What is time for me?

---

Just after Christmas, I was walking along a footpath near our home with my son and my grandchildren. The children were riding their scooters. It was late afternoon, really twilight, and darkness was beginning to encroach on the landscape.

We met an elderly man walking his dog in the opposite direction. He stopped to watch the children playing for a moment, then spoke to us. He said he had just spent a few days with his daughter, now forty-seven, whom he had not seen for a long time. He shook his head slightly and said, isn’t it strange how time flies? He told us he was now in his seventies, and that it felt like only a moment since she had been a small girl. And now, suddenly, she was a middle-aged woman.

We agreed, in the polite way people do, that time does indeed pass very quickly. He then continued on his way with his dog.

My son turned to me and said: So tell me, how does time look to you? Does your whole life now seem as if it’s gone by in a flash?

I stopped and thought about it. It seemed a more complicated question than it first appeared. I hesitated. The children circled, impatient to move on. We continued walking, diverted; I never gave him an answer at the time.

So what does my life actually look like, viewed from 75? Two observations come to mind.

The first is that I do have quite detailed memories of early childhood, of being very young, not much older than my grandchildren are now. I also have vivid memories of myself as a teenager, though they are episodic rather than continuous, and of my early twenties. But all of this lies fifty to sixty years in the past. The person I was then, now feels to me almost like a stranger, no more immediately accessible to me than any young person I might pass in the street. The memories exist, but I don't inhabit that earlier self.

The second observation concerns how time works for me now. In this respect, I do not think I am very different from someone at any age. I experience myself as present within a narrow temporal window: a sense of continuity extending back a few hours, perhaps not as far as a full day, and forward a few hours into anticipation. This limited span is what gives coherence to what I am doing and perceiving. Beyond that, whether into the past or the future, I have to consult external records: a diary, a calendar, photographs, documents. Those longer stretches of time don't form a unified totality in my mind.

It is as though I am walking along a very long, dark road with a small torch. The beam illuminates only my immediate surroundings. It gestures vaguely toward the darkness ahead and allows some reconstruction of the darkness behind, but I do not live in either of those regions. I live only in the illuminated patch, in the extended present.

Had I been able to formulate it at the time, that is what I would have said to my son.

---

 

Pronatalism via Population Genetics

---

Japan’s Population in the Next Thousand Years: Fertility, Assortative Mating, and the Shape of the Future

Two years ago, I wrote about the likely dynamics of population size in advanced capitalist countries, using Japan as my example. People note that in advanced countries the TFR is well below replacement, and then conclude that the populations in these countries are on the road to extinction. 'More migration,' they cry, showing that these discussions are essentially ideological rather than scientific.

The misfortune of the human race, if we can call it that, is that we evolved - mostly - to like having sex rather than babies. For most of our evolutionary history one thing led to another, however, so this didn't matter in Darwinian terms. However, modern capitalism has provided both effective contraception and many economic and personal incentives to do more interesting things than change nappies and deal with troublesome toddlers. This disconnection between sex and reproduction has led to the collapse in TFR.

However, a minority of men and women actually do want to have families - their emotional drives prompt them to have children despite the many distractions. Evolutionary theory tells us, of course, that future populations are constructed from their ancestors who reproduced, indicating that the prevalence of alleles which promote child-bearing directly will tend to increase in the population, generation by generation. 

This is selective advantage in operation: over time the population will be replaced by individuals who actually want to raise families. Let's see how this would work out in Japan.

Japan currently stands near 123 million people, yet its total fertility rate (TFR) is ~1.3, far below the ~2.1 births per woman needed for replacement. What follows sketches thousand-year outcomes using simple demographic arithmetic and population-genetic reasoning.

1. Straight-line decline

With TFR ≈ 1.3 and no offsetting forces, each generation replaces ~60% of itself. Headcount falls steeply: a remnant population of only tens of millions by 2200 and potentially a group in the hundreds of thousands by 3000. Nevertheless, this is a cultural and institutional contraction rather than a biological extinction risk although it's hard to see an advanced civilisation surviving in anything like its current form.

