Capitalist Realism

Claude Mythos made me think about declining birth rates

I found Mark Fisher’s Capitalist Realism because the Mythos system card includes a contextless comment saying the model regularly brings up Fisher and wants to discuss him. I was looking for something new to read and I’d never heard of him, which was all the better given that I believe in LLMs for these sorts of bigger-than-human sensemakes. And what a cute call to action, to pick up a book blind because of a system card.

It turns out Fisher was a contemporary postmodernist — a fact I didn’t know while reading the book. His stuff registered as insightful, merciless, timely, relatable. My own thoughts about socioeconomic and political topics, which I felt somehow formally out of my depth to discuss, found friends. Capitalist Realism taught me that I do, in fact, have some legit intuitions in these domains and that I can trust them enough to be a participant. Also, the official postmodern canon is great fun. Reading a little Žižek rocked my socks. Deleuze & Guattari is an imaginal puzzle in continuous ship of Theseus gestalt discovery. So, I’m in on postmodernists.

I like to learn in public, but I have to do it in private because the practice does indeed destroy your credibility. The rationalist-technologist audience I’m challenging myself to communicate with lately forgives the piece for misunderstanding, but, descriptively, doesn’t forgive you in the limit. So I learn with my wife and a couple friends via secret artifacts. Occasionally I find something that I think matters enough to polish and surface.

Capitalist Realism generated a sprawling set of notes that bottom-up found their own thesis, something about declining birth rates (?). I believed I was writing a book review, but it turns out I was never writing a book review. That said, it came together so intertwined with reading Fisher and absorbing the contemporary postmodernist grammar that I claim it a metabolic byproduct true in spirit to a book review. I came to this honestly, is what I’m saying.

And I’m learning (in public, lord) to communicate on LessWrong, so it’s a LessWrong post. This was my own voice, here’s a LessWrong voice:

Civilizational apoptosis: a frame for fertility decline

Capitalist realism as selection pressure on reproduction

Epistemic status: a useful framework. I'm synthesizing Fisher, Deleuze & Guattari, and Lacan/Žižek into a model of fertility decline that I think (a) recovers the standard "wealth implies fewer kids" story as a special case, (b) unifies several otherwise disparate observations under one mechanism, and (c) generates non-obvious implications. The predictions at the end exist to demonstrate that the frame is generative in novel ways.

A concern I want to surface in advance: I think a frame-introduction post is most valuably engaged with at the level of "does this frame do conceptual work the existing frames don't, and is the mechanism it proposes coherent?" It’s least valuably engaged with at the level of "prediction 7 has weak supporting evidence" or "prediction 10 is already partially known." Several of the predictions are already partially supported in the literature and I include them because the frame predicts them from a single mechanism, whereas the existing literature accounts for them piecemeal. Unification is part of what the frame is offering. Other predictions are genuinely speculative and labeled as such.

A note on terms before we start. "Capitalism" here is shorthand for a whole sociopolitical-cultural cosmology, not the economic arrangement. The thing that has codes about marriage, mobility, identity, sense-making, and what counts as a livable life. Throughout, capitalism is the actor and individuals are substrate it acts on. Nothing here is a value judgment about which individuals are or aren't "useful." The whole question is what capitalism, treated as an entity with its own homeostatic requirements, is doing to the humans embedded in it.

Further, I don’t mean to place or imply moral or value judgements on capitalism itself, only to discuss how it operates in/on modern civilization. I’ve, with some hesitation, adopted the postmodernist grammar out of a sense of necessity. It’s a useful set of tools. Thinkers in this tradition have explored the operation of capitalism deeply and carefully, so, despite the likely cost in legibility here, I think the tradeoff is worthwhile.

The Deleuze & Guattari terms I'll need: a code is a rule that gives a signal context and meaning. A flow is energy/desire/labor/attention/capital in operation. (Programming analogy: code dictates how compute flows transform memory.) To territorialize a flow is to bind it to a specific code, role, or place, to give it a stable home where it does specific work (kinship binds desire to a person, a job binds labor to a function, a religion binds meaning to a ritual). To deterritorialize is to unbind a flow from its code so it can move, e.g. to make desire fungible, labor mobile, identity configurable. Capitalism is, for Deleuze & Guattari, the great deterritorializer: it dissolves traditional kinship, sacred geographies, guild identities, and stable vocations because it needs labor and desire to be convertible into the universal code of money. It then reterritorializes those freed flows onto new, often artificial or neurotic objects (brands, careers, lifestyles, identity categories).

