📄 中文摘要
本指南为《铁路导向 TypeScript》的设置提供指导。随着新工具如 @railway-ts/pipelines 和 @railway-ts/use-form 的推出,AI 编码助手(如 Claude Code、Cursor 和 GitHub Copilot)面临挑战。这些助手在大量 Zod、react-hook-form 和 neverthrow 的示例上进行了训练,但对 @railway-ts 的训练数据几乎为零。当请求 Claude Code 构建一个使用 @railway-ts/use-form 的表单时,它会自信地写出 zodResolver、useForm 和 setError/clearErrors,而不考虑实际需求。这种现象反映了新工具与现有 AI 训练数据之间的差距。
📄 English Summary
Working With AI Tools on a New Library
This setup guide provides instructions for Railway-Oriented TypeScript. With the introduction of new tools like @railway-ts/pipelines and @railway-ts/use-form, AI coding assistants such as Claude Code, Cursor, and GitHub Copilot face challenges. These assistants have been trained on millions of examples of Zod, react-hook-form, and neverthrow, but they have almost no training data for @railway-ts. When asked to build a form with @railway-ts/use-form, Claude Code confidently writes zodResolver, useForm, and setError/clearErrors without considering the actual requirements. This phenomenon highlights the gap between new tools and existing AI training data.
Powered by Cloudflare Workers + Payload CMS + Claude 3.5
数据源: OpenAI, Google AI, DeepMind, AWS ML Blog, HuggingFace 等