• Activity
  • Votes
  • Comments
  • New
  • All activity
  • Showing only topics in ~tech with the tag "algorithms.recommendation". Back to normal view / Search all groups
    1. Which content-recommending algorithms are actually good?

      For the end-user, that is. I'm sure Meta and the like think their algorithms are fantastic at what they want them to do. I find myself routinely asking why I get so many suggestions I have no...

      For the end-user, that is. I'm sure Meta and the like think their algorithms are fantastic at what they want them to do.

      I find myself routinely asking why I get so many suggestions I have no interest in when using all types of websites. I haven't used social media since the early years of Facebook, but I imagine most recommendation algorithms are tuned much like the ones on those sites, i.e. to offer more of the same, whereas I'd prefer something to introduce different stuff I'd probably like. Maybe that differentiates me from the average user, but there should be enough people like me that it'd be factored in, no? Just because I watch a cat video doesn't mean I'm all in on cats.

      I mostly like Pandora's service but it feels like their music library isn't huge for my fav genres. Steam regularly tries to interest me in the most insipid games based on superficial commonalities to what's already in my library. Youtube can be good, but it can easily be echo-chambery. Shopping websites of all sorts are usually a crapshoot. What gives?

      28 votes