1kx: A history of generative manufacturing, web3 experiments and stacks

Author: Accelxr, 1KX; Translation: Golden Finance 0xjs

The future of consumer goods is generative.

Currently, blockchain-based generative algorithms are primarily used in the visual arts sector, with artists writing code to create digital and interactive works, animations and prints. However, art may only be the first suitable medium for this new blockchain-based creative process. We believe that blockchain-based generative media will become widespread across every other consumer goods and luxury vertical, and that this artistic creation process will generatively create unique new categories of physical products.

The appeal of generative collections is clear: consumers crave unique products that reflect their unique identities while connecting them to the larger community. With the 1/1/x model, generative algorithms achieve this by creating unique pieces within a larger collection with a unified aesthetic. These unique creations cater to the specific tastes of individuals, allowing for fine-grained expression within a tribe, and their success in this regard is reflected in the rise of the PFP market and the growth of niche communities that have emerged around specific generative characteristics.

Interestingly, generative algorithms and the 1/1/x rarity distribution also resolve the tension between mass production and customization. In traditional manufacturing, mass customization of products is often impractical and expensive. However, generative algorithms can be directly integrated into manufacturing hardware, such as 3D printers, CNC machine tools, laser printers, automatic looms, etc., providing not only the feasibility of production and distribution, but also scarcity and uniqueness.

The intersection between social dynamics and rarity, digital creation and physical production lays the foundation for new categories of consumer goods and luxury products that combine algorithmic randomness, end-user parameterization and verifiable uniqueness to Meet consumer needs.

The history of generative manufacturing

Artists have always used technology as a means to explore new dimensions of creativity. Over time, this relationship has clearly changed, from a purely artistic endeavor to the intersection of art and manufacturing.

1960s - Early Generative Art: Artists begin experimenting with algorithmic processes to create works of art. Using early computers and programming languages and tools like pen plotters, artists like Manfred Mohr, Vera Molnár, and Harold Cohen began creating algorithm-driven artwork.

1980s - Personal Computer and Software Revolution: The advent of personal computers makes digital tools more accessible. This allows more artists to try out these novel artistic processes.

1990s-2000s - The Birth and Expansion of Additive Manufacturing: As 3D printing technology emerged and developed, artists saw new opportunities. Generative artists are starting to experiment with these tools, creating sculptures and installations directly from their software-driven designs.

2000s-2010s - Digital art meets digital fabrication: As both fields mature, digital artists will collaborate with makers, architects, and designers to realize large-scale installations. Projects like The Living’s Hy-Fi Tower employ generative design principles in their conception and use modern manufacturing methods to create them. It was at this time that software tools like Processing, tailored for artists, enabled them to create complex procedural art without the need for in-depth programming knowledge.

2010s - Maturity of tools and methods: Generative art platforms and frameworks, such as openFrameworks and TouchDesigner, become increasingly popular. These tools are combined with more accessible and sophisticated 3D printing, laser cutting and CNC milling technologies to enable seamless production. For example, artists like Nervous use generative algorithms to design unique jewelry and clothing, which are then produced using 3D printing technology.

2020s - Convergence and collaboration: the boundaries between art, design and manufacturing become increasingly blurred. Art installations, architectural structures, and even everyday objects now demonstrate the unique aesthetics and capabilities this combination can produce. Notably, art on the blockchain at this time has sparked renewed interest in the field of generative art, using cryptographic inputs as random seeds for on-chain collections. Combined with new primitives in the digital physical space, we are approaching new areas where digital creation and physical production merge.

Today’s generative artists are not just making art, they are redefining consumer goods, combining aesthetic value with functional design, and pushing the frontiers of what’s possible in art and industry.

Web3 Experiment

In Web3, there were various early experiments with generative manufacturing.

Trame的Neolice Loom

Trame and CPG’s Craft Nouveau is a series of collectibles that focus on combining traditional craftsmanship with generative art, demonstrating the ability of generative code to preserve the styles of cultures and art from around the world. Alexis André’s Navette is the first collection from Craft Nouveau, in which Alexis wrote an algorithm to generate images that could be automatically woven by the Neolice Loom—an automated loom that could ingest code to weave physical pieces.

The fx(hash) ecosystem has a lot of experiments in manufacturing. This is likely due to its permissionless self-publishing approach. Klangteppich is an ever-evolving, dynamic NFT that provides instructions for weaving and allows collectors to obtain a physical piece of any frame generated from the code. Mini Dahlias includes instructions in the NFT’s metadata on how to create a 3.5-inch x 2.0-inch pocket sculpture from 14 layers of laser-cut alpha-cellulose mat. Nuages makes it possible to use the output of the code to recreate variations of Joanie Lemercier’s Cloud series in physical space by a plotter machine.

Nuages possible on fx(hash)

Alongside crafts and art, fashion is one of the most explored avenues of generative manufacturing. Iteration-002 produced by 9dcc is an early example of combining generative design with physical products. The Iteration-002 shirt was created in real time using a printer connected to SnowFro’s Squiggles algorithm. The printer relies on the algorithmic randomness of the source code to determine the design features printed on the shirt, following the same feature distribution as the original 10k collection.

9dcc ITERATION-002

Tribute Brand also recently remixed the Chromie Squiggle algorithm to create manufactured apparel. Chromie Squiggle holders can generate personalized sweaters using their unique Chromie Squiggle, while others can generate one-of-a-kind sweaters via the original Chromie Squiggle algorithm. The release includes digital and physical ODDS sweaters derived from Chromie Squiggle’s source code. The digital objects serve as blueprints for future versions of the sweater and can be used as skins in the immersive environment, and each unique ODDS digital object can be exchanged for a corresponding ODDS physical sweater, handcrafted by the Waste Yarn Project.

