Malk.
Follow people with taste — video discovery beyond the algorithm.
The algorithm decides what you watch.
YouTube's recommendation engine optimizes for engagement, not for taste. The best videos you've ever seen probably came from a friend, not the homepage — but there's no good way to share that signal at scale.
Profiles, follows, and human curation.
Malk lets anyone build a public profile of videos they love, organized into channels. Follow people you trust; discover what they're watching. No engagement metrics, no algorithm — just a feed of picks from humans whose taste you've opted into.
Profile-first
Every account is a curated taste-page, not a stream of activity noise.
No engagement metrics
No likes, no view counts. Just titles, channels, and who saved it.
Human-routed
Discovery happens through follows, not a black-box recommender.
How it came together
- 01
Wrote the manifesto
Started by writing what I believed about video discovery — clarity on values before code.
- 02
Designed the profile
Profiles are the product. Spent weeks on layout, density, and what fields earn their place.
- 03
Built the import flow
Pasting a YouTube URL pulls metadata, thumbnail, and channel — the lowest-friction save action I could ship.
- 04
Opened private beta
Invited 400 curators by hand to seed the network with real taste before opening discovery.
Lab Atlas →
The complete map of Spatial-Biology services.
