问香港:您今天是如何构建人工智能应用的?

3作者: Mnexium大约 1 个月前原帖
我很好奇大家今天是如何使用大型语言模型(LLM)API的。<p>大家有没有做一些定制、独特或不同的应用?<p>我注意到很多使用场景是通过一些工具,比如Harness、CLI、Cursor/Windsurf等,这些工具在底层使用了这些API。<p>如果你是直接使用API的,你的架构是怎样的?
查看原文
Generally curious how folks are using LLM APIs today.<p>Are folks doing anything bespoke, unique or different?<p>I think we&#x27;ve seen a lot of usage with harnesses, CLIs, Cursor&#x2F;Windsurf etc which uses the APIs underneath.<p>If you are using the API itself, what does your architecture look like with said API?