Which do we like better? Teams of agents or 1 executive agent w skills.
retained text
So when Anthropic launched the concept of Claude as like this 1 executor agent that would autonomously spin up and down subagents and leverage skills to “specialize the work” it seemed like a big pivot from the customGPT concept that had existed up until then - where basically you would build a specific agent for each thing. But now it seems like we might be heading back towards 1 agent for 1 job? Grokbot, Monday .com, every startup selling to enterprise a team of employee bots…..there are specific bots for specific jobs. Even if you essentially have a chief of staff agent that handles all the other agents, what’s really the technical advantage of that set up? Over an executor agent that just can call in skills, run loops, and spin up and spin down subagents to do stuff? I guess it’s the difference of a persistent memory per agent? But couldn’t you just house that in markdown files that Claude reads when a skill says to? Or….is it literally marketing and people just like the idea of being a boss to a team of agents? Someone please explain this to me ;)
document fetched 2026-10-31 00:19 · canonical https://reddit.com/r/AI_Agents/comments/1xj58uu/which_do_we_like_better_teams_of_agents_or_1