Presentation of Agent Designs & Scholarly Discussion

Overview

Each team presents what becomes newly possible, the language of action they designed, and a proof-of-concept conversation. The discussion is not critique for its own sake — it is to expose the assumptions hiding inside the conversation: about the task, the tools, what comes back from the world, and the cost of being wrong.

How the presentations run

Each team will present its problem, explain what becomes newly possible, describe the tool set and conversation protocol, and show a proof-of-concept exchange. The presentation should also identify initial evidence, unresolved risks, and the parts of the conversation that still need to be designed.

After each presentation, the other teams will have 5 minutes to discuss the proposal among themselves and produce a list of questions. The presenting team will then respond. The purpose is not simply to critique the idea, but to expose assumptions about the task, the tools, the user-side observations, the evaluation criteria, and the consequences of failure.