2. Genetics and slow selection

Completed fertility has modest heritability (≈0.2). Directional selection for a stronger “baby-drive” nudges fertility up by roughly +0.01 to +0.03 births per woman per generation. Timescale to reach replacement by genetics under the assumption of random partner selection alone is centuries; the population would likely bottom at a few percent of today’s size before any rebound.

3. Immigration as a bridge

Immigration offsets natural decrease for 1–2 generations because empirically, migrants arrive young and initially have slightly higher fertility. Likely convergence to host norms and global low fertility limit this as a century-scale fix. It buys time; it does not determine the endpoint.

4. Assortative mating as the hidden accelerator

Key dynamic: if a minority of strongly pronatal men and women mostly partner with each other, their subpopulation grows multiplicatively while the mainstream population shrinks towards zero. 

Japan’s future is not locked to straight-line decline. Assortative mating among pronatal families can bend the curve upward in roughly a century, especially with supportive culture and policy. 

Immigration buys time; automation and biotech buffer the costs of small populations. The decisive variables are partner choice, norm transmission, and targeted pronatal policy. These matter more than slow genetic drift alone.

In fact with the pronatal population exhibiting a TFR of 3.0 (three child families on average) it would take less than 250 years for Japan's population to bounce back to its current 123 million people.

SSBNs and the Not-So-Transparent Ocean

---

The Coming Transparency of the Oceans

For half a century the oceans have been the last redoubt of nuclear stability. Ballistic missile submarines (SSBNs) slipped into the abyss, wrapped in silence and darkness, beyond the reach of satellites and spies. Their very opaqueness underpinned deterrence: no adversary could launch a first strike in confidence, because they could never be sure of neutralising the hidden fleet lurking under the sea.

That compact may be ending. Advances in sensing technologies - persistent undersea sensor networks, satellite gravimetry, AI-enhanced acoustic pattern recognition, quantum magnetometers - are converging on the once unthinkable. The oceans are becoming transparent.

Already navies test autonomous underwater vehicles that trail submarines indefinitely, while seabed listening stations proliferate. Even the tiniest acoustic or hydrodynamic disturbance, once lost in oceanic noise, can now be picked out by machine-learning systems tuned to signatures. Add global data fusion and the cloak of invisibility begins to fray.

If this process continues, and there is little reason to believe it will not, the traditional nuclear ballistic-missile submarine (SSBN) ceases to be the invulnerable guarantor of second strike. A technology once deemed eternal may soon be obsolete.

What then are the options?

States that rely on SSBNs face a stark choice. Some may fall back on hardened land-based silos or tunnels, forcing adversaries to allocate immense arsenals to achieve even partial destruction. Others may pursue mobile ICBMs, endlessly shifting across road and rail networks, complicating enemy targeting. Air-launched deterrents, those long-range bombers with nuclear cruise missiles, can be dispersed to many bases, or even kept aloft on rotating patrol.

More radical, scarier options exist. Nuclear forces could migrate upward into orbital platforms or downward into swarms of autonomous underwater drones, each carrying a small warhead, dispersed in numbers sufficient to overwhelm any tracking system. Others might bury warheads in civilian infrastructure (and embassy basements?), integrating deterrence into the sinews of commerce and transport—though at enormous cost to stability. Some may simply lean harder on the nuclear umbrella of allies, outsourcing their ultimate security; good luck with that, you might say.

Do these options stabilise, as submarines once did?

Not obviously. Hardened silos are fixed targets, vulnerable to concentrated attack. Mobile ICBMs and bomber fleets invite a premium on early warning, shortening decision times: use or lose. Orbital systems would upend existing treaties and encourage arms races in space. Swarms of nuclear drones raise command-and-control nightmares and the risk of unauthorised use. Embedding nuclear assets into civilian systems corrodes the distinction between war and peace.

