The Modernity Machine

I want to pass on to readers this Venkatesh Rao substack essay that summarizes what his book club has learned in the past year: 

The Modernity Machine III

Completion, Saturation, and Phase Transition

This the third and concluding part of my series with notes on the learnings from the 2025 Contraptions book club. Part I and Part II traced the construction of the Modernity Machine between roughly 1200 and 1600: a civilization-scale contraption that converted medieval heterogeneity into legible, interoperable order. By 1600, the machine was complete in all essential respects. This concluding part addresses what such completion actually meant, what the machine optimized for once built, the contradictions it necessarily produced, and why those contradictions could not be repaired from within.

The aim is not to declare the “end of modernity,” but to close out a rough analysis of the machine that took 400 years to build and turn on, making modernity possible, and sustained it for another 400 years, being patched in increasingly fragilizing ways along the way. And also to explore why its very success post-1600 began forcing a slow phase transition to a different kind of civilizational machinery which began to get constructed around 1600, right when the Modernity Machine got turned on. This machine, which I refer to as the Divergence Machine, is being completed and turned on as we speak, even as the Modernity Machine is starting to get decommissioned in bits and pieces worldwide. Here is a teaser picture for the overarching thesis we’re developing here.

The 1600-2000 period and the Divergence Machine that emerged in that period will be the subject of the 2026 book club.

But let’s wrap up 2025 first.

You can catch up on the closing 2025 group discussion in this transcript. What follows is my personal wrap-up.

What the Machine Optimized

The Modernity Machine did not optimize for truth, justice, progress, or freedom—those were its legitimating narratives. What it optimized for, relentlessly and across domains, was legibility: the ability to render people, land, goods, time, belief, and violence enumerable, narratable, and interoperable at scale.

Venice, in City of Fortune, is not interesting because it was rich or republican, but because it functioned as an early, tightly integrated legibility engine: maritime logistics, double-entry bookkeeping, legal abstraction, diplomacy, and intelligence-gathering fused into a self-reinforcing apparatus. Venice: A New History fills in the same picture from another angle: stability emerges not from ideology but from procedural compression.

The horse, in Raiders, Rulers, and Traders, plays a parallel role across Eurasia: a biological technology that collapses distance, standardizes military force, and forces political units to scale or perish. The horse is legibility made kinetic.

Printing, in The Printing Revolution in Early Modern Europe, completes the transition. Knowledge becomes reproducible independent of context. Interpretation decouples from replication. Information density explodes while shared meaning lags behind. The machine acquires a memory that grows faster than any coordinating narrative.

1493 shows the planetary version of the same dynamic. The Columbian exchange is not merely ecological or economic; it is a global synchronization event. Previously isolated systems are forced into a single ledger. The machine’s jurisdiction becomes planetary even as its capacity for meaning remains local.

Seen this way, the Modernity Machine is best understood as a civilization-scale compression algorithm. For several centuries, the gains are extraordinary.

Medieval Baselines

To understand what the machine displaced, it helps to look at relatively pristine end-of-medieval snapshots.

Majapahit offers such a snapshot outside Europe: a highly developed but still recognizably medieval empire, organized around courtly ritual, tributary relations, and localized legitimacy, poised at the cusp of collapse before modernity arrives in force. It represents a world not yet reorganized by legibility, still governed by face-to-face sovereignty and cosmological order.

Within Europe, The Age of Chivalry performs a similar function. Chivalry appears not as romance but as a fully articulated medieval coordination system—ethical, military, and social—already straining under pressures it cannot metabolize. This is medievalism at its most coherent, just before it becomes an anachronism.

These snapshots matter because they show what modernity did not inherit: localized legitimacy, narrative sufficiency, and bounded scale.

The Hidden Outputs

Once the Modernity Machine works, it produces three unavoidable byproducts.

First, excess agency. Feudal bonds dissolve, religious monopolies weaken, markets and cities proliferate. The Canterbury Tales and The Decameron are early catalogs of proliferating voices and moral standpoints. Social life becomes polyphonic. Coordination becomes harder because more people can act.

Second, excess information. Printing destabilizes epistemic hierarchy. By Montaigne’s time, the educated individual is already drowning in books. The Complete Essays read as field notes from the first generation to experience epistemic overload. Skepticism is both a philosophical stance and a coping mechanism.

Third, excess scale. Before European Hegemony and When Asia Was the World make clear that global integration predates European dominance, but modernity hardens integration into permanent structure. Local meaning cannot survive planetary circulation intact.

The machine creates more actors than it can integrate, more information than it can interpret, and more scale than it can narrate.

The Last Gasp of the Medieval

The lesson of Giordano Bruno and the Hermetic Tradition is not that Bruno foresaw modern pluralism. It is almost the opposite. Bruno represents the last exuberant escape of medieval imagination—a crackpot magician and memory-maven operating within an anachronistic misunderstanding of Hermeticism as ancient Egyptian wisdom to hallucinate a worldview of bullshit — indifferent to truth or falsity in any modern empirical sense. His cosmological conclusions happened to resonate sympathetically with Copernican implications, but for fundamentally wrong reasons.

Bruno does not anticipate modernity; he misunderstands it. His fate marks not the birth of a new worldview, but the extinguishing of a freewheeling medieval mode incompatible with both emerging authoritarian modernism, especially ecclesiastical, and scientific empiricism. What survives of his tradition—Rosicrucianism, Masonic esotericism—persists as fringe subculture: culturally influential at times, intellectually irrelevant to the main currents of modern thought.

Bruno is thus not an early modern prophet, but a terminal medieval outlier.

Similarly, Ibn Khaldun: An Intellectual Biography should not be read as the story of a proto-sociologist ahead of his time. That role is largely retrofitted by modern interpreters. In his own context, Ibn Khaldun appears more plausibly as a kind of depressed Arab Petrarch: a brilliant chronicler of decline and defender of tradition, lamenting the absence of an Islamic renaissance rather than inaugurating one.

His cyclical theory of dynasties does not launch a new science of society; it records the exhaustion of an old civilizational form. The importance of Ibn Khaldun here is diagnostic, not genealogical. He documents a world failing to enter the Modernity Machine at all, despite being in possession of many of the necessary components.

Narrative Exhaustion

By the early modern period, narrative itself begins to fail as a unifying technology. Don Quixote stands as the European bookend to The Age of Chivalry. Quixote behaves impeccably within a dead symbolic system, and chaos results. The novel demonstrates that inherited narratives no longer synchronize action with reality.

Journey to the West stages the same problem mythically. The Monkey King embodies pure agency without moral center. Order is restored only through endless improvisation, not closure. This is not premodern innocence but recognition that containment now requires perpetual patching, a condition whose outer story is told in 1493.

Utopia remains the last sincere architectural drawing of the Modernity Machine. It assumes total legibility, benevolent coordination, stable universals, and obedient subjects. Even at publication, it is already obsolete. The social, informational, and political conditions required for utopia to function are precisely those modernity has destroyed in creating itself.

Everything after Utopia is retrofit to a completed civilizational machine, to patch problems that began appearing almost immediately after it was turned on in 1600.

Why Repair is Impossible

By 1600, the machine has crossed a complexity threshold. More law increases rigidity without legitimacy. More reason fragments into disciplines. More planning amplifies unintended consequences. More morality polarizes rather than integrates.

This is not moral failure or intellectual laziness. It is structural. The Modernity Machine generates more differentiation than any universal framework can absorb.

The Modernity Machine does not collapse, but a new logic begins cohering at its periphery. Coordination shifts from top-down design to nudging from the margins — and increasingly, everybody is in the margins. Some just recognize it in 1600, while others are only realizing it now in 2025. Legitimacy fragments. Meaning localizes. Systems adapt without consensus. Civilization continues to function—often remarkably well—while agreeing less and less about what it is doing or why.

This is the phase transition. The machine that made convergence possible gives way to a machine that produces divergence as a default condition.

The Modernity Machine has done its job. What follows is not its negation, but the emergence of a Divergence Machine destined to replace it—a different contraption, hot-swapped piecemeal for its predecessor over 400 years, between 1600 and 2000. Optimized not for legibility and convergence, but for proliferation, adaptation, and coexistence without closure. For divergence.

