I've been frustrated by the quality of our media, it often presents strong arguments for one side, and straw man for the other. So I built a tool where LLMs argue opposite sides of a debate as a way exploring a topic more deeply.
Each side is randomly assigned a model (pro or con), and the idea is to surface the best arguments from both perspectives.
Currently, it uses GPT-4, Gemini 2.5 Flash, and Grok-3. I’d love any reactions or feedback to improve it.
