Learn

The World of
Generative Art

History, movements, key concepts, tools, research papers, and community resources — a comprehensive guide to art made with systems.

800+
Years of History
14
Key Pioneers
11
Core Concepts
5
Art Movements

01 / Origins

Pre-Digital Generative Art

Algorithms existed long before computers. Rule-based systems for generating pattern and form stretch back centuries.

~1200 CE — 1950s

Before the Machine

Girih tile patterns at Darb-i Imam shrine, Isfahan
Girih tile patterns, Darb-i Imam shrine, Isfahan (~1453). Five polygon shapes generate infinite tessellations.
Jacquard loom with punched cards
Jacquard loom (1804). Punched cards automated complex woven patterns — a direct ancestor of computation.
Arnold Schoenberg
Arnold Schoenberg (1927). His twelve-tone technique was a rule-based system for musical composition.

Islamic Geometric Art (~1200 CE)

Girih tiles — a toolkit of five equilateral polygon shapes used to generate complex tessellated patterns in Islamic architecture. Roman Verostko argued these patterns are constructed using algorithms, making them direct ancestors of algorithmic art.

The Jacquard Loom (1804)

Joseph-Marie Jacquard's loom used punched cards to automate complex woven designs — a foundational link between rule-based pattern generation and computation. Ada Lovelace wrote: "The Analytical Engine weaves algebraic patterns, just as the Jacquard loom weaves flowers and leaves."

Serial & Stochastic Music (1920s–1950s)

Arnold Schoenberg developed the twelve-tone technique in the early 1920s. Iannis Xenakis (1922–2001) introduced stochastic music using probability theory. His piece Pithoprakta (1955–56) is based on Brownian motion — randomness as compositional material.


02 / Pioneers

The People Who Built the Field

From the first plotter drawings in 1965 to AI data sculptures at MoMA.

1960s — 1970s · The First Wave

Computer Art Pioneers

Georg Nees

Georg Nees

1926–2016

First person to publicly exhibit computer-generated art (February 1965, Stuttgart). Programmed in ALGOL on a Zuse Graphomat Z64 plotter at Siemens. His 1968 doctoral thesis "Generative Computergraphik" is the first on the subject.

First Exhibition ALGOL Plotter Art
Wikipedia
Frieder Nake

Frieder Nake

b. 1938

Mathematician and computer art pioneer. First exhibited November 1965 in Stuttgart alongside Nees. Deeply influenced by Max Bense's information aesthetics. Still active in both art and theory.

Information Aesthetics Mathematics
Wikipedia
A. Michael Noll at Bell Labs

A. Michael Noll

b. 1939

Bell Labs pioneer who created the earliest US computer art using FORTRAN. His "Computer Composition with Lines" (1964) was inspired by Mondrian. First US computer art exhibition at Howard Wise Gallery, April 1965.

Bell Labs FORTRAN First US Show
Wikipedia
Vera Molnár

Vera Molnár

1924–2023

First female digital art pioneer. Before gaining computer access, she developed the "machine imaginaire" — manually executing algorithmic procedures. Began working with computers in 1968 using FORTRAN and BASIC. Featured in the 59th Venice Biennale (2022).

Machine Imaginaire Paris Venice Biennale
Wikipedia
Harold Cohen with AARON

Harold Cohen

1928–2016

Created AARON — the earliest and longest-running AI program for art-making. Started as abstract drawings, evolved through representational imagery (1980s) to color (1995). Whitney Museum retrospective in 2024.

AARON AI Art Whitney
Wikipedia
Charles Csuri

Charles Csuri

1922–2022

Called the "father of digital art and computer animation" by the Smithsonian. His 1967 film Hummingbird is in MoMA's permanent collection. Founded ACCAD at Ohio State and one of the first computer animation companies.

Animation MoMA Ohio State
Wikipedia
Manfred Mohr

Manfred Mohr

b. 1938

Jazz musician turned algorithmic artist. First computer drawings in 1969. His 1971 solo show at Musée d'Art Moderne, Paris was one of the earliest solo museum exhibitions of computer art. Explored multidimensional hypercubes extensively.

