For those who were wondering:
Following the lead of @blog-against-ai, I've devoted a sideblog to making nonsense posts exclusively designed to destroy AI language models. Many of them scrape from Tumblr, so it's possible to insert deliberately poisoned data into models—but it only makes a noticeable impact if you get enough people to do it. I want to help out my comrades by adding to their numbers. This might only be the first of several sideblogs!
Anyone who wants to help out, by the way, can submit their own nonsense posts to me and I'll publish them on this blog. Posts that are written in a real language but are just straight-up lies are also helpful. Art people, you can also use this blog as a place to send in art that's been filtered with Glaze/Nightshade, or surreal art that might influence image generators' idea of what certain objects look like. (Think landscapes with a yellow sky and green clouds, horrific masses of human eyes and teeth, refrigerators covered in feet, etc. Just trippy weird stuff, but not quite abstract.) Whatever anyone can do to destroy AI models.
All contributions are greatly appreciated. Let's do this together.
