The Rise of Malleable Generative Art

Malleability references a material or object’s capacity to be molded or shaped. The word comes from the Latin verb malleare meaning ‘to hammer.’ The implication is that the shaping of a malleable object requires effort, and that the changes are durable and lasting. There’s a growing trend by generative artists to embrace malleability and incorporate these ideas into their generative art. The results of such explorations can extend beyond the aesthetic and allow connections with audiences in new ways.

This essay was inspired by Tyler Hobbs’ essay, The Rise of Long Form Generative Art. For fun, I’m using a very similar structure in this essay, and I’m also adapting portions of that essay here. If you haven’t recently read his essay, please take a few minutes now.

https://tylerxhobbs.com/essays/2021/the-rise-of-long-form-generative-art

Long Form Generative Art (LFGA) has eaten the generative art world, and artists aspire for their projects to join the hallowed halls of ‘grails’. To recap LGFA, we go to the source. Tyler describes how LFGA is distinct from earlier forms of generative art in two key ways:

“First, the script output goes directly into the hands of the collector, with no opportunity for intervention or curation by the artist. Second, the generative algorithms are expected to create roughly 100x more iterations than before.” -Tyler Hobbs

These differences from earlier forms of generative art create new requirements of the algorithm:

“In summary, long-form generative art introduces the new demands of consistent quality and high variety, while maintaining the existing need for unity across all output from a program.” -Tyler Hobbs

Fidenza by Tyler Hobbs is, obviously, a phenomenal example of LFGA. 999 high-quality outputs were generated with no artistic curation, and they are firmly rooted as NFTs minted on Art Blocks. These outputs will not change. The collection is final, and no others will exist…well, more specifically, no other official Fidenza NFTs will ever exist.

There are actually countless other outputs that the Fidenza algorithm can make, but these cannot be made into Fidenza NFTs. Many artists and collectors consider this kind of output an ‘out of band’ (OOB) or simply an ‘exploration’ output. They are outputs, but they’re not saleable and not part of the official collection.

The fact that exploration outputs cannot be sold doesn’t diminish the community interest in them. The popularity of the project inspires many to generate dozens (or thousands) of these outputs for fun. Instead of the traditional meaning of ownership, these outputs are ‘owned’ by whoever happened to generate them, i.e. “This is my favorite Fidenza OOB from this week.” Those who appreciate the artwork and explore the algorithm through OOBs feel a personalized sense of ownership over an output, even if they can’t literally own it.

These collectors are deeply exploring the output space of the algorithm – much more deeply than the Fidenza collection did with its predetermined quantity of saleable outputs. These collectors are creating value by engaging with the algorithm, but nothing they can do will change the 999 pieces in the Fidenza collection. Fidenza, is not malleable.

Experimental generative artists have asked, what if aspects x or y could change about a project after it is minted? What if an output, or the collection of outputs, was malleable?

I’ve explored this question in many projects. Most specifically, in both BLONKS on the EVM and Chainlife in JS I explore the possibilities of a project being more than one thing, of having dynamic outputs that change as they move across accounts, and of using different rendering algorithms selectable by the owners. Many other artists have explored similar and other opportunities as well, including Tyler Hobbs and Indigo with their collaborative, user-curated project, QQL. In this project, each token’s parameters are selected by collectors of a mint-pass, who can lock in those values by using the mint pass and add their preferences to the collection with their finalized QQL.

This project is malleable in that the collection is created as a result of owner action and because the collection isn’t yet finalized – it’s still taking form. As great as QQL is, many projects demonstrate more malleability. Examples include Mathcastles with the ‘state-shifting’ Terraforms. makio135 & clemsos with the mutable Shapes2048. Kim Asendorf with the evolvable Alternate. My own projects, and many, many others.

These all exemplify varying differences from the LFGA form, but consistently they incorporate direct user input. Not only can this art be owned in the literal sense, but the art is also ‘owned’ due to the personal manipulation of the malleable work by the owner. Regarding an owner-customized output: it isn’t just owned, it’s their own. This is new, and it’s a big deal. Malleable generative art enhances ownership.

Quality generative art algorithms now have new aspects on which they’re able to be gauged, and the most important distinction is that these kinds of works encourage owner interaction to shape the collection. With every action, an output changes, and the collection changes. Not only does this allow an individual to make their mark, it creates a time-based evolution of the collection as a whole. The evolution may be permanent (QQL, Alternate, sometimes Terraforms), or it may be infinitely temporary (BLONKS, Chainlife, Shapes2048, some Terraforms).

Algorithms that encourage owner interaction require a different developmental approach. No longer are some LFGA requirements firm, but with malleability comes a new requirement: engagement. On-chain actions have a literal cost, and that cost must be offset by the engagement created in the artwork’s audience. This can be through gamification (on- or off-chain) or through personal interest in the output space. Without engagement, a malleable project might as well be as fixed as Ether Rocks!

Gauging engagement is less straightforward than gauging the average quality of an algorithm’s output space. This is especially true when owner-curation is allowed, and the output space becomes massive. While technical quality will still be achieved between outputs, aesthetic quality may not – and this can be acceptable!

If a malleable algorithm/system maintains engagement, then its malleable nature will be explored. This allows the merger of both kinds of ownership as discussed earlier, the traditional ownership and the personalized ownership. Now, one can own something they personalized through their action. How much collectors value this engagement in personalization is how such a malleable project must be evaluated.

New platforms will lean into malleable generative art, and with them, artists will be better equipped to explore their creations. Intrinsic.art (launching with my project, Metta) allows for a new and unique malleability: artworks can be ‘unminted’ to reclaim their traits, which can then be mixed and reminted into new artworks. This allows a focused exploration of a project's generative system, while also allowing an incredible level of personalized engagement. As each artwork can forever evolve through the actions of each individual owner, so too can the entire collection. At any time, the collection may well represent the aesthetic tastes of the collection owners.

(strikethroughs are of the original text, italics are my additions)

In summary, long-form malleable generative art introduces the new demands of consistent quality engagement and high variety, while maintaining the existing need for unity across all output from a program. Right now, few artists are capable of navigating this balance, but I have no doubt that will change. The art form is just ramping up, and generative artists are starting to become familiar with the new dynamic. Short-form generative work [and long-form generative work] will continue to exist and will continue to be the best fit for many artistic visions, but right now long-form malleable is the fertile, unexplored space. I look forward to seeing all of the amazing work that will be created over the next few years!

PS: Tyler, I hope you interpret the above paragraph and the structure of this essay as a humorous homage – you’re brilliant, I’m a goof, see you around. :)