How to build a fediverse community when bots are indistinguishable from humans on applications to join?

  • tofu
    link
    fedilink
    English
    arrow-up
    5
    ·
    17 hours ago

    If nobody on your server follows them, they don’t cause any load to your server. Also admins can suspend the whole instance.

    It’s still going to waste electricity etc, but it can easily be cut off the fediverse.

    • Onno (VK6FLAB)@lemmy.radio
      link
      fedilink
      English
      arrow-up
      4
      arrow-down
      1
      ·
      17 hours ago

      I understand.

      My point didn’t state that all instances would be affected equally, just that there’s an effect everywhere.

        • Onno (VK6FLAB)@lemmy.radio
          link
          fedilink
          English
          arrow-up
          3
          arrow-down
          1
          ·
          11 hours ago

          Let me ask you this.

          If the LLM instance doesn’t talk to anyone, you’re right, but there seems little point in building an instance that talks to nobody.

          This leaves us with an instance that does talk to other instances, presumably responding to posts, making its own and subscribing to communities. This already costs money for each “touched” instance.

          At that point the administrator of an instance that doesn’t want to federated with the LLM instance, has to defederate from it, updating their instance and then still getting access requests from the LLM instance when it attempts to do the reply, post, community thing as described before.

          Even us discussing the phenomenon right now takes server resources across the fediverse.

          In other words, as I said, there is always a cost to everyone.

          • realcaseyrollins@hilariouschaos.com
            link
            fedilink
            English
            arrow-up
            1
            arrow-down
            1
            ·
            2 hours ago

            At that point the administrator of an instance that doesn’t want to federated with the LLM instance, has to defederate from it, updating their instance and then still getting access requests from the LLM instance when it attempts to do the reply, post, community thing as described before.

            That is a cost that, historically, the type of instance that would block an LLM instance just for having AI bots has no problem with expending fourfold.

            Since I presume that few users would try querying an LLM instance, and therefore few instances would likely even touch the LLM instance (again, remember, the posts would be unlisted), this is pretty much an imagined problem, even if the hypothetical LLM instance were to exist in reality.