IM: There is a moment where old forms stop to be operative or even become a hindrance. In comics, one could attempt to rewrite the entire history of the medium from the perspective of the remainder inspired for instance by Gilbert Simondon’s skeuomorphs; these are objects that retained attributes from original structures which lost their functionality only to now persist as affective assemblages. For instance, the blue pencil, a drawing tool that allowed early comics artists to finalize their preliminary sketches on the same substrate because its color tone was not captured by scanning devices, can still be found nowadays used by a younger, digital-native generation of cartoonists. The same could be described for the rather conventional museographical ways to exhibit comics originals- there is nothing obvious or necessary about putting “original” pieces of comics art on a wall setting. It’s not always easy to come to terms with the realization that a format for which we often have developed an affective relationship but which no longer is operational, carries already some folded violence, either in the forced division of labor expressed with a sub-editor tool such as the blue pencil, or the exhibition’s mode of address intended to isolate and deterriorialize art from its context and legitimize it by way of borrowed formats from more respected artistic forms.

The comics industry is a mediating environment for the production and the distribution of dense, multifaceted contemporary objects depending on particular forms of interdisciplinary labor disseminated in multiple time and space zones. Comics for me is not an old medium, and I am not speaking only in terms of how multimodal expressions are increasingly the basis of our information diet, or more specifically how text-to-image generation is becoming the cornerstone of synthetic media in a variety of visual domains. Comics, from the very early last century, have prefigured how artists are entangled in a hybrid object-human networked activity needing to speak a language of a machinic intensity. Comics are situated in a dense information economy and each step in the artistic process is technologically mediated and heavily coded: from the industrial adoption of the patented Ben Day photoengraving technique as a conversion tool for color separation in printing technologies and the optimal allocation of space for cartoons and daily strips in early newspapers, to the industry’s complex logistics funneled through global distribution sinews, transatlantic containers, warehouses and data centers, comics can be described as technical objects that continue to remain relevant for our computational age.

JB: The “conceptual” work that you practice is strongly rooted, at least at the start of your career (I know that this word must sound funny to you, but I am using it anyway since you do not always work outside the world of large artistic institutions), in the Oulipian notion of “constraint”. Could you specify some examples of your use of constrained writing (in the broad sense of the work) and explain how the notion of constraint is no longer sufficient to cover and explain what is at stake in your current work, which I would define as the attempt to bridge the gap between artistic production and AI (or, once again, synthetic cognition)?

IM: Yes! Oulipo had a great merit for me to point to artistic practices as always contingent to an artist’s limitations related to their talent, creative and inspirational means, but also to their position in relation to class, sex, race, and subsistence under any handicap; and then how to make of these very same limitations, a conscious, structural and strategized part of one’s practice (Queneau’s description of artists as laboratory rats constructing their own labyrinths). Unfortunately Oulipo came in comics when it has already morphed into a rather depoliticized exercise of a purely inspirational nature, taking the form of expensive writers retreats exploring a repertory of strategies in constrained writing that are no longer breaking any barriers. In comics, I found its application schoolish, suiviste, and symptomatic of a class complex. Oubapo (Ouvroir de bande dessinee pontentielle) is emblematic of a typical attitude of comics artists to bring forms and practices from other arts, too little and too late. Comics embrace literary experimentation only when a form has downgraded into a manner, and not without signposting a certain affinity to (more respected) literary histories – and here the same goes for other modes of address in comics such as autobiography, graphic reportage, etc. Comics should be able to create their own forms, strategies, their own legacies. The question remains always the same: How to get out of the sandbox? Refusing to tackle copyright issues when working with appropriated material in comics, is for example a common way to defuse a potentially disrupting gesture that comes with programmatic or constrained writing. Here, I contrast Oulipo or Oubapo with the much more inquisitive Uncreative Writing, a poetic practice conceptualized by Kenneth Goldsmith, Christian Bök and Darren Wershler, attempted to reinvent writing by mobilizing the expressive qualities of word processing, databasing, identity ciphering, and intensive programming. In Kenneth Goldsmith’s words, conceptual writing “is a poetics of the moment, fusing the avantgarde impulses of the last century with the technologies of the present”. Responding to the dominantly digital-technological landscape and the immersion in unparalleled amounts of text, Uncreative Writing leveraged newly or differently available tools of the digital economy in order to work language in new ways.

An example I would like to bring here from conceptual comics revists Uncreative Writing’s contextual approach to text and its known operations (appropriation, plagiarism, procedurality, passive indexing, decontextualisation, intertextual referencing, textual abundance to name just a few of the terms that used to define, and describe textual practices in early conceptualism) percolated through comics’ very own operational intensity in automation. Designed in collaboration with a team of computer scientists in 2020, the Neural Yorker is an unsupervised Twitter bot that posts algorithmic cartoons in the tradition of the famous weekly periodical. The production of its algorithmic cartoons relies on the aggregation of information flows from a multitude of software and hardware components, symbolic languages, deep learning models, computer vision algorithms, API queries, etc. Every evening thousands of Neural Yorker cartoons are generated by multiple AI conditional models that were trained on hundreds of thousands of cartoons and punchlines collected from a multitude of online repositories and databases with their own systems of classification and labeling. By the next day, a humor-detection algorithm informed by heuristics will select the highest ranking cartoon, and will post it on Twitter. In addition, the Neural Yorker offers a subscription-based service to media outlets, providing customized cartoons based on their media specificity (social news, financial press, sports, etc), geographic interest, territorial range and general position in a simplified political spectrum, thereby further evidencing computational media’s disruptive qualities. As an illustration, the Neural Yorker explored for instance multiple sets of unconventional operations, only some of which could be considered of artistic interest and most were rarely accounted for the production of comic books in general: programming headless browser applications, bypassing anti-scraping firewalls, parsing JavaScript code, classifying images, developing and fine-tuning computer vision algorithms, indexing and subsetting domain specific data, developing generative conditional models, maintaining, debugging, rolling out new design features, and growth hacking. The New Yorker is the result of an information-intensive online economy, where all the texts, the images and their correlation is a computational outcome of the medium’s very own digital ecosystem.