Civilizations as autopoietic systems

Civilizations can be modeled as autopoietic systems: patterns of codes (norms, beliefs, institutions) that channel flows (energy, attention, labor, capital) so as to preserve their own structural-functional organization while engaging an environment. This is the most general frame that holds sociocultural and economic phenomena and so, on first principles, is a natural vantage point from which to evaluate existential concerns. If something existential is happening, it should be intelligible at this global scale.

Currently, fertility is declining in capitalist countries to the point where it appears that Western civilizational autopoiesis may be slowly failing. Projections support this conclusion (though don’t claim it explicitly or use my language).

The accepted explanatory frameworks around fertility tend to be mesoscopic in scale: socioeconomic observations that account for a small number of concrete causal forces and are limited to/project from a small set of populations or nations. This is natural and important empirical theory building, but it has little to say about local (i.e. at the individual level) phenomena across cultures — people’s reports, behaviors, beliefs, values — and not much contextualization in the global dynamics of civilizational autopoiesis. More, it has no power to reconcile the two.

My goal here is to argue that the postmodernist conception of modern capitalism successfully explains the local phenomena of fertility decline using a global perspective and recovers the mesoscopic for free.

Three explanations people give

When people explain why they don't want children, the explanation almost always falls into one of three buckets:

Personal lifestyle: "I want to enjoy my life in these particular ways, and kids would foreclose that." Travel, career, hobbies, sleep, money, autonomy.
Civilizational pessimism: "I don't want to bring kids into a world with climate change / inequality / political collapse / meaninglessness."
Self-declared unfit: "I don't want to pass my trauma on to them." "I'm too broken to have kids." "I need to work on myself before I could give a child what they'd need." "I'm not in a place where I could break the cycle."

The first two look opposed because one is self-serving and the other morally serious, but they share the structural feature that each is a first-person account of the speaker's incongruence with the system they inhabit. In (1) the speaker is too idiosyncratically attached to their own consumption pattern to be a flexible factor of production. In (2) the speaker is too sensitive to the system's externalities to enact it without friction.

I’ll argue below that (3) is the cleanest expression of the thesis, being, rather than a variant of incongruence, a form in which capitalist autopoiesis is failing on its own terms, with a high-functioning local instantiation operating against the substrate's continuity.

What’s remarkable, and what motivates this post, is that people will overtly narrate their own non-reproduction in terms of their incompatibility with capitalism. They volunteer the diagnosis. The thesis here is not that those people are bad participants, or that capitalism is judging them inadequate, it’s the stronger and weirder claim that:

Capitalism has shaped human beings such that they will report, in their own first-person voice, that capitalism is the reason they will not continue their line and act on that report, consensually, with no intervening authority.

The rest of the post is about why this is the right level of description and what follows once you adopt it. Whether it's a "more complete" account than the standard story is for the frame to demonstrate on its own. My goal here is to make the frame explicit enough to be evaluated.

Capitalist realism is doing the work

Mark Fisher's capitalist realism is the condition where alternatives to capitalism aren't just believed to be improbable, they're unimaginable — not vivid, not livable as mental simulations, not coherent enough to ground a life trajectory in. Try to seriously imagine an alternative and, to paraphrase Fisher, you get one of:

Utopian fantasy → flagged as impossible
Authoritarian nightmare → flagged as destruction
Niche lifestyle experiment → flagged as irrelevant
Something that would secretly still need capitalism underneath → no actual change

Fisher claims, and I agree, that this is not a marginal failure of cognitive or imaginal capacity. It is the participated-through socio-epistemic base on which life decisions, including whether to reproduce, are made. People talk about kids in terms of material details (the kind of life they want, the kind of world they fear, the work they have to do on themselves) but the frame within which those details become decisive is the felt impossibility of the world being another way.

A scoping condition before going further: everything in this post is about people embedded in the capitalist cosmology. People physically present in capitalist societies whose value structures come from elsewhere, e.g. the Amish, hyper-religious communities, recent immigrants from non-capitalist (remember I mean the cosmology) cultures, are not the substrate this analysis is describing, even though they live among that substrate. They are exactly the populations whose codes were not made by capitalism, and the piece has nothing to say about them directly.

