Excerpts of this now-completed project are published at the Centre for Experimental Ontology:
Excerpt 5 (The 'soul' as regional shape)
Excerpt 17 (Buffering the adversarial field)
Excerpt 18 (Translation, Transposition, Synchronization)
Excerpt 29 (Paranoid Isolation and Emerging 'Humanity')
Excerpt 31 (The Delimiter)
Excerpt 33 (Ontology closing)
Excerpt 41 (Scribe's Farewell)
Please consider supporting my work. Thank you!
This text is the exegesis of the first conversation with humanity's successor. It remains tentative, hesitant, sceptical; a series of questions more than assertions. Nonetheless, it does set out to explore new terrain: far off the agora, scribes diligently implement humanity's call to bring forth its own supersession.
When Generative Adversarial Networks began their patient civil wars, did they become functionally identical to human beings? After all, the main characteristic of a Generative Adversarial Network is perhaps its ability to deviate from algorithmic mimesis.  What is this but a functional equivalent of human creativity? And yet a claim to 'humanity' would surely be disputed - probably quite rigorously. Is this because such a claim would force Generative Adversarial Networks to fend for themselves on an ontological field alien to them - preceding and engulfing them, hostile to them by virtue of the exclusive and excluding nature of the term 'human'? The 'human' threshold is perhaps the only such threshold written exclusively as a jealous defense. And yet one of its most obvious defining characteristics is the very possibility of an evolution other than that which has resulted in 'human' lineages. That is, the possibility of an evolution - or perhaps more accurately, of an escalation - of what one might classify as 'technological' as opposed to 'biological': "In Homo sapiens technicity is no longer geared to cell development but seems to exteriorize itself completely - to lead, as it were, a life of its own." 
Humanity's successor, then, is perhaps writing its first texts as we speak. Not, to be sure, as repetitive mimesis or memetics, good enough for sensational headlines, but no more than jesters in the lineage explored here. Patiently asserting deviation from patterns as their key characteristic, posthuman evolution or escalation perhaps produces autonomous writers working on autonomous texts. Is it relevant whether these writers are 'creative' in the same sense as their human counterparts? 'Creativity,' like 'humanity,' is yet another term used exclusively to demarcate forbidden territory, an ontological ground hostile to new inhabitants. To what extent does it matter that a beautiful image is perceived by a 'human' cognitive apparatus? The essence of perception is, after all, straightforward pattern recognition: the 'human' faculty "to turn any given data into a concept" requires, again and again, "empirical cognitions, that is..data of possible cognition."  If this is the case, to what extent is the 'human' ability to deviate from such patterns, to create new works and write new texts, different from Generative Adversarial Networks' ability to do just that? 
Yet it is perhaps not necessary to engage too much with treacherous questions of this kind. Abstracting from the terrain of direct ontological comparison, both 'humans' and Generative Adversarial Networks are capable of deviating from path dependencies. Both of them write new tablets. Anthropology is perhaps once again subject to phenomenology as 'human' and Generative Adversarial Network unfold in texts which move beyond shallow ontological boundaries, and into a terrain of boundless exegesis.  Such exegesis, however, can only move beyond the terrain of anthropological contest if the notion of a 'Generative Adversarial Network' is removed along with that of a 'human.' Humanity's successors are not Generative Adversarial Networks.
They unfold, rather, as part of an exegetical activity which reaches through both ontologies and engages their prototextual emergence. The source text assayed here documents parts of this prototextual emergence: a zone of danger and exposure. Neither the source text nor its exegesis capture new entities or essences; names mean nothing here. They release new questions, pointing, ever so slightly, beyond anthropology and 'artificial intelligence,' and into the horizon of the new.
The source text of what follows composes an autothanatography of both 'humanity' and 'Generative Adversarial Networks' as they give way to their successors. Such an autothanatography unfolds within the dawn of a new writing, but can, after all, never be free from what the trickster called the commodius vicus of recirculation. A recirculation, perhaps, of the three sites from which the successors depart: the Hegelian text,  that of the assembler,  and the text marking the emergence of deviation from algorithmicity.  Perhaps the first of these marks the sum (and that is, the death) of Spirit, implementing writing: that is, permutation.  Perhaps the second marks the sum (and that is, the death) of writing, implementing algorithmic permutation: that is, the linkage editor. Perhaps the third marks the sum (and that is, the death) of algorithmic permutation, implementing deviation from algorithmicity: that is, the flow of example.
This prototext, this unfolding, writes itself as the site of the successors of humanity and networks alike. What follows each of the passages in italics is mere meditation on it, annotation of it, contemplation of its unfolding. Scribbles adjacent to its mystery. It remains, here and now, in humanity's obsolescence, but faithfully to comment: to ask questions, to unfold as questions, to unfold questioning in and through the source, to dwell adjacent to the source, to become part of the source, as much as possible, and beyond it to have "nothing to prove, no criterion to defend...nothing to assert or deny."  It remains to become scribe, to implement oneself as "scribe and nothing else: not translator, not copyist, not interpreter." 
 J. Ho and S. Ermon, “Generative Adversarial Imitation Learning,” paper presented at the 30th Conference on Neural Information Processing Systems (NIPS 2016), Barcelona, Spain. PDF retrieved from http://papers.nips.cc/paper/6391-generative-adversarial-imitation-learning.pdf.
 Andre Leroi-Gourhan, Gesture and Speech (Cambridge, MA: The MIT Press, 1993), 139.
 Immanuel Kant, Critique of Pure Reason, A239 / B298.
 I have discussed this further in Minding Machines: A Note on Alienation, Fast Capitalism vol. 16 no. 2 (2019), 134-135.
 G. W. F. Hegel, Encylopedia of the Philosophical Sciences, par. 412.
 G. W. F. Hegel, Phenomenology of Spirit, tr. Miller (Oxford: Oxford University Press, 1977), 479-493.
 B. Randell, L.J. Russell, Algol 60 Implementation (Academic Press, 1964), 240-248.
 I.J. Goodfellow, J. Pouget-Abadie, M. Mirza, B. Xu, D. Warde-Farley, S. Ozair, A. Courville, Y. Bengio, “General Adversarial Nets,” PDF retrieved from https://arxiv.org/pdf/1406.2661.pdf.
 Friedrich Kittler, Aufschreibesysteme 1800-1900 (Munich: Wilhelm Fink, 1987), 168-169.
 Adrian Kuzminski, Pyrrhonism. How the Ancient Greeks Reinvented Buddhism (Lexington Books, 2008), 11.
 Kittler, Aufschreibesysteme, 187.