I'm wondering, will the companies be able to efficiently expand their datasets in the future, given how much generated slop is now floating around, which is also poisonous to the neural networks. But the models that are already trained and already existing datasets will remain, so we'll inevitably be living in a sea of shit, if we are currently in a sea of diluted piss.
Yeah, I was referring to my experience on lemmy.
I'm wondering, will the companies be able to efficiently expand their datasets in the future, given how much generated slop is now floating around, which is also poisonous to the neural networks. But the models that are already trained and already existing datasets will remain, so we'll inevitably be living in a sea of shit, if we are currently in a sea of diluted piss.