• VitoRobles@lemmy.today
    link
    fedilink
    English
    arrow-up
    104
    ·
    2 days ago

    I got super stoned thinking about how much wasted data (storage, traffic) occurs from sharing memes.

    For example: the same meme may be hosted in a thousand places.

    Multiply that by billions of images and videos.

    • nixukty@lemmy.zip
      link
      fedilink
      English
      arrow-up
      5
      ·
      14 hours ago

      The fediverse also wastes a lot of data because each server duplicates every post and comment.

      • OpenStars@discuss.online
        link
        fedilink
        English
        arrow-up
        2
        ·
        5 hours ago

        Although ostensibly that’s not so much a bug as a feature - so that whenever one goes down, its content remains elsewhere.

        That thought will need revisiting as we get larger though, imho. Like many “downstream” instances could set their own chosen expiration dates as they receive content from the more “primary” ones… or sth.

        • nixukty@lemmy.zip
          link
          fedilink
          English
          arrow-up
          2
          ·
          5 hours ago

          I think all instances should have a sort of cache, while fetching data from the original server. then if a server is detected to go down, the instance “immortalizes” the cache. i think this reason something like this isnt done though is because it doesnt fit with the intention of activitypub

          • OpenStars@discuss.online
            link
            fedilink
            English
            arrow-up
            1
            ·
            3 hours ago

            The Lemmy devs primarily make software that they want to have working on their home instance, Lemmy.ml. How that federates outwards to other instances, something more desirable to the Western world, is a much lower priority. Which is still quite generous to offer as FOSS to the entire world.

            An instance admin can choose to immortalize content if they want, but it happens so very rarely that it’s probably not worth automating.

    • rucksack@feddit.org
      link
      fedilink
      English
      arrow-up
      88
      ·
      2 days ago

      I’m not sharing the exact same meme as others, I add my own layer of deep-fried jpeg compression to it, as it has been a tradition since ancient times.

      • rumba@lemmy.zip
        link
        fedilink
        English
        arrow-up
        1
        ·
        4 hours ago

        IPFS, it would be deadly slow.

        Let’s make a DHT meme service where everyone can share the meme torrent and you can just extract the single mem you want out of a multi-terabyte archive.

        • Sir. Haxalot@nord.pub
          link
          fedilink
          English
          arrow-up
          6
          ·
          22 hours ago

          Fun fact: A NFT is only guaranteeing a URL to an image, because the image data would take too much space on the blockchain. There is no guarantee this URL stays up or even stays the same, or that the image data is unique for that matter.

          • James R Kirk@startrek.website
            link
            fedilink
            English
            arrow-up
            3
            ·
            16 hours ago

            This was my favorite fun fact to point out during the NFT hype. Even skeptics would derisively say “you’re buying a jpeg” and I would always say “Ackhually, not even that!”

    • atomicbocks@sh.itjust.works
      link
      fedilink
      English
      arrow-up
      17
      ·
      edit-2
      2 days ago

      Back around 2010 the peoples over at Google did some math and figured out that if you took all of the electrons used for online data it would weigh about as much as a strawberry. I’ve always wondered how much that has changed.

    • SapphironZA@sh.itjust.works
      link
      fedilink
      English
      arrow-up
      1
      ·
      1 day ago

      I think these social media providers have an algorithm that can encode the screenshot from the other social media and reconstruct it on demand.

      They are doing the former for data-mining anyway. So why not do the reverse to save on stotage?

  • mesa@piefed.social
    link
    fedilink
    English
    arrow-up
    40
    ·
    2 days ago

    Sometimes I think about this.

    Im old enough to remember when reddit didnt have image access and you had to click the link to go to the other resource. like imgr and such. It acutally forced traffic onto other websites. It wasnt a bad thing to simplify the process, but eventually reddit started…owning the content. Like even images were “owned” by the algorithm. Which had copyright strikes etc…etc…

    Now we on the fedi are capable of directly posting from other platforms and interacting with them. Feels nice.