Hypercubes SIGGRAPH Award Golden Nica
Wikipedia
Herbert W. Franke

Herbert W. Franke

1927–2022

Austrian physicist, artist, and co-founder of Ars Electronica. Generative photography from 1953, analogue computer work from 1954. His "Computer Graphics — Computer Art" (1971) is the earliest comprehensive text on the subject.

Ars Electronica First Textbook
Wikipedia
1980s — 2000s · Tools & Platforms

The Democratizers

The demoscene, Processing, and the creative coding movement brought generative art from research labs to anyone with a computer.

John Maeda

John Maeda

Aesthetics + Computation Group, MIT

Established the Aesthetics and Computation Group at MIT Media Lab (1996). Created Design By Numbers — a simplified programming language for artists and designers that directly inspired Processing.

MIT Media Lab DBN
Wikipedia
Casey Reas

Casey Reas & Ben Fry

Processing, 2001–present

Created Processing — the most influential creative coding platform ever built. Initiated spring 2001 at MIT; alpha released February 2004. Won the 2005 Golden Nica at Ars Electronica. Democratized algorithmic art for an entire generation.

Processing Golden Nica
processing.org
Zach Lieberman

Zach Lieberman

openFrameworks & SFPC

Co-founder of openFrameworks (C++/OpenGL creative coding toolkit) and the School for Poetic Computation (2013, NYC). His EyeWriter project won Design of the Year and a Golden Nica in Interactive Art.

openFrameworks SFPC EyeWriter
Wikipedia
2010s — Present · Contemporary

The New Wave

On-chain generative art, AI data sculptures, and creative coding communities push the field into mainstream consciousness.

Tyler Hobbs Fidenza

Tyler Hobbs

Fidenza, Flow Fields

Quit his engineering job for full-time generative art. Fidenza (999 NFTs, Art Blocks, June 2021) uses flow field algorithms for organic curves. Fidenza #313 sold for 1,000 ETH (~$3.3M). Later created QQL, a collaborative project.

Fidenza Art Blocks Flow Fields
tylerxhobbs.com
Refik Anadol

Refik Anadol

b. 1985

Turkish-American media artist. "Unsupervised" at MoMA (2022–23) was trained on 138,151 collection records — first generative artwork in MoMA's permanent collection. Machine Hallucinations at the Las Vegas Sphere was the largest AI artwork at 580,000 sq ft of LED.

MoMA AI Art Data Sculpture
Wikipedia

Erick Calderon (Snowfro)

Art Blocks, 2020–present

Founded Art Blocks in November 2020 — the platform that brought on-chain generative art to the world. Launched with Chromie Squiggles. Peak month: ~$600M in sales (August 2021). Total secondary market exceeded $1.3 billion.

Art Blocks On-Chain Chromie Squiggles
artblocks.io

03 / Exhibitions

Landmark Shows

The exhibitions that defined and legitimized computer-generated art.

Year Exhibition Venue Significance
Feb 1965 ComputergrafikGeorg Nees Studiengalerie, TH Stuttgart First-ever computer art exhibition
Apr 1965 Computer-Generated Pictures — Noll & Julesz Howard Wise Gallery, NYC First US computer art exhibition
Nov 1965 Nake Exhibition Galerie Wendelin Niedlich, Stuttgart Early plotter art alongside Nees
Aug 1968 Cybernetic Serendipity
Curated by Jasia Reichardt. 130+ contributors incl. Xenakis, Cage. 40,000–60,000 visitors. Toured to DC & SF.
ICA London First major international computer art show — all aspects of computer creativity
1971 Computer Graphics: Une Esthetique ProgrammeeMohr Musée d'Art Moderne, Paris One of the first solo museum shows of computer art
1979– Ars Electronica
Co-founded by Herbert W. Franke. Prix Ars Electronica from 1987.
Linz, Austria Longest-running festival for art, technology & society
2022–23 Refik Anadol: Unsupervised
First generative artwork acquired for MoMA's permanent collection (Oct 2023).
MoMA, New York Gen art enters the permanent canon

04 / Concepts

Core Ideas & Techniques

The mathematical and algorithmic foundations of generative art.

🎲

Randomness vs. Determinism

The fundamental tension. Pseudorandom number generators provide controlled randomness — same seed, same output. The artist designs the system and constraints; randomness provides variation within those constraints.

