CONCEPT Cited by 1 source
Amorphous computing¶
Amorphous computing is a distributed-systems model in which many simple, unreliable participants following local rules collectively produce coherent global behavior. There is no distinguished leader that holds the complete plan; the intended result emerges as locally visible state spreads and participants react to it. (Source: sources/2026-08-11-aws-scaling-patterns-for-self-organizing-multi-agent-clusters-with-kiro)
Agent-cluster application¶
Kiro-flock applies this model to LLM agents. Each agent sees the shared direction and a bounded local neighborhood of append-only logs, chooses work independently, and leaves an artifact and a trace for later neighbors. A global synthesis and eventual idle state emerge from those local decisions rather than from an orchestration graph.
Distinctions¶
- It differs from gossip because the source uses a deterministic ring neighborhood rather than random peer exchange.
- It differs from stigmergy in scope: stigmergy names coordination through environmental traces, while amorphous computing names the broader local-rule-to-global-behavior computational model.
- It is inappropriate when every transition requires a centralized verification gate, strict order, or global agreement before progress.
Seen in¶
- sources/2026-08-11-aws-scaling-patterns-for-self-organizing-multi-agent-clusters-with-kiro — AWS explicitly frames the bounded-ring cluster as an application of amorphous computing to AI-agent collaboration.