Compared to other models that halt the loop on intermediate steps, or to ask further clarification, even if it's not the human equivalent of proactive, you see the similarity, right?