Emergence

Complex patterns arising from simple rules. The output exceeds what the rules alone would predict. The artist sets up conditions for emergence rather than dictating outcomes.

🌊

Perlin Noise

Developed by Ken Perlin in 1982 for Disney's Tron. Produces smooth, continuous pseudorandom values — the backbone of organic-looking generative art. Won an Academy Award for Technical Achievement (1997).

Paper: "An Image Synthesizer" (SIGGRAPH 1985)
Variants: Simplex noise, Value noise, Worley noise, fBm
🚜

Fractals

Self-similarity across scales. Benoit Mandelbrot coined the term and published The Fractal Geometry of Nature (1982). Mandelbrot set, Julia sets, Koch snowflake, Sierpinski triangle — found throughout nature and art.

Key work: The Fractal Geometry of Nature (1982, ISBN 978-0716711865)
🌿

L-Systems

Parallel string-rewriting systems invented by Aristid Lindenmayer (1968) to model plant growth. Start with an axiom, apply production rules, interpret as turtle graphics. Generates trees, plants, fractal curves, and branching structures.

Key ref: The Algorithmic Beauty of Plants (free online)

Cellular Automata

Grids of cells updated by local rules. Conway's Game of Life (1970) demonstrates extraordinary complexity from four simple rules. Wolfram's Rule 110 is proven Turing-complete. Produces patterns that feel alive.

Key works: Martin Gardner (1970), Wolfram's A New Kind of Science (2002)
💨

Flow Fields

Vector fields assigning direction to every point in space. Particles follow these vectors to create organic, flowing compositions. Tyler Hobbs' Fidenza (2021) is the most celebrated example.

Approaches: Noise-based, analytical, image-based
🧬

Reaction-Diffusion

Alan Turing proposed this as a mechanism for biological pattern formation (1952). The Gray-Scott model produces spots, stripes, and labyrinthine patterns resembling animal markings, coral, and lichen.

Paper: "The Chemical Basis of Morphogenesis" (Turing, 1952)
🐦

Agent-Based Systems

Multiple autonomous agents following simple rules produce emergent collective behavior. Craig Reynolds' Boids (1986) — separation, alignment, cohesion — simulates flocking with three rules.

Paper: "Flocks, Herds, and Schools" (SIGGRAPH 1987)
🌀

Recursion & Self-Similarity

A function that calls itself. Fundamental to fractals, L-systems, tree structures, and many generative algorithms. Produces self-similar patterns at different scales — a hallmark aesthetic.

🎰

Stochastic Processes

Systems where randomness evolves over time. Random walks (Brownian motion), Markov chains, Monte Carlo methods. Xenakis used stochastic processes in music; equally applicable to visual art.


05 / Tools & Languages

The Creative Coder's Toolkit

From FORTRAN punch cards to browser-based shaders — the tools that shaped generative art.

1960s — Historical Languages
FORTRAN The first computer art language. Noll used it with custom subroutines at Bell Labs, output to CalComp 565 drum plotters.
ALGOL Georg Nees programmed his graphics in ALGOL to control the Zuse Graphomat Z64 plotter at Siemens.
Logo (1967) Turtle graphics by Feurzeig, Papert & Solomon. Made computational thinking visual and accessible. Based on Lisp.
1980s — Desktop Publishing
PostScript (1984) Adobe's Turing-complete page description language. Enabled precise vector graphics, foundational for digital typography.
1990s–2000s — Creative Coding Revolution
Design By Numbers (1999) John Maeda's simplified language at MIT Media Lab. Direct precursor to Processing.
Processing (2001) Java-based, by Casey Reas & Ben Fry. The single most important tool in democratizing creative coding. Simple syntax lowered barriers dramatically.
openFrameworks (2005) C++/OpenGL toolkit by Zach Lieberman, Theo Watson & Arturo Castro. For performance-critical interactive installations.
TouchDesigner (2000) Node-based visual programming by Derivative. Real-time installations, VJing, projection mapping, immersive experiences.
2010s–Present — Web & Shaders
p5.js (2013) JavaScript reinterpretation of Processing by Lauren Lee McCarthy. Runs in the browser. The dominant tool for web-based generative art.
Shadertoy Browser-based GLSL shader community by demoscene veteran Inigo Quilez. Centre of the shader art world.
The Book of Shaders Open-source guide to fragment shaders by Patricio Gonzalez Vivo & Jen Lowe. Democratized GPU programming for artists.
three.js WebGL 3D library by Mr. Doob (Ricardo Cabello) — a demoscene alumnus. Dominant for 3D web graphics.
nannou Modern creative coding framework in Rust. A newer alternative to Processing for systems-level performance.
Also Notable
Houdini Procedural 3D with powerful L-system tools, particle systems, VEX scripting. Industry standard for procedural effects.
Max/MSP/Jitter Visual programming for audio/visual processing. Strong in interactive installations and live performance.
Hydra Live-coded visuals in the browser. Inspired by analog video synthesis. Great for live performance.

