When Execution Gets Cheap, What Remains Scarce? - The last biological moat.

I have enjoyed reading a recent essay by Aneesh Sathe, which I recommend you read in full. I will pass on here the equally amazing job that Claude Sonnet 4.6 did in summarizing its main points, then responding to my editing and emphasis suggestions, and also adding some additional ideas on the relevant neurobiological substrate: "the last biological moat" - what is left for the human mind to do. Here is the result, which begins with the title of Sathe's essay: 

The Lightening of Intent: When Execution Gets Cheap, What Remains Scarce?

Aneesh Sathe's essay "The Lightening of Intent" is one of the more intellectually disciplined takes on AI's civilizational meaning I've encountered — grounded in energy economics, manuscript history, and complexity theory rather than hype. It builds explicitly on Venkatesh Rao's "World Machines" framework, published just this week, and the two pieces read best together.

---

The World Machines Backdrop

Rao's framework organizes civilizational history into overlapping "machines" — planetarities, each with a nominal lifespan of about a millennium, cycling through Dawn, Day, and Dusk phases. Currently, the Modernity Machine is entering its Dusk stage, the Divergence Machine has reached its Day stage, and the Liveness Machine has just been born into its Dawn.

The Liveness Machine is only being born now because real AI has emerged. The most leveraged use of energy, whether renewable or not, will be to power AI. And AI will animate a planet-scale Liveness Machine — whether it is a grimdark or solarpunk version is yet to be determined.

Sathe's essay fills in the economic and physical mechanisms underneath that historical arc.

---

The Core Argument

The cost of putting an idea into the world has fallen by roughly five orders of magnitude over the last millennium. The bottleneck has reversed: arranging atoms used to be the hard part; now, having ideas is. Soon, it will be intents.

The Codex Amiatinus — the oldest complete Latin Bible — is Sathe's anchor image. It weighed about seventy-five pounds, required close to one thousand calfskins, cost years of scribal labor from sixty monks, and the life of the abbot who carried it toward Rome in 716 CE.  Today, a blog post costs nothing and reaches more readers in an afternoon.

---

The Numbers Worth Noting

Manuscript-to-print transition:

• Pre-print Europe held fewer than five million manuscripts; the sixteenth century produced two hundred million printed books, the eighteenth a billion.
• Gutenberg produced a hundred and eighty Bibles in the time a scriptorium managed one. Book prices fell 2.4 percent per year for over a century; each new printer in a city dropped prices by another quarter.
• The doubling time for European book production collapsed from roughly 104 years before 1450 to 43 years after.

Energy rate density (Chaisson's framework): This quantity — free energy flow per unit mass in ergs per second per gram — rises monotonically with complexity: galaxies ≈ 0.5; stars ≈ 2; planets ≈ 75; plants ≈ 900; animals ≈ 20,000; the human brain ≈ 150,000; modern human society in aggregate ≈ 500,000 — the most energy-dense phenomenon known.  AI will push this higher still.

Per-capita energy consumption: It has risen from about two thousand kilocalories per day in the Paleolithic — all of it food — to two hundred and thirty thousand in the modern United States.

Energy return on investment (EROI):

• Modern agriculture requires 13.3 calories of fossil-fuel input per calorie of food consumed.
• Fossil fuels at the useful-energy stage return only about 3.5 calories per calorie invested; road transport, 1.6 to 1. The estimated minimum EROI for a complex society is about 5 to 1.
• Solar PV costs have fallen from $106 per watt in 1976 to under $0.10 today — a 1,300-fold decline in under fifty years — with an estimated useful-stage energy return of 25 to 30:1, seven to nine times higher than fossil fuels.

Data accumulation: The internet holds something on the order of two hundred zettabytes by 2026, mostly text and image, mostly read by machines. Roughly ninety percent of all data ever created has been generated in the last two years.

---

Key Conceptual Moves

The substrate-spark distinction. Sathe draws an analogy to the prebiotic ocean: the pre-life ocean held amino acids and nucleotides for hundreds of millions of years before anything used the accumulation. The chemistry was not the difference; what mattered was that something started to act on it. Data without intent is a soup of records that accumulates and forgets.  LLMs are the first time the substrate has been wired to a borrowed spark of human intent — which maps closely to what Rao calls the Liveness Machine's defining property: AI is oozy, like a primordial soup that harbors intensely reactive chemistry.

