A definition for generative systems
Inputs to Outputs
According to the Tate:
Generative art is art made using a predetermined system that often includes an element of chance—is usually applied to computer-based art.
For years this held true. But with the introduction of LLMs into our everyday lives, we’ve had to expand what the word “generative” means. I’ve seen discussions suggesting that artists should use “algorithmic art” or “serial art” to better differentiate themselves from the leviathan nature of our current crop of language models. Our systems for making art can feel rudimentary next to the sheer malleability and infiniteness of these models.
But I think we must resist this retreat. Instead of fleeing, we can expand the tent pole and welcome these new tools into the “generative” fold. We can look around us and give a second consideration to tools that before might’ve seemed to fall outside the label.
With that in mind, I suggest a definition for all generative systems:
A set of steps that can be followed to turn inputs into outputs.
With this, here is a in-complete list of generative systems:
A recipe
Sheet music
Algorithms
Sol LeWitt wall drawings
Generative art
LLMs
Wind chimes
Kaleidoscopes
Madlibs
Tarot cards
Knitting patterns
Cloud formations
Synthesizers
This definition creates room for the work we have been doing all along: instruction-based making that spans digital and physical, deterministic and chance-based, simple and complex. There’s a world of generative systems around us, each with their own character and constraints.
-Mello