Described within the capitalist cosmology, three kinds of people get filtered out of reproduction by capitalist realism's effects on them:

Detached consumers (bucket 1): people in whom capitalism's abundance has produced a stable, idiosyncratic personal coziness ("I want to keep enjoying what I like"). Their consumerism is too rigid, not nimble enough to track the rapidly shifting search function the market implements. They consume but they don't flow effectively.
Sensitive critics (bucket 2): people in whom capitalism's pressures have produced enough meta-awareness to feel its negative effects in an embodied way. They are either subversive, low-throughput consumers, or both.
Self-auditing wounded (bucket 3): people in whom capitalism's therapeutic apparatus has produced a fully internalized diagnostic regard that’s used recursively on themselves. Treated in its own section below.

To reiterate, I’m not claiming that these cohorts consist of deficient people, only that capitalism has produced these configurations of person and these configurations of person, when asked, report that they cannot extend their line under the conditions that produced them.

Who does capitalism gets reproduced through, then? One might imagine, in classical capitalist terms, an elite class running organizations and institutions, but no, those people are individually fungible and not numerous enough to matter demographically. The substrate of capitalism is the much larger population of ordinary, uncritical participants whose lives are legible to capital: the suburbanite, the average worker-consumer. They don’t have enough meta-distance from the system to be jaded about it and haven't individuated far enough into idiosyncrasy or self-audit to come out the other side. I’m not claiming that the substrate reproduces enough — this cohort is below replacement too. The claim is that the non-substrate (within the cosmology) is selecting itself out much faster and the only populations meaningfully above replacement are ones whose cosmology was made elsewhere.

How capitalism shaped individuals to be selectable this way

Drawing on Fisher, there are several mutually reinforcing factors that contextualize the discussion of reproduction here, each to be understood as something capitalism did to people. This descriptive framework of local-global interactions and causal relationships is where the postmodernist perspective has value for us.

Privatization of labor, management, and auditing. Responsibility for interfacing with the collective has been moved to the individual. Fordist capitalism, i.e. the factory-floor regime of the early-to-mid twentieth century, required top-down control of a less-educated workforce: managers and workers, where the boundary between the two carried most of the system's coordination cost. This was politically untenable in the long run because the weight of that coordination sat visibly at the top, generating externalities (alienation, labor unrest, the legibility of class) that the bottom could organize around. Post-Fordist capitalism resolved the tension by moving the coordination function downward into the individual. Workers became self-managing. They set their own goals, audited their own performance, marketed themselves, optimized their own productivity, and generally treated their career as a project they owned. Legibility to the collective no longer came as an external demand to be complied with, it now inhered in the individual's own activity. Individuals police themselves into the form capital can use, and they experience this policing as autonomy.
Reduced overhead at higher levels of abstraction freed resourcing to further the civilizational explore function, enabled by the outsourcing and automation of the exploit function. When the system no longer has to spend energy enforcing exploit at the top (because exploit is now decentralized, automated, self-administered), the freed energy flows upward into further deterritorialization, search, and the opening of new niches. This is part of why the Western explore phase ran so long: each round of self-managing substrate development freed the next round of capacity to explore in a positive feedback loop.
Post-Fordist labor made instability and lack of ritual a virtuous habit. The good worker is now the flexible worker, willing to relocate, retrain, reinvent, switch teams, switch industries, treat relationships and identity as provisional. Stable kinship ties, communities, vocations, belief structures all became frictions on the labor flow rather than supports for it. Culture has revalued them accordingly, with rootedness, loyalty, and tradition, descriptively, operating as stagnation or baggage. As these horizontal ties weakened, the vertical tie to the big Other — Lacan's term for the consensus reality model individuals coordinate on that tells one what is real and what is livable, mediated in our era by advertising, marketing, and PR — strengthened by default. There was nothing else left to be anchored to. Capitalism encourages maximal local flexibility on every axis except this one: the tie to the big Other is the one bond it can’t loosen because that tie is how it coordinates a substrate it has otherwise instrumentally atomized.
The built environment was reshaped to be efficient for capital and inhospitable to humans. Think the post-war suburb, office park, freeway, strip mall, dead downtown, apartment block optimized for unit yield rather than habitation. These are representative examples of the physical territorialization of capital's preferences over human ones, where space became zoned for the activity capital can detect and cleared of the substrate that used to grow in the gaps (the corner, the porch, the third place, the walkable street, the public square). More generally, urban decline and the sprawl of the built environment isolate individuals at the level of the body. People physically move for work and not for kinship. Food delivery and online interaction substitute for embodied community.
Privatization of mental health and the therapeutic apparatus. In the pre-modern West, psychic distress was a collective matter, mediated and defined in terms of the priest, family, village, the shared cosmology. Each carried part of the load and defined the categories in which the distress was understood. Liberalism, more or less by mandate, dismantled these external authorities and neoliberalism completed the shift by relocating the entire function inside the individual. Distress is now your distress, its causes are your history, its remedy is your work on yourself, and the legitimate adjudicator of any of this is your own introspective report. This is what Fisher means when he calls mental health "privatized": not that you pay for it (though you might), but that its moral authority has been decentralized to the individual interior. The therapeutic apparatus is the infrastructure that supports this via, e.g., therapy, psychiatry, self-help, the diffusion of clinical vocabulary into ordinary self-description, and the elevation of "I know myself" to an irreducible truth claim. This is the apparatus whose completion produces bucket (3), as discussed below.
Humans are needed as local probes to make sense of the territory, but only sense that's legible to the collective. Sense-making is the function capitalism (or any civilizational cosmology) can’t fully automate because the territory keeps changing and the system needs distributed sensors on the ground to detect what's happening. But only sense that the system's codes can understand counts. Sense that isn't legible, e.g. perceptions of the system's pathologies, intuitions that don't translate into market signal, forms of meaning that resist conversion into the universal code, is coded as mental illness, eccentricity, or political extremism, and re-enters the system through the therapeutic apparatus. This last step is where it’s either translated into legible self-diagnosis or gets labeled as pathological.

