@davidrevoy@framapiaf.org @netbsd@mastodon.sdf.org @atraidez@infosec.exchange I don't think it was intended as a scam; the original research behind it is logical, and its initial release was able to defeat the current state of image2image generators.

Unfortunately, it's fundamentally incapable of being a long-term solution to the problem; it doesn't matter if it can block all of the current LLMs, because as soon as even one person finds a way to defeat it, every image that had previously been glazed is effectively now completely unprotected, and this exact scenario has happened several times.

I think the developers behind it mostly just have a lot of hubris and think that they will be able to come up with an unbreakable solution if they try hard enough. (I am also a little biased because I tried to reach out to them about helping make a Linux port in the early days of it, and I was fairly rudely rejected)