Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

Imagine following some people who talk about houseplants. Maybe you like some tweets with the kind of information you consider of above-average usefulness for you. Maybe you praise someone for a beautiful pic of their monstera variegata.

The algorithm decides to show you more high-quality tweets and prettier photos.

If you consider that an echo chamber then all right, but I don't think it's a useful definition. Is going to a scifi-focused bookshop an echo chamber?




Maybe it’s because houseplants aren’t a controversial topic. Take any topic where people hold strong opinions (politics, religion, ...) and start following people you agree with. I guarantee you’ll have to very deliberately look for people on the other end of the spectrum at some point because your feed and your follow recommendations will indeed be an echo chamber.


I agree, that's exactly my point — online communities might be (for sure often are) an echo chamber, but that depends on what content you subscribe to. Specifically on how opinionated-ness-y and on-topic that content is.


If it only were that well. In reality, when I follow people who posts on various niche subjects, it turns out the algorithm believes I'm mostly interested in American Politics because quite a few of those posters are into that and that overshadows all the niche subjects.

Looking at the list of things Twitter believes I'm interested in is quite enlightening about the limits of recommendation engines.


I think the proper analogy is a bookstore whose titles rearrange themselves on the shelves based on your liminal reactions to the covers.




Consider applying for YC's Fall 2025 batch! Applications are open till Aug 4

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: