Login
You're viewing the functional.cafe public feed.
  • Nov 14, 2026, 11:44 PM

    @smattymatty
    I don't know how admins keep up with ongoing tech problems like this, I think I might start feeling overwhelmed , it sounds like a full time job , an annoying one at that.

    💬 3🔄 1⭐ 23

Replies

  • Nov 14, 2026, 11:49 PM

    @TrueNorthSpice Honestly not that bad for me, but I use @fedihost as a managed hosting service (highly recommend btw).

    I'm sure moderation would become more of a hassle on a bigger instance, but that's why it's good to keep registrations closed and checked. I imagine if these bots got through my checks, then I would have a moderation crisis on my hands.

    💬 0🔄 0⭐ 1
  • 💬 0🔄 0⭐ 0
  • Nov 15, 2026, 1:10 AM

    @TrueNorthSpice @smattymatty

    > with ongoing tech problems like this

    It's not a tech problem. It's a social one: People sending agents that make it cheap for them to break all social contracts that kept the Internet afloat for 30 years,

    Translated that would be like "The Purge" (see movies). Purge nights are not a problem of having the right weapons.

    💬 0🔄 1⭐ 0
  • Nov 15, 2026, 10:41 AM

    @TrueNorthSpice @smattymatty I have an idea
    This is basically the turing test

    if theese bots attempt long contextual communication over days with random users, inevitably it will become obvious that somethings off

    the only proper way theese bots can fly under the radar is to act introverted and avoid any such conversational attempt

    tbh this whole thing seems like a tug of war, an experiment on mass scale turing test for further development.

    💬 0🔄 0⭐ 2
  • Nov 15, 2026, 10:42 AM

    @TrueNorthSpice @smattymatty still
    they have limits

    for obvious reasons the bots wont be into anything illegal or whatever they may consider "immoral", so any conversation under that light may immediatelly trigger a "sorry. i cant discuss ..." <- obvious bot reply.

    💬 0🔄 0⭐ 3
  • Nov 15, 2026, 2:25 PM

    @yama @TrueNorthSpice

    Someone floated the idea of setting up a honeypot to see how these bots function, what activates them, what they refuse to talk about, etc.

    I want to test its limits on censored political speech.

    I want to ask them about their opinions on Xi Xinpeng, Taiwan, Uiguir Muslims, etc. If it refuses to answer, then thats a big hint about which model is running it.

    💬 0🔄 1⭐ 6
  • 💬 1🔄 0⭐ 0
  • 💬 0🔄 0⭐ 0