Massive

A content app for intellectually curious people. Massive surfaces evergreen, snackable pieces — the kind of writing you'd save to read later but actually want to read now. Personalised recommendations adapt to what you've read and what you've lingered on.

The recommendation engine uses the same vector embedding approach as my other projects. Each piece of content is embedded and stored in pgvector. As you read, your implicit preferences build a taste profile that the system matches against new content. No explicit "like" buttons — the signal comes from behaviour.

The editorial challenge is harder than the technical one. Curating content that's genuinely stimulating without being clickbait, evergreen without being stale, short enough to read in five minutes but substantial enough to be worth it. I've been sourcing from long-form journalism, academic papers, and independent essays.