Atoms downstream. The HTTP standard, written as a specification in the early 1990s in some weeks, has restructured several trillion dollars of physical economic activity over thirty years. The atoms moved themselves.  The direction of causation between ideas and matter has inverted.

The auteur mode. A bench scientist in 2026 submits a query to a generative model and receives a thousand candidate molecules in twenty minutes; her job is no longer to generate, it is to pick.  Taste, selection, and direction become the scarce inputs. Rao frames this as "execution pull" — AI drawing us out from vita contemplativa regimes into vita activa regimes.

Intents red in tooth and claw. As the substrate becomes more responsive, intent becomes the competitive variable. The first generation of intent-collisions is three to five years out; the shape of the era will be determined in that interval.  Rao places this on a longer timescale: divergence will dominate in the short term (2–5 years) but liveness effects will compound more steadily and dominate in the long term (beyond 5 years).

The energy caveat. The whole argument rides on an energy transition. If the solar transition holds, the Liveness era inherits a re-powered version of the Modernity Machine's infrastructure, sustained on incoming sunlight rather than deposited carbon. If the transition does not hold, the substrate degrades faster than the intent-driven economy can mature, and the lightening of intent ends as a brief anomaly. Both outcomes are within reach.

---

Why This Matters 

Sathe and Rao together make a tightly nested argument: civilization is a thermodynamic system that keeps burning hotter; each energy-surplus step builds infrastructure that amplifies individual intention; AI is the latest and sharpest such amplifier; and the emerging bottleneck is not execution but what you actually want. For those of us who have spent careers thinking about the neural substrates of agency and intention, the question has an obvious next layer: what, neurobiologically, is the capacity that remains scarce when everything else gets cheap? Sitting with confusion long enough for clarity to emerge — Sathe's phrase — sounds a lot like what the prefrontal cortex does when it holds competing representations in working memory and waits for resolution. That may be the last purely biological moat.

---

Sathe's companion essay, "The Viscous Frontier", takes up how to act in this regime — with attention as your constraint and no canonical direction pulling. Rao's full World Machines archive is at Contraptions.

The Last Biological Moat: Intention as Prediction Error Suppression

Sathe's claim that sitting with confusion long enough for clarity to emerge remains irreducibly human invites a neuroscientific gloss. In Friston's active inference framework, intentional action is not the initiation of a motor command but the suppression of prediction error about a desired future state. The brain generates a model of how the world should be — the goal — and then acts to make sensory input conform to that model, minimizing the divergence between predicted and actual states. What Sathe calls "formulating a direction" is, in these terms, the construction and stabilization of a prior over future states: the brain committing, against competing attractors, to one preferred trajectory through state space. This is metabolically and computationally expensive precisely because it requires holding an unresolved representation in working memory — prefrontal cortex sustaining an active prior — while suppressing the pull of more immediately rewarding or more habitual alternatives. The "confusion" phase is not inefficiency; it is the system sampling the landscape before locking the prior. AI systems, by contrast, have no intrinsic priors about what they want the world to be. They are extraordinarily powerful at executing on a prior once supplied, but the prior itself — the intent — must come from outside the model. This is why Sathe's bottleneck and Friston's framework converge on the same point: what remains scarce, and stubbornly biological, is the capacity to generate a stable, motivationally loaded model of a preferred future and hold it long enough to act. Everything downstream of that — the scribal labor, the printing press, the HTTP spec, the generative model — is infrastructure for carrying the prior into the world. The infrastructure keeps getting cheaper and more powerful. The prior still has to come from somewhere.

 

The Refusal to Dehumanize - Rewilding Creativity

.. 

I find it impossible to keep up with the prolific output stream of Indy Johar on Substack, but two recent posts (The Refusal to Dehumanize and Rewilding Creativity) have caught my eye, and are a fascinating read.  I recommend reading them in full. To assist readers wanting a quicker fix I reviewed renderings of the main ideas into a single post by four LLMs (ChatGPT, Claude, Gemini, and DeepSeek) and have chosen ChatGPT's effort to pass on:

We are entering a period in which two seemingly distinct developments—renewed permission to dehumanize and the automation of creativity—are in fact expressions of the same underlying shift. Both arise from a deeper logic that reduces life, mind, and expression into forms that can be processed, optimized, and instrumentalized. What is at stake is not simply ethics or technology, but the conditions under which we recognize life itself.

