HarryLime [any] to the_dunk_tank • 2 months agoWhat the fuckimagemessage-square61 fedilinkarrow-up170
arrow-up170imageWhat the fuckHarryLime [any] to the_dunk_tank • 2 months agomessage-square61 Commentsfedilink
minus-squareHestia [comrade/them, she/her]hexbear21·2 months agoWe're gonna be several levels deep when AI starts using AI for its reference material... link
minus-squareUlyssesT [he/him]hexbear18·2 months ago We're gonna be several levels deep when AI starts using AI for its reference material... It's already doing that right now. link
minus-squareAlisu [they/them]hexbear11·2 months agoImagine ai feeding on itself as it also shapes and influences our languages, how would that end up in 10 years link
minus-squareUlyssesT [he/him]hexbear12·2 months agoI don't see any reason why that isn't already happening, considering the Tiktok model of being recommended something else, then something else, then something else, all with social cues and expectations built right in. link
minus-squarefootfaults [none/use name]hexbear11·2 months agoLike the tower of babel where it's just random strings of words link
minus-squareCysioland@lemmygrad.mlhexbear3·2 months agoThey'll probably have to build in some sort of steganography into every AI generated response (then we'll be able to make decent AI detection tools) linkfedilink
minus-squareUlyssesT [he/him]hexbear21·2 months agoThat is a very interesting question! Indeed, here is a list of points and counterpoints to balance out the question. I hope this answer was helpful. link
We're gonna be several levels deep when AI starts using AI for its reference material...
It's already doing that right now.
Imagine ai feeding on itself as it also shapes and influences our languages, how would that end up in 10 years
I don't see any reason why that isn't already happening, considering the Tiktok model of being recommended something else, then something else, then something else, all with social cues and expectations built right in.
Like the tower of babel where it's just random strings of words
They'll probably have to build in some sort of steganography into every AI generated response (then we'll be able to make decent AI detection tools)
(I wonder if AI wrote this...)
That is a very interesting question!
Indeed, here is a list of points and counterpoints to balance out the question.
I hope this answer was helpful.