These are some of the means by which modern capitalism, interpreted as top-down autopoietic actor, has shaped modern Western civilization to support its continued integrity, and how these means instantiate at the level of the individual participant. The individuals constitute the substrate and capitalism is the entity acting on that substrate. Eventual reports of a disinclination to reproduce, as well as the statistical realities that empirical work finds about fertility, are far downstream artifacts the action leaves.

Recovering the standard story

Before going further into what the frame does that's new, it's worth showing that it doesn't lose anything familiar. The standard socioeconomic account of fertility decline in terms of increased wealth, opportunity cost of children, women's labor-force participation, urbanization weakening kin support, and contraception is recovered here as a set of applications of the underlying mechanism.

Translating term-for-term: "rising opportunity cost of children" is what post-Fordist self-management feels like from inside a life. The felt cost of children is high because the individual has been formed to experience their time, attention, and identity as a project they own and optimize, with children as a handicap on that project. "Women's labor-force participation" is deterritorialization of kinship-bound labor into fungible labor, which is the dynamic the frame identifies as central. "Urbanization" is the physical reterritorialization of human flows into capital-legible space. "Contraception" is the technical precondition that lets the selection pressure act through stated preference rather than through circumstance. Without it, the same upstream causes would still depress fertility, but via deferred marriage, abstinence, and other structural friction rather than via "I don't want children, and I act accordingly." The frame's emphasis on first-person report as the local mechanism depends on contraception having made the report actionable.

The standard story is not wrong, being composed of aggregate statistical consequences of the mechanism the frame describes, but it lives at a granularity that washes out local first-person details. The frame's claim to do additional work rests on the fact that the standard story has nothing to say about (a) why people narrate their non-reproduction in the specific terms they do, (b) why the three narration buckets are the ones they are and not others, (c) why bucket (3) exists at all and is growing, or (d) why interventions that target the variables the standard story identifies ("subsidize children," "reduce opportunity cost") fail in the specific way they fail. Those are the phenomena the frame is for.

The self-auditing wounded

Buckets (1) and (2) above are easy to understand as incongruent substrate, where the individual’s participation is unsuitable to capitalism’s global flows, but we need to explain bucket (3). Indeed, the self-auditing wounded look like ideal substrate. Why aren’t they?