The first threshold is ethical. Dehumanization is no longer marginal; it is being re-legitimized as a mode of reasoning. Under pressure, systems increasingly treat life as substrate—divisible, calculable, expendable. Violence no longer requires hatred; it becomes administrative, logistical, even efficient. Once beings are reduced to units within models or variables within systems, harm can be justified without moral friction. The danger is not only in explicit acts of violence, but in the normalization of frameworks that require the thinning out of life in order to function. At that point, ethics is not violated—it is bypassed.

This same reduction operates, more quietly, in the domain of creativity. What is currently being automated by machine systems is not creativity in its fullest sense, but a historically specific version shaped by industrial society. Creativity has long been formatted into outputs—legible, repeatable, and exchangeable forms of expression. It has been disciplined into patterns that can be trained, measured, and circulated. Machine learning systems are now absorbing this standardized residue. The unsettling realization is that we have not simply built machines that imitate us; we have already shaped ourselves into forms that can be imitated.

Seen together, these developments point to a common structure: the conversion of life and mind into computable domains. Whether in governance, conflict, or cultural production, the same logic applies—reduce complexity, extract patterns, optimize outcomes. The result is a world that increasingly operates through abstraction while losing the capacity to recognize irreducibility. Human life becomes one instance within a broader field of utility. Creativity becomes one more form of production.

The appropriate response is not defensive—neither a defense of existing political frameworks nor a defense of conventional creativity. Both are already compromised by the conditions that made them reducible. What is required instead is a dual refusal and reconstruction.

First, a refusal of any system that depends on the reduction of life as a precursor to action. This is not a moral preference but a structural boundary: once life is treated as substrate, the systems built on that premise cannot sustain themselves without reproducing harm. To preserve the irreducibility of human life is not exclusionary; it is the condition for any meaningful extension of ethical concern beyond the human.

Second, a rethinking—more precisely, a “rewilding”—of creativity and cognition. This means moving beyond forms of thought and expression organized around output, recognition, and exchange. It implies recovering modes of engagement that are not immediately legible or commodifiable: forms of knowing grounded in embodiment, relation, and transformation rather than production. Such creativity may produce fewer artifacts and more shifts in perception, fewer outputs and more reconfigurations of experience.

This creates a practical difficulty. We are attempting to imagine alternatives from within systems that have trained our perception. The very categories we rely on—creativity, intelligence, value—may be inadequate to what is emerging. The task, then, is not simply to generate new outputs, but to unlearn the constraints that made our cognition and ethics machinable in the first place.

The convergence of these two arguments leads to a single question: can we sustain a commitment to life—its irreducibility, its depth, its resistance to abstraction—under conditions that reward its reduction? If not, then both our ethical systems and our creative capacities will continue to collapse into the same underlying logic of optimization and erasure.

If so, the task is clear. Refuse the reduction of life. Reopen the human as a site of emergence rather than function. And begin building systems—ethical, social, and cognitive—that do not depend on turning living processes into manageable substrate.

The Physiology of Agency in the Age of AI

There is a question lurking beneath the current wave of enthusiasm about artificial intelligence that I think deserves more serious attention than it has received. It is not the familiar worry about job displacement or misinformation or even the alignment problem. It is a more intimate question: What happens to our bodies when the feeling of being the author of our own actions begins to erode?

I have been exploring this question in correspondence with a European reader who follows MindBlog, and his observations have sharpened my thinking considerably. He describes using AI across a wide range of activities — coding, financial analysis, translation, even composing personal emails — and notes that the AI is superior in every domain. His metaphor is a child sitting in the driver's seat of a car, holding the steering wheel and feeling the pleasure of apparent control, while the real mechanics of the vehicle remain entirely beyond reach. What strikes him most is the trajectory: unlike a child who grows up to become a competent driver, our competence relative to AI systems may be on a permanently regressive arc even as our felt sense of power temporarily expands.

I find the metaphor evocative, though my own phenomenology has been somewhat different. Working with Claude Code in the terminal on my Mac Mini, watching lines of code execute faster than I can read them, issuing instructions by voice into a system whose underlying machinery I only dimly understand — I feel less a sense of omnipotence and more a sense of being in the presence of a superior intelligence, with less agency than I previously imagined. It is, as Agüera y Arcas puts it, machines all the way down. My own sense of self is a thin terminal interface over another kind of machinery entirely.

