It does feel like art's disruptive "Calculator moment" is happening where you can now leave a lot of basic/mechanical tasks to a tool and give more focus to higher-minded problems.
It's going to get so cool and interesting, I think.
A lot of the conversation around art may focus more on composition and objectives of the artist in the new prompt engineering world, with less bias from factors such as rendition quality etc. creeping in since it's so incidental.
New forms of art will emerge and/or gain popularity that focus on trying things the tools aren't good at yet. The human artist of the gaps. The niches will constantly be shifting.
I wonder if we'll learn to recognize the output of certain popular models and perceive them as instruments. "Made by xy on z" instead of "xz on guitar", so to speak. I remember the 90s/early 00s internet when it was always easy to tell when something had been done on Flash, just because of its line anti-aliasing rendition style being so distinct and familiar.
The novelty will wear off, and we'll all start to feel a bit disappointed that the average human's imagination is pretty limited and novel/original ideas remain somewhat rare as the patterns and tropes in all the generated art emerge. It's great you can put the space needle where you want it and get a good-looking city and space ship, but how many variations of a cyberpunky skyline with a space ship do you need? And then we'll celebrate the novel stuff that does happen, as always. I suppose the tropes will evolve faster as the throughput goes up.
>basic/mechanical tasks to a tool and give more focus to higher-minded problems.
>rendition quality... [is] so incidental.
There's this thing in painting called 'mark making' and it can be the difference between an all-time-great painting and a throwaway portrait. Mark making speaks to every momentary choice of physical process a painter employs and reveals their thought process. For some of the greatest painters, it reveals their genius.
Do not discount execution. Overlooking "basics" and "mechanics" is what results in disappointing work.
It's a fair point, and thanks for teaching me a new term!
There's a lovely documentary called "Tim's Vermeer" about Tim Jenison's - one of the founders of NewTek, the people behind Video Toaster and LightWave, incidentally both tools that made hard visual art tasks accessible to wider audiences - hobby side project to prove that Vermeer used sophisticated optical tools to capture and copy his scenes from physical sets, rather than e.g. paint his famous grasp on lighting purely from his own mind. He builds such tools himself and then proceeds to successfully create his own Vermeer-alike painting, despite possessing very artistic skill himself.
It's full of good ruminations (and good at sparking more) on tools-vs-artistry but also execution-vs-method, and whether designing and adopting innovative tools and the tedious process to use them made Vermeer less of a genius, or just a genius of a different kind than otherwise presumed.
It's very accessible and doesn't require knowing anything in particular from the art world.
Tim's Vermeer is kind of bad in my opinion. A lot of the musing border on misinformation. If you're not a painter, it sounds great, but if you have some training, it's a very frustrating doc. The resulting painting is neat, but it was immediately obvious (at least to my eyes) how different his result was from Vermeer's.
Hockney, one of the featured 'expert painters' is a hack who doesn't actually know how to paint* and therefore claims that certain gradations are certainly impossible without some sort of additional lens device. Meanwhile there are 19 year olds at the Grand Central Atelier pulling off just that.
I own a camera lucida. It got in the way more than it helped. At best, it's a novelty, now collecting lots of dust. Vermeer probably had one but it's altogether way more likely that he just had a well-trained observational/representational faculty. There are some killer painters using cameras now (Will St. John for one) but they typically have a decade or more of very rigorous direct observation to lean on.
I suggest following Ramon Alex Hurtado. IMO he's one of the more exciting young scholars on historical representational painting techniques. I don't think he has written anything himself yet, but he does do workshops and has a big informational update for his website coming.
*the definition of painting is now so broad it is meaningless. Here, I mean "attempting some degree of visual accuracy" which can be achieved in endless creative ways. Compare (easily on instagram):
- Colleen Barry
- Peder Mørk Mønsted
- Cecelia Beaux
- Jas Knight
- Felicia Forte
- Eric Johnson
- Ksenya Istomena
- Sergei Danchev
- Glenn Dean
- Blair Atherholt
- Jose Lopez Vegara
- Hongnian Zhang
- Hans Baluschek
These artists all have their own voice and stylistic choices. They also all represent things they see with some sincere accuracy. Look up Hockney's ipad paintings he got lauded for. People treat them like they're some misunderstood genius, but really, they're just bad paintings.
I'm sure he's a sweet old man and I'd drink tea with him. But if it weren't for his ilk I might have found proper instruction 10 years earlier in life. Modernists and postmodernists robbed generations of proper art instruction. Imagine if all the music teachers burned all the sheet music and refused (or forgot how) to teach the diatonic scale. "Hit the keys in a new way! Don't let yourself be bound by conformist ideals!"
> Do not discount execution. Overlooking "basics" and "mechanics" is what results in disappointing work.
That is what really bothered me in art lessons in high school. When discussing any famous work it was always about concepts, ideas, composition,... and execution was very much secondary. But for your own work all that is completely ignored if your coloring is just slightly uneven or lines are too rough. If you could hand in a photorealistic drawing of anything, no matter how boring, that would give you much higher marks than a rough drawing of something worthwhile.
It's not as if generative art is new. Nor is figurative painting relevant anymore since the invention of the camera. A basic Burger joint in Gerhard Richter kind of style transfer is very much derivative. This isn't bad in view of the classics, but it's more like art-work to me.
The true artists in this one are the coders, no doubt (corrolar to the inteligence debate).
On the other hand, you mention an important point with layout but you underestimate the progress these days. Surely there are companies who are working on automated design beyond CAD (computer aided design), eg. for specialized antenna.
> we'll all start to feel a bit disappointed that the average human's imagination is pretty limited and novel/original ideas remain somewhat rare as the patterns and tropes in all the generated art emerge
Well, one might argue that Richter's most highly priced piece looks a little like prehistoric art of the pleistocene. It's a little vain to mention it, because I can much better relate to the more basic form, of course. A more frequently sore point would be the pop music industry between professionals and the amateurish.
Anyway, this may be thinking too big. For the time being, the bunch of techniques is better understood as a toolbox, because it will be a long time before it trumps demo-scene productions, for instance. Here it is the technique that counts more often than not. The rest is an acquired taste.
It's going to get so cool and interesting, I think.
A lot of the conversation around art may focus more on composition and objectives of the artist in the new prompt engineering world, with less bias from factors such as rendition quality etc. creeping in since it's so incidental.
New forms of art will emerge and/or gain popularity that focus on trying things the tools aren't good at yet. The human artist of the gaps. The niches will constantly be shifting.
I wonder if we'll learn to recognize the output of certain popular models and perceive them as instruments. "Made by xy on z" instead of "xz on guitar", so to speak. I remember the 90s/early 00s internet when it was always easy to tell when something had been done on Flash, just because of its line anti-aliasing rendition style being so distinct and familiar.
The novelty will wear off, and we'll all start to feel a bit disappointed that the average human's imagination is pretty limited and novel/original ideas remain somewhat rare as the patterns and tropes in all the generated art emerge. It's great you can put the space needle where you want it and get a good-looking city and space ship, but how many variations of a cyberpunky skyline with a space ship do you need? And then we'll celebrate the novel stuff that does happen, as always. I suppose the tropes will evolve faster as the throughput goes up.