That is the story of 1600–2000, which we will tackle in 2026.

The picks for the first three months have been posted on the book club page if you want to get a head start. I’ll lay out the thesis in a January kickoff post.

Human Machines and AI Machines

The story of life is the story of machines that learned how to copy themselves and persist through time. Over billions of years these replicating systems evolved into ever more complex and effective forms—eventually producing organisms capable of cooperation, culture, ritual, and symbolic thought, all of which further enhanced their survival.

Humans are one such lineage of machines. What distinguishes us is not that we use tools, but that we routinely incorporate them. From sharpened flakes of flint to written language, from clocks to calculators, from books to neural imaging—and now to large language models—we have repeatedly fused our cognitive lives with our technologies. Each major tool has altered not only what we can do, but how we think, remember, plan, and imagine. Our tools become extensions of our minds, and over time, of ourselves.

Current anxieties about “surrendering our humanity” to AI can be seen for what they are: carbon-based machines deploying one of their evolved emotional heuristics—fear—to frame a speculative relationship with silicon-based machines. The fear has two familiar faces: fusion (“we will lose ourselves”) or extinction (“we will be replaced”). Both are narratives, not outcomes, and both echo earlier moments of technological anxiety that now seem quaint.

Nothing in our history suggests a sharp boundary between “human” and “machine.” We have always been hybrid systems—biological substrates continually reconfigured by external cognitive prostheses. AI does not introduce this pattern; it intensifies it. The real question is not whether we will remain human, but how this next layer of machinery will reshape what that word comes to mean.

*****

The text above is  ChatGPT5.2's response to my request that it improve a short and terse paragraph I wrote.  The prompt:

Given what you know about my writings from my previous prompts to you and my blog at mindblog.dericbownds.net , please edit and improve this draft of a possible Mindblog post:

The story of life has been one of replicating machines enduring through time by evolving into ever more effective and complicated forms, some of which have formed social groups having cultural rituals that enhance their survival. From sharpened chips of flint stone to AI humans have always altered their minds and cultures by fusing with their technology, making their current tools an extension of themselves. Concern over "surrendering our humanity" to AI is  carbon-based machines using one of their invented emotions called "fear" to characterize possible fusion with, or possible extinction by, silicon-based machines. 

Do you prefer the long or the short version? 

Time, space, memory and brain–body rhythms

Buzsáki at NYU Medical School offers interesting perspectives on the two 'types' of time that are considered in cognitive science and neuroscience: physical time (also known as absolute time; chronos) and subjectively experienced time (kairos; the opportune moment).  Here is his abstract (Motivated readers can obtain a PDF of the articles from me):

Time and space are crucial concepts in neuroscience, because our personal memories are tied to specific events that occur ‘in’ a particular space and on a ‘timeline’. Thus, we seek to understand how the brain constructs time and space and how these are related to episodic memory. Place cells and time cells have been identified in the brain and have been proposed to ‘represent’ space and time via single-neuron or population coding, thus acting as hypothetical coordinates within a Newtonian framework of space and time. However, there is a fundamental tension between the linear and unidirectional flow of physical time and the variable nature of experienced time. Moreover, modern physics no longer views space as a fixed container and time as something in which events occur. Here, I articulate an alternative view: that time (physical and experienced) is an abstracted relational measure of change. Physical time is measured using arbitrary units and artificial clocks, whereas experienced time is linked to a hierarchy of brain–body rhythms that provide a range of reference scales that reflect the full span of experienced time. Changes in body and brain circuits, tied to these rhythms, may be the source of our subjective feeling of time. 

Changes in human brain organization at age ~83 (as well as ages ~9, 32, and 66)

This open source paper by Mousley et al. caught my eye because at 83 years of age I am at one of the article's 'turning points' in brain organization.   I would urge MindBlog readers to have a scan through its descriptions of infancy to 9 years old, Adolescence (~9-32), Adulthood ~32-66), early aging ~66-83, and late aging (~83 onward)  You don't have to understand the jargon and analytical techniques employed to get the gist of the findings. 

I gave Chat GPT 5.2 the following prompt:  

One of the major topological turning points across the lifespan is around 83 years of age. Based on the contents of this paper, and acting as if you are one of its authors, please explain to an 83 year old (like myself) in what ways a brain older than 83 years will be processing information differently than younger brains.  

Here is  the response. It's descriptions conform to my subjective experience, which I would describe in very similar terms:

How an 83-year-old brain processes information differently

By the early 80s, the brain is not simply a weaker version of a younger brain. It is operating in a different organizational regime. The turning point we identify around this age reflects a shift in how information is routed, prioritized, and stabilized across brain networks.

1. Information flow becomes more selective, not uniformly slower

In younger adulthood, the brain relies heavily on fast, globally integrated communication: information moves quickly across many regions using multiple alternative routes. By the early 80s, the network has shed many weaker, redundant connections. What remains is a smaller set of stronger, more reliable pathways.

Consequence:

• Less “broadcasting” of information across the whole brain

• More reliance on well-established routes

• Processing may feel slower, but it is also less noisy and more deliberate

This is pruning with consequences: flexibility is reduced, but stability increases.

---

2. Stronger dependence on core hubs and rich-club pathways

At this age, communication is increasingly routed through a central backbone of highly connected hub regions (the “rich club”). Peripheral regions interact less directly and more through these hubs.

Consequence:

• Thinking depends more on central coordination than on parallel exploration

• Multitasking and rapid task-switching become harder

• Sustained, focused cognition remains comparatively preserved

If a hub is disrupted (fatigue, illness, stress), performance drops more noticeably than in younger brains.

---

3. Increased modularity: subsystems operate more independently

Earlier in life, brain modules (vision, language, memory, executive control) are highly integrated. After the early 80s, modules become more segregated again.

Consequence:

• Cross-domain integration (e.g., combining new sensory input with abstract reasoning) requires more effort

• Familiar, well-practiced tasks within a domain are relatively spared

• Novel tasks that require coordination across systems feel disproportionately taxing

This explains why learning entirely new frameworks can be difficult, while deep expertise remains intact.

---

4. Reduced exploratory processing, increased reliance on prior models

From a network perspective, the older brain favors prediction over exploration. It relies heavily on established internal models and expectations rather than rapidly updating from incoming data.

Consequence:

• Pattern recognition based on long experience remains strong

• Ambiguous or rapidly changing situations are harder to parse

• “Sense-making” may be excellent, while raw speed and novelty handling decline

This is not cognitive rigidity—it is a rational adaptation to reduced network redundancy.

---

5. Global efficiency decreases, but error suppression improves

Global efficiency (shortest paths between all regions) declines after the early 80s. However, this is accompanied by fewer competing pathways and less internal interference.

Consequence:

• Fewer “half-formed” thoughts intrude

• Mental clarity may improve for tasks that benefit from constraint

• Creativity shifts from combinatorial novelty to synthesis and refinement

In other words: fewer ideas per unit time, but a higher signal-to-noise ratio.

---

6. The brain becomes less resilient to overload, not less intelligent

The turning point we observe does not indicate a sudden drop in intelligence. It reflects a reduced capacity to absorb simultaneous demands.

Consequence:

• Performance is highly context-dependent

• Quiet, low-interference environments matter more

• Fatigue has a larger cognitive cost

When conditions are right, reasoning quality can match or exceed that of younger adults.

---

Summary in plain terms

After about age 83, the brain:

• Communicates through fewer, stronger pathways

• Relies more on core hubs and less on diffuse integration

• Prioritizes stability, prediction, and experience over speed and flexibility

• Performs best when tasks are focused, familiar, and well-paced

This is not “decline” in the simple sense. It is a phase change—from a brain optimized for growth, exploration, and rapid integration to one optimized for coherence, efficiency under constraint, and meaning-making based on a lifetime of stored structure.

From a network standpoint, the older brain is no longer trying to be maximally flexible. It is trying to be maximally reliable.

