Søren Pold's riPOSTe to last month's "Thoughts on the Textpocalypse" by Davin Heckman.
In “Thoughts on the Textpocalypse” Davin Heckman discusses Matthew Kirschenbaum’s short essay in The Atlantic, where Kirschenbaum launches his idea that we should “Prepare for the Textpocalypse” (Heckman; Kirschenbaum). Heckman’s main concern is, how Artifical Intelligence (AI) and Machine Learning (ML) will devalue language, including how it will be used by corporate powers to capture even more of our cultural activity in ways we can’t even yet imagine. As he writes in his thrilling (if not chilling) clear-cut manner:
“In other words, there will be a boom followed by a bust. As the saying goes, talk is cheap. But it’s about to get cheaper. Soon, you’ll see people leaving the libraries with wheelbarrows full of words worth less than the paper they are printed on.”
Heckman’s points are clearly worth discussing, including his call for a more critical left-wing understanding of digital media, and especially his point about the “pervasive exploitation” from the corporate monopolies that also control most of the LLMs (Large Language Models) is important. However, I believe Heckman risks overlooking two things in his critique.
1) It already happened:
Text is already technologized, at least since print according to (Ong) and many others. Just think of what has happened with academic writing within the last 20-30 years, where too much is written by publish-or-perish precarious academics, and much more than anybody can ever read is available. Actually, already (Bush) saw this problem, but the technology he helped probe only multiplied the problem. This is why so much academic writing is added into schematic formats, making it easier to do literature reviews and measure quality quantitatively e.g., as h-index, while not having to actually read the papers.
2) Even though Heckman knows his Bernard Stiegler, perhaps the old bank robber can still help us from beyond the grave. Heckman points to Stiegler’s idea of “transindividuation” where the individual is “grammatized” by “tertiary retention,” which Stiegler defines as “technical supports of various kinds” (Stiegler
Organology of Dreams and Archi-Cinema 10). The idea is, as also pointed out by Heckman, that “for Stiegler, the role of the technical apparatus as the place where tertiary retentions (recorded, archived, and recirculated memory) weave one into the cultural framework of history and the future is where a broad, generalized sense of belonging is brought into contact with the individual’s singular existence” (Heckman). We internalize the memory and sensation of the technical apparatus, and it becomes part of our identity, culture and ways of perception. The technical apparatus becomes part of (critical) reflection when used properly and things go well, and we avoid falling for propaganda or manipulation. However, this balance “collapses” when confronted with social media. With social media tertiary retention is disrupting “transindividuation in favor of automated processes” according to Heckman and Stiegler.
However, did Stiegler really understand social media? And did he understand digital subjects, as they are explored by e.g., Olga Goriunova (Goriunova; cf also Pold)? In another text, Stiegler argues that Horkheimer and Adorno’s criticism of the culture industry is “at once lucid (if not prophetic) and erroneous (if not reactionary)” because they “ten years after Benjamin” and his famous text (Benjamin) still fail to think how their own thinking is already based on the technical culture industry that they condemn. In other words, they don’t include an understanding of tertiary retention and how it changes our perception and transindividuation, even if their concept and critique of the culture industry is spot on. But perhaps Stiegler is also sometimes “erroneous (if not reactionary)” measured by his own yardstick (Stiegler Technics and Time, 3 : Cinematic Time and the Question of Malaise 37-40; Cf also Andersen and Pold 74-80)? If so, this is probably not the place to find the seeds for a new left-wing critique, but rather presents a risk of joining the reactionary. Even the techno elite wants to hide their kid's smartphones.
Let me be clear: There is definitely something deeply worrying about the Textpocalypse and the “collapse of the language, the proletarianization of culture, and the destruction of the social knowledge base” that Heckman points to, not least the speed by which it is currently introduced and injected into platforms and even my writing tools such as Microsoft 365’s introduction of the AI tool Copilot. Heckman’s sharp and well-written text clearly points to matters of concern. But language as a “shire, a familiar place of comfort where we dwell” (Heckman) - isn't this a paradise myth from before Apple was introduced? Isn’t language always already technologized (as anybody reading this journal will know)?
I would argue, that there is another question, that we need to answer while trying not to drown in the flood of GPT-generated textual waste: How do we learn to read this technologized Textpocalypse, if all the structures and interfaces that we usually use to navigate our trust in a text – from references to enunciative structures and the idea (however fraught) of intentionality – are simulated to (almost) perfection by AI and ML? And to add, this is done on purpose by an industry that earns its enormous wealth by manipulating us through the misuse of our own data? To again reference Stiegler, how do we understand the many ways this tertiary retention grammatizes us? This is the question I’m left with after reading Heckman’s brilliant text. For this, we need electronic literature - more than ever - to develop writing and reading of technologized textuality.
Andersen, Christian Ulrik, and Søren Pold. The Metainterface: The Art of Platforms, Cities and Clouds. Cambridge, Massachusetts. London, England: MIT Press, 2018. Print.
The Work of Art in the Age of Its Mechanical Reproducibility. Selected Writings. Eds. Jennings, Michael William, et al. Vol. 4. Cambridge, Mass. ; London: Belknap Press of Harvard University Press, 2003. 251-83. Print.
As We May Think. The New Media Reader. Ed. Montfort, Noah Wardrip-Fruin & Nick. Cambridge, Massachusetts & London, England: The MIT Press, 2003 (1945). 35-47. Print.
The Digital Subject: People as Data as Persons. Theory, Culture & Society 36.6 (2019): 125-45. Print. https://journals.sagepub.com/doi/abs/10.1177/0263276419840409. 10.1177/0263276419840409.
Thoughts on the Textpocalypse. Electronic Book Review May 7 (2023). Print. https://doi.org/10.7273/6hjd-mp06.
Kirschenbaum, Matthew G.
Prepare for the Textpocalypse. The Atlantic 8 March 2023 (2023). Print. https://www.theatlantic.com/technology/archive/2023/03/ai-chatgpt-writing-language-models/673318/.
Ong, Walter J. Orality and Literacy - the Technologizing of the Word. London & New York: Routledge, 1988. Print.
Pold, Søren Bro.
Critical Attention and Figures of Control: On Reading Networked, Software-Based Social Systems with a Protective Eye. Electronic Book Review (2020). Print. https://electronicbookreview.com/essay/critical-attention-and-figures-of-control-on-reading-networked-software-based-social-systems-with-a-protective-eye/. https://doi.org/10.7273/gp2w-c620.
Organology of Dreams and Archi-Cinema. The Nordic Journal of Aesthetics 24.47 (2014): 7-37. Print. http://ojs.statsbiblioteket.dk/index.php/nja/article/view/23053/20141.
---. Technics and Time, 3 : Cinematic Time and the Question of Malaise. Meridian. Stanford, Calif.: Stanford U.P., 2011. Print.