If I send you links to 2-3 rare videos and you view them, I can probably uniquely identify you (or restrict the list of potential users to ~5 users then recognize you based on other videos you would have watched). I then have access to your full video watching history.

Even without this, I could just check videos you shared on Twitter, follow the same process, and have a decent chance of identifying you.

Anonymization is very very hard. Maybe federated learning is more promising here, keeping all your data privately stored on your own devices.