Raw LLM Responses

​@Ratoons_wasTaken I am so sorry in advance, this is going to be a big response, I tried to make it as clear as possible and apologize if its a bit overwhelming: It didnt make it worse, I did not elude to this at all. i tested it individually with both the unpoisoned version and the poisoned versions on various intensities and poison times. The strongest and the weakest and it didnt stop the AI from doing stuff with it, the results were pretty much the same. I did not say anything about it still making the AI worse or give any indication of that. The problem with AI messing up hands is not because of image poisoning but a lack of refinement in the AI software itself, or lack of more data. They are getting better and many of them know how to draw things way better than ever before and even draw hands as well with out a hitch of a problem, and they are going to keep improving. Please I'm on your side about this, I'm only telling you this so that you don't blindly believe something works to protect you when its not doing as much as you hope. I am also not trying to be a doomer about it. I know its really hard right now in the state of things right now, and the threat seems massive, and we want to cling onto hope wherever it comes for us to make things a little easier but I really just want to bring light to this so that we can be accountable about it. Not just in how much we are putting ourselves out there to the risk but also to hold accountable the software developers giving this hope that we are safer than we are. If we are honest about its vulnerabilities we can use that information to make better poisoning software, to improve things so that we can absolutely make sure we are all safe. However I also wanted to address that there is a fundamental issue with some of this poisoning work. I'm going to try to explain it as clearly as possible. AI programs get developed at a very fast rate and eventually learn how to reverse engineer or get around poisoning, if the poisoning software does not update recently enough it leaves work potentially vulnerable. And just because you poison one work a bit ago and then had uploaded it at the time, it stays on the internet, it is there even after AI learn how to reverse engineer it. Then it is vulnerable again. And also if nightshade and glaze finally catch up to the AI and are able to challenge it again, those works that you uploaded with the previous version poisoning will always still be at risk. Each image would have to be reprocessed and reuploaded every single time nightshade/glaze is finally able to challenge them again. This is all besides the original point I tried making before which is as it currently stands, I tested glaze and nightshade poisoned works with stable diffusion and LM studio Yesterday. Right now. At This Moment In Time, nightshade and glaze's ability is unable to properly poison the images enough for it to deter the AI's ability to work with it. I've still yet to test it with other software but that is the conclusion I was able to make just by running it through myself. I promise I am only sharing this information because I too am a scared artist, and I too am afraid of my works, my voice, my image being stolen. I wanted to share this information so that everyone knew. I encourage everyone to please test their works with other AI software as well to make sure its safe. And please be careful where you are uploading your works, like on social media websites that automatically lift peoples posts and sell them to 3rd party or train them on their own house made software. Thank you for anyone who took the time to read this and try to understand where I am coming from. I really am hoping the best for all of us and anyone else who wants to continue being a public artist on the internet, I am sending my best regards to everyone. Truly.