Even if it will plateau, same was said with moorrs law, which held up way longer than expected. There are so many ways to improve this. Open source community is getting to the point where you can actually run decent models on normal private hardware (talking about 70-120b model)

Portal buyers would be suspicious

You can run a 7b model on cpu really fast even on a phone.

doesnt work anyway lol

you need to learn the difference between opinion and fact then

Still its neither implemented in a1111,fooocus and comfyui. I acknowledge, it exists and certainly works, but was not adapted by any tool i know of

thats almost a year old. and their github 2 years. also i did train multiple loras with only ai generated images with sd1.5 and XL without any issue

Screenshot and its gone

Thats not how it works at all

Not really. Check midjourney v6 generated images. I found many images, which look undistinctable from real images. So i dont see, why image generation should get worse. What matters is the dataset and only dataset. It doesnt matter if the model is trained on ai images, as long as the dataset is good

Sooo how long until theres a plugin in a1111?