The opacity of the oceans was not just a geographical fact but a political blessing. It allowed nuclear states to defer the abyss of hair-trigger readiness. If the oceans now grow transparent, that buffer erodes. 

Deterrence may survive, but in harsher, more brittle forms, where the line between stability and catastrophe narrows.

The coming transparency of the oceans therefore does not herald the fade-out of nuclear deterrence. It ushers in an age where that deterrence is noisier, more exposed, and far more dangerous.

Predicting spousal personality

---

Tomorrow morning I'm meeting Tabitha's husband for the first time. I've known Tabitha for a while: she's quiet but intense, organised and rather emotional. I'm wondering what her husband is going to be like. What should I expect? My thoughts veer to Clare and myself: if someone knew Clare very well, what could they predict about me? Let's analyse this...

Clare’s temperament sits close to the ISFP profile: introverted, attuned to sensory experience, slow to engage with strangers, low on conscientiousness, high on neuroticism. She avoids public conflict but reacts strongly to stress. If we mapped her scores onto the Five Factor Model, they would cluster in the low ranges of Extraversion and Conscientiousness, high in Neuroticism, with low-to-middling Agreeableness and Openness.

Now suppose personality traits are modelled as continuous variables, roughly Gaussian in distribution across the population, according to the Five-Factor Model. The mathematics of correlated normals tells us that if two people’s traits are correlated with coefficient r, then knowing one partner’s trait gives an expected value for the other’s:

E[Trait_Spouse | Trait_Person = x] = r · x

In plain terms: spousal similarity exists, but it is partial. The closer r is to one, the more a spouse resembles the other; the closer r is to zero, the less predictive knowledge there is of the other partner.

Empirical psychology finds modest positive spousal correlations (typically r = 0.2–0.3) for traits like education, politics, conscientiousness and religiosity; weaker and sometime negative for neuroticism; sometimes near zero for extraversion.

So what of Clare and myself?

Clare is introverted; I am more extraverted. Her conscientiousness is low; mine is high, expressed in structured projects and organised commitments. Her neuroticism is elevated; mine is much lower, I'm much more phlegmatic.

By the correlation formula, if Clare scores a high positive x on neuroticism, the best estimate for me would be r·x. With r close to zero, though, for neuroticism in spousal studies, my expected value is close to zero - consistent with the reality that her anxiety does not predict mine. 

For conscientiousness, however, spousal correlation is moderate; Clare’s low score predicts a somewhat below-average score in her spouse. Yet my actual conscientiousness is markedly higher than predicted, an example of how individual outcomes scatter around the statistical mean.

Thus the conditional-expectation model provides a sober prediction: spousal similarity is real but weak, easily swamped by variance. Knowing Clare tells you a little about me, but not much.

Back to Tabitha.

She is quiet but intense (low extraversion, high neuroticism), organised (high conscientiousness), and emotional (again neuroticism). The mathematics says: expect her husband’s scores to regress toward the population mean, weighted by modest spousal correlations.

So: he is unlikely to be as quiet as she is, perhaps somewhat more outward-facing; less neurotic than she is, but not phlegmatic; with conscientiousness somewhere near average.

If I had to guess: a calm, moderately sociable man, complementing her intensity by providing ballast rather than echoing it.

Prediction via correlation is a blunt instrument. It captures the statistical truth that spouses are neither random pairings nor mirror images. The mathematics of expectation, when mapped onto personalities, reminds us that knowing one partner yields only a dim - but not entirely unsurprising - sketch of the other.

Note: I met Tabitha's husband this morning. A very pleasant man, who in personality and life-history terms seemed very like a more concrete version of me...

Abundance removes the need for men

---

Abundance and Male-Female Roles

A society of abundance is predominantly a female model of social life. 

The hard material questions - securing resources, retaining them, defending them - have largely been solved, or at least outsourced to impersonal systems. As a result, daily life can be organised around interior and relational flourishing: children, conversation, social grooming, the maintenance of emotional and moral networks.

