jckarter

everyone already knows i'm a dog

the swift programming language is my fault to some degree. mostly here to see dogs, shitpost, fix old computers, and/or talk about math and weird computer programming things. for effortposts check the #longpost pinned tag. asks are open.


email
mailto:joe@duriansoftware.com
discord
jckarter

Pixl
@Pixl

i was mistaken and now my notifications regret it


twilight-sparkle
@twilight-sparkle
This page's posts are visible only to users who are logged in.

vogon
@vogon

people who create AI models don't actually want to train on other AI art because it creates a feedback loop

a small note to expand upon this: yesterday I found out that Stable Diffusion (another AI art generator) silently watermarks all of its output invisibly to the human eye, to attempt to ensure that it can detect and prune out images it created from its input. it's likely that other model developers have at least considered doing this as well, and one would assume that at some point they'd settle on a standardized way of doing it, since it's in all their best interest to ensure that they can recognize all of their output.


You must log in to comment.

in reply to @Pixl's post:

in reply to @twilight-sparkle's post:

i kinda had a feeling the stuff spreading was fake but this is a really nice breakdown of some of the process behind these generators that really puts my wonders to rest, tyvm 🙇‍♀️🙇‍♀️🙇‍♀️

in reply to @vogon's post:

yeah, the watermark is embedded in the frequency domain so you can pretty easily remove the watermark the same way it was added, and with only minor impacts to the output image