5 ms·
When you describe subagents, are those single-tool agents, or are they multi-tool agents with their own ability to reflect and iterate? (i.e. how many actual LL
by pamelafox 1y ago
When you describe subagents, are those single-tool agents, or are they multi-tool agents with their own ability to reflect and iterate? (i.e. how many actual LLM calls does a subagent make?)
- imsh4yy 1y agoSo I have a main agent that is responsible for streering the overall flow, and then there are the sub-agents that, as I mentioned, are stateless functions that are called by the main agent. Now these could be anything really: API calls, pure computation, or even LLM calls.