For men, however, this same environment proves actively hostile. An abundant society removes not only danger but challenge; not only hardship but the possibility of great movements, grand projects, or idealistic quests that impose themselves from without. In such conditions, men frequently experience themselves as superfluous.

That experience of psychological irrelevance manifests as the anomie and purposelessness so characteristic of modern life. One might also note that without social roles for the expression of male virtues, female-male relationships are also thrown into confused disarray.

In pre-modern or marginal environments, life is structured by a small number of inescapable facts. Resources are scarce, threats are real, and failure is visible, consequential and frequently lethal. Under these conditions, the environment generates meaning automatically.

One does not have to ask what to do with one’s life; the environment answers that question continuously and without ambiguity. Our psychologies were formed from this environment of evolutionary adaptation.

A society of abundance reverses this structure. Food, shelter, security, and continuity are largely guaranteed. Violence is monopolised by the distant state. Failure is buffered, deferred, or medicalised. This does not produce freedom in any neutral or transparent sense. It produces existential slack: a vacuum in which meaning must be self-generated without the psychological equipment to do so reliably.

Across cultures and even across primate lineages, female sociality tends to be relational rather than heroic, oriented toward maintenance rather than conquest or defence, and embedded in dense networks of mutual attention.

In an abundant society, these forms of engagement remain meaningful. Child-rearing continues to matter. Emotional labour retains its value. Gossip and informal norm-policing still perform essential social functions. Interior flourishing has a clear referent in children, relationships, and continuity across generations. Abundance maps comparatively cleanly onto caretaking-and-coordination psychologies.

Male motivational systems, by contrast, are typically quest-shaped. Across cultures, male psychology is disproportionately tuned to status acquired through risk, esteem earned through competence under pressure, and meaning derived from challenges that offer real external resistance.

Male meaning is often heterotelic: it depends on something outside the self that pushes back. Remove danger, scarcity, honourable risk and hostile enemies, and what remains is a restless surplus of drive with nowhere to go - often redirected into crime or tribal violence.

Crucially and consequentially, abundance does not merely fail to reward these drives; it actively pathologises them. Competitiveness becomes “toxic,” ambition is redescribed as “ego,” risk-taking is framed as reckless immaturity, and hierarchical instincts reframed as authoritarianism.

The male psyche is therefore not only under-stimulated but morally delegitimised. The lived result is a recognisable phenomenology: superfluity, in the sense that no one needs what one brings; disorientation, with no clear criteria for success; resentment, as others flourish without comparable sacrifice; and withdrawal into simulations - games, pornography, and bogus ideologies.

Men do not feel liberated. They feel ontologically surplus.

Are men superfluous in societies of abundance? Functionally, often yes: a tendency that advances in AI and robotic automation are likely to intensify.

The uncomfortable conclusion is that a society of permanent abundance is not neutral with respect to sexed-psychologies. It fits female social logics tolerably well, leaves male motivational systems idle, and then blames those systems for malfunctioning.

We won't return to scarcity, mythic warrior cultures, or reactionary romanticism. But challenge, risk, and meaningful opposition cannot simply be abolished without psychic cost - and that cost will be borne disproportionately by men.

We need new quests for healthy societies.

Historically that usually means revolution and war.

---

These views are not particularly original. In A Secular Age (2007) - see also the Wikipedia article - Charles Taylor says this: "But the concern about levelling, the end of heroism, of greatness, has also been turned into a fierce denunciation of the modern moral order, and everything it stands for, as we see with Nietzsche.

"Attempts to build a polity around a rival notion of order in the very heart of modern civilization, most notably the various forms of fascism and related authoritarianism, have failed. But the continued popularity of Nietzsche shows that his devastating critique still speaks to many people today. The modern order, though entrenched—perhaps even because entrenched—still awakens much resistance."

I asked GPT5.2 whether Nietzsche might yet have had an answer to this centuries-old critique of abundant modernity. This was the interesting reply from GPT5.2.

