这个周末我没有只是随便看看演示，而是真正对几款AI写作工具做了压力测试。很惊讶它们处理长提示时差异这么大—一个越来越有创意，另一个却陷入重复。基准测试说明不了全部情况。
Original
Spent the weekend actually stress-testing a few AI writing tools instead of just skimming demos. Surprised how differently they handle long prompts—one got more creative, another collapsed into repetition. Benchmarks don't tell the whole story.