Hardware

Pen Plotters (1960s) CalComp 565, Zuse Graphomat Z64, HP 7475A — the original output for computer art.
AxiDraw (modern) By Evil Mad Scientist. The modern pen plotter beloved by generative artists.
Projectors / LED Projection mapping and LED arrays for building-scale generative installations.

06 / Papers

Key Papers

The research that built the field — from Turing's morphogenesis to latent diffusion.

Foundational (1950s–2000s)
1952
The Chemical Basis of Morphogenesis
Alan Turing · Philosophical Transactions
Proposed reaction-diffusion as a mechanism for biological pattern formation.
1968
Generative Computergraphik
Georg Nees · Doctoral thesis (advisor Max Bense)
First doctoral thesis on computer-generated graphics.
1968
Mathematical Models for Cellular Interactions in Development
Aristid Lindenmayer · Journal of Theoretical Biology
Introduced L-systems — parallel string-rewriting for modelling plant growth.
1972
Shape Grammars and the Generative Specification of Painting and Sculpture
Stiny & Gips · IFIP Congress
Formal grammar systems for generating visual compositions.
1974
On Purpose: An Enquiry into the Possible Roles of the Computer in Art
Harold Cohen · Studio International
Early philosophical framework for machine-made art by the creator of AARON.
1982
The Fractal Geometry of Nature
Benoit Mandelbrot · W.H. Freeman
Coined "fractal" and demonstrated self-similarity across natural scales.
1985
An Image Synthesizer
Ken Perlin · SIGGRAPH 1985
Introduced Perlin noise. Won Academy Award for Technical Achievement in 1997.
1987
Flocks, Herds, and Schools: A Distributed Behavioral Model
Craig Reynolds · SIGGRAPH 1987
The Boids flocking algorithm — separation, alignment, cohesion.
1990
The Algorithmic Beauty of Plants
Prusinkiewicz & Lindenmayer · Springer
Definitive text on L-systems for botanical simulation.
1991
Reaction-Diffusion Textures
Witkin & Kass · SIGGRAPH 1991
Applied Turing's reaction-diffusion to procedural texture synthesis.
1993
Complex Patterns in a Simple System
John Pearson · Science
Classified the full range of reaction-diffusion pattern types (spots, stripes, waves).
1994
Evolving Virtual Creatures
Karl Sims · SIGGRAPH 1994
Evolved 3D creatures with morphology and neural-net brains via genetic algorithms.
1998
Painterly Rendering with Curved Brush Strokes of Multiple Sizes
Aaron Hertzmann · SIGGRAPH 1998
Non-photorealistic rendering that simulates painted artwork from images.
Algorithmic Techniques
2002
Improving Noise
Ken Perlin · SIGGRAPH 2002
Improved noise with better gradient selection and quintic interpolation.
2003
What is Generative Art? Complexity Theory as a Context for Art Theory
Philip Galanter · Generative Art Conference
The most-cited definition of generative art: systems "set into motion with some degree of autonomy."
2010
Cyclic Symmetric Multi-Scale Turing Patterns
Jonathan McCabe · Bridges 2010
Multi-scale reaction-diffusion producing incredibly intricate organic patterns.
Computational Aesthetics & Creativity
1998
Computing Aesthetics
Machado & Cardoso · Brazilian AI Symposium
Early framework for computationally evaluating visual aesthetics.
1998
Creativity and Artificial Intelligence: A Contradistinction
Margaret Boden · Artificial Intelligence
Framework for exploratory, combinational, and transformational creativity in AI.
2005
Open Problems in Evolutionary Music and Art
Jon McCormack · EvoMUSART 2005
Survey of fundamental challenges in computational creativity and evolutionary art.
2010
Formal Theory of Creativity, Fun, and Intrinsic Motivation
