Anything can be anything
The infinite repurposing of content forever and ever
Hi there! This month Georgia and I made a video as part of our ‘Generating High-Weirdness’ series, where we sit down and try to make something very specific with generative AI tools in two hours. This time, Georgia asked if we could make a promotional website for a sequel to the movie Clueless — as inspired by 90s web relic, the Space Jam website from 1996. If you can bare the attack on the eyes, our finished website is here.
In this video we used Claude Artifacts, which is a fairly new feature of Claude that allows you to create content (like an image, a document, or an entire website) and iterate on it in chat while it’s displayed in a separate panel. You can also ask it to make something based on existing content you give it — e.g. “make me an action RPG based on this blog post I wrote”.
In experimenting with Claude, Georgia and I got into a discussion about how the addition of consumer generative AI tools makes it easier than ever to turn one thing into something else: to make a magazine article into an 8-part Netflix series; to make a spoken word performance into an escape room. It’s the idea that anything can be anything.
It was a short discussion, and it didn’t really fit with the rest of the video, so Georgia has drawn out the main points below. Enjoy!
Our eternal struggle with The Infinite
— By
Even before the ubiquity of generative AI tools, there had been a huge transformation in how we make and consume content over the last decade or so. I wrote about this on my own blog last year — we are most certainly in a kind of Content Universe Renaissance right now. It’s not enough just to make a movie or write a book and let that be it; nowadays an entire multiverse of content, media, and experiences is created around a single concept.
You see this with the Avengers Campuses that have been developed for the Disneylands in California, Paris, and Hong Kong — these campuses are essentially immersive experiences where visitors are trained up to become heroes and be part of the story. These three experiences have distinct themes, but are all connected as part of the Marvel Cinematic Universe; they explore alternate timelines where certain events in the movies didn’t occur, and there are congruent nuances that you can only appreciate by visiting all three campuses (I’m sure there are people out there that would do that!). These are complex content universes that deep fandoms follow and obsess over.
On the lighter side, we also now have TV shows that are essentially comprehensive character studies of popular and/or complex characters from decades ago, such as the shows Ripley or Picard. It feels as though we’re making a lot of ‘original’ content by reforming and rehashing pre-existing assets.
How does generative AI fit in? In our discussion Jeremy pointed out how new AI tooling really lowers the barrier to refactoring content: where before, it would take immense effort to write a movie script based on a book, now you can probably do it in a matter of hours. What happens to intellectual ownership when you can, for instance, repurpose a coffee advert into a romantic comedy?
For simplicity, let’s look at an old book-to-movie journey: Mrs. Dalloway is a novel by Virginia Woolf. It’s set over the course of one day of a woman’s life, and is written as a stream of consciousness. In 1998 Michael Cunningham writes The Hours, as a tribute to Mrs Dalloway. Cunningham actually wanted to do a straightforward modern retelling of Mrs Dalloway, but couldn’t make it work — so instead he explored the themes of the book by telling the story of three women at once (still over the course of one day): we have Virginia Woolf herself, while she writes Mrs Dalloway, a 1950s housewife reading Mrs Dalloway, and then a modern day woman sort of living Mrs Dalloway.
Then in 2002, Stephen Daldry made The Hours into a movie, and so the story is refactored once again. We expect that Cunningham’s novel probably took months and months of teasing out themes and working them into a new story — and the same goes for when the novel was then adapted into a screenplay. It’s funny, because in 2024, this deep exploration of themes and characters feels like a very gen AI thing to do, purely because you can conduct those explorations in much less time.
That reduction of effort allows us to ask ‘okay but what was Virginia Woolf going through when she wrote Mrs Dalloway? How did women feel having read it? What would happen if someone today lived through Mrs Dalloway’s rollercoaster of a day?’ — and get answers very quickly. The possibilities are, in a sense, endless. You could fall down several deep rabbit holes in trying to explore what happened to every character in each period of time covered in The Hours: from post WWI England, to post WWII LA, to present day in New York.
What does this say for the future of content? Does everything have to be a universe now, or can we personalize these experiences and sit down and explore our favourite worlds on our own? Maybe it’s a little of both. Let us know what you think in the comments!