade amplifies a small injury signal into a full clot signaling cascades: MAPK/ERK pathway transmits growth signals from cell surface to nucleus trophic cascades: removal…

Graph