The core difference here seems to be white-box vs. black-box thinking. In the latter mode, we’re considering what pieces of the environment we’d term “agents” if we knew nothing about their internals. In the former mode, we’re thinking about internal structures — what mechanisms would we term “agents”.
The core difference here seems to be white-box vs. black-box thinking. In the latter mode, we’re considering what pieces of the environment we’d term “agents” if we knew nothing about their internals. In the former mode, we’re thinking about internal structures — what mechanisms would we term “agents”.
Relevant problem: Does Agent-like Behavior Imply Agent-like Architecture?