Antikythera
Generative Art Tools

Developing algorithms for converting media into creative outputs

Context

Antikythera is a philosophy of technology think tank. The organization frequently collaborates with MIT Press on publications of books and online journals.

Challenge

For the digital release of Blaise Agüera y Arcas’ “What Is Intelligence?”, Antikythera sought digital graphics to anchor each chapter of the publication. The organization wanted artworks reflecting a connection between source inputs (like videos) and programmatic outputs generated by algorithms.

Solution

No Replica developed three generative art tools capable of producing a wide range of outputs. The first two tools utilize dithering algorithms to process source media into dynamic representations featuring only two colors. Each of these tools features the option to provide an input (video or image) and customize parameters to generate a new piece of algorithmic art. The outputs can then be saved or directly embedded into websites as code. Meanwhile, the third tool implemented a Turing pattern algorithm that produces an output without the need for a source image.

No Replica built these tools as websites with a graphic user interface such that the in-house team could easily generate new graphics without needing to modify code.

Try the custom tools ↗

Client
Antikythera ↗
Collaborators
Blaise Agüera y Arcas, author and development consultant
Minkyoung Kim, digital book design
James Goggins, physical book design
Services
Computational art
Custom tool development
Year
2025
No Replica built a site to host three custom tools capable of generating a wide range of dynamic outputs. The site’s homepage offers a glimpse into each tool’s functionality.
The first tool, “Dither,” converts a source file into a dynamic output consisting of only two colors. We achieve this output by looking at the original file, determining if a pixel is light or dark, and then converting that pixel to one of two colors. To make the output dynamic, we add some randomness into our calculations that leads to the distinctive noisy effect. The tool’s controls include options for output resolution, animation speed, pixel colors, and blending with the original image.
The second tool, ”Floyd-Steinberg,” is an implementation of a specific dithering algorithm that achieves a smoother effect. Instead of randomly deciding each pixel’s color, the image is smoothed over to achieve a result that is more visually cohesive over long periods of time. The tool’s controls are similar to that of the first tool, with more granular controls for how different base colors (red, green, and blue) are treated.
The third tool, ”Turing Pattern,” involves a completely different process that generates an output without taking an inputted media file. Instead, this tool features an implementation of an algorithm meant to simulate nature, showing how dark and light spots interact almost like living creatures. The tool’s controls allow the user to change the intensity and style of these interactions, producing unexpected results through experimentation.

Say hi! We’re always open to new projects, ideas, and collaborations.


Reach out ↗