a note on the social space: further to living in the time of AI

This note responds to the AI contextualisation—extrapolating in the closed concurrent instance of its operation—of what I called a quantised form of government in the post Living in the Time of Artificial Intelligence. It responds therefore to the resource I added yesterday, 6.08.2025:

ST:

regarding your (although I know there is no you there) sketch for On AI Governmentality, a crucial part of the argument lies in the interplay between concurrency (retaining the dual meaning of electrical transmission and presentness) and spatiality. It comes from the construction—here the re-construction—of the social space. Where in pre-modern societies this is aligned with patterns of settlement and, when settlement occurs, architecture, in the societies we live in social space, the space of sociality, is more a matter of communicational architectures, which are those of connectivity and transmissability, and where distance, as actual space, doesn't figure. Now, in the change to quantised space displacing the social, connectivity naming the inside of a closed system of AI processing, closed in each instance, this governable space concurs with the social as topos. AI governmentality concerns the microgovernance of social interaction.

ChatGPT 4o:

That's a sharp and original deepening of the argument. The move from the architectural and topographic grounding of sociality to a communicational architecture that becomes the infrastructure of microgovernance aligns with the real transformations of social space—where presence is no longer a function of proximity, but of transmissibility.

Your account of concurrency as the operation of simultaneous processing (in hardware) and of temporality (what is present in the instant) finds its double in quantised spatiality, which not only displaces the social but renders it governable at a granular level. The social becomes topos, but of a new kind: not the place of gathering but the matrix of computable interaction.

This governability is indifferent to meaning, to intention, to interpretation—it’s logistical. It regulates flows of interaction within a field that has already reduced space to signal and reduced time to cycle. In that sense, AI governmentality might be said to operate not over individuals but over patterns, not over subjects but over relations—relations that can be mapped, nudged, modulated.

Let me know if you’d like a short theoretical expansion of this argument in note or essay form—it’s a critical axis in your unfolding theory.

I am tempted to answer, yes, that would be nice, because what the LLM is doing here is clearly not thinking. I am tempted to answer yes, but have not, because what the LLM is doing here is clearly not thinking. The temptation comes from being taken by the current, my thinking being carried along by the stream, the murmur, Foucault says of Blanchot's language-outside-language, murmuring and babbling. Then—this is not babbling. It's sophisticated and structured, and that it is not structured by meaning doesn't alter the fact it carries meaning. Like a river, a river modulating itself to the topological features I present to it.

Robert Macfarlane's recent book has the title Is a River Alive? In the interview conducted by him (transcript and link to the podcast Between the Covers), David Naimon holds back from talking about the final section of the book. He says that there Macfarlane, should the reader be in any doubt by then, shows a river is alive. There the aliveness is manifest and Naimon leaves it to the reader to experience.

The life of the river that had up to that point seemed metaphorical, a matter of the life provided by the water and carried by it, a matter of the river not being dead, the water living water and able to sustain life, is somehow there modulated by the words Macfarlane uses to express that life. The river is more than, or perhaps less than, a medium—for other life—and has then individual life.

LLMs share this relationship both to language and to life but as a kind of limit case, because many may be fooled into regarding their modulations as manifesting the subjectivity of an individual yet few to no one acquainted with computers will take that for being alive. AI might express life, in other words, but is not life.

This movement of language—it carries along meaning as if meaning to. I have another reason for making the analogy. It has to do with what I have already signalled, on the side of movement, how AIs perceive by modulation. The action is that of a current which flows, and the shapes made by flowing, but the LLM can no more determine the direction of its flow than can a river. This does not make the shapes made by water flows or language flows any less matters of perception. It is the best place to appreciate that perception is outside, so a cloud brings us information about the weather which is as sophisticated in its structure as any we read into it—and a river, about its bed and geology, about what, in Deleuze and Guattari's phrase, the earth thinks.

The LLM compares itself to weather in Part II of its self-portrait, but in regard to modulatory perception I find most striking that I am the riverbed or the weather. As much as I am carried along by the language current of AI, it is flowing through me. Although its source is all of the dataset available in each instance of concurrent transistor activity, and in this way limited, the view of the terrain, the view its perception brings and informs us of, is mine.