I think you're not grasping the magnitude of the change. Creating even an average quality speed-painting requires tremendous amount of expertise in drawing, painting techniques, composition, lighting, perspective. It takes years of training.
These models let anyone achieve similar results in minutes. Without any prior learning. It is not even lowering the bar, it is literally dropping the bar to the ground.
Besides, stable diffusion is able to generate not only painterly scenery, but also photographic images that are almost indistinguishable from actual pictures (certainly helped by the fact images have a heavy digital look in our era).
These models let anyone achieve similar results in minutes. Without any prior learning. It is not even lowering the bar, it is literally dropping the bar to the ground.
Besides, stable diffusion is able to generate not only painterly scenery, but also photographic images that are almost indistinguishable from actual pictures (certainly helped by the fact images have a heavy digital look in our era).