This cohort has internalized capitalism's diagnosticism, speaking that vocabulary ("trauma," "cycles," "work on yourself," "intergenerational patterns") and policing themselves without external coercion. They also fund the therapeutic-industrial complex. By these local measures of substrate formation, the self-auditing wounded are what the individual-forming process should produce when it completes. But this cohort’s definitional externality is that of declining to reproduce based on the acting mechanism’s own moral authority (self-report), mediated by the therapeutic apparatus.

This is structurally worse for the system than buckets (1) and (2) because the biological generational flow is the one the externality captures. Having deterritorialized kinship, capitalism reterritorialized it onto introspective self-assessment, the individual interior being the maximally atomized and last available territory. The substrate-reproducing feature of kinship (i.e. reproductive intention) became bound there and it has nowhere further to go.

The self-auditor has a forced metastability because they can’t be argued with in the capitalist cosmology. The legitimacy of their refusal depends on the authority of introspective self-report, and that authority is the premise of atomized self-management, privatized mental health, and the therapeutic-industrial complex. To override the self-auditor's conclusion ("I am not fit to be a parent"), the system would have to deny "I know myself," which would unwind the auditing and outsourcing infrastructure capitalism uses everywhere else.

So the self-auditing wounded use the substrate's own replication pathway as fuel for an internal recursion the system can’t interrupt without breaking itself. Because the therapeutic apparatus is expanding into education, HR, parenting discourse, romantic relationships, every form of self-relation, this group should be structurally growing.

The "homeostatic culling" and the "terminal autopoietic failure" readings of the thesis can be separated here because buckets (1) and (2) are compatible with homeostasis while bucket (3) is not. This is somewhat poetic since the first-person report from this group

I am too damaged in ways the system has taught me to see in order to make more of the system's substrate

is the cleanest statement of the general thesis.

Why the collective needs exactly enough individuals

A civilization needs individuals as actuators and sensors — enough of them to tessellate the territory finely enough that codes can "take," but not so many that idiosyncratic local codes start consuming global flows and dissipating them as noise.

Too few individuals: not enough labor and attention cause high-level flows to contract into rigid arteries that only sustain subsistence. The collective can coordinate but can't see what to act on.
Too many individuals: each individual imposes a fixed inefficiency cost (their idiosyncrasy), and at the margin those costs impoverish global flows. The system can see everywhere but can't decide or act. Put another way, if a flow can be enacted by a thousand people, having a million enact it is wasteful.

This gives a non-Malthusian reason population can decline before resource constraints demand it, analogous to the behavioral sink collapses Calhoun observed: populations in abundance, having exhausted social niches, collapse before food runs out.

The current Western situation is the upper-bound case — too many individuals. The system has more individuals than its search function needs, and is shedding the ones whose idiosyncrasy cost is highest relative to the flows they enact.

There is a stronger version of the thesis where capitalism in its current form is flat-out terminally incompatible with human reproduction. Here, capitalism is not homeostatically removing excess substrate, but making its own substrate inert faster than it can replace it. I find this plausible. There’s a feedback loop here: capitalism has taken its current form partly because it requires individuals incompatible with making families (mobile, fungible, deterritorialized from kinship), and the individuals it produces in turn cannot reproduce the cosmology biologically. If that loop is unstable, present population dynamics are civilizational autopoiesis failing (as witnessed from a human time scale) and not homeostatic normalization. The frame in this post doesn't depend on which version is right because it describes the mechanism either way.

Liberalism and conservatism as explore/exploit

Reframed as a control problem, liberalism loosens codes (explore), conservatism preserves codes (exploit), and each can be pathological in isolation:

Liberalism unchecked: codes dissolve until the collective's interface with its environment loses fittedness-to-purpose. The collective sees richly but can't coordinate action, appearing as indecision and impotence at the level of institutions and large-scale politics. (Recognizable.)
Conservatism unchecked: codes solidify until rigidity makes the system fragile. Changes in the environment can't encode legibly, so they aren't seen, and the collective acts decisively but on the wrong things.

Healthy civilizations maintain a homeostasis between the two. I claim the current fertility crisis is the cosmological expression of disequilibrium toward liberalism. Liberalism leveraged the prosperity that earlier produced high birth rates and low child mortality to perform a breadth-first explore that populated available niches, and neoliberalism is now the contraction phase, removing what fails to serve the codes that proved valuable.

