📄 中文摘要
在没有终止策略的情况下，将两个模型放入迭代优化循环中会产生意想不到的后果。一种模型负责生成API文档，而另一种模型则对其进行评估和批判。这个过程不断循环：生成、批判、改进，直至达到一个理想的结果。然而，这种无限循环可能导致资源浪费和效率低下。因此，设计有效的终止策略至关重要，以确保模型在达到预定目标后能够停止运行，从而避免不必要的计算和时间消耗。
你的智能体永远运行——如何让我的停止
出处: Your Agents Run Forever — Here's How I Make Mine Stop
发布: 2026年3月11日
📄 English Summary
Your Agents Run Forever — Here's How I Make Mine Stop
When two models are placed in an iterative refinement loop without a termination strategy, unexpected consequences can arise. One model generates API documentation while the other critiques it. This process continues in a cycle of generation, critique, and improvement until an ideal result is achieved. However, without a proper termination strategy, this infinite loop can lead to wasted resources and inefficiencies. Therefore, designing an effective termination strategy is crucial to ensure that models can stop running after reaching their intended goals, thus avoiding unnecessary computation and time expenditure.