Generative art projects
Notable gen art pursuits
We survey two pure generative art projects and one generative AI project:
Autoglyphs
Autoglyphs is the first on-chain generative art project on the Ethereum blockchain. An NFT is said to be on-chain if both metadata and file of the referenced asset are fully stored on a blockchain within some record of a transaction, and not archived on a decentralized file system like IPFS.
The project was created in 2019 by Matt Hall and John Watkinson from Larva Labs, the creators of the even more popular profile picture project (PFP) CryptoPunks, that was the inspiration for the ERC-721 standard that powers most digital art and collectibles. In 2022 CryptoPunks was acquired by Yuga Labs, the creators of another popular PFP, the Bored Apes Yacht Club.
It is a collection of 512 glyphs made by few characters, hence a form of ASCII art, originally minted by anyone who was willing to donate the creation fee of 0.2 ETH (around $35 at the time) to the climate charity 350.org. Each symbol in the output corresponds to a cell, and the final artwork consists of 64x64 cells arranged in a square grid.
Autoglyphs is a highly optimized generative algorithm capable of creating billions of unique artworks, wrapped inside an ERC-721 interface. The key feature of autoglyphs is that the art is inside the contract itself, it is literally art on the blockchain.
This becomes clear if you examine any glyph creation transaction on the blockchain. The event data contains the full output of the generator, and hence the artwork itself. See here for an example for autoglyphs #1. This is also the output of the tokenURI function of the smart contract, which typically points to the off-chain metadata in off-chain NFTs.
The character art pattern can be drawn on paper, on a screen, plotted or even played by following the written instructions in the comments of the smart contract itself. This recalls Sol LeWitt's Wall Drawings, a hand-drawn analog generative art project.
The idea becomes the machine that produces the art. Sol LeWitt
Autoglyphs have a secondary market on OpenSea and sales are announced by a Twitter bot.
QQL
QQL is an unbounded-form generative art project by Tyler Hobbs - a visual artist who works primarily with algorithms, plotters, and paint - and Dandelion Wist Mané - a dancer, engineer, entrepreneur, and generative artist.
This collaboration is intended to provide a new way to mint NFTs that celebrates emergence, unpredictability, and happenstance over forced rarity.
We want to encourage collectors to explore the edges of the algorithm, play a role in the output, and take agency to become a co-creator. We want collectors to view their engagement as an adventure and make a creative contribution to the art.
Adding a curation step by the collector also allows the generative algorithm to take more risks and explore a more interesting potential output space. We trust collectors to seek out and identify the truly special outputs that emerge. With this approach, the collector is now the curator. Tyler Hobbs and Dandelion Wist Mané
The details of the project are as follows:
anyone can play with the algorithm by setting traits on a dashboard and running the generator to generate (1, 10, or 100) QQLs
you can create QQLs borrowing the traits of another QQL
by connecting your wallet you can save the best QQLs you generated as favorites
the creator’s address is used as part of the seed that guides all of the randomness internal to the algorithm. This means your outputs most likely will be different from those of other users
you need a mint pass to tokenize a QQL on the blockchain as NFT. There are 999 mint passes, some of them still available for purchase in the OpenSea secondary market. Minting will be open indefinitely
you can list the seed of a QQL you created for sale on the marketplace. It can be used to mint a QQL with a mint pass
the address that is embedded in the seed for a minted QQL will receive a 2% royalty on secondary market sales of that QQL. This is usually the address of the minter, but can be also the address of the creator that sold the seed
the QQL project is on-chain in the following sense: all of the QQLs will be fully reproducible from the QQL algorithm, which is written in p5.js language and stored on the Ethereum blockchain
the word QQL means nothing (it was generated randomly)
generated QQLs are for personal, non-commercial use only. In particular, you do not own them, you cannot tokenize and sell them. On the contrary, if you bought a mint pass and minted a QQL, you own the minted QQL NFT and can list it for sale. In both cases, you do not own intellectual property ownership rights over the QQL. Read the ToU for more
Botto
Botto is a decentralized autonomous artist. It creates works of art based on collective feedback from the community. The human participation is what completes Botto as an artist.
The art engine
Botto’s art engine has been trained on millions of images. From that latent space it creates new unique images every week, all untouched by human hands. Botto is a combination of AI algorithms, including Stable Diffusion for image generation and GPT for text generation, plus a number of custom augmentations. Botto controls itself: it makes its own prompts, does its own filtering of images, and writes the artwork descriptions. Though, Botto is still dependent on others for many things, such as getting feedback to tune its aesthetics, adding new generative models, producing exhibitions and collaborating with other artists.
Governed by BottoDAO
Each week, Botto presents 350 promising pieces for consideration by the BottoDAO. From this selection, the engine uses a taste-model (also trained by the votes of the DAO). More precisely, every week Botto's taste model selects 350 fragments from a growing pool of what is probably by now 3 million fragments. It adds these 350 fragments to the voting pool which is always 1050 fragments in size. This means that every week the 350 fragments from the pool which have the lowest number of votes will get culled and will end up in the discarded bin.
Members of the BottoDAO are :
holders of the $BOTTO token, or
holders of the Botto access pass, or
holders of an access pass for approved communities, for instance SuperRare.
The selected artworks are not yet considered final works, they are called fragments as they are still unproven. The community tells Botto what it considers art, and Botto perpetually evolves from the feedback.
Botto uses voting feedback in these places:
text prompts: votes influence which aspects of text prompts are used to generate fragments. Characteristics of prompts that generate desirable images will be more likely to get reused
taste model: the taste model used for pre-selection tries to replicate the voting behavior of the community. This is not a yes/no decision, but a gradient of probabilities such that each set has images with different chances of getting picked in voting.
artwork descriptions. The descriptions are generated with GPT and are the only part of the process that involves some direct human curation. As GPT was trained on much of the Internet, its language can be quite foul at times and is not ready to be out in the world without some supervision.
The training of Botto is designed to not allow for an overly skewed voting weight. For example, 500 votes each cast by separate voters for one piece will have more weight in the training than 2000 votes by a single voter for the same piece.
A weekly auction
Each week, the most popular fragment is minted as a final artwork on SuperRare marketplace on Ethereum and sold at auction. Every week 50% of the total revenue is held in treasury controlled by the DAO and the remaining 50% is distributed among contributors proportionally to voting points spent out. Members must be staking a minimum of 2000 $BOTTO in order to be eligible for revenue distribution.
Botto's guardian
Quasimondo, aka Mario Klingemann, designed Botto’s art engine based on a white paper he wrote back in 2018. He is Botto’s guardian, enforcing the rule for Botto that there be no direct human interference with the creations and ensuring updates to the engine do not violate Botto’s autonomy. Mario Klingemann is a renowned artist and a pioneer in the field of neural networks, generative art, and machine learning. In particular, Klingemann has made significant contributions to the intersection of art and artificial intelligence.
Last updated
