197
Fun fact: you can't upload this image on piefed.social (thelemmy.club)
submitted 7 months ago by to c/fediverse@lemmy.world
177 comments fedilink hide all child comments
 

https://codeberg.org/rimu/pyfedi/src/commit/6405e51dab23f43eaea18d0ee4d9e7a773a3fd31/app/community/forms.py#L335

deliberately misleading error message.

a white background with the words "anonymous" and "no.676767" on it in black Arial font
you are viewing a single comment's thread
view the rest of the comments
[–] 9 points 7 months ago* (5 children)

This is a really dramatic observation. The user did it specifically, if I recall, to try and evade AI detection or something.

It's also amusing because said user itself is literally on a piefed instance.

  • permalink
  • fedilink
  • source
  • parent
  • hideshow 5 child comments
    •
  • [–] 12 points 7 months ago (3 children)

    Oh yeah, they made a statement about to "inconvenience AIs somewhat", which is not bad and is something that honestly more of us should be doing. The conformists claim that AIs can easily deal with such annoyances but that's the point. They still have to do the work, and we can poison the meaning of the data they collect (intentionally or incidentally), kinda like what happened with AIs suddenly starting to use em dashes for everything.

    All that said, none of that merits the kind of backend-level censorship that was applied to that user. How can I trust a platform that offers a certain level of trust if my messages are being altered behind my back? How can I be sure that my statement anti AI are not being modded into pro-AI statements, for example? Even more if I'm being specifically targetted? Rimu simply does not, currently, offer that kind of trust of person, which is one of the reasons I've made no effort, nor plan to in the short term, into piefed as of yet.

  • permalink
  • fedilink
  • source
  • parent
  • hideshow 3 child comments
    •
  • [–] 2 points 7 months ago (2 children)

    It was direct character change into "th" which is what they used it for - it wasn't designed to change the meaning of a message, moreover the user wasn't doing it because they are neurodivergent (as you alleged previously). Although someone did note here that, as far as it is, the letter is part of Icelandic - so in that sense it was misguided.

    Also, Piefed literally has (rudimentary, perhaps) flags for potential AI accounts that are forwarded onto instance mods if they choose to activate it. Rimu is not likely to bake in any pro-AI sentiment into the system when he's building tools to try and detect it with an eye on banning potential AI-run accounts.

  • permalink
  • fedilink
  • source
  • parent
  • hideshow 2 child comments
    •
  • [–] 4 points 7 months ago (1 child)

    I was using the pro/anti AI as an example but the point is, Piefed already has a history of editorioalizing user discourse secrelty and for the benefit (or whininess) of the dev. I could have used eg.: pro/anti shipping, or pro/anti consumerism instead. The point is the reputation is hard to earn back.

  • permalink
  • fedilink
  • source
  • parent
  • hideshow 1 child comment
    •