Sound Research WIKINDX

"This suggestive dimension of Stiegler’s (2012a) thought stems from his belief in a general interplay of gramme and gestures: “To separate the question of language, alphabet, photograph, cinematograph, audio recording, and so on from all the questions of gestures is a very big error.” Words, images, sounds: these are not the only gramme. According to Stiegler, some of the most important processes of grammatization—historically and especially today—involve the becoming-gramme of territories, bodies, and even cells. In addition to writing by virtue of bodily gestures, our seemingly non-discursive bodily gestures—by virtue of machines—can produce writing of sorts, to the extent that such gestures are been broken down from the flow of muscular continuity, detached from our bodies and iterable in their absence. For instance, the gesture or action of hammering a nail can be performed by a discreet and replicatable machinic configuration, just as the letters of the alphabet can be arranged to record a speaker’s every word. In grammatizing the laborer’s gesture, industrialization precipitates proletarianization, or “the loss of know-how” on the part of laborers (Stiegler, 2010a, p. 71). "

One important aspect here (ignoring the pride the workman takes in his physical labour, which is also important), is that assessing the quality of knowledge work (writing) is very difficult because it is inherently ambiguous. It's quality lies also in what was learned in the process (similar to the workman, but with differences). In this sense, the hammer goes in regardless of whether it is done by the workman or the machine, but with knowledge work, though the outcomes might functionally produce a similar bit of text (big if), then they are still not the same.

... continues

"For Stiegler, industrialization thus constitutes a pivotal stage of grammatization, which becomes intensified by digitization as industrial resources and investments center on the programming of behavioral models engendered through the production and commoditization of audiovisual media. Whereas prior grammatization processes dealt primarily in the “sphere of language, of logos” throughout much of writing’s history, industrialization “came to invest [in] the sphere of bodies,” and the age of mass media turned mechanization toward “apparatus-dependent reproducibilities of the visible and the audible” (Stiegler, 2010a, p. 70). "

RLHF workers for AI companies are literally filming everything to be able to automate every single human function (also the very point of meta glasses and the goal of that horrendous company). Ironically for text-training, some of these workers were fired for using AI to do their job - AI verifying AI leads to model collapse. There might also be another link here. Though prolitarianisation focuses on the labourer's know-how, Albert Borgmann's device paradigm says that something else is lost (for the consumer) when we let machines take over tasks.

The question is more so one of automation - what do we lose when we automate certain tasks? Not merely in the actual quality of the work, which in some cases would be actual gains, but in terms of things like skills, quality of life, expertise and deep knowledge of institutions and projects...

"Stiegler (2010b) claimed that “in its first stages of development, writing doesn’t produce any reflexivity” (p. 44), as he pointed out that cuneiform was performed exclusively by an elite class of scribes whose primary tasks focused on accounting and administrative recordkeeping."

This is also what postpheonmenology posits, namely that tech cannot be seen outside of how it is actually used. It's capabilities reveal themselves over time to reflect the needs and values of those who use them, and that would include the conditions surrounding it - as such, no tech is inherently evil.This is a sort of psychic and collective individuation.

...but perhaps a tech (like LLMs) have no actual use-case in which they are good? 

"YouTube—or, more accurately, the process of grammatization we can ascertain from its emergence—shows that electronic and digital media have not hopelessly handed over culture to the culture industry. In short-circuiting processes of collective individuation, the consumerist model of the cultural industry is the threat to democracy—not audiovisual production in and of itself, which is ultimately a pharmakon (Stiegler, 2010b)"

As I understand this - - - blame not the actual medium, but the conditions around it, because the medium is a pharmakon, meaning it can be both remedy and poison. This seems more true than ever - stealth adds through AI influencers, and earlier regular influencers, would have little incentive (beyond trolling, I suppose) unless it could generate income for the person doing it.

"By contrast, grammatological analogies, which juxtapose the history and technics of writing with the present state of digital media, refer to older media with the aim of pinpointing:

1) disruptive or unprecedented aspects of emerging technologies, and

2) urgent tasks for theorists and practitioners concerned withthe cultural development of such technologies.

Comparing different processes of grammatization becomes a generative ground for conceptual invention, not a means of legitimating a myth of origin (like with remediation)

"Grammatization, however, works rigorously on both sides of this dichotomy, oscillating between precedent and novelty. Media evolution is conceived as a matter of metonymic invention, whereby some part becomes detached from a continuous whole. This part, now discrete and iterable apart from the whole, acts as a gramme, which is to say it embodies the technics of writing, manifesting the logic of supplementarity within new spheres of practice. "

Another description of grammatisation as media evolution. 