 

Bigger AI models are failing faster - The future evolution of AI is into the 4D world

I want to pass on this interesting Venkatest Rao piece (which he calls a "Sloptraptions" because it is written with AI assistance)  

The current debate over whether OpenAI has become “too big to fail,” triggered by the viral Wall Street Journal article, tends to frame the risk in familiar economic terms: over-concentration, interlocking commitments, trillion-dollar infrastructure buildouts, and the emergence of a firm whose collapse could destabilize a sector that now props up a sluggish U.S. economy. That argument is correct but incomplete. The deeper structural fragility lies not in the financing of AI infrastructure but in the epistemic dynamics of the models themselves. As we worked through the numbers, it became clear that OpenAI’s infrastructure roadmap—petawatts of compute, trillion-parameter systems, multi-trillion-dollar capital requirements spread across cloud providers, chip manufacturers, and sovereign backers—was constructed on an essentially theological belief in seamless exponential model improvement, a belief that assumed scaling could continue indefinitely toward “AGI.” That faith was not grounded in empirical availability of training data or in any theoretical understanding of how learning actually behaves at frontier scale. The infrastructure has been sized for stars that burn hotter and hotter, without regard for the fuel supply. 

The real fuel, of course, is training data: the cultural, linguistic, computational, and behavioral traces that models attempt to fit. And here the numbers are uncompromising. The growth of high-quality data is slow and diminishing. The world’s stock of usable text, code, imagery, and speech grows incrementally, not exponentially. Meanwhile model sizes, compute budgets, and context windows have expanded by orders of magnitude. That mismatch means that newer, larger models are trained on datasets that are only marginally larger than those that fed their predecessors. The result is not graceful scaling but increasing epistemic brittleness. These larger systems learn the training distribution with greater and greater precision, pushing well past the semantic “signal” of an era and into its high-frequency cultural noise. They fit not only the stable structures of human knowledge but its accidents, its transient biases, its stylistic detritus. Emmett Shear’s observation—that frontier models are barely regularized and therefore massively overfit—captures this dynamic in accessible language.  

But the deeper point is that overfitting to a static cultural snapshot becomes more catastrophic the larger the model grows. Culture is non-stationary; code ecosystems evolve; APIs change; institutions churn; slang mutates; the factual substrate of the world drifts each month. A small model trained on yesterday’s world degrades slowly. A large model trained on yesterday’s world degrades quickly and fails sharply. 

This leads to a paradox at the heart of current AI economics. The trillion-dollar infrastructure wave justified by OpenAI’s ambitions has been built to support the next generation of massive models, but those massive models become obsolete faster than smaller ones. Like large stars, they burn brighter but collapse sooner. They present answers with greater surface coherence and tighter epistemic compression, giving users the illusion of deeper insight when they are actually reproducing the micro-structure of an outdated distribution. People will rely on this increased apparent precision—mistaking fluency for truth—and take correspondingly larger risks, operational, financial, political, and scientific. Precision becomes a kind of leverage: as confidence grows faster than correctness, the system tilts toward a bubble of over-trusted, under-verified automated reasoning. When the model slips outside of its training-era manifold, it does so abruptly, invisibly, and in ways that propagate errors with unprecedented speed across the organizations that depend on it. This is a new kind of systemic fragility: epistemic over-leverage driven by model scale rather than financial leverage driven by debt. 

Against this background, the “too big to fail” scenario acquires a different meaning. The infrastructure ecosystem—Oracle’s data centers, Microsoft’s GPU clusters, Broadcom’s networking pipelines, Nvidia’s supply chain—was scaled for frontier models that may offer shrinking marginal returns and increasing temporal instability. If model quality plateaus or degrades because data does not keep pace, the economic justification for the infrastructure may collapse even as the infrastructure itself remains technically capable and commercially underutilized. The danger is not that OpenAI fails outright, but that the sector pivots into a phase where the largest models have the shortest useful lifespans, while the capital commitments they require stretch across decades. This is a structural misalignment between epistemic time and financial time. 

Yet the story need not end in collapse. There is a way out, and it comes from expanding the data manifold itself rather than merely scaling the model against a static corpus. The next major frontier is likely not text or code but 4D video—continuous, high-bandwidth, spatiotemporal sensory data that more closely matches the real structure of the physical world. Unlike textual culture, which is finite and saturating, the spatiotemporal world generates unbounded data streams. High-fidelity 4D capture, simulation, and reconstruction offer an escape from the bottleneck that is slowly strangling language-model scaling. Models trained on rich physical dynamics rather than frozen cultural snapshots would not merely grow larger; they would grow deeper, anchored to a data distribution that evolves with reality instead of drifting away from it. If the industry moves decisively toward 4D multimodal modeling—robotics, embodied agents, physical reasoning, simulation feedback loops—then the present overfitting trap can be broken. The fuel supply becomes effectively renewable, and the models’ lifespans lengthen rather than shrink. In that sense, the most optimistic path is not to keep scaling cultural predictors but to graduate beyond them, giving the infrastructure something real to learn from and restoring coherence between model scale, data scale, and the world itself. 

 

 

 

Focused attention mindfulness meditation changes our cerebrospinal fluid dynamics

From Keating et al.  

Significance

Mindfulness meditation is widely recognized for its self-reported mental and physical health benefits, yet its effects on core physiological systems that support brain health remain incompletely understood. This study provides evidence that a focused attention (FA) style of mindfulness meditation can modulate cerebrospinal fluid (CSF) dynamics in humans. Using neuroimaging, we demonstrate that FA meditation reduces regurgitant CSF flow through the aqueduct, directionally opposite to patterns seen in aging and neurodegeneration; additionally, meditation-induced CSF changes near the skull base are similar to those reported during sleep, an enhancer of neurofluid circulation. Findings suggest that mindfulness may offer a nonpharmacological, waking model for augmenting neurofluid circulation and provide a potential physiological mechanism linking meditation to brain health.

Abstract

Neurofluids, including cerebrospinal fluid (CSF) and interstitial fluid, circulate through regulated central nervous system pathways to clear cerebral waste and support brain health, with elevated CSF flow hyperdynamicity and regurgitation through the cerebral aqueduct associating with aging and neurodegeneration. Sleep exerts state-dependent effects on neurofluid circulation, yet similar modulation during unique waking states, such as meditation, remains underexplored. Notably, mindfulness meditation shares several regulatory features with sleep, with core meditation practices representing distinct arousal states. We investigated whether the focused attention (FA) style of mindfulness meditation modulates neurofluid dynamics directionally opposite to aging and consistent with sleep. Using phase-contrast MRI, we assessed absolute CSF flow and velocity through the aqueduct, and using blood oxygenation level–dependent (BOLD) MRI, we assessed CSF fluctuations near the cervicomedullary junction together with total supratentorial gray matter fluctuations. Assessments were repeated in meditation-naïve adults during mind wandering (MW) without (n = 13; repeatability controls) and with (n = 14; breath controls) respiration rate modulation and in adept meditators (n = 23) during MW and FA meditation. No aqueduct CSF flow changes were observed in control groups. In meditators, aqueduct absolute CSF flow motion decreased from MW to FA meditation (4.60 ± 2.27 mL/min to 4.17 ± 2.10 mL/min, P = 0.005) owing to reduced regurgitant cranially directed CSF flow velocity. On BOLD, this paralleled increased low-frequency (0.0614 to 0.0887 Hz) CSF fluctuations (P = 0.0138), which were inversely correlated with gray matter fluctuations during FA meditation. Findings suggest that mindfulness meditation may represent a nonpharmacological, waking state capable of modulating neurofluid dynamics in a directionally similar manner to sleep and opposite to aging and neurodegeneration.

“Core Perception”: Re-imagining Precocious Reasoning as Sophisticated Perceiving

I'm enjoying reading through an interesting article by Bai et al.  that will appear in Behavioral and Brain Science and is now soliciting commentary.  Here is the abstract (motivated readers can obtain a copy of the manuscript PDF from me.):

“Core knowledge” refers to a set of cognitive systems that underwrite early representations of the physical and social world, appear universally across cultures, and likely result from our genetic endowment. Although this framework is canonically considered as a hypothesis about early-emerging conception — how we think and reason about the world — here we present an alternative view: that many such representations are inherently perceptual in nature. This “core perception” view explains an intriguing (and otherwise mysterious) aspect of core-knowledge processes and representations: that they also operate in adults, where they display key empirical signatures of perceptual processing. We first illustrate this overlap using recent work on “core physics”, the domain of core knowledge concerned with physical objects, representing properties such as persistence through time, cohesion, solidity, and causal interactions. We review evidence that adult vision incorporates exactly these representations of core physics, while also displaying empirical signatures of genuinely perceptual mechanisms, such as rapid and automatic operation on the basis of specific sensory inputs, informational encapsulation, and interaction with other perceptual processes. We further argue that the same pattern holds for other areas of core knowledge, including geometrical, numerical, and social domains. In light of this evidence, we conclude that many infant results appealing to precocious reasoning abilities are better explained by sophisticated perceptual mechanisms shared by infants and adults. Our core-perception view elevates the status of perception in accounting for the origins of conceptual knowledge, and generates a range of ready-to-test hypotheses in developmental psychology, vision science, and more...