Other noteworthy generative fashion projects include mmERCH and RSTLSS, both of which plan to experiment around algorithmic randomness and design.

Generative Manufacturing Stack

The generative manufacturing stack for generative products can be divided into 5 layers:

Creation: The initial stage of generating a design or concept using an algorithm or artificial intelligence process.

Curation: The process of selecting and fine-tuning generated designs to achieve desired results or specifications.

Translation: The process of converting a digital design into machine-readable instructions or code for use by manufacturing equipment.

Fabrication: The physical production or manufacturing process that transforms virtual designs into physical objects. Utilize different materials and equipment, such as 3D printing, CNC milling, laser cutting, machine weaving and automatic looms, etc., to create objects of different shapes and materials.

Certification/Linking: Verify the authenticity of a manufactured product and link it to the digital twin to ensure its provenance.

Creation layer

The creation of generative products starts with code. Libraries like p5.js and Processing provide artists and designers with powerful tools for creating generative art. These libraries extend randomness on the blockchain with seeds generated from transaction hashes, token data, block headers, and more. Blockchain art engines like ArtBlocks Engine and fx(hash) allow artists to easily insert these random seeds into their code and mint artwork directly on the blockchain.

For AI artists, this layer focuses on the development and fine-tuning of models to create the desired aesthetic effect. They usually choose one from existing AI models, such as generative adversarial networks (GAN) as a basis. Through backpropagation, the model weights are gradually improved to produce artwork consistent with the desired style. Artists provide feedback by curating the most engaging output and incorporating it into the training dataset. This iterative process continues, constantly improving the performance of the model, allowing the artist to explore different possibilities. In addition to custom models or Stable Diffusion LoRA, etc., there are tools that can simplify this process, such as Scenario.gg.

Curator level

After the authoring layer, the output of the code can be further refined to suit the user’s preferences. In the context of creative coding, this is often done in the form of multi-person parameterization, such as fx(params) of fx(hash) provides such functionality.

In the context of generative models of AI, curation is often accomplished through a broader community of token holders, as is the case with Botto’s generative algorithms and Deep Objects’ community design process.

Studio or self-publishing is the final link in the curatorial process. This is where generative studios, like Trame and ArtBlocks, showcase their work to the public, or fx(hash) as a self-publisher.

Translation layer

Once the algorithm and design are determined, the generative goods must be translated into machine-readable instructions suitable for manufacturing hardware. Translation is a relatively simple process that aims to recreate a work in physical space as accurately as possible.

Translation can be done in several different ways, including:

Artist/Collector Interpretation. It’s easiest to leave the physical design specifications to the artist or collector to translate the object. They will decide how a piece will be made, the materials to use, the exact dimensions, etc.

Embedded features. A more scalable and interesting approach is to embed the physical information required for manufacturing into the NFT itself. The features in the NFT’s metadata define the domain of the translation (e.g., fabric texture, thread size, weaving instructions, etc.).

Direct instantiation. A third approach is to generate interpretable assets directly: the generative algorithm has been adapted to the manufacturing hardware, or the output of the algorithm is a file that can be 3D printed or the vertices of a 3D mesh.

Manufacturing layer

After translation, the resulting goods will be manufactured. The manufacturing phase is a critical step that involves converting virtual designs into physical objects. Use different technologies such as 3D printing, CNC milling, laser cutting, machine printing and automatic weaving to create objects in different materials and shapes.

For the aforementioned first release of Trame with Alexis Andre, Neolice Loom was used as the manufacturing hardware. Neolice Loom accepts an artist’s custom script and reinterprets the code into 3D space through weaving. Trame is also branching out into new media, with the image above highlighting an experiment in generating pottery.

While generative art production today is specific to generative art, Artmatr highlights what advanced manufacturing tools can do for the physical production of digital objects. Artists work with the Artmatr team to submit a variety of digital file formats such as code, 3D models, PSD files (Photoshop), vectors and animations. Next, they define the physical “thread,” including medium (oil, UV, acrylic), substrate, dimensions, and more. Finally, it is realized by using machines such as robotic arms and 6-axis printers. Using different techniques such as inkjet printing, jet spraying and extrusion, the generated topology can be 2D, 2.5D or 3D.

Authentication/Linking Layer

After a physical object is created, it needs to be associated with its digital twin. This is similar to digital physical processes in other fields, such as fashion. Near field communication chips, steganography and QR codes manufactured by Kong and IYK are among the technologies that connect the digital with the physical and provide authentication of provenance.

Future Possibilities

Going forward, we expect existing generative artworks on the blockchain to be used as derivatives. We’ve seen this with various fashion projects using Squiggles. Another early example is Terraflows, which is built on top of the Terraforms art program. This kind of networked art can generate interesting effects in physical spaces. Reinterpret. For example, Fidenza Art Script can be used to create architectural layouts for 3D printed houses.

Another interesting future possibility is the tokenization of decentralized manufacturing facilities for the mass production of generative goods, forming a kind of physical infrastructure network. Hobbyists and commercial makers with the appropriate equipment can bid to print or produce their published works for collectors or artists. Tokens can meter the hardware network and help with the initial cost of launching a manufacturing facility. This fits particularly well with the CC0 paradigm used with blockchain code.

Looking further forward, synthetic biology and/or chemical fabrication may also be interesting avenues for generative properties: for example, generative code could algorithmically determine the characteristics of lab-grown crystals, the phenotype of plants, and so on.