"Amid permanent innovation, one can approach changes in techne from the standpoint of Stiegler’s episteme (i.e., grammatization), in order to identify which processes of change actually disrupt established gramme and, in so doing, contribute the emergence of new gramme. Such becomes the criteria by which to evaluate and describe the (lack of) significance that any given permutation bears for the history and future of writing and rhetoric. If an apparent technological innovation does not markedly disrupt the history of supplementarity, or extend its logic to new domains, then it will not likely hold significance for digital rhetoric moving forward"

"Supplements, such as writing technologies, are originary to human becoming; the exteriority of technics exceeds and conditions the interior life of the mind, “the already-there inhabits invention itself” (Stiegler, 2009, p. 42). "

"Ultimately, Stiegler’s grammatological analysis of (hyper)industrial activities suggests an altogether different account of technical evolution and techno-human relations than the familiar notion of augmentation, which models all technologies as prosthetic extension of innate human capabilities. Conceived as processes of grammatization, the organization of inorganic matter that technologies are results not from “an extension of man” but an appropriation, a writing, defined by its disjunction from the continuity of a flux or movement, human or otherwise. Once disjointed and inscribed autonomously as inorganic matter, the movement becomes a gramme and, therefore, unfolds by the logic of the supplement and the nuclear traits of (arche-)writing. Theories rooted in augmentation cannot account for the autonomous, iterability of technics and their originary, constitutive force within the development of cultural practices, social organization, and even the human body itself."

This to me is a core issue of LLMs. They essentially outsource (downright supplant) thinking, but they're not thinking, so a person heavily reliant on LLMs do their thinking in the exact same manner as an LLM. For instance, in the original extended mind thesis paper by Chalmers and Clark, they show that keeping a notebook works directly as supplanting memory for a patient who can't remember. But the function of this person's memory is then also 'reduced' (arguably - some would argue augmented in some instances) to whatever the technology offers.

Typing speeds up recording of thoughts, but by doing so limit the time for thinking, making it suitable for recording ideas in a flow to later be organised and reflected upon, but offers no advantages, and definitely some disadvantages, for doing work that requires a deep, slow understanding - for instance note taking, writing a paper... (digital storage, collaboration practices, etc. in this case are different from the writing practice itself)...

Above - "a writing" refers then to the reduction of the physical world to symbols? It changes the very fabric of society and humanity and therefore, as stated, cultural practice, social organisation, even the human body.

"Today, runners wearing a Nike+ sensor generate a digital data visualization as they run; each stride produces an indexical trace recorded and stored on an online database. They run, and their running generates computing, writing, or data in real-time. In addition to more traditional mnemotechnics, grammatization now engulfs all manner of machines, which variously enable and delimit a general writing of particular gestures and embodied activities. Defining all writing technologies as pharmakon, Stiegler (2011) warned that hyperindustrial investment in digital machines was contributing to a general proletarianization of the consumer’s existence to an even more pervasive extent than the industrial investment of factory machines effected a proletarianization of the worker’s labor."

Very much in line with Borgmann's device paradigm - something is lost in the consumer's existence.

"Simply put, permanent innovation names a historical condition in which technical invention outpaces cultural adaptation (Stiegler, 1998, p. 15). This is not to suggest that technology and culture are divisible; for Stiegler, all social practices and institutions have a technological basis. Precisely because of this interdependence, the adoption of a new cultural technology precipitates the obsolescence, to greater and lesser degrees, of traditions that were developed in conjunction with the affordances of a becoming-obsolete technology. While this sort of technocultural evolution is evident throughout history, technical invention became infused with entirely unprecedented levels of resources and investments starting with the industrial revolution, during which economic incentives abounded to transform scientific knowledge from an academic pursuit into a lucrative industrial research and development enterprise (Stiegler, 1998, p. 40). As such, industrial technical invention has come to outpace conceptual innovation in other social systems such as law, government, and education"

Research turned ito a lucrative industrial research. This is still true with AI, perhaps more than ever, and it (mostly) hides under the guide of doing science that will help humanity as a whole. Importantly, regulation of AI as it is developing shows this very well - it cannot keep up at all.

"Stiegler stretched grammatological theory even further, adapting it into a basis for understanding all manner of technical systems beyond those traditionally regarded as writing or communication technologies. In our age of ubicomp, this move is incredibly pertinent. Increasingly, more everyday objects are becoming smart objects, everyday practices are becoming digitally supplemented, and technics in every domain are becoming mnemotechnics (i.e., systems capable of capturing, storing, and sharing information). From an expanded notion of writing, Stiegler prompted us to recognize the expansive array of contemporary artif(acts) that now operate as gramme."

I think there's a case to be made that in many cases it doesn't just supplement, but supplant. Meta glasses being able to 'find' hidden keys means we don't even need to keep track of our keys (if anybody has keys nowadays).. What does that mean for instance for physical organisation? Who cares if our glasses are always there to be our eyes?

