• Ben Matthews
    link
    fedilink
    English
    arrow-up
    2
    ·
    edit-2
    9 hours ago

    Whilst I like the concept, doesn’t such algorithm need access to a huge amount of data (i.e. a large set of potential posts from which to choose), if so how could this run client side?
    Maybe could envisage something running in a parallel distributed federated way (vaguely recalling projects like seti@home or climateprediction.net), grouping results for sets of people with similar preferences, with minor local tweaks and short-term updates - but that implies many interacting layers, rather complex.

    • AItoothbrush@lemmy.zip
      link
      fedilink
      English
      arrow-up
      2
      ·
      9 hours ago

      Yeah i get what youre saying. Thats why i said if its possible. Idk how they algorithms work as that really isnt my field but someone can def figure something out in the future, i hope at least.