Jürgen Schmidhuber · IEEE Trans. Autonomous Mental Development
Creativity as compression progress — seeking novel patterns that improve predictive models.
2012
The Painting Fool: Stories from Building an Automated Painter
Simon Colton · Computers and Creativity
Case study of an AI painting system designed to be "taken seriously as a creative artist."
2016
The Howard Wise Gallery Show: A 50th-Anniversary Memoir
A. Michael Noll · Leonardo, Vol. 49 No. 3, MIT Press
Firsthand account of the first US computer art exhibition.
2018
Can Computers Create Art?
Aaron Hertzmann · arXiv
Argues computers are tools, not artists — creativity requires intentionality and social context.
Neural & AI Art (2014–2022)
2014
Generative Adversarial Nets
Goodfellow et al. · NeurIPS 2014
The GAN framework — generator vs. discriminator producing realistic outputs.
2015
A Neural Algorithm of Artistic Style
Gatys, Ecker & Bethge · arXiv
Neural style transfer — separating and recombining content and style representations.
2015
Inceptionism: Going Deeper into Neural Networks
Mordvintsev, Olah & Tyka · Google AI Blog
DeepDream — visualizing what neural networks see by amplifying patterns.
2017
Image-to-Image Translation with Conditional Adversarial Networks (pix2pix)
Isola et al. · CVPR 2017
General-purpose image translation via paired training — edges to photos, day to night.
2017
Unpaired Image-to-Image Translation using Cycle-Consistent Adversarial Networks (CycleGAN)
Zhu et al. · ICCV 2017
Style transfer without paired data — horses to zebras, Monet to photo.
2017
CAN: Creative Adversarial Networks
Elgammal et al. · ICCC 2017
Modified GANs that maximize deviation from established styles to produce novel art.
2020
Denoising Diffusion Probabilistic Models
Ho, Jain & Abbeel · NeurIPS 2020
Foundational diffusion paper — iterative denoising as a generative process.
2020
Analyzing and Improving the Image Quality of StyleGAN (StyleGAN2)
Karras et al. · CVPR 2020
State-of-the-art face synthesis with fine-grained style control at multiple scales.
2021
Taming Transformers for High-Resolution Image Synthesis (VQGAN)
Esser, Rombach & Ommer · CVPR 2021
Combined vector-quantized codebooks with transformers for high-res image generation.
2021
Learning Transferable Visual Models From Natural Language Supervision (CLIP)
Radford et al. · ICML 2021
Connected text and images in a shared embedding space — the bridge enabling text-to-image.
2021
Zero-Shot Text-to-Image Generation (DALL·E)
Ramesh et al. · ICML 2021
Autoregressive text-to-image generation using discrete visual tokens.
2022
High-Resolution Image Synthesis with Latent Diffusion Models (Stable Diffusion)
Rombach et al. · CVPR 2022
Diffusion in latent space — made high-quality image generation accessible to everyone.
2022
Photorealistic Text-to-Image Diffusion Models with Deep Language Understanding (Imagen)
Saharia et al. · NeurIPS 2022
Showed that scaling language models matters more than image model size for text-to-image.
Experimental / Cutting-Edge (2020+)
2020
Growing Neural Cellular Automata
Mordvintsev et al. · Distill 2020
Self-organizing systems that learn to grow into target patterns via differentiable CA.
2020
NeRF: Representing Scenes as Neural Radiance Fields for View Synthesis
Mildenhall et al. · ECCV 2020
Neural 3D scene representation from 2D images — photorealistic novel views.
2022
DreamFusion: Text-to-3D using 2D Diffusion
Poole et al. · ICLR 2023
Generate 3D objects from text prompts by distilling a 2D diffusion model.

07 / Bibliography

Essential Reading

The books and resources that define the field.