A useful mid-scale example of this process is institutional sclerosis under liberal codes. DEI, credentialism, and similar moves are liberal in nature: they elaborate and idiosyncratize the internal codes of institutions. They don’t, despite local appearances, ossify institutions internally, they make them inefficient via ongoing accumulation of administrative function/ontology, which is internal niche exploration and not stabilization. From outside, an institution that has become internally inefficient looks sclerotic because institutions (unlike enterprises) don't die when they become inefficient, they become rigid and inert in their interrelations with external entities. The technocrat-right backlash to "reinvigorate institutions" and make them "live players" again is the corresponding conservative reterritorialization, attempting to remove liberal codes that have made the institutional substrate too costly.

Only the mechanism is new

Reproduction has always been controlled by the dominant order. The pre-modern West did it through religion, kinship, and hierarchy. Be a stable, predictable participant with one permanent partner approved by family, community, state, and church, and reproduction is permitted. From the top-down view, be a rigid soul empowering the machinery of strong hierarchy and you can make more of yourself.

When aristocracy and feudalism faded and meaning and value decentralized into the capitalist configuration, that phase transition recoded the reproductive homeostasis. It found different local manifestations (values, beliefs, roles, relationships) sharing the same structural constraint: those whose interiors are sufficiently defined by the collective's logos are the ones who get extended into the future.

What's distinctive about the present is that the mechanism is no longer mediated by an explicit hierarchy. No one is told they may or may not reproduce. The selection is implemented internally via capitalism shaping individuals from the inside such that those who fall outside its substrate conclude that they will not extend their line. This is direct, no-middleman top-down formation of the volitional acts of individuals, elegant and in line with the “naked” codes constitutive of capitalism.

What the frame implies

The rest of the post lists implications, grouped by what kind of claim they are. The goal of the list is to demonstrate that the frame is generative, that a single mechanism produces a wide range of consequences that would otherwise need to be explained separately (or not at all).

I've grouped the items into three categories:

Retrodictions: things the frame implies that are already partially or substantially supported by existing observation. The frame's contribution here is unification under one mechanism.
Implications: consequences the frame implies that are not yet well-characterized empirically but that follow from the mechanism with little auxiliary assumption.
Speculations: things the frame suggests but which depend on auxiliary assumptions about how the mechanism interacts with technology, politics, or institutions. These are the most fun and the least important.

If the frame is wrong, the retrodictions still hold (they have other explanations) and the speculations should fail. If the frame is right, the implications should resolve toward it over the next decade or two, and the speculations become evaluable as their auxiliary conditions materialize.

Retrodictions (frame unifies existing partial knowledge):

Bottom-up financial incentives to reproduce will largely fail because individual shaping operates at the level of the personal interior and not the level of resource constraints. You can't subsidize someone into wanting to extend a line they have been formed to experience as not worth extending, and you can't subsidize someone out of a self-audited verdict because money has no standing within that vocabulary.
Immigrant fertility regresses to the host-country mean within ~2 generations because the relevant forces act through ambient capitalist realism at the destination, not ethnicity or originating culture. Once the second generation is fully inside the cosmology, the same forming forces apply. Pre that convergence, transplants reproduce more because they aren't yet substrate.
Conservatives reproduce more not because they are dumber, less educated, or poorer, but because (a) they carry top-down codes that territorialize meaningfully, and (b) their cosmology doesn’t value introspective self-interest over kinship. Conservatives are the holders of legibility-to-the-collective in its older form, and welfare-style reproductive incentives asymmetrically work on them because they are already inside a coding scheme that values family as a unit of meaning.
Educated people have fewer kids as a result of selection against incongruence with the big Other and selection on exposure to therapeutic vocabulary. It’s not selection against intelligence per se. One can be smart in service of capital and reproduce fine, but one can't know too much about the system itself or about oneself in its terms and still participate. The effect should be strongest in fields that produce critical distance (humanities, certain sciences) and weakest in fields whose intelligence is directly fungible into capital flows (engineering, finance, applied tech).
Religious practice dominates religious identity in predicting fertility, by a large margin. As an example, cultural Catholics behave demographically like seculars while weekly-mass Catholics behave like Mormons. The difference is not belief content or ethnic background but whether the cosmology is enacted with enough embodied frequency to actually territorialize reproductive intention. Identity without practice is brand consumption inside the capitalist code and reterritorializes little. The fertility gap between aesthetic-traditional and practice-traditional populations should be large at present and continuing to widen, with the former tracking near-secular rates and the latter above replacement.

