a visual course

How Your Feed
Knows You

(Recommender Systems)

How apps pick, from millions of posts, the handful you actually see — in about half an hour.

A free, visual course on recommender systems — how the apps you use rank millions of items to fill your feed, learn your taste from clicks and watch-time, find people like you, and create the feedback loop behind filter bubbles.

in the real world

It's what fills your TikTok, YouTube, Spotify and Netflix — the system deciding, out of millions of options, exactly what you see next.

Start with your feed →

Part 1 — What a feed really is

1

Your feed is a ranking

Out of millions of posts, the feed scores every candidate and shows you the top few.

2

You are what you tap

Every click, watch-time and skip is a quiet vote that builds your taste profile.

Part 2 — How it learns your taste

3

People like you

Find users with similar taste, then recommend what they loved that you haven't seen.

4

Taste as coordinates

Users and items become points in a hidden 'taste space' where closeness predicts what you'll like.

Part 3 — The catch

5

The new-user problem

With no history to learn from, the feed falls back on what an item is about — and on what's popular.

6

The loop you're in

Your feed shapes what you do, which shapes your next feed — and why a little randomness matters.

Part 4 — Out in the world

7

Where this shows up

Social feeds, streaming, shopping, news, matching and ads — choosing what billions see.