Until one can produce work that makes an impression with some precision one has to have experience in the medium though - and different media are different regarding to what that means.
With illustration and representative art it starts with something ‘reading’ correctly, i.e. whether the intended representation even gets to the recipient. And then there are more layers on top of that getting ever more meta.
Someone who can put a urinal on a pedestal and cause an uproar in whichever direction has a lot of experience - but if a picture is just a picture or a urinal is just a urinal, it’s not worth looking at much, except for its engineering. Good art doesn’t have to be on that level though, entertainment can also be good art (but a lot of it isn’t) - there, it’s about resonance.
You’re right that craftsmanship alone cannot produce good art, there is something else driving the desire to hone craftsmanship, which is maybe to better be able to express what was impressed on the artist through life. Something that resonates with the artist is made with the hope it also resonates with other people, art is a social endeavour.
But I also feel that to a large extent, honing the craft also hones the intuition (and some knowledge as far as it can be distilled) for what makes things resonant with others. I make myself into the diffusion model to resonate with what I’m making while making it, you feel each curve you put to paper or canvas, you feel the tension in a pose, the impact of a composition - the resulting art is what’s there when that process is abandoned.
I feel like a vegan about the currently available models - once there is something made from public domain art only I’ll experiment. But right now I’m sitting in front of them like a vegan in front of sausage: For others the result is food but for them, they just see the process turning individuals into sausage.
Star Trek TNG had it pretty right in terms of what’s moral or what is desirable