RGB@group.lt to TechnologyEnglish · 1 year agoGenerative AI Goes 'MAD' When Trained on AI-Created Data Over Five Timeswww.tomshardware.comexternal-linkmessage-square42fedilinkarrow-up1134arrow-down110file-textcross-posted to: machinelearning@lemmy.world
arrow-up1124arrow-down1external-linkGenerative AI Goes 'MAD' When Trained on AI-Created Data Over Five Timeswww.tomshardware.comRGB@group.lt to TechnologyEnglish · 1 year agomessage-square42fedilinkfile-textcross-posted to: machinelearning@lemmy.world
minus-squaredystop@lemmy.worldlinkfedilinkEnglisharrow-up33·1 year agoThe best term I’ve heard to describe this this “Hapsburg AI”.
minus-squareZILtoid1991@kbin.sociallinkfedilinkarrow-up7·1 year agoSomeone should make a dataset like that called “Habsburg Diffusion” or something like that.
minus-squaree033x@lemm.eelinkfedilinkEnglisharrow-up3arrow-down1·1 year agoHah, that is brilliant, and I am stealing it.
minus-squaredystop@lemmy.worldlinkfedilinkEnglisharrow-up1·1 year agoI stole it from another random lemmy user too (lemming? Lemur? Lemmyer?)
The best term I’ve heard to describe this this “Hapsburg AI”.
Someone should make a dataset like that called “Habsburg Diffusion” or something like that.
Hah, that is brilliant, and I am stealing it.
I stole it from another random lemmy user too (lemming? Lemur? Lemmyer?)