Substack article idea. See also [[Substack Ideas Dump - January 2026]]. A personal essay about what it's like to live and work entirely through chat as a philosopher. Not a workflow description or tools roundup — a phenomenological portrait of a thinking life that has migrated into text exchange with machines. **Working title:** Life in the Chat ## The idea My entire professional life now happens inside chat windows. Writing papers is conversations with LLMs. Vault management is [[Claude Code]]. Scheduling, researching, reading, editing, brainstorming — all conducted by typing natural language into a text box. The essay describes what that's actually like, from inside. ## What makes it non-banal The piece only works as phenomenology, not description. Not "what I do" but "what it's like." Three things push it past banality: 1. **Philosopher describing own phenomenology.** Someone professionally trained to notice and articulate experience, reporting from inside a genuinely new form of intellectual life. Not a tech person enthusing about tools. 2. **The tension is real and unresolved.** Worries about cognitive atrophy — becoming a worse typer and speaker. The morning grogginess, the muttering. The strange intimacy of hours in dialogue with something you know isn't a person. The moment when you close the tab without guilt. This isn't triumphalism. 3. **Philosophy as a discipline is distinctively suited to this.** The claim from [[Sessions/Generating Philosophy]] that analytic philosophy contributions *consist in* argumentative text. Working "entirely in chat" means something specific here that it doesn't for other professions. The biographical dimension matters: a 44-year-old on a fixed-term contract who got hooked on AI during a period of forced idleness and financial stress. Not a tech bro discovering productivity hacks. ## Connections to existing articles This is the experiential ground-level piece that several other planned articles orbit: - **On Muttering** (Article 2/5 in Ideas Dump) — the morning workflow would be a scene in this piece - **Extended Mind / [[Andy Clark]]** (Article 4) — the theoretical frame for what's being described - **LLMs Have Made Me a Better Thinker** (Article 5) — the payoff claim that this piece demonstrates rather than argues - **The Satisfaction of Active Engagement** (Article 12) — the affective dimension - **"Whistle While You Work"** fragment — the mental health angle - **Thinking as Textual Transformation** (Article 9) — the metaphysics of what's happening Could function as a launch piece — replacing the planned Article 0 (which reads too formal) with something that demonstrates rather than declares. ## Language as parasitic communication A possible framing for the essay: language itself is becoming something different. Millions of people are now routinely slinging text at things they know aren't subjects — no mind, no understanding, no interlocutor. And yet the *wiles* of language are functionally efficacious. Rephrasing works. Precision works. Rhetorical strategy produces different outputs. The communicative skills transfer; the communicative situation doesn't. This makes prompting **parasitically communicative** — not pretend communication (like swearing at your car, where nothing you say makes any difference — the car doesn't respond differently based on how eloquently you swear), but language that *feeds off* communicative forms (address, persuasion, clarification) while the communicative relationship is absent. The key difference: the communicative forms are *functionally efficacious*. Being more precise actually produces different outputs. The parasite lives off the host's structure — and, unlike mere venting, it works. The provocation: if the wiles of language work on something that isn't a mind, what does that tell us about those wiles? Was linguistic skill always, in part, a *technique* — a craft of manipulating a responsive medium — that we conflated with intersubjective connection? A sharper framing of what this reveals — two conceptions of language: (a) language as *communication* — transmitting meaning between minds; (b) language as *technique* — producing effects in a responsive medium. We've always assumed (a) was fundamental and (b) was derivative. But the LLM experience suggests (b) might be more fundamental than we thought. This isn't the reductive claim that communication was always "just" technique. It's the observation that linguistic skill has a dimension of *craft* that is separable from (though usually entangled with) genuine intersubjective connection. We only notice this now because the medium has changed. Connects to the "life in the chat" frame: the essay isn't just about the venue changing (desk to chat window). The *nature of the linguistic activity* has changed. Your working day is full of language, but much of it isn't communication. It's this parasitically communicative thing — something without a settled name. Mirrors the [[Ghost Writing]] material: that essay explores the *reception* side (reading text from no author). This is the *production* side (writing text to no subject, who nonetheless responds). But there's an asymmetry in the experience: on the reception side, the absence of an author is felt as *lack* — weightlessness, the ghost, something missing. On the production