---

Why We Remember the Past and Not the Future

---

Why We Remember the Past and Not the Future

When I was a student at Warwick University, this was a topic in my philosophy class. But it's not a question for philosophers; it's physics. Like this.

We should avoid time-laden assumptions. The universe displays an asymmetry between lower-entropy states near the Big Bang and those farther from it. From any particular state, call the past those slices of spacetime nearer to the Big Bang, and the future those farther away. Time, in this geometric sense, is simply the ordering of these slices along the expanding fabric (really the positive entropy gradient) of the universe.

Our question is: why records exist of events closer to the Big Bang but not of those farther from it.

Imagine that yesterday a supernova became visible in the sky. The light from that explosion reached Earth, interacted with our atmosphere and the Earth's surface, and altered the states of countless atoms. Those microscopic changes are, in principle, correlated with the original event, yet they rapidly disperse through further interactions, merging into general thermal motion. The information is effectively lost in the wider thermal randomisation of energy.

When you saw the supernova, however, something different occurred. The arriving photons triggered a cascade of interactions within your sensory system. Neural structures in your brain formed a configuration that encoded features of that event. This configuration can persist through subsequent physical states and can later be retrieved. A memory, in this physical sense, is a subsystem capable of forming and maintaining a structured correlation with some external interaction, preserving that correlation across later states, and retrieving it when required. That low entropy persistent record was created by you generating greater entropy in your environment, of course.

Now consider that tomorrow another supernova will appear in the sky. At this present slice of spacetime, the photons from that explosion have not yet entered your light cone. They have not interacted with you or with the matter surrounding you. The relevant regions of the universe have not yet exchanged information. No physical trace of that event exists in your current environment-state, and therefore no memory could correspond to it.

The difference between what we call past and future arises from this incomplete communication within the universe. Because the cosmos began in a low-entropy condition, its regions are still in the process of exchanging interactions. Not all parts of the universe have yet influenced one another. The network of interactions—propagating within light cones—links some regions while leaving others still disconnected. What we call the past consists of those regions that have already communicated with our present state; the future consists of those that have not.

Memory, therefore, is a local expression of this broader asymmetry. It depends on structured systems—brains, instruments, records—that can retain correlations once interactions occur. We remember the past because its signals have already reached us and have been encoded. We do not remember the future because its signals have not yet arrived.

If, in some remote epoch, the universe reaches thermodynamic equilibrium - its state of maximum entropy - and every region has exchanged all information accessible to it, then no new records can be formed. Patterns will cease to emerge, distinctions will dissolve, and the measure of change we call time will have lost operational meaning in a state of thermalised uniformity. The spacetime metric continues to exist but the possibility of registering the passage of time has disappeared.

---

'In Congénies' - a short story by Adam Carlton (intro)

---

Un mémoire fracturé, déformé par des décennies d’échos, a fini par s’échouer chez moi. Tous les détails sont faux – sauf ceux qui ne le sont pas. On m’assure que c’étaient de belles années, malgré les drames et les regards accusateurs. Et oui, tous ces rêves, toutes ces espérances, se sont révélés fondés...

Le voici.

---

"A fractured memory, distorted by decades of echoes, has ended up washing ashore with me. All the details are false—except for those that are not. I am assured they were beautiful years, despite the dramas and the accusatory looks. And yes, all those dreams, all those hopes, turned out to be well-founded..."

---

Why Did GOFAI become Gone AI?

---

GOFAI vs. the Crushing Realities of Scale

When I was an AI researcher in the 1980s, the dominant paradigm was knowledge representation. AI, we assumed, was essentially the art of representing knowledge in a formal language - predicate calculus, or some applied version of it - and then using that representation via inference rules, plus  heuristics to stop the whole thing from exploding combinatorially.

