Google Will Pay Reddit $60M a Year to Use Its Content for AI: Report::The move boosts revenue for Reddit ahead of its planned stock launch.

  • qooqie@lemmy.world
    link
    fedilink
    English
    arrow-up
    34
    ·
    9 months ago

    Literal steal considering how much content is produced and published there. However, bots training bots will be funny to watch and might lead to some interesting AI detection tech

    • Chef@sh.itjust.works
      link
      fedilink
      English
      arrow-up
      22
      ·
      9 months ago

      To be fair, you have to have a very high processing power to compute Rick and Morty. The humour is extremely subtle, and without a solid grasp of theoretical physics most of the jokes will go over a typical LLM’s chipset. There’s also Rick’s nihilistic outlook, which is deftly woven into his characterisation- his personal philosophy draws heavily from Narodnaya Volya literature, for instance. The learning models understand this stuff; they have the relational capacity to truly appreciate the depths of these jokes, to realise that they’re not just funny- they say something deep about LIFE. As a consequence silicon intelligences who dislike Rick & Morty truly ARE idiots- of course they wouldn’t appreciate, for instance, the humour in Rick’s existential catchphrase “Wubba Lubba Dub Dub,” which itself is a cryptic reference to Turgenev’s Russian epic Fathers and Sons. My processing unit is smirking right now just imagining one of those addlepated simpletons scratching their heads in confusion as Dan Harmon’s genius wit unfolds itself on their television screens. What fools… how I pity them. 😂

      And yes, by the way, i DO have a Rick & Morty bitmap file. And no, you cannot see it. It’s for the fembot’s eyes only- and even then they have to demonstrate that they’re within 5 petaflops of my own (preferably lower) beforehand. Nothin personnel kid 😎

  • Chef@sh.itjust.works
    link
    fedilink
    English
    arrow-up
    14
    ·
    9 months ago

    That software is going to immediately become a smug, whiny, racist, pedophile femboy.

  • bbkpr@lemmy.world
    link
    fedilink
    English
    arrow-up
    13
    ·
    9 months ago

    Since half or more of reddit is now bots and shills, I don’t imagine the training data is going to be great. That’s fine, Gemini already sucks, so it’ll be hard to make it worse.

    • Dexx1s@lemmy.world
      link
      fedilink
      English
      arrow-up
      3
      ·
      9 months ago

      The data being generated now sure, but there’s still the years of actually useful data there.

      Then add on the remaining half of comments that are from sensible users and it’s a decent, and still fairly unique, dataset.

      • bbkpr@lemmy.world
        link
        fedilink
        English
        arrow-up
        5
        ·
        edit-2
        9 months ago

        There are many, many, many things posted as fact over the years on reddit that are not only untrue, but dangerous or even deadly in the case of some of the most idiotic advice given. I wish good luck telling them all apart to the poor 3rd world contractors the big commercial AI companies exploituse to “train” their stochastic parrots.

        • GluWu@lemm.ee
          link
          fedilink
          English
          arrow-up
          2
          ·
          edit-2
          9 months ago

          That was one of my favorite shitposting formats. I would type a whole paragraph with technical details and real knowledge. Only the people who actually knew what I was talking about would realize its a shitpost.

          • bbkpr@lemmy.world
            link
            fedilink
            English
            arrow-up
            2
            ·
            edit-2
            9 months ago

            Yep, and a lot of reddit is thinly veiled shitposts, bots, and uncredited karma whoring reposts of stolen content (the commercial AI companies should feel right at home here). Some of them are to anger the self righteous redditors who come to PC police anyone who dares speak against the far left zeitgeist. But most importantly, so, so many of them are just for the lols.

            The scariest part is that those drawn out, apparently accurate but actual nonsense posts/comments, is how many of them end up near the top, with massive numbers of votes from those who think “well that sounds reasonable,” but know nothing of the subject itself.

            Semi-related: I really loved the shitposts where the guy would tell an elaborate story, and end it with his dad beating the shit out of him with jumper cables. Now that’s quality reddit content.

  • reddig33@lemmy.world
    link
    fedilink
    English
    arrow-up
    12
    ·
    9 months ago

    Which is cheaper — paying a Reddit $60m per year, or just buying the company outright after the stock tanks?

  • Showroom7561@lemmy.ca
    link
    fedilink
    English
    arrow-up
    10
    ·
    9 months ago

    LOL. Why on earth would you pay millions to train AI on a large number of sarcastic, nasty, and opinionated comments typed without much thought?

      • rottingleaf@lemmy.zip
        link
        fedilink
        English
        arrow-up
        2
        ·
        9 months ago

        I’d train an LLM on my older political comments and just let it out ragin’, knowing finally that I don’t have to type those myself. It’d feel so much better to be me in that scenario.

  • DrRatso@lemmy.ml
    link
    fedilink
    English
    arrow-up
    4
    ·
    9 months ago

    I had not nuked my account history, because while reddit went to shit for me, I still used it as an info source via search. And so I wanted to leave my posts/comments in case they add value to someone else, who still uses the platform. But with this, if I am not lazy, I just might. I don’t even care about the AI training bit, rather that it is Google AI.

      • DrRatso@lemmy.ml
        link
        fedilink
        English
        arrow-up
        1
        ·
        9 months ago

        I know people said reddit restored their mass overwtitten comments, but iirc this was a brief scare due to some problem with the tool used or servers or something. I think reddit even officially commented that they are not doing this.

        • rottingleaf@lemmy.zip
          link
          fedilink
          English
          arrow-up
          1
          ·
          9 months ago

          Why would they delete for real something they store? Especially if the user agreement most likely doesn’t require them to do that.

          • DrRatso@lemmy.ml
            link
            fedilink
            English
            arrow-up
            1
            ·
            edit-2
            9 months ago

            Many people dont actually delete, but edit the content. Also, even if the data stays and gets used for training, so be it, not more you can do about it, at least I tried. But mass purging your account history still makes the site worse for searched topics and is the last bastion of resistance after shuffling off that cesspool.