Darren Aronofsky’s AI creation studio Primordial Soup was announced in May 2025 at Google’s annual developer conference. The launch was aimed at simultaneously enhancing the pioneering reputations of both companies. Google released a new version of its text-to-video conversion model Veo and an AI video tool called Flow, while Aronofsky, like many other alarmist filmmakers, sought to forge a new and bold path into uncharted technological territory by partnering with one of the world’s largest technology companies. “Filmmaking has always been driven by technology. After the breakthrough inventions of the Lumière brothers and Edison, filmmakers unlocked the camera’s hidden storytelling power,” Aronofsky said in a classic, hyperbolic statement praising the dawn of a new art form driven by hardware rather than human creativity.
The first Primordial Soup project was a short film called “Primordial Soup”. ancestordirected by Eliza McNitt, combined AI-generated imagery, live-action photography, and computer-generated animation to create what is essentially a silly eight-minute version of Terrence Malick. tree of life. Some of the macro photographs of cells, microbes, fish, and nature are accompanied by a heavy, whispered narration from the mother to the fetus. If you start watching this movie without any context, you could be forgiven for thinking you’re watching a drug commercial.
Already infamous, Aronofsky’s pretentious and brazenly named “Primal Soup” appeared to be on hold until the end of 2025. In January, the company collaborated with the nation’s most relevant and cutting-edge news organizations to release a series of videos exploring the founding of the United States in honor of the nation’s 250th anniversary. time magazine. The series is called This day… 1776 It’s been steadily making waves online, mainly in YouTube comment sections. “Wow, this is awesome if you like shit!!!” reads one representative entry. “It looks like poop coming out of my ass” is a bit corny, but “This is the most beautiful thing I’ve ever seen since I drank all the mercury and spilled acid in my eyes” vividly sums up the experience of watching such pointless shit.
Kudos: It’s weird to use AI to celebrate the founding of a country full of incredible suckers, like asking Grok to summarize the Wikipedia page about the Declaration of Independence rather than skimming it. But it’s a better joke on paper than in reality. Perhaps the most confusing aspect of most AI-generated videos, which strive for professionalism and legitimacy, is the drive toward photorealism. This is believed to be because those who want to legitimize AI as a filmmaking or creative tool believe that the only way to impress skeptics is to fool them with something that looks like it’s not AI. One problem, in particular, is that motion is not an area where AI excels at anything longer than a minute, and even then, hallucinations such as defying the laws of physics (objects merging, floating, glitching) or accidentally adding appendages to objects are prevalent. This is true no matter who uses which program. The fact that Google’s partnership with Primordial Soup (and more recently A24) includes the use of Google DeepMind, which is responsible for accelerating Gemini’s popularity, doesn’t help matters.
the very first on this day The video “January 1st: Flag” begins with a speech from King George III. The voice comes out too loudly from the mouth of a character model whose lips aren’t moving at the right speed. Wanting to prove that their studio was doing AI the right way, Primordial Soup announced that real actors would provide voice work in the video series. This inadvertently acknowledged that the main reason for such an ugly act, in addition to fulfilling contractual obligations, was to reduce the effort and expense of actually filming a historical documentary.
Most AI videos like Primordial Soup seem more interested in working well when you look at a paused frame. Pores, individual hairs, irises, dirt and grit, clothing, wrinkles, and more are rendered with the convincing detail most people expect from a digital camera. It’s not until you run the video that you see all the now recognizable artifacts of the AI-generated clip. Foreground and background elements shift from highly detailed fidelity to soft fuzziness. Very short shots with little continuity between shots. Dead eyes. A sudden change from one expression to another. And that weird lag in body movements that makes it seem like the character is loading.
None of this even mentions the fact that these videos, which are less than five minutes long at most, have freeze-frame stingers introducing key figures like George Washington and John Adams, royalty-free Renaissance Faire music, and no consistent visual style. Extreme close-ups, slow motion, camera movements that simulate gonzo, and uniform orange lighting contribute to extremely boring content, an eerie spectacle in which the information conveyed is secondary. That is, whoever was in charge of the uploads after the first few installments. on this day Go to video time‘s YouTube channel abandoned the concept of the calendar day and tried to do some vile SEO farming like “British Plan to End America in One Operation,” “How Ordinary People Were Forced to Participate in the King’s War,” and “John Adams Had to Turn a Drunken Argument into a Revolution.”
Intentionally or not, the vision of colonial America presented through these videos is both incredibly idyllic and frightening in its homogeneity. The Thomas Kinkade quality of the natural scenery clashes with the repetition of the same facial expressions, generated to look more like modern actors than plausibly weathered historical figures. The tone of these videos is meant to be melodramatic and humorous, but whatever story they’re trying to tell is sketchy and complex. Not to mention the fact that each video attempts to portray America as built on the courage and defiance of colonial settlers, with parades of squinting white men fighting British tyranny, and Native American, African, and Indian equivalents of this history sprinkled like extras in the background. This is, of course, an old and familiar kind of alienation, but it’s interesting to see it summarized in such a strange and perhaps authorless form.
in financial times Interview with Aronofsky and Demis Hassabis, CEO of Google DeepMind. Hassabis is doing the same thing that AI evangelists ultimately do when touting the potential of their favorite technologies. My emphasis:
”the currentthese tools cannot give you a fresh start. They are extrapolations of what is already known. Can these systems actually come up with new guesses, rather than just solving existing ones?The answer is right now No. obviously there is still What these systems lack is the out-of-the-box thinking and true invention that great creators, whether scientists or artists like Einstein or Picasso, are capable of. ”
This kind of thinking erases everything interesting and beautiful about human creativity in favor of a mindset that only looks at input and output. The marriage between a fellow engineer and an artist, in this case Hassabis and Aronofsky, is meant to show the implicit dependence of two parts of a whole, one on the other.
Except no one needs that kind of garbage This day… 1776or whatever else Primordial Soup eventually produces. At another point, F.T. In an interview, Aronofsky compares AI to hip-hop sampling, which is clearly an example of nonsense given that sampling is a product of human curation, with artists intentionally choosing which samples to use and how. AI has no perspective or sense. There is no instinct or intuition. In that sense, it reflects the personality of the person who uses it avidly.
There are fewer and fewer moments of levity when it comes to making fun of AI. This shit hits us from all sides, not just our devices, ads, and emails. Even if the project is ultimately unpopular and frivolous, on this dayThis 40-second trailer still has less than 250,000 views, the highest in the series, and it’s no wonder it’s considered insidious. This patronizing of myths is never a good thing. But if it’s doomed to be bad, I’d like to see what kind of evil humans can come up with.
