Media and search engines nowadays need a flag system and a filter for AI junk results

  • Hubi@feddit.org
    link
    fedilink
    English
    arrow-up
    41
    arrow-down
    2
    ·
    2 months ago

    Why are you listening to the youtube recommendations? Why are you looking at the recommendations? What are you expecting there in the first place?

    There was a time when the YouTube recommendations were actually pretty good. I discovered a number of great artists just from suggested songs and playlists. That no longer seems to be the case though, my recommendations have been garbage for more than 5 years at this point.

    • iAmTheTot@sh.itjust.works
      link
      fedilink
      English
      arrow-up
      20
      arrow-down
      3
      ·
      2 months ago

      Dunno what y’all did to your algorithms but mine is still working just fine. When something odd comes up that I don’t like, I tell it not to continue recommending that.

      • Klear@sh.itjust.works
        link
        fedilink
        English
        arrow-up
        9
        ·
        2 months ago

        Yeah. You gota curate that shit. Then it works.

        Well, I make maps for a rhythm game and like to rewatch my showcase videos, so it started offering me that (meaning it’s music I enjoy but with a lot of shooting in the background).

        But other than that it’s solid.

    • Dizzy Devil Ducky@lemm.ee
      link
      fedilink
      English
      arrow-up
      8
      arrow-down
      1
      ·
      2 months ago

      It can definitely work if you’ve trained it right. It’s how I discovered many of my now favorite vocaloid/utau/deepvocal/whatever songs, amongst other groups outside of that vocal synth genre. It definitely isn’t doing me super dirty in the music department.

      • KeenFlame@feddit.nu
        link
        fedilink
        English
        arrow-up
        3
        arrow-down
        1
        ·
        2 months ago

        It could work when the algorithm was like that but they changed it to consider using users as test subjects for more lucrative users