请问HN：你们是如何应对“提示疲劳”和懒惰的语言模型输出的？
我非常依赖大型语言模型（LLMs）来帮助我编写副项目和撰写文案，但最近我遇到了提示疲劳的问题。
在上大学和进行销售工作的时间之间，我的实际开发时间非常有限。我开始注意到，我花了20分钟只是与模型争论，以获得我真正想要的结果。如果我每次不写一个庞大且结构完美的系统提示，人工智能就会默认给我半成品的代码（// 在这里插入剩余逻辑）或者用那种冷冰冰、通用的语气来包装一切（总是使用“深入探讨”或“强大”等词汇）。
我厌倦了保持一个杂乱的Notion文档，里面充满了“负约束”，以便复制和粘贴，因此我最终构建了自己的轻量级包装器（约束引擎），在模型处理之前先加载所有格式规则。
但我真的很想知道这里的高级用户们现在是如何处理这个问题的。
你们是否只是保留大量的系统提示Markdown文件以供复制/粘贴？
你们使用了哪些具体的约束或框架来强制模型在第一次尝试时编写完整、可投入生产的代码？
查看原文
I rely heavily on LLMs to help me code side projects and write copy, but lately, I’ve hit a wall with prompt fatigue.<p>Between my college classes and working my sales shifts, my actual dev time is pretty limited. I started noticing that I was spending 20 minutes just arguing with the models to get what I actually asked for. If I don't write a massive, perfectly structured system prompt every single time, the AI defaults to giving me half-finished code (// insert remaining logic here) or wraps everything in that sterile, generic voice (always using words like 'delve' or 'robust').<p>I got so tired of keeping a messy Notion doc full of "negative constraints" to copy and paste that I ended up just building my own lightweight wrapper (a constraint engine) to front-load all the formatting rules before it hits the model.<p>But I'm really curious about how power users here are handling this right now.<p>Are you guys just keeping massive markdown files of system prompts to copy/paste?<p>What specific constraints or frameworks are you using to force models to write complete, production-ready code on the first try?