How does this work? Have you tried training anything on images processed with this and verified that whatever you're doing actually glitches up the resulting datasets?
Well, I've tried making an automated crawler for this, and it doesn't crawl the images from this properly - so yeah, it does work. And no, it's not because it's a PHP script. You can email me at info@kendasi.com if you want to hear how it actually works - I'm really proud of this idea, so I'd rather not just share out here how it works.
Is it possible because the URL is a php script, the crawler will skip it? I would think OpenAI could figure that level of obfuscation out.