[Tanoc]

This is terrifying to us, but not in the way you think. Tools like this place selective pressure so that increasingly only the best of the best get paid to make stuff, and their stuff influences the data sets in the future. The rest of us get forced out simply due to cost unless we're doing it as a hobby. Some artists know all too well their limits, and those that do are the most vulnerable to things like this, because these learning machines use the ever higher ceiling of human ability to determine the floor of their ability. You can't go back to the funky weird pasta men of Stable Diffusion's 2021 iteration as the default for example, but someone who started painting in 2002 might still have the same level of skill now as they did in 2010.

Text was paralleled first. Then sound was paralleled next. And now image is being paralleled. It will be on level with highest percentile of human ability just as text and sound were before it. Game devs and comic artists are already replacing texture and background artists with AI generated images, just as they used AI to create hundreds of thousands of lines of fluff dialogue before then.