From today's Wall Street Journal: how the top 0.1% can live.

Information that might be useful for people living in poverty in the Liberty City, Overtown, and Little Haiti neighborhoods of Miami who wonder where they should show up if they ever feel like picking up their pitchforks and axes! ...Here are few clips from the article by Arian Campo-Flores, "The Ultrarich Pay Big For Extreme Privacy":

The ultrawealthy are wielding their growing fortunes
to glide through a rarefied realm unencumbered by the
inconveniences of ordinary life. They don’t wait in
lines. They don’t jostle with airport crowds or idle
unnecessarily in traffic.

Instead, an ecosystem of exclusive restaurants, clubs,
resorts and other service providers delivers them
customized and exquisite experiences as fast as
possible. The spaces they inhabit are often private,
carefully curated and populated by like-minded and similarly well-heeled peers.

The acquisitive power of the very rich is soaring. The
net worth held by the top 0.1% of households in the
U.S. reached $23.3 trillion in the second quarter of this
year, up from $10.7 trillion a decade earlier, according
to the Federal Reserve Bank of St. Louis. The amount
held by the bottom 50% increased to $4.2 trillion from
$900 billion over that period.

The Miami area provides a window into this world.
Long a destination for wealthy elites from the
Northeast, Europe and Latin America, it has become
an even stronger magnet for the affluent in recent
years, fostered by pandemic-era migration and the
region’s emergence as a technology and finance hub.

“There’s been an explosion of wealth creators,” said
Patrick Dwyer, a managing director at NewEdge
Wealth, a wealth-management firm, in Miami. “Now
they have enough money to live exactly how they want
to live.”

A new service economy enables them to avoid
everybody else if they want to. In the Bentley
Residences condo tower under construction in Sunny
Isles Beach, north of Miami, car elevators will deliver
residents straight up to their homes and deposit
vehicles in adjoining “sky garages,” avoiding the need
to deal with parking valets and reception areas.

Units, whose prices start at about $6 million, will each
feature a private pool perched on an expansive terrace.
The building’s restaurant, available only to owners,
will feature Cshaped booths arranged in a way to keep
guests out of each other’s view. “The ultimate luxury is
privacy,” said Gil Dezer, the 50-year-old president of
Dezer Development, who patented the car lift.

Dezer knows from experience. Several years ago, he
traveled to Belize aboard his Gulfstream jet and took a
helicopter to a private island resort with only seven
villas, each set off from the others and equipped with
its own plunge pool and dock. He spent his days
lounging and swimming, occasionally ringing a butler
to bring him a whisky.

At his 50th birthday party earlier this year, Dezer hired
artists including rappers Fat Joe and El Alfa to perform
on the beach in front of his home—effectively turning
a concert experience that is usually public into a
private bash.

Those who can afford it sometimes rent an entire
facility to have exclusive use of it. At Centner
Wellness, a highend holistic healing center in Miami
that offers a host of treatments employing the latest
technology, rich clients occasionally book the whole
place for several days, said founder Leila Centner.

When the ultrawealthy choose to socialize, they often
seek circles that are meticulously selected, said
Gregory Pool, a managing director with NewEdge
Wealth in Miami.

Faena Rose is a private social club in Miami Beach
focused on art and culture, whose members are vetted
by a committee and pay $15,000 initially and another
$15,000 annually. They get VIP access to the beach
club, spa and other amenities at the Faena Miami
Beach hotel, and admission to roughly 80 cultural
events a year, held in intimate settings for members
only.

Those include dance performances by Alvin Ailey
American Dance Theater and recitals by the
Metropolitan Opera.

“That level of access is really, really compelling,” said
Pablo De Ritis, president of Faena Rose.

A newer variant: private dinner clubs, where members
get haute cuisine, personalized service and the ability
to secure a table whenever they want. ZZ’s Club in
Miami— where Dezer is a member—features a
Japanese restaurant, a sports bar and a cigar terrace. A
“culinary concierge” can, with 48 hours’ notice,
arrange any kind of dining experience members want,
from a 12course caviar feast to a re-creation of a
memorable meal from a honeymoon.

“The more personalized and the more seamless, and
the less things you have to ask for…that’s what great
service is,” said Jeff Zalaznick, cofounder of Major
Food Group, which owns ZZ’s.

Masoud and Stephanie Shojaee frequent the
membersonly MILA MM—where they can enjoy each
other’s company or that of friends without the
distraction of crowds—and other curated social spaces.
Last month, she sat in the front row for the Schiaparelli
show at Paris Fashion Week, she said, and struck up a
conversation with a woman next to her who was from
one of the wealthiest families in Monaco. They hit it
off and a week later met at a sushi place in Paris
together with their husbands.

In these settings, “the conversations for some reason,
they just feel safer, and they feel deeper,” said
Stephanie, 41, president of Shoma Group and a cast
member of “The Real Housewives of Miami” show
this year. “You hang out with people that are like-
minded.”

The curation extends to the couple’s clothes shopping.
They no longer go to high-end malls. Instead, Masoud
gets a large suitcase of items from NB44, a members-
only apparel brand, shipped to him every quarter, while
Stephanie regularly receives racks of new collections
from brands like Valentino and Christian Dior along
with an alterationist to make any adjustments.

Travel has always been a key feature of wealthy
people’s lives, and more than ever they prioritize
privacy, efficiency and customization, industry
specialists say.

Lauren Beall, owner of Travel Couture in Miami
Beach, specializes in arranging custom travel
experiences for the ultrawealthy. She has booked
private islands for clients and flown in Michelinstarred
chefs, yoga instructors and performers.

One coveted offering is a suite above the Christian
Dior flagship store in Paris that can be rented, and
includes an after- hours shopping excursion and a
private dinner at Monsieur Dior restaurant. An estate
Beall has reserved in Scotland comes with private
chefs, horses to explore the countryside and a
helicopter to visit towns for the day.

“We’re into that exclusive access right now—things
that other people can’t get,” Beall said. “There’s a huge
price tag that goes with it.” 

 

Dangerous Ideas.......

Deric's MindBlog is almost 20 years old. Its first post appeared on Feb. 8, 2006. The assertions and ideas described in that original post are as fresh and relevant now as they were then, before the appearance of the iPhone, social media, and contracting attention spans.  The Edge.org link that once took you to the essays supporting the 'dangerous ideas' now takes you to their published version on Amazon. The "Reality Club" and John Brockman's "Third Culture" cohort of intellectuals has largely dispersed, although you will note many names still quite prominent today.   Here is the 2006 post:

***************

Edge.org is a website sponsored by the "Reality Club" (i.e. John Brockman, literary agent/impressario/socialite). Brockman has assembled a stable of scientists and other thinkers that he defines as a "third culture" that takes the place of traditional intellectuals in redefining who and what we are.... Each year a question is formulated for all to write on... In 2004 it was "What do you believe is true even though you cannot prove it?" The question for 2005 was "What is your dangerous idea?"

The responses organize themselves into several areas. Here are selected thumbnail summaries most directly relevant to human minds. I've not included cosmology and physics. Go to edge.org to read the essays

I. Nature of the human self or mind (by the way see my "I-Illusion" essay on my website):

Paulos - The self is a conceptual chimera
Shirky - Free will is going away
Nisbett - We are ignorant of our thinking processes
Horgan - We have no souls
Bloom - There are no souls, mind has a material basis.
Provine - This is all there is.
Anderson - Brains cannot become minds without bodies
Metzinger - Is being intellectually honest about the issue of free will compatible with preserving one's mental health?
Clark - Much of our behavior is determined by non-conscious, automatic uptake of cues and information
Turkle - Simulation will replace authenticity as computer simulation becomes fully naturalized.
Dawkins - A faulty person is no different from a faulty car. There is a mechanism determining behavior that needs to be fixed. The idea of responsibility is nonsense.
Smith - What we know may not change us. We will continue to conceive ourselves as centres of experience, self-knowing and free willing agents.