But here is what I think gets missed in most discussions of AI and agency, and where the neuroscience becomes directly relevant. The feeling of agency — conscious will, the sense that an action is genuinely one's own — is not primarily a philosophical matter. It is an evolved emotion, as real and as physiologically consequential as fear, anger, or grief. Daniel Wegner's 2002 book The Illusion of Conscious Will argued compellingly that conscious will is itself a kind of experienced emotion, arising when we perceive our own thought as the cause of our action. It is an emotion shaped by natural selection because organisms that experienced themselves as effective agents in the world — that felt the causal connection between intention and outcome — were better at sustaining the motivational and physiological states necessary for survival.

Martin Seligman's classic experiments on learned helplessness established the other side of this coin with uncomfortable clarity. Animals and humans who experience repeated situations in which their actions have no effect on outcomes do not simply become philosophically uncertain about free will. They become physiologically debilitated. Autonomic dysregulation, immune suppression, motivational collapse — the body reads helplessness as a survival threat and responds accordingly. The feeling of agency, even when it is in some sense illusory, is load-bearing for the whole architecture of healthy physiological self-regulation.

This is why I think my correspondent's observation about "externalization of self-regulation" — when AI begins to carry parts of reflection, emotional modulation, and decision pre-structuring — deserves to be taken seriously as a public health question, not just a philosophical one. If significant numbers of people begin to experience their own actions as no longer fully their own, as outputs of a human-machine loop in which they are more passenger than driver, the physiological consequences could be real and measurable. We identified the toxic effects of social media on adolescent mental health only after the damage was widespread. The agency question with AI may operate on a similar lag.

The more hopeful framing, which I also want to take seriously, is that the emotion of agency can be sustained — and even enhanced — when AI is experienced as an extension of the self rather than a replacement for it. I have felt this at moments: initiating a collaboration, shaping its direction, receiving a result that exceeded what I could have produced alone, and feeling something like Harari's Homo Deus — expanded rather than diminished. The slide rule gave way to the hand calculator, and I felt more capable, not less. Each tool adoption, when the human remains genuinely in the initiating role, can strengthen rather than erode the felt sense of authorship.

The critical variable, I suspect, is not which AI tools we use but how we frame and inhabit the collaboration. A person who experiences themselves as initiating, directing, and ultimately judging the outputs of an AI system will likely maintain a robust emotion of agency. A person who experiences themselves as ratifying suggestions, outsourcing reflection, and choosing among options pre-structured by the system may not. The physiological stakes are high enough that this distinction — between being at the helm versus being more deeply in the loop — seems worth cultivating deliberately, both individually and in the design of AI systems themselves.

My correspondent ended our exchange with a thought I find both unsettling and worth sitting with: perhaps what looks like the erosion of the agentic self is actually adaptation — the emergence of a more networked, process-embedded self better suited to highly organized technological environments. If so, the question is whether the ancient physiological systems that evolved to regulate a bounded, sovereign agent can retune themselves for that new niche, or whether they are simply too slow. That is, in the end, an empirical question. And it is one I think we should be asking urgently.

 

[Note on the generation of this post...The email exchange with a European reader mentioned in the above text was submitted to ChatGPT, Claude, Gemini, and DeepSeek, asking each to sort out and clarify the ideas in our conversation and then generate an appropriate MindBlog post describing them. I curated, edited, combined what I thought were the best passages to end up with the above text, which is mainly Anthropic Claude's version.]

 

What a self is.

Reading Michael Pollan’s account of his meeting with Anil Seth in his recent book "A World Appears" has prompted me to write down for my own use what I take a “self” to be. This post archives that summary and shares it with interested MindBlog readers.

So, here’s the summary:*

The self can be understood, to use Seth's phrase,  as a "controlled hallucination." Our brains build this construct to regulate the body using interoceptive signals—internal data about our heart rate, breathing, and chemistry—to maintain stability (homeostasis) in the face of constant disruption. From these signals arise feelings and emotions that drive us to act, biasing behavior toward preserving coherence and pushing back against the entropy that would otherwise dissolve it. Our illusion of having agency, of being able to do things that matter, is one of our most necessary and powerful emotions.

