>>105883951
nothing to do with every new model inserting a LLM to "refine" your prompts btw, purely a coincidence