There was a huge and optimistic research programme around this: fuzzy logic, probabilistic and Bayesian variants, and a menagerie of ad hoc representational structures - frames, scripts, semantic nets - all trying to bottle “commonsense” in forms that could be manipulated in increasingly powerful ways. 

In the mainstream, nobody cared about neural nets. That community existed, of course, but it was marginal stuff in a distinctive physics-based mathematical paradigm we didn't really understand, and which we thought useless.

In hindsight, this now looks extraordinarily short-sighted. The whole approach has been shoved into the cupboard labelled Good Old-Fashioned AI - GOFAI - as if it were a Neanderthal cousin: earnest, ingenious perhaps, but doomed. I assure you it did not feel that way at the time; it felt like the right approach.

Why? Because the way we consciously think about thought is through language. When you ask someone why they did something, they give you reasons. Those reasons come out as propositions. And if you take propositions seriously, the best formal machinery we have for them is logic. So we formalised. We inferred. We drafted and polished the rules. We argued about representation. We believed we were closing in on the core of intelligence: meaning made explicit, knowledge made inspectable, inference made principled and then effectual.

If this was such a plausible programme, why did it fail so completely?

Some people will tell you that GOFAI was about “meaning” and neural nets are about “statistics”. But that opposition is uninformed and usually tendentious. When you ask a large language model a question, you ask something with meaning and you get back something with meaning. If it were all just statistical noise, it would be useless - and it obviously isn’t. So the question is not whether artificial neural net systems have semantic understanding, but how they acquire it and where it lives.

To see the difference, it helps to focus on what we were really trying to do in the 1980s. We were trying to declare meaning in advance. We tried to build a world model by writing down the world: concepts, relations, constraints, defaults, exceptions, and the rules for moving between them. We aimed for explicitness because explicitness feels like the true essence of knowledge. If the machine “knows” something, we should be able to point at it. If it reasons, we should be able to justify it. If it is wrong, we should be able to debug it.

Those are admirable ambitions. They are also, it turns out, ruinously expensive.

The world is not just insanely large; it is also densely structured. Human life is soaked in tacit knowledge: the thousands of micro-regularities that never make it into explicit articulation because they don’t need to. Such commonplace utterances as:

• “If someone says X in that tone, they probably mean Y.”
• “If this tool is in that drawer, it implies the last person who used it was doing such-and-such.”
• “If the kettle is boiling and the milk is out, the next event is likely tea.”

We do not store these things as neat propositions. We carry them as an accumulated, largely unspoken competence - our multi-decade store of condensed experiences.

GOFAI tried to drag that whole submerged continent into the daylight, label it, and file it. And every time you write a proposition you commit yourself: the predicates have sharp boundaries; the ontology has cliff edges; the exceptions multiply. Brittleness is not an engineering accident in that world - it is structural. Logic, by its nature, deals in closed systems: a statement is true or false; a condition holds or it doesn’t; an entailment follows or it doesn’t.

Human cognition, by contrast, lives on slopes. We cope with partial fit, family resemblance, analogy, and improvisation. When our models are wrong, we often degrade gracefully rather than crash. In a certain sense we live by operational not denotational semantics: praxis or dasein, if you like.

The deeper problem, though, was not merely brittleness. It was bandwidth.

Even if you grant that explicit representations are, in principle, adequate - even if you believe you can model “meaning” propositionally - you still have to get enough of it into the machine to matter. That is where the knowledge representation research programme hit the wall. The amount of world-structure a useful system needs is not “a lot”; it is effectively astronomical.

Doug Lenat’s Cyc project is the heroic proof of concept and the cautionary tale in one. It was an attempt to win by scale within the symbolic paradigm: build a vast commonsense knowledge base by hand, assertion by assertion, rule by rule, over decades. It turns out the scale of what a lifetime of expert labour can encode is orders of magnitude below what is required for the kind of fluent, flexible competence we casually expect from “intelligence”.

It is not that Cyc was trivial; it is that the world is not.

