Reddit’s data is hugely helpful for training AI models.
  • Pissmidget@lemmy.world
    48·
    3 days ago

    I could give a well thought out comment about AI, scraping the web, giving your end users a worse experience in order to copyright their content for your own benefit, or any number of subjects.

    However, all I can muster is this:

    lol

      • Damage@feddit.it
        14·
        3 days ago

        Since multiple Lemmy instances show the same, federated, content, I wonder if our posts will have more weight in the model, for a normal scraper it would be as if many people repeated the same thing over and over on different sites.

        • Tangent5280@lemmy.world
          4·
          3 days ago

          It would be trivial to remove duplicates. With a little bit of foresight they could just as easily avoid the duplicates in the first place.

          • Damage@feddit.it
            9·
            3 days ago

            They could also avoid to re-crawl the whole internet every day, but here we are, so who knows.

      • MyBrainHurts@piefed.caEnglish
        6·
        3 days ago

        I’d imagine they’d be doing it for the whole fediverse. I mean from their perspective, why not? The whole open nature, in my limited understanding, seems to make that easier.

  • Jimbabwe@lemmy.world
    17·
    3 days ago

    lol it’s not Reddit’s content to paywall in the first place.

    “Bulletin Board sues passersby who took a picture without giving money first”

  • fakeplastic@lemmy.dbzer0.comEnglish
    9·
    3 days ago

    I was trying to think up a droll simile about how trying to build intelligence out of Reddit content is like… something. But I couldn’t get any further because the premise is so stupid to begin with that there’s no better way to illustrate it than let the absurdity of it stand on its own.