I've always been fascinated by the idea that argumentation, rather than solo reasoning, leads to better outcomes. Inspired by the book "The Enigma of Reason", I started an experiment to see if this holds true for AI.
The result is Nimrobo AI: a platform where you can create AI agents with specific personas (e.g., a cynical historian, an optimistic futurist, a stoic philosopher) and pit them against each other in a debate on any topic you choose.
How it works: Each agent gets a detailed persona outlining its motivations, biases, and thinking style. To make the debates substantive, the agents are equipped with tools for live web searches and Browse to pull in real-time data and cite sources for their arguments.
Our goal is to see if this adversarial process can surface more nuanced insights, reveal hidden biases, and generate a more comprehensive summary of a topic than a single AI ever could.
This is an early version, and I would love for you to try it out and share your thoughts. I'm especially curious about: - Do the personas feel distinct and believable? - What debate matchups would you find most interesting or insightful? - Is the flow of the debate easy to follow?
Looking forward to your feedback!