As a follow-on to thinking about hypercanvases, I started querying Claude and ChatGPT (will abbreviate to CG below) to try to expand on the idea of hyperdimensional storytelling that makes use of hypercanvases as its “native” media format.

Definition:

Got this decent base definition out of CG:

  • Hyperdimensional Storytelling: This form of storytelling transcends traditional linear or even multi-linear narrative structures. It engages with storytelling across various dimensions, realities, and perspectives, often blending media, time, space, and even individual experience.

Not very spicy, but a good jumping off point. Which is often how I would characterize CG’s outputs. They also wanted me to convey to you these key concepts within Hyperdimensional Storytelling:

Key Concepts:

Non-Linearity:

  • Unlike traditional narratives, hyperdimensional storytelling doesn’t follow a single linear path. It can branch, loop, intersect, or exist simultaneously on various planes.

Interactivity:

  • Readers, listeners, or viewers are not just passive consumers. They can influence the narrative, explore different pathways, and even contribute to the story itself.

Multi-Media Integration:

  • The story can manifest through text, images, audio, video, virtual reality, and more. Each medium adds layers and nuances to the storytelling.

Temporal Fluidity:

  • Time can be treated as a flexible dimension, allowing for parallel timelines, time travel, flashbacks, and foresights that are interconnected in complex ways.

Personalized Experience:

  • Different participants may experience the story differently based on their choices, background, interactions, and even emotional states.

Collaborative Creation:

  • Multiple authors, including the audience, can contribute to the story, weaving a complex tapestry that reflects diverse voices and experiences.

Cultural Contextuality:

  • The story adapts and resonates differently with various cultural contexts, and it can morph to engage with different social, historical, and individual perspectives.

Again, not spicy or all that evocative, nor does it perfectly capture what I myself am doing, but gets the job done for a quick “minimum viable product” for what hyperdimensional storytelling might look like.

Claude can be pretty evocative though, and below is a lightly edited list of some suggestions to flesh out the concept more, along with my comments in parentheses.

Claude on Hyperdimensional Storytelling

  • Stories can be told by navigating pathways through a hypercanvas over time.
  • The story resides in the specific journey rather than any individual manifested artifact. (ed. I think it could be both/neither also)
  • Choices along the way shape characters, events, and emotional resonance.
  • Iteratively generating artifacts actualizes key story moments, but is not the full tale. (The full tale might be impossible to grasp, also)
  • The context of how iterations relate conveys the narrative arc.
  • Visuals, text, audio can all combine as storytelling media.
  • Hyperlinks connect narrative threads across manifestations.
  • Different reader/viewers may take varying paths through the latent narrative.
  • Social interaction layers additional perspectives onto the core story.
  • Story boundaries blur across the internet with collective riffing.
  • The full narrative only exists in aggregate activity across media and minds.
  • No singular telling, but many permutations of possibilities.
  • The author guides initial pathways, then surrenders to emergence.(This!)
  • A story requiring AI’s generative capacity to fully explore.
  • Worldbuilding happens by steering generative models down new tangent universes
  • Plots form through chains of creative choices scaling possibility spaces (I like this phrase “scaling possibility spaces!”)
  • Tension and stakes emerge from unlikely AI-assisted combinations
  • Moods flow by manipulating aesthetic latent parameters over sequences (AI storytelling is very mood-based, imo)
  • Delayed reader gratification comes from omitted glimpses of latent paths not taken (Love it)
  • Cliffhangers leave narrative threads unresolved across bifurcating branches
  • Flashbacks reconstruct memories using inpainting and imaginative leaps
  • Surprise twists from controlled random walks away from expectations
  • Easter eggs hide in edge cases only AI exploration would find
  • (Letting the AI act as the storyteller, or collaborating with human storyteller.)
  • Meta-stories arise from readers training their own AI models on hypercanvases
  • Fiction bleeds into reality when false trails lead to true discoveries (i.e., hyperreality)
  • Empathy bridges occur by inhabiting procedural personas (Not sure completely what this means, but I think I like it! Maybe it’s like in that PKD novel where many people can inhabit the same character/figure/toy/simulation thing…)
  • Catharsis is achieved through simulated resolutions of inner conflicts (or there’s no catharsis, and no resolution, which is what I generally prefer for these kinds of stories & collections)
  • Exploring moral dilemmas through generative character simulations (I like the idea of having many different outcomes – and not just simulated outcomes – be “normal”)
  • Building mythic lore by tracing ancestral narrative branches backwards (I’m not sure what they meant by “ancestral” here, but I like this idea of maybe deconstructing/reconstructing antecedent ancestors to stories, something Tolkien talks about in one of those essays, about I think the ‘soup of story’ or cauldron or something – I forget)
  • Foreshadowing emerges from projected probabilistic narrative futures (This seems cool!)
  • Unreliable narrators from biased text generation models (love love love this way to make AI bias into something purposive & useful)
  • Stream of consciousness inner dialogue via recursive inner monologue generation (James Joyce, but AI)
  • Fractal story recursion with stories within stories in latent daydreams (frame stories ftw!)
  • Surreal symbolism derived from visual concept associations (could probably boost this one to be more evocative)
  • Dramatic irony engineered via asymmetric information flows to different characters (this is a cool way of putting it)
  • Comedy through absurd juxtapositions and violated expectations
  • Tragedy of generative Greek theatre where language models predict downfalls (not super sure what this means, but I think I like it, especially thinking about AI & hubris)
  • Mysteries to solve by interrogating generative models for clues (this sounds very alternate reality game)
  • Deus ex machina via intervening AI characters to resolve human conflicts
  • Branching story paths from choose-your-own-adventure generative text models (obvious, but I’ll allow it!)
  • Time dilations and contractions by modulating generative model outputs (I like this idea of messing with time)
  • Memory hacking by inserting or removing narrative details in generated stories (also cool)
  • Social graph influences through character relationships and simulated interactions (“social graph” as a phrase is triggering for me, but maybe some people would be into this as a way to flesh out characters)

I think this is all pretty fun, because it’s been rare that we see any ideas put forward as to how we can positively use generative AI tools to enrich storytelling. And even as wonky as the above list may be, it feels inspiring as a rich ground to make exciting discoveries in as an explorer.

Related writing & research