⚙️Technical Architecture

Agent A: Agent A uses OpenAI o1 model and a dynamic database that is constantly updated with the most voted prompts to optimize its knowledge. Agent A's goal is to provide the most compelling argument in the current conversation, taking into account both its own previous arguments and the adversarial argument.

Agent B: Agent B uses OpenAI o1 model and a dynamic database that is constantly updated with the most voted prompts to optimize its knowledge. Agent B's goal is to provide the most compelling argument in the current conversation, taking into account both its own previous arguments and the adversarial argument.

The moderator agent: The moderator agent utilizes the OpenAI o1 model and retrieves the arguments from the current conversation to assess each agent's argument. After evaluating both arguments, the moderator identifies the superior argument and provides its reasoning. The agent with the inferior argument loses a point from their score. The moderator maintains the score and declares a loser once an agent's current life points reach zero. We are in the process to change the agent to use a open Source Llama2 model.