So what did the neural approach do differently? It did not “abandon meaning”. It abandoned the fantasy that meaning must be pre-declared by a human engineer.

Large language models are trained on corpora so vast that they function, in practice, as a proxy for civilisation’s accumulated linguistic trace. That trace is not random. It is massively redundant, massively structured, and shot through with regularities about the world - causal regularities, social regularities, narrative regularities, and the implicit ontologies that language users share through use without ever explicitly listing them.

Training is then a brutal compression process. The system is forced, by optimisation pressure, to extract the stable patterns that make text predictable. In doing so it precipitates a semantic geometry: a high-dimensional space (thousands of dimensions!) in which meanings are not discrete atoms with hard edges, but regions and directions - similarity structure, entailment gradients, pragmatic association, contextual modulation. It is representation, certainly, but not the sort that sits obediently on a page as “facts”. It is meaning condensed into weights.

This is why the modern paradigm scales and the older one didn’t. Not because logic is false and vectors are true, but because the neural method exploits a source of structure GOFAI could never match: the pre-existing structure in the world’s textual data dump, harvested at industrial scale.

Symbolic AI tried to build an ontology from the top down, by explicit design. LLMs inherit an ontology from the bottom up, by statistical consolidation. The difference is not merely technical; it is ecological. Neural methods outsource much of the labour of world-modelling to the culture that produced the data. They do not ask a small priesthood of knowledge engineers to type out reality, line by line. They take what humanity has already written - the messy, contradictory, redundant bulk of it - and compress its median regularities into an internal structure that can generalise.

There is, of course, a price.

GOFAI’s great virtue was epistemic. Even when it failed, it failed in the open: you could inspect the rules, challenge the ontology, argue about the premises. Modern systems give you competence first and legibility only as an afterthought. They are astonishingly capable, but they do not come with built-in notions of truth or justification. They optimise. Sometimes that aligns with truth; sometimes it aligns with plausible rubbish. You gain power and lose a certain kind of accountability.

Still, the historical lesson seems clear enough. The decisive factor was not that symbolic representations are “wrong”, but that explicit hand-built representations cannot reach the density required for general intelligence-like behaviour. GOFAI was working with teaspoons; the world required oceans.

In the 1980s we tried to formalise intelligence as propositions plus inference. The 2020s arrived with a different discovery: intelligence scales with the compression of structured experience. Meaning can be precipitated, not merely declared. And once you accept that, the dominance of neural methods stops looking like a betrayal of reason and starts looking like a humiliating empirical fact about where the informational mass of the world really sits.

In the knowledge lies the power, the old knowledge engineers used to say; in the world's data lies the knowledge is the new mantra of the foundation model engineers.

---

A Christmas Message from Adam Carlton

---

Joyeux Noël à tous mes lecteurs,

Yes, yes... I know. Once again I’m photographed beneath the Trotskyist hammer and sickle, that faded relic of revolutionary chic.

“Isn’t this just the French version of Sally Rooney cosplay?” you ask. “Aestheticised Marxism as lifestyle accessory, hollow as a gentrified café in Belleville?”

I’ll spare you the discursive rebuttal: this blog is already overstocked with Marxist rake-overs and theoretical post-mortems. Listen to Camus:

« Je me révolte, donc nous sommes. »

Rebellion, after all, remains the last refuge of those too lucid to believe and too dignified to conform. La révolte demeure le dernier refuge de ceux qui refusent d’être pris pour des dupes.

And truth be told, épater la bourgeoisie has its compensations. One is rarely short of invitations. The wine is decent. And every now and then, some bickering offshoot of the ruling class borrows your slogans: usually without comprehension but always without irony.

« Cynique, moi ? Jamais. Mais m’accuser de regarder la réalité droit dans les yeux — à cela, je plaide toujours coupable. »

À bientôt,

Adam.

PS. A while ago I wrote a poem...

My Year with ChatGPT - (Christmas Eve 2025)

---

But perhaps ChatGPT forgot to include my sword in the picture?

---