Joe Reeve describes a prototype that photographs a statue, researches its identity and creates a conversational voice agent. He says the demo connects existing APIsAgent tool use lets an agent call bounded external capabilities so it can retrieve information or change state beyond generating a response. and would need curator involvement and further work before museum deployment.
In discussion with the audience, Reeve considers how cultural institutions might choose voices and narratives, how visual interfaces could complement speech, and why interruptions and feedback remain difficultTurn-taking coordinates when each participant speaks in a conversation.. He also shares lessons from making short videos about prototypes.
Watch on YouTube
