Why am I tired, and what am I tired of?
I’m tired of trolls flooding uploads with smiley and frown emoticons - and devs doing nothing about it.
I’m tired of bad reviews left just for the fun of it, or because someone couldn’t figure out how to use something and blamed the model instead. This has actually happened very rarely to me, but I've seen other great, even fantastic models flooded with bad reviews.
LoRA trainings being paused without giving any reason or explanation - ok, I'll wait, no problem, but at least show me, tell me where the problem lies - I know that my datasets are clean of TOS violations, or at least I firmly believe so.
Potential celebrity tags? Ok, I understand that CivitAI mods are in no-win scenario in this case - and they chose better safe than sorry approach. I can appreciate that, I respect that. But think about this: there are over 8 billion people in the world - and if you take any photorealistic AI image with humans in it, and you’ll find dozens of real people that resemble people in that image. That is simply unavoidable statistical reality, at least unless photorealistic generation is completely disabled. And as I just wrote, this is a Kobayashi-Maru test, a no-win scenario, and I don't envy CivitAI mods, because those are tough challenges to face.
And, let's face it: there are no viable alternatives to CivitAI right now.
Tensorart? That place is Mos Eisley multiplied by a factor of 1000 - I mean, images that look like pedoshit are not only tolerated there, they’re encouraged by being placed on the front page! That’s not an alternative. That's not normal.
I think that AI image generation has hit a huge roadblock, and I doubt that there is anything good left on the horizon.
So, this might just be it. I think I'll wait until those LoRAs that are currently paused finish training, and then give away the remaining buzz.