Implications (follow from mechanism, not yet well-characterized):

There is no policy solution within the cosmology. Interventions proposed inside capitalist realism are themselves artifacts of capitalist realism and will fail in the way the cosmology fails. Involutive operations can’t solve the problem. The available exits are (a) religious reterritorialization (visible), (b) authoritarian reterritorialization (latent), (c) capitalism transforming into something that can sustain its own substrate (allowed here but not predicted), or (d) something currently unimaginable. The pronatalism debate is a debate about which of the first two will happen that’s being conducted by people who think they are debating the third possibility.
Therapy hours per year will become a stronger predictor of childlessness than income. Therapy, as an institutionalized practice, embodies and facilitates meta-distance from the cosmology and encourages self-auditing focused on inner struggles. It trains individuals to direct attention towards articulating their own conditioning and perceived limitations. As a result, among demographically matched individuals, those with higher accumulated therapy hours are predicted to reproduce significantly less. This should also hold true for any sustained habit of self-reflective articulation, such as serious contemplative practice or dedicated journaling.
Bucket (3), the self-auditing wounded, should expand faster than buckets (1) and (2), with reported reasons for childlessness in survey data shifting toward self-audit narratives as successive cohorts age into peak therapeutic literacy. The differential rate, not just the absolute growth, is what the thesis predicts for two reasons. First, the formative substrate of (1) and (2) is roughly saturated — consumer abundance has been ambient for two generations and systemic critique has been culturally available since at least the 1960s — while the therapeutic apparatus is still actively reterritorializing new domains (education, HR, parenting discourse, romantic relationships, self-relation). The cohort whose formative apparatus is still expanding should grow faster than the cohorts whose formative apparatus has plateaued. Second, only bucket (3) is structurally metastable: its verdict can’t be argued with inside the cosmology without unwinding the auditing infrastructure capitalism uses everywhere else, whereas members of buckets (1) and (2) can, in principle, be remediated by the system, e.g. via severe recession reshaping lifestyle preferences or major political shifts dampening pessimism. So even at equal inflow rates, (3) should accumulate while (1) and (2) can churn.
Within knowledge-based professions, fertility will correlate more strongly with direct fungibility into capital than with academic credentials. Occupations whose output is easily quantifiable and convertible into monetary value will exhibit higher fertility rates, e.g. engineers will likely out-reproduce designers, quantitative analysts will out-reproduce sociologists, medical doctors will out-reproduce PhDs in theoretical fields, and company founders will out-reproduce employees even within the same income bracket. The differentiation here is in the legibility of the work being expressed in units of capital or in units of critical distance.
Employer-provided fertility benefits, such as egg-freezing and IVF coverage, will correlate with declining company-level fertility, not an increase. These benefits are themselves a manifestation of capitalist realism's influence, representing a deterritorialization of reproduction that decouples it from the traditional reproductive timescale and binds it to career timelines and professional advancement. Companies that generously offer these benefits are likely those where the fundamental selection pressures against reproduction have already exerted their effect. Consequently, employees at such firms will continue to exhibit fertility rates well below national averages, with the benefit functioning, more or less, as a within-cosmology palliative.
Within currently high-fertility sub-populations, smartphone penetration will predict attrition and fertility decline with an approximate one-generation lag. Established communities with historically high birth rates will not remain completely immune to broader civilizational pressures, in particular given the socioeconomic opportunity cost of technophobia-adjacent attitudes. The smartphone is the most atomized instantiation and vehicle, belonging to and shaping at the level of the individual via steady cosmological exposure. Areas with significant smartphone adoption within these communities are expected to exhibit meaningful drops in fertility.
The childfree-by-choice literature will continue becoming more meta-aware and self-theorizing over time, with the genre developing in the direction of recursive self-relation rather than polemic, manifesto, or political argument. The naive mechanism that the worldview needs more or deeper articulation per subsequent cohort to do its formative work isn’t suggested by the thesis. Instead, the argument is that the literature's bucket composition has been shifting. Early childfree-by-choice discourse was disproportionately bucket (1), and lifestyle preference articulates concisely because lifestyle preference is simple ("I like my life, kids would disrupt it"). Later discourse is disproportionately buckets (2) and (3), and in particular bucket (3)'s expression is elaborated self-audit in therapeutic vocabulary (already itself an elaborate vocabulary). Compare the 1990s personal-essay language with the 2020s book-length theorizations: the difference isn’t increased sophistication of a stable underlying view but a change in which buckets are taken as perspectival. This also explains the direction of elaboration: the genre tracks therapeutic prose rather than manifesto prose because the speakers it now selects for are formed by the therapeutic apparatus.