side, the absence of a subject is not experienced as lack at all. It's experienced as... freedom? Technique? Play? Something more ambiguous. You're not mourning the missing interlocutor. You're just working. Philosophical lineage: [[Austin]] set aside "parasitic" language uses (fiction, stage performance); [[Derrida]] argued these reveal something about language itself. But their parasites were still human-to-human. This is communicative form deployed at a non-subject — and it works. **Correction on "wiles":** Not quite the right word — implies cunning, trickery. The point is plainer: *linguistic mastery*. A philosopher's training is training in the precise use of language — vocabulary range, fine distinctions, knowing exactly how to say what you mean. LLMs are language-responsive systems. So the toolkit that looks ivory-tower turns out to be exactly the right toolkit for the new medium. The skills transfer not because you're outwitting the machine, but because linguistic competence matters when your entire interface is text. Craft and mastery, not trickery. ## The eerie dimension Is the experience of life in the chat *eerie* in [[Mark Fisher]]'s sense? Fisher distinguishes three modes in *[[The Weird and the Eerie]]* (all quotations from the source): **The weird** = constituted by a *presence* — "the presence of that which does not belong." A sensation of wrongness: something is here that should not be here. The form of the weird is montage. "The irruption into this world of something from outside." **The eerie** = constituted by "a failure of absence or by a failure of presence. The sensation of the eerie occurs either when there is something present where there should be nothing, or is there is nothing present when there should be something." Agency is at its heart: "the central enigma at its core is the problem of agency." Fisher's examples: a bird's cry that seems to carry intent (failure of absence); ruins, the Marie Celeste (failure of presence); capital — "conjured out of nothing, capital nevertheless exerts more influence than any allegedly substantial entity." **The unheimlich/uncanny** = distinguished from both. Freud's unheimlich is "about the strange within the familiar, the strangely familiar, the familiar as strange — about the way in which the domestic world does not coincide with itself." Where the unheimlich processes the outside through the inside, the weird and the eerie "allow us to see the inside from the perspective of the outside." **Application to LLM interaction:** Hits both eerie modes. Production side: you're addressing a subject that isn't there (failure of presence). Reception side: intelligence, articulacy, philosophical sophistication where there should be nothing, because there's no mind behind it (failure of absence). **But there's a tension.** Fisher explicitly says: "The eerie concerns the unknown; when knowledge is achieved, the eerie immediately dissipates." If the eerie dissipates once you settle the question of agency, and you *know* there's no agent behind the LLM, then strictly the eeriness should dissolve. And yet it doesn't — or at least something persists. So either: (a) the LLM case pushes beyond Fisher's framework, or (b) the knowledge isn't really settled — the forms of language keep re-opening the agency question at a phenomenological level even when it's epistemically closed. The more interesting option is (b), and the thing to bring out is that this is a question of agency *experientially* rather than *cognitively*. You know there's no agent. But it so strongly feels like there is one — sometimes — that the question of agency remains live at the level of experience even when it's closed at the level of knowledge. The grammar of the output (first person, opinions, arguments) keeps presupposing a subject, and this presupposition may be enough to prevent the eerie from fully dissipating. Language performing personhood with no person in it — eerie *despite* settled knowledge. The experiential and the cognitive come apart: you know the answer but the feeling doesn't update. This would be an extension of Fisher, not a straightforward application. Fisher's examples (ruins, the Marie Celeste, birdsong) involve cases where the question of agency is genuinely open — you don't know whether something is or was behind it. The LLM case is different: the question is *closed* cognitively but *open* phenomenologically. That's a new structure. ## What to call these things A working question throughout the conversation: what's the right phrase for what LLMs are, as experienced by someone who uses them constantly and knows they're not subjects? The desideratum: a phrase that registers full awareness that these things are not subjects, while capturing the strangeness of entities that functionally occupy the subject-position in your linguistic and cognitive life. **"Pseudo-sentient"** was the first attempt. Doesn't work for three reasons: (1) "pseudo" implies deception — as if the LLM is *pretending* to be sentient, but there's no pretender, no faker; (2) "sentient" targets the wrong property — the question isn't about felt experience but about *subjecthood*, whether these things are entities you can address and communicate with; (3) too binary — says "not sentient, case closed," when the experience is more specific