essay · article
Why I'm Building a Personal AI Operating Site
Why a static portfolio is no longer enough when personal knowledge needs to be reusable by both people and AI agents, and what an AI-readable knowledge product looks like instead.
Key takeaways
- Typed content beats a flat blog archive for both readers and agents.
- A public knowledge layer needs an explicit publishing boundary.
- AI-readability is an operating structure, not a gimmick.
The thesis
Most personal sites are display cases. They answer “what has this person done?” but not “what here can I actually reuse?” When both people and AI agents read your site, reuse is the real product.
From portfolio to operating site
An operating site treats every piece of content as a typed object — article, prompt, template, lab note, project, reference — with metadata that tells a reader (human or machine) what it is, what it is for, and whether it is safe to reuse.
The boundary is the trust mechanism
None of this works without a visible publishing boundary. Every item here is classified as public, generalized, review, or private-only before it ships. The private knowledge base never reaches this layer.