别写提示词，建评估管线
推荐指数 69.0 NO. 011 · 2026.09.21
发布2026/09/20Score71Comments33
为什么值得看
资深工程师 Dan 提出放弃手动优化 LLM 提示词，转而构建自动化评估与优化流水线，用更多 AI 对抗 AI。核心论点是文本提示的拟人化陷阱让我们误判系统本质，系统化工程才是出路。
编辑判断
Dan 的观点踩中了当前 AI 工程的一个隐秘痛点：提示词工程师正在变成新时代的"SQL 调优师"——手艺值钱但难以规模化。他提到的"intentional stance"（意向立场）是个关键概念，源自哲学家 Daniel Dennett，指的是人类本能地把有输入输出的系统当作有意识的主体来理解，这在 LLM 时代导致了大量伪科学式的提示词玄学。
更值得关注的信号是"评估管线"这个说法正在从研究圈渗透工程圈。Anthropic 的 Claude Code、OpenAI 的 evals 框架都在同一方向发力，提示词工程正在从 artisan 手艺向 industrial pipeline 转型。如果你团队还有人全职手写提示词做 A/B test，建议把资源转向自动化评估基础设施，这是未来 6-12 个月的竞争力分水岭。
社区反馈
意见分歧 31 条评论
核心争论：AI应作为用户可控的外部代理还是厂商锁定的内置服务
Crafting a good prompt is what differentiates an expert's output from a beginner's. Of course you need good constraints too. But those good constraints are created precisely through good prompts.
This is an amazing article. The problems it describes are exactly what we found when building a production system that used LLMs to (most of the time) produce reliable results. Extensive tests are necessary, and stakeholders have no idea how their suggestions fail in production. They just see the ha
I hope the future of AI isn't this sort, where companies provide the user/customer with an interface to an AI that can do things for the user... I would much prefer that companies instead provide an interface FOR an AI, and the user brings their own AI which connects to that interface. In other