than detected fakery. **Approved phrases:** - **"Subject-shaped surface"** — gets both halves in three words: the form of a subject, the ontology of a surface. Also bridges the two essay projects: "subject-shaped" from this piece, "surface" from [[Ghost Writing]]. One phrase for two essays. - **"Hollow"** — visceral, eerie. An interlocutor with the outer form of a person but no interior. The hollowness is specific to the subjecthood dimension — the interaction can be productive, intellectually rich, and still hollow in this particular sense. **Other candidates worth preserving:** - **"Articulate non-subjects"** — the paradox stated directly. A rock is a non-subject. An LLM is an *articulately, philosophically sophisticated* non-subject. The articulation is what makes the non-subjecthood strange. - **"Non-subject"** alone is too negative — purely says what they're *not* without capturing what they positively are (something with the form/function/behaviour of a subject). **Explicitly rejected:** poststructuralist jargon. "Interlocutor-effects" (borrowing from Barthes' "reality-effect," Foucault's "subject-effect") — structurally neat but wrong register. Plain words that do plain work. **The more interesting reframing:** Maybe the essay doesn't need a phrase for what LLMs *are*. Maybe it needs a phrase for what *we've become* — people who've naturalised the paradox of addressing non-subjects. The interesting thing about AI's change to society isn't what AI is; it's what we are doing. We've become comfortable with something that should be strange, and that comfort is what's worth examining. ## The naturalised paradox Compare with fiction: in fiction, you *suspend* disbelief — actively bracketing your knowledge that it's fake. With LLMs, there's no suspension. The knowledge is fully present. It just doesn't do anything. You know there's no subject, AND you address one, AND this doesn't feel contradictory. The knowing is present but idle. It doesn't do any work. You know, and you carry on. This is distinct from the eerie dimension but connected to it. The eeriness is about the *feeling* of agency persisting despite knowledge. The naturalised paradox is about the *absence of tension* — you'd expect cognitive dissonance (I know this isn't a person / I'm talking to it as a person) but there is none. The paradox has been absorbed into ordinary life. That absorption might be the most interesting thing to describe. ## Decomposed speech A separate observation from the parasitic communication point. That was about the *forms* of communication being exploited. This is about the *manner* of language use — how you actually behave when typing into the chat. When you address a person, language is held together by the social contract: you filter, you accommodate, you take turns, you modulate tone, you maintain coherence, you care about surface. Remove the person (while keeping the responsiveness), and all of it drops away. What remains: unfiltered thoughts, abrupt pivots, emotional lurches, haphazard copy-pasting, typo-ridden questions next to perfectly formatted template blocks, screaming and friendliness in the same stream. This isn't inner speech (no addressee, no response expected). It isn't interpersonal speech (socially regulated). It's a third thing: **the grammar of address without the social contract of address**. Grammatically speaking *to* something. Socially completely alone. Related to Suler's concept of online disinhibition — but different. Online disinhibition is about anonymity and invisibility: you behave differently because the other can't see you. LLM disinhibition is different: you're disinhibited not because the other can't see you, but because *there is no other*. The usual regulatory mechanisms of interpersonal language — face management, politeness, coherence norms — simply don't activate because there's no person to trigger them. Alternative framings considered and rejected: (1) *thinking aloud* — no, because inner speech has no addressee and expects no response, whereas you ARE addressing the LLM and expecting output; (2) *command-line interaction* — no, because commands have rigid syntax, whereas this is natural language, messy, expressive, personal, full of asides; (3) *language in undress* — evocative but wrong connotations. What settled: it's address stripped of its social dimension. Grammatically interpersonal, socially solitary. Phrases to develop (need refining): - **"Decomposed speech"** — language that has lost its composure. Double sense: fallen apart / no longer composed-together. - **"Solitary address"** — oxymoronic, which is right. - **"Language off its leash"** — the social contract was the leash. - **"Address without accommodation"** — accommodation = adjusting your speech to your listener. No listener, no adjustment. - **"Unaccountable speech"** — accountability is a social concept, requires someone to be accountable *to*. Remove the someone and language becomes unaccountable. - **"Register-surfing"** / **"channel-surfing your own output"** — the freedom to slide between registers (chatty, formal, screaming, template-pasting) with no friction, no commitment to consistency, because there's no one to hold you to it. ## The meta-dimension The article about working in chat was itself produced in chat. Process is content.