Books

1969
Generative Computergraphik
Georg Nees
Siemens · First doctoral thesis on computer graphics
1971
Computer Graphics — Computer Art
Herbert W. Franke
Phaidon · Earliest comprehensive text
1982
The Fractal Geometry of Nature
Benoit B. Mandelbrot
W.H. Freeman · 978-0716711865
1990
The Algorithmic Beauty of Plants
Prusinkiewicz & Lindenmayer
Springer · 978-0387972978
1999
Design By Numbers
John Maeda
MIT Press · 978-0262632447
2002
A New Kind of Science
Stephen Wolfram
Wolfram Media · 978-1579550080 · 1,280 pages
2007
Processing: A Programming Handbook
Casey Reas & Ben Fry
MIT Press · 978-0262182621
2010
Generative Art
Matt Pearson
Manning · 978-1935182627
2014
When the Machine Made Art
Grant D. Taylor
Bloomsbury · 978-1623568849
2018
Generative Design
Benedikt Gross et al.
Princeton Arch. Press · 978-1616897581
2021
Code as Creative Medium
Golan Levin & Tega Brain
MIT Press · 978-0262542043

Online Resources


08 / Timeline

800 Years in Generative Art

Every major milestone from Islamic tessellation to MoMA's permanent collection.

Pre-Digital
~1200 CEGirih tile patterns in Islamic architecture — algorithmic tessellation
1804Jacquard loom patented — punched cards automate woven patterns
1920sSchoenberg develops twelve-tone technique — rule-based music composition
Early Computer Art (1950s–1960s)
1953Herbert Franke begins generative photography experiments
1955–56Xenakis composes Pithoprakta — stochastic music based on Brownian motion
1964Charles Csuri & A. Michael Noll create first digital artworks
Feb 1965Georg Nees: "Computergrafik" — FIRST computer art exhibition (Stuttgart)
Apr 1965Noll & Julesz at Howard Wise Gallery — first US exhibition
1967Logo created; Csuri's Hummingbird (now in MoMA)
1968Cybernetic Serendipity at ICA London — first major international show
1968Lindenmayer invents L-systems; Molnár begins computer work
1969Nees publishes Generative Computergraphik; Mohr's first computer drawings
Consolidation (1970s–1980s)
1970Conway's Game of Life published in Scientific American
1971Franke's first textbook; Mohr solo show at Musée d'Art Moderne, Paris
1975–82Mandelbrot publishes fractal books
1979Ars Electronica founded in Linz, Austria
Early 80sDemoscene emerges from software cracking culture
1982Ken Perlin develops Perlin noise (for Disney's Tron)
1986Craig Reynolds creates Boids flocking algorithm
Digital Tools Revolution (1990s–2000s)
1990The Algorithmic Beauty of Plants published
1996Maeda establishes Aesthetics + Computation Group at MIT
1997Ken Perlin wins Academy Award for Technical Achievement
2001Processing initiated by Reas & Fry at MIT
2004Processing Alpha released; openFrameworks initiated
2005Processing wins Golden Nica at Ars Electronica
Contemporary Era (2010s–Present)
2013p5.js created by Lauren Lee McCarthy; SFPC founded in NYC
2020Demoscene recognized as UNESCO cultural heritage (Finland)
Nov 2020Art Blocks founded by Erick Calderon
Jun 2021Tyler Hobbs releases Fidenza on Art Blocks
Aug 2021Art Blocks peak — ~$600M in sales in one month
Nov 2021fxhash launches on Tezos blockchain
2022Vera Molnár at 59th Venice Biennale
Oct 2023MoMA acquires Anadol's Unsupervised — first gen art in permanent collection
Dec 2023Vera Molnár dies at age 99
2024Whitney Museum retrospective: "Harold Cohen: AARON"

09 / Movements

Contemporary Art Movements

The five paradigms reshaping computational creativity — from immutable on-chain systems to live coded performances.

01

On-Chain & Smart Contracts

The defining distribution revolution of the 2020s. By deploying generation algorithms directly onto Ethereum and Tezos, artists created systems where the artwork itself is produced and permanently linked to the collector's transaction hash at the moment of minting — immutable, verifiable, and uncurated at scale.