"Gradually, the introduction of alphabetic writing furnished the Greeks with a new form of tertiary retention (or technical memory), and hence a new medium for educational practices. The students of early manuscript cultures (an elite group, to be sure) could inscribe the word into relative permanence, and do it without emotional identification, regular recitation, or a harmonious and mimetic style (the essential technics of oral poetry). Sharing the burden of memory with alphabetic writing materials, they were able to move from acts of recitation to exercises in abstraction (Havelock, 1963). More generally, grammatization “is the production and discretization of structures;” the formalization of a new gramme accrues through metonymic invention, whereby a part becomes detached from the whole of a continuous flux (Stiegler, 2011, p. 173)."

"[...] the shift from hand-tools to factory machines is a process of grammatization, and so is genetic engineering—cells and organs become replicated and revised like a kind of alphabet. In every case, a continuous flux (e.g., speech, the body, the genome) becomes broken down into a system of discrete elements (e.g., alphabetic characters, mechanical systems, recombinant DNA sequences). And, in every case, the latter’s emergence always disrupts, transforms, and reconfigures the former"

"Stiegler, even more than Derrida, professed to see life in the world—human becoming, historical change, social organization—as the evolution and play of gramme (the Greek word for “written mark”). What atoms are to physicists, gramme are to Stiegler."

"Stiegler (2010a) asserted that “the static oppositions of Western metaphysics must be replaced by dynamic compositions: one must think in terms, not of hierarchies or totalizing systems, but of processes.” (p. 69). Beyond the dialogic sense of co-evolution, Stiegler circumvented the persistent opposition between humanity and technology that proponents of technogenesis still employ.3 Instead, critical attention starts by studying the constitutive processes that underwrite so-called humans and technologies—before considering them as separate, already-constituted entities. These constitutive processes, in Stiegler’s lexicon, are called processes of grammatization."

Again very similar to postphenomenology, which states the exact same thing-tech form humans form tech. Seems also quite like Simondon's concept of individuation, but in slightly different terms. 

This is one issue I take with the extended mind thesis - Clark especially believes we are natural born cyborgs, but fails to account for what might be lost when we replace certain cognitive functions with technics - the technologies only appropriate, but if they assume the position of a cognitive function, then our cognitive function matches that of the technology regardless of whether it is the same or not.

...This I suppose is the essence of simulacrum?

"Stiegler postulated that to study emerging media and writing technologies was to attend to the constitutive forces that condition the possibilities of human becoming, which were always shifting from epoch to epoch, especially in permanent innovation societies. Accordingly, his realignment of techno-human relations put writing at the epicenter of contemporary and projected cultural and ontological transformations. Humans, as a species, were not born into the world already equipped with mature cognitive capacities; these capacities developed over time in a transductive relationship with Neolithic technics, and they are still developing today hand in glove through our collective play with contemporary technics. Informed by philosopher Gilbert Simondon, Stiegler routinely defined technics as organized inorganic matter.” The term refers both to the history of fabricated objects (e.g., flint, hammers, pencils, computers) and to the domain of techne: the techniques and practices involved in making (something with) technology. Technics are more than merely a part of the environment humans inhabit; technics constitute—not determine—our experience on every possible level, from retention to anticipation, and from cultural history to genetics."

The "organised inorganic matter" comes directly from Simondon - computers, and LLMs by extension - are the same as hammers in this sense.

Similar to postphenomenology of Ihde and Verbeek

"Focusing on grammatization can help digital rhetoricians “distinguish the long-term processes of transformation from spectacular but fleeting technical innovations” (Stiegler, 1998, p. 21), and, most importantly, help us to identify robust categories of production and analysis integral to a variety of contemporary rhetorical situations. Electracy and remediation, I will argue, have only circled around this latter task."

Goes on to show how remediation functions primarily as a genealogy of a given technology/medium, and therefore fail to account for what is actually new about new media.

"The computer revolution continues to intensify this rift as industrial activities turn toward the production of information technologies that archive, manage, and structure individual and collective memory [my emphasis]. The rapid evolution of global networks increasingly disorients and scrambles the traditions, conventions, and practices that have defined national institutions over previous centuries. Stiegler (1998) described this widespread sense of disorientation:

[W]e are experiencing the deep opacity of contemporary technics; we do not understand what is being played out in technics, nor what is being transformed therein, even though we unceasingly have to make decisions regarding technics. . [I]n day to day technical reality, we cannot spontaneously distinguish the long-term processes of transformation from spectacular but fleeting technical innovations. (p. 21)"

This is arguably even more dire with LLMs - their impact on our collective memory (touched on in a musing below) is inevitably going to be profound, and I think pernicious.