The Art is in the Work
I recently read a published screenwriter who commented on the ubiquity of the “where do you get your ideas” question and it reminded me of AI enthusiasts. The truth is ideas are easy, because there are very few original ideas. Imitative AI hype people don’t seem to recognize that simple truth, and as a result, miss what makes art.
I have a section on OneNote (because I am a dork and everything is digital in my life) that lists nothing but ideas. As of now, it has 102 items in it. Some of the more “elevator pitch” ones include:
Zombies taste of chicken (a zombie apocalypse where people coral zombies for food)
Fuck English Magic book (a cross between Luddite history and Johnathon Strange and Mr. Norell)
Raising the dead and giving them the choice to join an eco-revolution.
A magical healer finds he can actually use magic to kill — the only person in the world who can do that.
I am not afraid to list these because ideas are easy. Shakespeare stole every plot he ever had form history and fairytales. Execution is hard. I have not failed to be published because I don’t have ideas but because I write for shit. Or at least, write for not good enough to get an agent, which is a difference that makes no difference. Better writers than I have likely used similar ideas to get published. And that is why I think that imitative AI has severe limits artistically.
Art is about the process, not the idea, not the result. What makes Shakespeare Shakespeare, what makes LeGuin LeGuin, what makes Octavia Butler Octavia Butler are not the ideas but the writing. And the writing is unique to that person. They bring their emotions, their perspective, their history, and their way of looking at the world to every piece they create. Imitative AI cannot do that.
Imitative AI merely picks the most likely word (to oversimplify the math) that comes next from their training data. They cannot bring any perspective to work — at best they can produce the mushy middle consensus or the median of a specific style. That might be okay for summarizing emails or rote responses to simple business tasks, but it won’t produce anything resembling good art. The math has nothing interesting or unique to add to the piece. It merely regurgitates what has come before, adding nothing to it.
The people who push imitative art as artistic are very much like the people who tell published authors that they have a great idea and will split the proceeds fifty-fifty and all the author has to do is write the book. They don’t actually understand what makes a piece of art good. I have no doubt that they will make some headway in pushing Hollywood and publishing execs to use their tools to “start” written works and let writers come in and finish them. I am also confident that even the schlockiest of movies will generally do better than these pieces, assume the writer just doesn’t redo the whole thing, because ideas don’t really matter.
The execution is what makes something good, and the execution is inextricably tied to what makes a person that person. A word or pixel calculator simply cannot imitate that.
Nice essay. I would only caution against the 'writes like sh*t' internalization/characterization. As you allude to, its much more of a sociological dynamic than that. In other words, its *them*, not *you*.
A recent review (of Hecht's works) from David Orr in the latest New York Times Review of Books touches on this topic. In pertinent part, Orr wrote:
The question for poets is always, "How do I write poetry?" — and for a long time, the answer, provided in part by Eliot, was, "By knowing a lot about poetic tradition and making a show of it." It's not the most obvious answer (the more one thinks about it, the stranger it seems). But for Hecht's generation, it was a reliable answer that led to measurable rewards and it did so because a cohort of poets, editors and critics agreed that it would...
...Yet the question "How do I write poetry?" and the question "How can I be seen and respected as a poet?" aren't the same. In fact, they're frequently in tension, because the preferences of the "club" are so twisted by that group's tiny size and self-dealing that to satisfy them often says more about acceptability than artistry.