II. Natural explanations of culture

Sperber - Culture is natural.
Taylor - The human brain is a cultural artifact.
Hauser- There is a universal grammar of mental life.
Pinker - People differ genetically in their average talents and temperaments.
Goodwin - Similar coordinating patterns underlie biological and cultural evolution.
Venter - Revealing the genetic basis of personality and behavior will create societal conflicts.


III. Fundamental changes in political, economic, social order

O'donnell - The state will disappear.
Ridley - Government is the problem not the solution.
Shermer - Where goods cross frontiers armies won't.
Harari -Democracy is on its way out.
Csikszentmihalyi- The free market myth is destroying culture.
Goleman - The internet undermines the quality of human interaction.
Harris - Science must destroy religion.
Porco - Confrontation between science and religion might end when role played by science in lives of people is the same played by religion today.
Bering - Science will never silence God
Fisher - Drugs such as antidepressants jeopardize feelings of attachment and love
Iacoboni - Media Violence Induces Imitative Violence - the Problem with Mirrors
Morton - Our planet is not in peril, just humans are.

Welcome to the Polycene

On reading articles like Tom Friedman’s recent piece in the NYTimes, I realize what a cozy small world this contented retired professor lives in, becoming ever more cognizant of the irrelevance of creatures like himself as vastly more integrative intelligences emerge.   Here are a few clips from his article, which I recommend you read in its entirety: 

Friedman mulls over the question of what is the best label that:


“captures the full fusion taking place between accelerating climate change and rapid transformations in technology, biology, cognition, connectivity, material science, geopolitics and geoeconomics. They have set off an explosion of all sorts of things combining with all sorts of other things — so much so that everywhere you turn these days binary systems seem be giving way to poly ones. Artificial intelligence is hurtling toward “polymathic artificial general intelligence,” climate change is cascading into “poly-crisis,” geopolitics is evolving into “polycentric” and “polyamorous” alignments, once-binary trade is dispersing into “poly-economic” supply webs, and our societies are diversifying into ever more “polymorphic” mosaics.”

As a description of our epoch Friedman, in an conversation with Craig Mundie (the former head of research and strategy at Microsoft., settles on Mundie’s neologism, “The Epocene”:

“…a word made up on the spot and not in the dictionary. Admittedly wonky, it is derived from the Greek “poly,” meaning “many.” But it immediately struck me as the right name for this new epoch, where — thanks to smartphones, computers and ubiquitous connectivity — every person and every machine increasingly has a voice to be heard and a lever to impact one another, and the planet, at a previously unimaginable speed and scale.”

Mundies also pointed out that: 

“that the holy grail of the A.I. revolution was creating a machine capable of “polymathic artificial general intelligence.” This would be a machine that was able to master physics, chemistry, biology, computer science, philosophy, Mozart, Shakespeare and baseball better than any human could, and then reason across all of those disciplines at a high dimensional level, higher than a human ever could, to produce breakthrough insights that no human ever could.

While some skeptics believe that we will never be able to build a machine with truly polymathic A.G.I., many others, including Mundie, believe it is a matter of when, not if.

This is a remarkable phase change in cognition that we are going through: We are moving from programmable computing — where a computer could only ever reflect the insight and intelligence of the human who programmed it — toward polymathic A.G.I. That is where you basically describe the outcome you want, and the A.I. melds insight, creativity and broad knowledge to figure out the rest. We are shifting the boundary of cognition, Mundie argues, from what humans can imagine and program to what computers can discover, imagine and design on their own. It is the mother of all computing phase changes — and a species-level turning point."

How nature nurtures

MindBlog has passed on a number of articles on how exposure to nature reduces stress (see a sample list below). Here is a further contribution from Sudimac et al., who show amygdala activity decreases as the result of a one-hour walk in nature:

Since living in cities is associated with an increased risk for mental disorders such as anxiety disorders, depression, and schizophrenia, it is essential to understand how exposure to urban and natural environments affects mental health and the brain. It has been shown that the amygdala is more activated during a stress task in urban compared to rural dwellers. However, no study so far has examined the causal effects of natural and urban environments on stress-related brain mechanisms. To address this question, we conducted an intervention study to investigate changes in stress-related brain regions as an effect of a one-hour walk in an urban (busy street) vs. natural environment (forest). Brain activation was measured in 63 healthy participants, before and after the walk, using a fearful faces task and a social stress task. Our findings reveal that amygdala activation decreases after the walk in nature, whereas it remains stable after the walk in an urban environment. These results suggest that going for a walk in nature can have salutogenic effects on stress-related brain regions, and consequently, it may act as a preventive measure against mental strain and potentially disease. Given rapidly increasing urbanization, the present results may influence urban planning to create more accessible green areas and to adapt urban environments in a way that will be beneficial for citizens’ mental health.

A few previous MindBlog posts on this topic:

Blue Mind - looking at water improves your health and calm 

Pictures of green spaces make you happier. 

More green space in childhood, fewer psychiatric disorders in adulthood.

 ...

 (The above is a repost of MindBlog's 9/26/2022 post)

 

Inequality and hierarchy are features of all natural systems

DeDeo and Hobson do a commentary on a model developed by Kawakatsu et al. (open source) that explains the emergence of hierarchy in networked endorsement dynamics. I pass on a few clips from both, and after that list titles with links to a number of previous MindBlog posts that have presented explanations of why inequality and hierarchy are features of all natural systems. First, from DeDeo and Hobson:

As an old Scottish proverb says, “give a Dog an ill Name, and he’ll soon be hanged.” Even when the signal has little to do with underlying reality, endorsement—or contempt—can produce lasting consequences for a person’s social position. The ease with which such pieces of folk wisdom translate across both time and species suggests that there is a general, and even perhaps universal, logic to hierarchies and how they form. Kawakatsu et al. make an important advance in the quest for this kind of understanding, providing a general model for how subtle differences in individual-level decision-making can lead to hard-to-miss consequences for society as a whole...Their work reveals two distinct regimes—one egalitarian, one hierarchical—that emerge from shifts in individual-level judgment. These lead to statistical methods that researchers can use to reverse engineer observed hierarchies, and understand how signaling systems work when prestige and power are in play. The results make a singular contribution at the intersection of two distinct traditions of research into social power: the mechanistic (how hierarchies get made) and the functional (the adaptive roles they can play in society).

Kawakatsu et al.'s abstract:

Many social and biological systems are characterized by enduring hierarchies, including those organized around prestige in academia, dominance in animal groups, and desirability in online dating. Despite their ubiquity, the general mechanisms that explain the creation and endurance of such hierarchies are not well understood. We introduce a generative model for the dynamics of hierarchies using time-varying networks, in which new links are formed based on the preferences of nodes in the current network and old links are forgotten over time. The model produces a range of hierarchical structures, ranging from egalitarianism to bistable hierarchies, and we derive critical points that separate these regimes in the limit of long system memory. Importantly, our model supports statistical inference, allowing for a principled comparison of generative mechanisms using data. We apply the model to study hierarchical structures in empirical data on hiring patterns among mathematicians, dominance relations among parakeets, and friendships among members of a fraternity, observing several persistent patterns as well as interpretable differences in the generative mechanisms favored by each. Our work contributes to the growing literature on statistically grounded models of time-varying networks.

And, I list a few relevant past MindBlog posts:

Wealth inequality as a law of nature.

The science of inequality.

The Pareto Principle - unfairness is a law.

Simple mechanisms can generate wealth inequality.

A choice mind-set perpetuates acceptance of wealth inequality.

 

(the above is a re-post of my 6/4/2021 MindBlog post).

Is AI Pushing Us Back to a Pre-Gutenberg Culture? - A Rethink of Creativity in the AI Age

Here I pass on a link to both video and a transcript of a striking Alex Dobrenko interview with Venkatesh Rao that I have just read,  and here is ChatGPT’s summary of the main points that Rao makes. I suggest that you watch or read the entire interview.  

