A home for unfinished thinking

Most personal sites I see are either a résumé with extra steps or a landing page that is trying to convert a stranger. I wanted neither.

I wanted a place that can hold a long essay, a short note, a project that is still awkward, and an experiment that may never become a product — without asking a database for permission.

Content first, software second

The useful architecture is almost boring:

Content lives in Git. Collections validate frontmatter. Layouts turn that into pages. The host serves static files.

That is not a lack of ambition. It is a bet that the scarce resource is attention and honesty, not infrastructure. A CMS is useful when many people edit the same corpus. This is one person, thinking in public.

Static generation also matches how I want the site to age. URLs stay stable. Drafts never ship. There is no runtime that can go stale while the writing is still true.

What the site is for

I use four shapes of work on purpose:

• Writing is the slow, edited argument.
• Notes are observations I do not want to inflate into essays.
• Projects are things that shipped for someone else, described without résumé metrics.
• Lab is “I wondered what would happen if…”, which is how most of my actual learning starts.

If everything looks like a case study, the site starts lying. Informal work needs a room of its own.

Trade-offs I accepted

No search in v1. No newsletter backend. No comments. Those are easy to add and hard to stop thinking about. A static search index can sit on top of this later. An email provider can own subscriptions. The site does not need to become a product in order to be a home.

JavaScript is reserved for things HTML cannot do: theme preference, a mobile menu, copying a code block. The article itself should render without a framework hydrating it.

What I am optimizing for

I want a recruiter, a collaborator, or a future version of me to read a page and understand how I think — not just which logos I have stood next to.

That means the writing has to be specific. Dates stay out of URLs so a piece can be revised without pretending it was always published today. Tags exist to cluster related work, not to invent a taxonomy department.

If this site works, it will look slightly under-designed and slightly over-written. That is the point.