Hacker News Reader

Comment by _6fmb | Hacker News Reader

As said in some other comments, TikTok does not use a "follow" model, new content is algorithmically recommended to you. Therefore there is no problem with the content creators you're following burning out or your follow list getting stale. For example my highschool friends are no longer that active on facebook, and in 2020 I'm no longer asking random people to add me to FB, therefore facebook is now kind of dead to me. My youtube follows like Ryan Higa etc are burned out, hence youtube is less useful for me. That'll never happen with TikTok because my follows list is irrelevant. I see pretty much only fresh faces with content I'm interested in.

View on HN

Youtube does a pretty good job of recommending new content though.

Yeah, it's one of the ones that's doing well and don't seem to suffer much from people getting bored and leaving for fresher networks.

I feel like YT used to be better at recommending stuff to me. Now it just sends me more what I've already seen. It seems like it's falling into the "you just bought a lawnmower. do you want to buy 500 more lawnmowers???" recommendation engine trap.

I haven't installed TikTok, so I'll ask you instead.

If there's no follow model, how do I see stuff that I want to see? Say there's a particular guy I like, I want to see a bunch of his content. Do I have to like it and hope the algo gives me more?

there's a follow model, it's just secondary. when you log into tiktok, the first page you see is called the For You Page, where content is pushed to you algorithmically. to see the content you actually subscribed to, you'd need to swipe left.

The follow mode is secondary as others said. But aside from that it figures out what stuff you like by tracking things like time spent on videos and dozens of other micro creative analytics that I'm sure would piss off privacy advocates. The end result is a very fine tuned recommendations algorithm that's really adept on its task.

Gotcha - is this where Vine fell apart? By relying too heavily on follows?

Reply on news.ycombinator.com