koe

I don’t write code, and I came to AI the way many readers of this book will: needing to understand systems that other people build. I learned the fundamentals first, then RAG, then everything that kept turning up around it, and along the way I kept explaining it all to non-engineers.

My working rule is to trust nothing secondhand. Before a claim goes on the page, I trace it back to the paper, the CVE record, or the official documentation. I write in English and Chinese, and I keep both versions of everything in step.

You can reach me at koe@llmknowledge.org.