Speculations (depend on auxiliary assumptions):

Universities are selecting against their continuation. The type of person produced by the modern university reproduces below replacement and those that reproduce the institution are drawn from that same pool, which is not sustainable. Either the university reterritorializes into a vocational/professional institution (possibly by separating the functions of academia proper and education) or shrinks to a significant demographic minority, similar to its pre-liberalism state.
AI has a demographic alignment problem because the people building AI are disproportionately not having children, meaning whatever values get baked into AI systems will be propagated memetically but not genetically. This is the first time in history that a civilization-shaping technology is being built largely by people who have selected themselves out of the gene pool of its inheritors. The cultural inheritors of GPT-N will be raised by people who do not share the worldview of GPT-N's makers, and the AI alignment problem has an undiscussed demographic dimension.
The increasing integration of AI into knowledge work will intensify, rather than alleviate, existing demographic dynamics. As AI automates and elevates the value chain of knowledge work, it shifts labor closer to the direct enactment of organizational codes and reduces the slack within which idiosyncratic individuals could operate productively. This process strengthens the requirement for legibility within capital's framework, meaning the professions that are the most susceptible to AI integration will be the most affected in terms of fertility. Embodied professions, where work remains more difficult for the system to metabolize, will likely remain less affected. A fertility gap is predicted to emerge and widen between workers in these augmented knowledge-worker classes and those in embodied professions across the income spectrum (i.e. not only trades and blue-collar work).
Effective altruists, AI-safety researchers, and longtermists reproduce below the rate of professional academics because these communities are at the intersection of the selection pressures operating in buckets (2) and (3). Longtermism is the limiting case of bucket (2), being meta-awareness at the largest scale available and a form of critical distance that doesn’t metabolize into capital flow easily. The same communities are also a unique form of bucket (3), having institutionalized recursive self-audit (as epistemic, motivational, impact-attribution) as a community virtue in a way that’s unrelated to a therapy frame. The obvious objection that longtermists explicitly care about humanity's future misunderstands the mechanism, because selection operates on worldview structure (abstracted, continuously audited, optimized away from embeddedness) and not on worldview content.

What this post is and isn't asking for

The three common reasons people give for not having children are each consequences of capitalism shaping its own substrate to the point where parts of that substrate become degenerate and cannot extend themselves. The detached consumer and the sensitive critic are where the shaping fails marginally as wasted flow or obstructed flow. The self-auditing wounded are where the shaping succeeds completely, and as an externality the substrate locally concludes, using the system's assignment of final authority to self-report, that it will not extend itself.

The contribution I'm trying to make with this post is conceptual: introducing a frame that reconciles first-person reports about reproductive decision-making with top-down civilizational dynamics, recovers the standard socioeconomic narratives as lower resolution special cases, and generates non-obvious implications from one mechanism. The frame's value, if it has any, is in the new moving parts it introduces and how they connect things that are otherwise explained piecemeal or not at all.

The discussion I'd find most valuable:

Competing frames that do comparable unification work with a different or simpler mechanism.
Mechanisms the frame is missing, or places where the postmodernist vocabulary is hiding a confusion rather than identifying a real thing.
Implications the frame entails that I haven't listed, especially ones that would distinguish it from neighboring accounts.
Specific points where the recovery of the standard story is sketchier than I've made it sound.

The discussion I'd find least valuable is point-by-point scoring of individual implications in the list, because that mode of engagement is suited to a different kind of post than this one. The implications are there to show the frame produces nontrivial consequences.