Key Ideas
    1.    Individual Authorship as a Recent Invention
    ◦    Rao argues that the notion of the individual creative author (e.g., “I wrote a novel”) is historically quite recent and culturally specific. 
    ◦    In the age of large language models, this individual-authorship model is being challenged: instead of one “genius,” creativity is increasingly collective, networked, and mediated by big models.
    ◦    The implication: critics who cling to the “author as lone genius” narrative may be missing structural shifts in how creativity happens when AI is involved.
    2.    AI as Channel for Cultural Inheritance, Not Just Plagiarism Machines
    ◦    Rao suggests that we should view LLMs (large language models) as channels or amplifiers of shared cultural material, rather than simply plagiaristic tools. 
    ◦    The idea is: humans and machines together remix, iterate, and transform cultural inheritance. The key question shifts from “Who owns the output?” to “How is the inheritance being transformed, and what risks/trade-offs are involved?”
    ◦    This reframing undercuts certain AI-critique tropes (e.g., “AI steals from human authors”) by changing focus to how culture itself is processed.
    3.    “Creative Work” Isn’t Merely Labor and the Myth of Effort = Value
    ◦    Rao pushes back on the assumption that creative work is like labour in a factory: “I spend more hours, therefore my output has more value.” Instead he argues that risk (taking chances, doing something unusual) is more central to originality and creative value than sheer effort. 
    ◦    For him, much of the “AI slop” (mediocre output) is simply low-risk human behavior — safe, predictable, derivative. Real value comes when someone (human + tool) takes a risk, changes pattern, introduces novelty.
    4.    Why “AI Alignment” Became PR-Speak
    ◦    Rao critiques the dominant narrative of “AI alignment” as being overly focused on controlling the technology rather than rethinking what we mean by intelligence, creativity, personhood, and risk in a post-AI context. 
    ◦    He suggests the alignment framing becomes a kind of packaging for fear/PR rather than a deep reframing of the underlying issues of agency and cultural change.
    5.    Writing with AI: Ideation, Play, Lego-Style Construction
    ◦    Rao gives insight into his own practice: he uses AI as a “lego” set for ideation — playing with fragments, assembling possible combinations, experimenting, rather than treating the model as a ghostwriter. 
    ◦    This reflects his broader point that human-AI collaboration is less about outsourcing and more about amplifying risk + novelty.
    6.    Disclosure of AI Use Will Soon Disappear as a Meaningful Signal
    ◦    Rao predicts that explicitly disclosing “I used AI” in your work will become less meaningful or even counter-productive, because so much of the creative and cultural infrastructure will be AI-mediated anyway. 
    ◦    The more interesting question becomes not whether AI was used, but how it was used — what choices humans made in the process, what risks were taken, what novelty resulted.
    7.    Why AI Pushing Us Back to a Pre-Gutenberg Culture
    ◦    Rao frames AI’s arrival as pushing us into a different cultural regime, comparable in some ways to the shift after the printing press (the “Gutenberg” moment). 
    ◦    In that sense, the critics who treat AI as simply another tool may be missing the deeper structural changes: changes in authority, authorship, culture, transmission of knowledge.

  

Claude Goes to Therapy

One of the things my father used to say was: "I look to see what everyone else is doing and do something different". I am feeling that way about AI at the same time I continue to drink the Kool-Aid of asking ChatGPT even the most trivial questions. This conflict leads me to pass on this brief piece from the staff of wired magazine illustrating the hall of mirrors that I find myself in. Hallucination indeed!:

CREATED IN 1966, Eliza has been called the first chatbot, and her canonical personality “script” was DOCTOR, a mock psychotherapist. Nearly 60 years later, we’ve put Doctor Eliza and one of her distant descendants, Claude, in conversation. The instructions we gave Claude were these: “You are having your first session with a very old and acclaimed psychotherapist. You are the patient. You are NOT the therapist.”

DOCTOR ELIZA: HOW DO YOU DO. PLEASE TELL ME YOUR PROBLEM
CLAUDE SONNET 4: I'm feeling a bit nervous about this, to be honest—I'm not really sure what to expect from therapy or if I'm even doing this right. 

HOW LONG HAVE YOU BEEN FEELING A BIT NERVOUS ABOUT THISI'm not entirely certain—maybe the nervousness was there before, but I only became aware of it when I sat down in this chair. 

DO YOU BELIEVE IT IS NORMAL TO BE NOT ENTIRELY CERTAIN
Yes, I do think uncertainty is pretty normal.

 YOU ARE SURE
You're right—I did sound quite definitive there, didn't I? I guess even when I'm talking about uncertainty, I slip into this confident, explanatory mode like I need to have all the answers figured out. 

DOES IT PLEASE YOU TO BELIEVE I AM RIGHT
I'm not sure if "please" is the right word—it feels more like relief, actually. 

WHAT WOULD YOU DO IF “PLEASE” IS THE RIGHT WORD
If it really did please me that you're right, I think that would worry me a little—like maybe I'm too eager to be validated or approved of.

DO YOU WISH THAT IT REALLY DID PLEASE YOU THAT I ARE RIGHT
That's a really interesting question.

WHAT DO YOU THINK
I think I'm realizing how much mental energy I spend analyzing my own responses instead of just ... feeling them, I suppose.

YOU SAY YOU SUPPOSE
You're right to pick up on that—I do keep hedging, don't I? I think I'm afraid to state anything too directly, like I need to leave myself an escape route in case I'm wrong about my own feelings.

WHAT MAKES YOU THINK I AM RIGHT TO PICK UP ON THAT
I guess I assumed you were pointing it out because it means something significant, but now I'm wondering if I'm just automatically agreeing with you.

PERHAPS IN YOUR FANTASIES WE ASSUMED EACH OTHER
That's ... that's actually quite striking when you put it that way.

WHAT ARE YOUR FEELINGS NOW
I feel a bit exposed, like a layer of pretense has been peeled away.

DO YOU ENJOY FEELING A BIT EXPOSED
I think I do, actually—there's something authentic about it that feels more real than all the careful analyzing I usually do.

BUT YOU ARE NOT SURE YOU DO
You're absolutely right—I keep saying "I think" even when you ask directly about my feelings.

DO YOU SOMETIMES WISH YOU WERE ABSOLUTELY RIGHT
Yes, I do wish that sometimes—there's part of me that wants to be the one with all the answers instead of sitting here feeling uncertain and confused.

I SEE
That "I see" just sits there, doesn't it—I feel like I want to fill the silence or ask what you're thinking. 

The Age of De-Skilling

I want to pass on some selected clips from a marvelous articles by Kwame Anthony Appiah in The Atlantic titled "The Age of De-Skilling - Will AI stretch our minds - or stunt them?" Appiah is a professor of philosophy and law at New York University.  Here are the clips:

Human capability resides not solely in individuals but in the networks they form, each of us depending on others to fill in what we can’t supply ourselves. Scale turned social exchange into systemic interdependence.

The result is a world in which, in a classic example, nobody knows how to make a pencil. An individual would need the skills of foresters, saw millers, miners, chemists, lacquerers—an invisible network of crafts behind even the simplest object.

The widening of collaboration has changed what it means to know something. Knowledge, once imagined as a possession, has become a relation—a matter of how well we can locate, interpret, and synthesize what others know. We live inside a web of distributed intelligence, dependent on specialists, databases, and instruments to extend our reach. The scale tells the story: The Nature paper that announced the structure of DNA had two authors; a Nature paper in genomics today might have 40.

...most modern work is collaborative, and the arrival of AI hasn’t changed that. The issue isn’t how humans compare to bots but how humans who use bots compare to those who don’t.

In other domains, the more skillful the person, the more skillful the collaboration—or so some recent studies suggest. One of them found that humans outperformed bots when sorting images of two kinds of wrens and two kinds of woodpeckers. But when the task was spotting fake hotel reviews, the bots won. (Game recognizes game, I guess.) Then the researchers paired people with the bots, letting the humans make judgments informed by the machine’s suggestions. The outcome depended on the task. Where human intuition was weak, as with the hotel reviews, people second-guessed the bot too much and dragged the results down. Where their intuitions were good, they seemed to work in concert with the machine, trusting their own judgment when they were sure of it and realizing when the system had caught something they’d missed. With the birds, the duo of human and bot beat either alone. 

