Jameson Lopp · 6d

We're gonna need a micro LLM that acts as a proxy, analyzes each prompt, and determines the best model to use to execute the task.