THE WORD AFTER US: An AI poetry unreading

nathaniel stern Sasha Stiles500/500

“Where in the world will words go,
after us? How will they be arranged,
and what will they sound like
when they arrive at their destination?
What will they be doing?
What will the world be doing?
Will the words be quiet or loud?
Will they be full of themselves,
or full of longing,
or full of longing’s opposite?”

An ars poetica on the future of literature and humanity.

This project is a one-time generative poetry performance of a transhuman text by Sasha Stiles and her AI alter ego, Technelegy – read and unread by an algorithm authored by Nathaniel Stern. Every hash transforms one of 353 lines of AI-written poetry from legible letters into wordless language, generating a unique textblock from millions of possible combinations; each algorithmic translation unwrites itself, forever. Press "s" to overlay and toggle the original line of verse, and watch its shapes and meanings degenerate – or evolve – into an infinite visual poem.

All minters will receive an airdropped broadside (mp4) of the entire original poem after the performance is complete. A 1/1 media-rich audiovisual version of THE WORD AFTER US will also be made available.

Poetic traits:
- 353 possible lines of generative poetry produced by a GPT-3 language model fine-tuned on Stiles’ and Stern’s writing and research.
- 28 possible color palette combinations selected from Stern’s performative scanner art, with rare palettes including RGB, Hacker, Charcoal, and Chalkboard (black background).
- 6 possible fonts, including the premiere of Stern and Stiles’ custom AI font, NNFT (Neural Network Font Type), developed for their ongoing collaboration, “Mother Computer.”
- millions of layout and layering possibilities, including a rare “spiral” effect.
- billions of animation attribute combinations, which carry on forever.

This page has been generated using fx_hash public API, to display an overview of a creator's collection from The computation of "rarity" is not the official computation and therefore can differ. Dev by @zancan.