  • Long-form generative art — algorithms designed to produce thousands of distinct, high-quality outputs across a full mint run without hand-selection
  • Transaction hash as seed — the collector's wallet determines the output, making provenance and generation inseparable
  • Trustless provenance — the blockchain is both the distribution system and the permanent record of origin
  • Feature rarity systems — artists define probabilistic trait distributions that collectors can verify against the on-chain algorithm
By the numbers
1000 typical edition size for long-form series

The hash is the seed. The algorithm is the artist. The collector is the minter.

determinism immutability long-form provenance ERC-721
02

AI & Machine Learning

A fundamental shift from writing explicit rules to directing data-driven neural networks. Artists navigate the vast, high-dimensional latent spaces of GANs, VAEs, and diffusion models — bending statistical distributions of reality to synthesize post-photographic imagery, organic simulations, and entirely new visual ontologies.

  • Latent space navigation — interpolating between billions of learned parameters to find images that have never existed
  • Post-photographic synthesis — generating flawless realities through text conditioning, style transfer, and diffusion guidance
  • Artistic curation of process — the artist's role shifts to dataset curation, prompt design, and output selection
  • Biological and ecological themes — leading practitioners often explore non-human intelligence and natural systems
Scale of possibility
512M+ parameters in a typical diffusion model

The network doesn't know what it knows. The artist's job is to ask the right question.

latent space GAN diffusion style transfer prompting
03

Shader Art & Real-time GPU

A move away from CPU-bound drawing functions toward the massive parallel processing power of the GPU. Using fragment shaders written in GLSL, artists define the color of every pixel simultaneously through pure mathematics — enabling photorealistic 3D worlds, infinite fractals, and fluid simulations packed into a few hundred lines of code.

  • Raymarching & signed distance functions — rendering infinite, analytically precise 3D geometry without a single polygon
  • Extreme code efficiency — entire scenes compressed into kilobytes, a discipline celebrated in demoscene culture
  • Real-time interactivity — outputs respond to time, mouse, audio, and camera at 60fps with no pre-rendering
  • Mathematical aesthetics — domain repetition, noise functions, and procedural texturing as primary visual vocabulary
GPU parallelism
geometric resolution via analytic SDFs

Every pixel is computed independently and simultaneously. The canvas is a mathematical function, not a raster.

GLSL raymarching SDF procedural demoscene
04

Live Coding & Algoraves

A performative, improvisational practice in which the code editor is exposed to the audience on a projector screen. Artists build and modify algorithms in real time — generating music through TidalCycles and visuals through Hydra — in club or gallery environments that celebrate the process over the product.

  • Radical transparency — the TOPLAP manifesto declares that the audience should see the code, the mistakes, and the thinking
  • The algorave — algorithm plus rave: dance music built live from first principles in front of a crowd
  • Glitch as aesthetic — errors are not hidden but embraced as part of the performance's texture
  • Feedback and synchrony — audio and visual streams modulate each other through shared signal routing
Philosophy
Live no pre-renders, no playback — only now

Show us your screens — TOPLAP Manifesto, 2004

TidalCycles Hydra algorave TOPLAP improv
05

Physical Plotter Revival

As digital art becomes increasingly ephemeral, a counter-movement translates code back into physical form. Artists program vectors to drive mechanical pen arms and CNC machines — uniting exact mathematical structure with the irreducible imperfections of ink, paper, and physical motion. The AxiDraw plotter has become the movement's instrument of choice.

  • The hand of the machine — bleeding ink, paper texture, and the physics of watercolor introduce irreproducible variation into deterministic code
  • Algorithm to physics — translating logical structures into kinetic movement; the path matters as much as the output
  • Unique multiples — each plotted print carries physical variations that make editions genuinely distinct despite identical code
  • Community and craft — the #plottertwitter community shares pen tests, paper experiments, and toolchains openly
Medium
Ink on paper — irreproducible, physical, permanent

The code is identical. The paper is not. Every print is the first print.

AxiDraw SVG pen plotter vpype physical

10 / Communities

Communities

Where generative artists gather, share, and collaborate.


11 / Repos

Notable Repos

Open-source code, toolkits, and curated lists for creative coding.


12 / Learning

Learning Resources

Tutorials, books, interactive guides, and expert blogs.