When do multi-agent systems beat single-agent?
hardAnswer
- (1) Specialized roles: a 'coder' agent + 'reviewer' agent produce better code than one agent doing both — parallel expertise.
- (2) Adversarial verification: 'writer' vs 'critic' loops catch errors a single pass misses.
- (3) Divide-and-conquer: distribute independent subtasks across agents.
- (4) Domain isolation: each agent has domain-specific tools + system prompt.
- Costs: more tokens, coordination complexity, potential message-passing loops.
- Frameworks: CrewAI, AutoGen, LangGraph, MetaGPT.
- Often overhyped — a well-prompted single agent handles most tasks.
Check yourself — multiple choice
- Always multi-agent
- Multi-agent wins on specialized roles / adversarial verification / divide-and-conquer / domain isolation; often overhyped for simple tasks
- Never useful
- Same as single-agent
Multi-agent: specialized roles + verification + divide-and-conquer — overhyped for simple tasks.
#agents
Practise LLMs & GenAI
214 interview questions in this topic.