周末我真的对几个AI工具做了压力测试，而不是只看演示。在我小规模的工作流测试中，提示词质量比模型选择重要得多。很好奇大家在做真实任务（比如总结文档或起草邮件）时是否也有同样的感觉？
Original
Spent the weekend actually stress-testing a few AI tools instead of just watching demos. Prompt quality mattered way more than the model choice in my small workflow tests. Curious if others see the same when doing real tasks like summarizing docs or drafting emails?