Does that imply that with enough effort spent on fine-tuning the configuration, other LLM runners can achieve similar results?