Interesting analogy. The obvious reason so far is that GenAI has very generic output, and struggles massively with specific prompts. Errors aside, even when it gets it right it tends to be like some kind of average of the art style, character or scene. This still has its uses but for any interesting art project or product, the expectation is something unique and novel.
If GenAI can overcome its generic output issue and somehow can tap into some algorithm that gives it more creativity, then I think it will not be like polyester, and it will revolutionize art like photography or film did in the early 20th century.
These past 2 days I have seen some very entertaining short videos made with veo3 in my feed, I'm surprised. Brazilians are creating some really funny mashups of old historic moments with current brazilian culture and slangs. I can see a creative editor making a top quality movie with it.
pirate speak was only fun for a very particular small subset of LLM users, the early adopters. Those videos I'm talking about are as good as any high quality movie scene.
True, it's definitely another tool for creative people already and will at least create a new genre in many categories. The question is whether it will become the dominant force for creative works across the board.
Reminds me of the Transcendental Painting Group in the early 20th century around Santa Fe, New Mexico, whose members explored the nature of substance and form and spirituality within the total abstract.
It was a reaction to the fact that too many contemporary painters in the region focused on landscapes, capturing the beauty of the area through rigid photorealism. This was seen to TPG as derivative and completely missing the essence of painting; after all, cameras were now able to take increasingly vivid shots of natural landscapes, and as such the value of such paintings began to decline.
An excerpt from their manifesto:
The Transcendental Painting Group is composed of artists who are concerned with the development and presentation of various types of non-representational painting; painting that finds its source in the creative imagination and does not depend upon the objective approach.
The word Transcendental has been chosen as a name for the group because it best expresses its aim, which is to carry painting beyond the appearance of the physical world, through new concepts of space, color, light and design, to imaginative realms that are idealistic and spiritual. The work does not concern itself with political, economic, or other social problems. Methods may vary.
Your observation is astute: The recent revolution in generative art and human interfaces is here to stay, and it is the next disruptive and contemporarily misunderstood evolution in art, just as film was to painting. Regardless of how current-generation artists feel, next-generation artists will be born into these tools and adopt them without question, whether directly or through subversion.
One future is in realtime hypercontextualization... Art installations which prize subjectivity more than TPG could ever hope to achieve in their time, creating the abstract not from the mind of the artist, but the observer. Art which is not just observed subjectively, but created subjectively, the observer being able to fully experience themselves from new angles, guided by the hand of the artist.
These installations may be physical or digital, and will use all sorts of signals as input. Local signals, remote signals, colors, shapes, sounds, brain waves, weather patterns, HUMINT, data dumps, trending topics... you name it. Any information will be fair game for integration and resynthesization. Observers will weep, will walk away with a new feeling or realization about themselves, will stay up that night staring at the ceiling and contemplating deep, unearthed aspects of themselves. And for some installations, the observer may continue to participate with the installation over a period of time, whether in person or digitally, and sometimes in a way which incorporates the interactions between observers. This kind of experience is only possible at scale with generative art.
And while still a pipe dream 10 years ago, it's become an increasingly viable reality, especially with the recent upgrade to GPT 4o's generation capabilities, or tools like Sora, or the incredible community tooling around Stable Diffusion, etc.
If GenAI can overcome its generic output issue and somehow can tap into some algorithm that gives it more creativity, then I think it will not be like polyester, and it will revolutionize art like photography or film did in the early 20th century.