Underneath Hollywood, a native creator scene grew up on these tools, and it has its own canon, its own stars, and its own sense of humor.
It even has a benchmarkBenchmarkA standardized test used to compare models on a task, such as prompt adherence or image quality.. In March 2023 someone posted a grotesque clipCLIPA text encoder developed by OpenAI in 2021 that learns to align text and images in a shared embedding space. Foundation of most text-to-image models from 2022 onward. of Will Smith eating spaghetti, made with an early open model, and it became the community's running yardstick: every new video model gets a Will Smith spaghetti re-test, and the side-by-side against the 2023 original is how the scene narrates its own progress. The other early landmarks were memes too, the Balenciaga pope that briefly fooled the internet, the 'Harry Potter by Balenciaga' format, and in 2025 the fully AI-native absurdism of 'Italian brainrot.'
The scene's biggest mainstream moment was the Ghibli trend of March 2025, when OpenAI's new image generation let anyone turn a photo into a Studio Ghibli frame and demand melted OpenAI's servers, with Sam Altman posting that 'our GPUs are melting.' It also reignited the ethics fight, because the resurfaced clip of Hayao Miyazaki calling an AI animation demo 'an insult to life itself' became the anti-AI movement's banner, even though the 2016 clip was about a different technology entirely.
And it produced real artists, not only memes. The most cited is Neural Viz, a single creator who builds an entire mockumentary universe of post-human aliens, writing, performing, and directing every character, and who is held up across the discourse as proof that authorship survives the tools. PJ Accetturo made a fully AI advertisement for the prediction market Kalshi that ran during the NBA Finals for a claimed two thousand dollars, and became the face of AI advertising. Curious Refuge, founded by Caleb and Shelby Ward after their viral 'Star Wars by Wes Anderson' trailer, turned AI filmmaking into a school. Runway's AI Film Festival grew from a few hundred submissions in 2023 to thousands, and screened its winners in IMAX theaters.
What the scene believes, by mid-2026, is remarkably coherent, and it rhymes with this book's thesis. Direction beats prompting: the human choices, script, performance, and edit, are what separate a film from slopSlopAI content shared without human review or intention, defined less by the tool that made it than by the missing step of someone curating it before it goes out.. Consistency is the craft frontier. Taste is the scarce skill, knowing which one generation in a hundred serves the story. And the best work designs around the models' weaknesses rather than fighting them, which is exactly why Neural Viz's characters are talking-head aliens whose lack of full-body consistency never shows.
Check your understanding
pass: 5 of 7
Answer at least 5 of 7 correctly to unlock the next chapter.
1. What became the community's running benchmark for video model progress?
2. What was the scene's biggest mainstream moment, in March 2025?
3. Why did the Ghibli trend reignite the ethics fight?
4. Who is the most cited real artist to emerge, building a mockumentary universe of post-human aliens?
5. What did Curious Refuge, founded by Caleb and Shelby Ward, grow out of and become?
6. What does the scene believe is the scarce skill by mid-2026?
7. Why is Neural Viz's design of talking-head aliens a smart choice?
The weekly briefing
Get the week's moves in your inbox.
A short, sourced digest of what actually moved across generative AI, every week. Free.
Free. One email a week, no spam, unsubscribe anytime. Prefer a reader? RSS.