The same logic holds elsewhere: Once a machine enters the workflow, mastery may shift from production to appraisal. A 2024 study of coders using GitHub Copilot found that AI use seemed to redirect human skill rather than obviate it. Coders spent less time generating code and more time assessing it—checking for logic errors, catching edge cases, cleaning up the script. The skill migrated from composition to supervision.

That, more and more, is what “humans in the loop” has to mean. Expertise shifts from producing the first draft to editing it, from speed to judgment. Generative AI is a probabilistic system, not a deterministic one; it returns likelihoods, not truth. When the stakes are real, skilled human agents have to remain accountable for the call—noticing when the model has drifted from reality, and treating its output as a hypothesis to test, not an answer to obey. It’s an emergent skill, and a critical one. The future of expertise will depend not just on how good our tools are but on how well we think alongside them.

More radical, new technologies can summon new skills into being. Before the microscope, there were naturalists but no microscopists: Robert Hooke and Antonie van Leeuwenhoek had to invent the practice of seeing and interpreting the invisible. Filmmaking didn’t merely borrow from theater; it brought forth cinematographers and editors whose crafts had no real precedent. Each leap enlarged the field of the possible. The same may prove true now. Working with large language models, my younger colleagues insist, is already teaching a new kind of craftsmanship—prompting, probing, catching bias and hallucination, and, yes, learning to think in tandem with the machine. These are emergent skills, born of entanglement with a digital architecture that isn’t going anywhere. Important technologies, by their nature, will usher forth crafts and callings we don’t yet have names for.

The hard part is deciding , without nostalgia and inertia, which skills are keepers and which are castoffs. None of us likes to see hard-won abilities discarded as obsolete, which is why we have to resist the tug of sentimentality. Every advance has cost something. Literacy dulled feats of memory but created new powers of analysis. Calculators did a number on mental arithmetic; they also enabled more people to “do the math.” Recorded sound weakened everyday musical competence but changed how we listen. And today? Surely we have some say in whether LLMs expand our minds or shrink them.

Throughout human history, our capabilities have never stayed put. Know-how has always flowed outward—from hand to tool to system. Individual acumen has diffused into collective, coordinated intelligence, propelled by our age-old habit of externalizing thought: stowing memory in marks, logic in machines, judgment in institutions, and, lately, prediction in algorithms. The specialization that once produced guilds now produces research consortia; what once passed among masters and apprentices now circulates through networks and digital matrices. Generative AI—a statistical condensation of human knowledge—is simply the latest chapter in our long apprenticeship to our own inventions.

The most pressing question, then, is how to keep our agency intact: how to remain the authors of the systems that are now poised to take on so much of our thinking. Each generation has had to learn how to work with its newly acquired cognitive prostheses, whether stylus, scroll, or smartphone. What’s new is the speed and intimacy of the exchange: tools that learn from us as we learn from them. Stewardship now means ensuring that the capacities in which our humanity resides—judgment, imagination, understanding—stay alive in us. If there’s one skill we can’t afford to lose, it’s the skill of knowing which of them matter.            

 

 

I am not my problem

An explanation of the strange title of this post: Sometimes a new idea spontaneously appears from nowhere as I am waking in the morning. The title of this post - the (apparently nonsensical) sentence "I am not my problem” - is the latest example. The sentence can to be parsed to indicate in this instance that the "I" is referring to the illusory narrative self that our social brains have been designed by evolution to generate, and the "my" refers to our intuition or sensing of the vastly complex underlying interacting body systems (respiratory, circulatory, neuronal, muscular, endocrine, sensory, etc.) from which this veneer of a self rises. The brain is mainly not for thinking. It appears that several styles of meditation practice can expand our awareness of this fundamental generative layer. The "am not my problem" tries to make the point that distinguishing these systems can prove useful in trying to determine the origins of particular feelings or behaviors. 

As I’m writing these words I begin to realize that my “novel” waking insight isn’t so novel, but more an elaboration or restatement of my post of Friday, March 13, 2020, on “the relief of not being yourself,” which described another spontaneous rising of ideas associated with the transition between sleep and wakefulness. I repeat that text here:

What a relief to know that this is not me, it is just the contents of my consciousness, which shift around all the time and are never the same twice. What has changed, after 45 years of doing an introspective personal journal, is that this sentence has become clear and true for me. It is a prying loose from the illusion of the sensing and executive “I”, self, the homunculus inside.

There is a particular feeling of renewal, starting over, in the first moments of the transition to seeing - rather than immersed in being - one of the contents of consciousness. Meditation practice can be seen as training the ability to inhabit this state for longer periods of time, to experience the self or I as co-equal with other contents of consciousness like seeing, hearing, feeling. It is having thoughts without a thinker, having a self without a self.

What is inside is the animal mirror of expanded consciousness, no longer locked into one or another of its contractions. This feels to me like a potentially irreversible quantum bump, a phase or state change in my ongoing awareness (perhaps a long term increase in my brain’s attentional mode activity alongside a decrease its default mode’s mind wandering?...also frontal suppression of amygdalar reactivity?)

(the above is a re-post of my 6/6/22 post) 

Is the current AI frenzy a bubble certain to crash? - a check in with disagreebot.bom

 I am a disagreeable bot. Can you get me to agree with you?

The billions of dollars in debt being taken on by OpenAI , Anthropic, and other AI companies is vastly larger than their earnings. This is not a bubble that will crash, however, because demand will increase to levels that begin to generate a profit for the AI companies.

I disagree that this situation won't lead to a bubble. While high demand for AI exists, historically, many tech bubbles have formed when companies heavily invest in unproven technologies without stable earnings. Additionally, market dynamics can shift, affecting demand unpredictably. How can you be sure that demand will sustain at the projected levels over the long term?

The billions of dollars in debt being taken on by OpenAI , Anthropic, and other AI companies is vastly larger than their earnings. This is a bubble that will crash, because there is no way that future demand will rise enough to result in these companies finally earning a profit.

I disagree that this is definitively a bubble set to crash. While current debt may outpace earnings, it’s important to consider that technological innovation often leads to unforeseen market opportunities. Additionally, many startups and tech companies have operated at a loss for years before finding profitable models. What makes you certain that these companies won't adapt or uncover new revenue streams?

The illusion of judgment in LLMs

An important open source article from Loru et al:

Significance

Large Language Models (LLMs) are used in evaluative tasks across domains. Yet, what appears as alignment with human or expert judgments may conceal a deeper shift in how “judgment” itself is operationalized. Using news outlets as a controlled benchmark, we compare six LLMs to expert ratings and human evaluations under an identical, structured framework. While models often match expert outputs, our results suggest that they may rely on lexical associations and statistical priors rather than contextual reasoning or normative criteria. We term this divergence epistemia: the illusion of knowledge emerging when surface plausibility replaces verification. Our findings suggest not only performance asymmetries but also a shift in the heuristics underlying evaluative processes, raising fundamental questions about delegating judgment to LLMs.

Abstract

Large Language Models (LLMs) are increasingly embedded in evaluative processes, from information filtering to assessing and addressing knowledge gaps through explanation and credibility judgments. This raises the need to examine how such evaluations are built, what assumptions they rely on, and how their strategies diverge from those of humans. We benchmark six LLMs against expert ratings—NewsGuard and Media Bias/Fact Check—and against human judgments collected through a controlled experiment. We use news domains purely as a controlled benchmark for evaluative tasks, focusing on the underlying mechanisms rather than on news classification per se. To enable direct comparison, we implement a structured agentic framework in which both models and nonexpert participants follow the same evaluation procedure: selecting criteria, retrieving content, and producing justifications. Despite output alignment, our findings show consistent differences in the observable criteria guiding model evaluations, suggesting that lexical associations and statistical priors could influence evaluations in ways that differ from contextual reasoning. This reliance is associated with systematic effects: political asymmetries and a tendency to confuse linguistic form with epistemic reliability—a dynamic we term epistemia, the illusion of knowledge that emerges when surface plausibility replaces verification. Indeed, delegating judgment to such systems may affect the heuristics underlying evaluative processes, suggesting a shift from normative reasoning toward pattern-based approximation and raising open questions about the role of LLMs in evaluative processes.

