An interesting paper about hierarchies of agents: (Perhaps) Beyond Human Translation: Harnessing Multi-Agent Collaboration for Translating Ultra-Long Literary Texts

In this paper they put agents in a simulated hierarchy.

The same LLM can do translation better with a hierarchy of agents made out of the LLM than using the LLM itself.

If you organize agents into corporate hierarchies and give them personalities you get better results

What this implies to me is partially that dialogue really does create insights, that it is a catalytic process.

What if the way we run organizations of people today is actually useful and the thing you go to even in organizations of agents?

Hierarchies show up partially because boundaries show up naturally to contain the superlinear information cacophony growth, and hierarchies are nested boundaries.