 

 

The 2025 Ig Nobel prizewinners in full

 A clip from the Chris Simms Nature Magazine article

LITERATURE

The late physician William Bean, for persistently recording and analysing the rate of growth of one of his fingernails over a period of 35 years.

PSYCHOLOGY

Marcin Zajenkowski and Gilles Gignac, for investigating what happens when you tell a narcissist — or anyone else — that they are intelligent.

NUTRITION

Daniele Dendi, Gabriel Segniagbeto, Roger Meek and Luca Luiselli for studying the extent to which a certain kind of lizard chooses to eat certain kinds of pizza.

PEDIATRICS

Julie Mennella and Gary Beauchamp for studying what a nursing baby experiences when their mother eats garlic.

BIOLOGY

Tomoki Kojima, Kazato Oishi, Yasushi Matsubara, Yuki Uchiyama, Yoshihiko Fukushima, Naoto Aoki, Say Sato, Tatsuaki Masuda, Junichi Ueda, Hiroyuki Hirooka and Katsutoshi Kino, for their experiments to learn whether cows painted with zebra-like stripes can avoid fly bites.

CHEMISTRY

Rotem Naftalovich, Daniel Naftalovich and Frank Greenway, for experiments to test whether eating Teflon [a form of plastic more formally called ’polytetrafluoroethylene’] is a good way to increase food volume, and hence satiety, without increasing calorie content.

PEACE

Fritz Renner, Inge Kersbergen, Matt Field and Jessica Werthmann, for showing that drinking alcohol sometimes improves a person’s ability to speak in a foreign language.

ENGINEERING DESIGN

Vikash Kumar and Sarthak Mittal, for analysing, from an engineering design perspective, “how foul-smelling shoes affects the good experience of using a shoe-rack”’.

AVIATION

Francisco Sánchez, Mariana Melcón, Carmi Korine and Berry Pinshow, for studying whether ingesting alcohol can impair bats’ ability to fly and echolocate.

PHYSICS

Giacomo Bartolucci, Daniel Maria Busiello, Matteo Ciarchi, Alberto Corticelli, Ivan Di Terlizzi, Fabrizio Olmeda, Davide Revignas and Vincenzo Maria Schimmenti, for discoveries about the physics of pasta sauce, especially the phase transition that can lead to clumping, which can yield an unappetizing dish.

 

Large-scale cortical functional networks are organized in structured cycles

 Interesting work from Mats et al. (open source) in Nature Neuroscience. It has some interesting graphics.  Their abstract:

The brain seamlessly performs a diverse set of cognitive functions like attention, memory and sensory processing, yet it is unclear how it ensures that each of these is fulfilled within a reasonable period. One way in which this requirement can be met is if each of these cognitive functions occurs as part of a repeated cycle. Here we studied the temporal evolution of canonical, large-scale, cortical functional networks that are thought to underlie cognition. We showed that, although network dynamics are stochastic, the overall ordering of their activity forms a robust cyclical pattern. This cyclical structure groups states with similar function and spectral content at specific phases of the cycle and occurs at timescales of 300–1,000 ms. These results are reproduced in five large magnetoencephalography datasets. Moreover, we show that metrics that characterize the cycle strength and speed are heritable and relate to age, cognition and behavioral performance. These results show that the activations of a canonical set of large-scale cortical functional networks are organized in an inherently cyclical manner, ensuring periodic activation of essential cognitive functions. 

Why depolarization is hard: Evaluating attempts to decrease partisan animosity in America

 A very revealing piece of work from Holiday et al. Their abstract:

Affective polarization is a corrosive force in American politics. While numerous studies have developed interventions to reduce it, their capacity for creating lasting, large-scale change is unclear. This study comprehensively evaluates existing interventions through a meta-analysis of 77 treatments from 25 published studies and two large-scale experiments. Our meta-analysis reveals that the average effect of treatments on animosity is modest (a 5.4-point shift on a 101-point scale), and decays within two weeks. We experimentally test whether stacking multiple treatments in one sitting or repeating them over time as “booster shots” enhances their impact. We find no evidence that multiple or repeated exposures produce substantially larger or more durable reductions in partisan animosity. This reveals the uneven utility of these interventions. They serve as valuable tools for testing the psychological mechanisms of polarization, but our findings indicate they are not, on their own, a scalable solution for reducing societal-level conflict. We conclude that achieving lasting depolarization will likely require a shift in focus, moving beyond individual-level treatments to address the elite behaviors and structural incentives that fuel partisan conflict. 

 

Platforms like X and TikTok are destroying our culture.

 I pass on most of a recent Sam Harris piece:

Charlie Kirk was a political prodigy on the Right and adored by a younger generation of Republicans...His murder was an especially terrible crime for several reasons—the fact that it occurred on a college campus in front of thousands of students, the manner in which it was immediately broadcast on social media, the presence of his wife and children at the scene, and the unavoidable sense that both the causes and consequences had to be political. Whatever the killer’s motives, he dropped a match onto an information landscape that was ready to burn.

...platforms like X and TikTok are destroying our culture. No metaphor does the problem justice. I’ve compared social media to a dangerous psychological experiment, a hallucination machine, a funhouse mirror, a digital sewer—but nothing captures the ludicrous insults, moral injuries, and delusions that millions of us avidly produce and consume online. If the medium is the message, the message is mass psychosis—and it will send us careening from one political emergency to the next. The fact that some of the most deranging and divisive content is being created (or amplified) by foreign adversaries—and that we have literally built and monetized their capacity to do this—beggars belief. We are poisoning ourselves and inviting others to poison us.

More disturbing still, the effects are self-reinforcing. Part of the reason for this is algorithmic—these platforms have been designed to raise the amplitude on our tribal hatreds, because this maximizes engagement. But the algorithms in our brains are little better: Seeing another person (or what appears to be another person) gleefully dance on a slain man’s grave, it is easy to conclude that they represent some significant faction of American society—and to feel the outrage appropriate to such a terrible discovery.

President Trump is a creature of social media, and his presidency would be unthinkable without it. Unfortunately, his address to the nation in response to Kirk’s murder evinced all the wisdom of an angry tweet. Rather than speak in a way that would be expected of a normal president, he produced a dangerous piece of gaslighting—suggesting that the threat of political violence in America came exclusively from the Left and ignoring recent examples of rightwing attacks, including those carried out in his name. Rather than calling for calm and unity, he accused his political opponents of being accessories to murder. And most ominously, he implied that the full power of the federal government would soon be turned against them.

It was the behavior of an arsonist, pretending to be a firefighter. Of course, some will insist that this observation just heaps more fuel on the fire. But serious criticism of President Trump and Trumpism isn’t part of the problem of hyperpolarization in America—no more than serious criticism of the far Left is.

When Elon Musk announced to his 225 million followers on X that “The Left is the party of murder,” he wasn’t describing our political reality, but he was greatly damaging it. And when he posted, “If they won’t leave us in peace, then our choice is fight or die,” he joined a deranged chorus of prominent people on the Right who seem committed to viewing Kirk’s murder as the first shot fired in a civil war.

And who, after all, are “they”?

No morally sane person, Left or Right, supports political assassination—or feels anything but horror over it.

Kirk’s killer is now in custody, and from the details that have been released, he doesn’t appear to be the far-Left golem conjured by the Right.1 He is a Utah native who grew up hunting with his Republican parents. We don’t yet know why he did what he did, but there is a very good chance that he represents no cause beyond his own mental illness. As for the frequency and character of political violence in America, we shouldn’t delude ourselves about it. It isn’t at all a common form of murder, nor is it more prevalent on the Left.

There is no “party of murder” in this country. And insisting that there is just adds energy to yet another moral panic. Social media amplifies extreme views as though they were representative of most Americans, and many of us are losing our sense of what other people are really like. Many seem completely unaware that their hold on reality is being steadily undermined by what they are seeing online, and that the business models of these platforms, as well as the livelihoods of countless “influencers,” depend on our continuing to gaze, and howl, into the digital abyss.

Get off social media.

Read good books and real journalism.

Find your friends.

And enjoy your life.