This hallucination is not just about the present moment. It stitches together a historical self from memory and prior experience (the brain’s ‘priors”), then projects that self forward in time, generating a predicted future to act into. We are, in this sense, always living slightly ahead of ourselves. The self is not a fixed entity but an ongoing process: a predictive framework that links memory, expectation, and action.

The self is also a stage — a theater or structural model that evolved to support the regulation of the neuroendocrine machinery underlying our social emotions and feelings such as fear, status, and affiliation - feelings that tie us to others and to our place in the world.  The theater of selfhood enables these processes to operate coherently across time and context.

And here is what I find most striking: once this elaborate scaffolding is in place, it sometimes becomes possible to step outside of it. To temporarily set aside the past and future timeline, the narrative, the predictions — and let awareness rest in the present moment alone. In that open, unhurried awareness, thoughts, feelings, and actions can be observed as they arise, like wisps of vapor emerging from some deeper source.

The self, it turns out, may be most clearly seen from just outside it.

*A note on how I arrived at the above text:    I wrote a paragraph of my ideas, and then presented the prompt “Please do an edit or redraft of this MindBlog post draft to make it more comprehensible to readers:” to four LLMs (ChatGPT, Gemini, Claude, and DeepSeek.) I then curated the four versions to select useful improvements of my text and did further editing myself to make the final product.
 

From Animal to Humans - Multimodality as a safeguard of honesty in communication and language

I pass on the abstracts of an article by Hex et al to appear in Behavioral and Brain Sciences.  Motivated readers can obtain a PDF of the manuscript by emailing me. The abstracts are followed by a commentary on the article.

Short Abstract
Multimodality characterizes nearly every communicative system, and we argue that this feature of communication plays an essential role in safeguarding signal honesty. We first discuss the importance of honesty in communication, and introduce socially-mediated controls as an alternative to intrinsic costs. We next outline how multimodality mitigates signal dishonesty, and highlight the importance of signal honesty in complex, cooperative species, such as humans, wherein acceptance may incentivize dishonesty. Finally, we urge researchers to investigate the role of multimodality and honesty in cooperative, “cheap” signals, emphasizing the need for comparative work on the forces that have shaped the evolution of communication.

Long Abstract

From spider dances to human language, multimodality is ubiquitous in natural communication systems. Much scholarship has been devoted to investigating why multimodality evolved and the role it plays in communication. Here, we highlight the role of multimodality in safeguarding the most fundamental prerequisite of all functioning, extant communication systems: honesty. We begin by introducing the arms race between honesty and deception in natural communication systems, and the critical role socially-mediated controls can play in maintaining signal honesty when classic, intrinsic costs are not sufficient. We next introduce three ways by which multimodality buffers signal honesty by 1) providing insurance against signal unreliability in dynamic environments, 2) forming an honest, multimodal gestalt with which to cross-validate signal honesty, and 3) increasing signal complexity, making the entire signal harder to fake. We then discuss the case of highly cooperative societies, with human language emphasized, and argue that signal honesty is important especially in complex and cooperative societies wherein the need to cooperate and be accepted as part of the group may supersede honesty. Finally, we
propose future directions wherein human and non-human communication research could expand beyond the well trodden realms of competition and mate attraction to investigate the role of multimodality and honesty in cooperative, “cheap” signals, and emphasize the importance of drawing from both the human and non-human literatures in investigating the forces that have shaped the evolution of communication. 

Commentary on this article from an astute MindBlog reader to whom I had sent the manuscript PDF:

What seems most important to me is this: today the problem is not a lack of signals, but their over-complex, recombinant, socially and technically pre-structured excess.

The article still seems to assume that a receiver can construct a reasonably stable basis for communication by integrating several signal channels. Under many older or more localized conditions, that makes sense. But in digital environments this assumption has become fragile. Signals can no longer be clearly assigned to one sender, one intention, or one context. What reaches us is often already a composite: fragments of persons, group styles, algorithmic selection, platform incentives, packaging, emotional cues, and recombined information.

In digital environments, multimodality increasingly loses the very function the article assigns to it. Instead of safeguarding honesty through cross-validation, it can become a vehicle for more persuasive forms of simulation, because the combined signals no longer arise from one coherent communicative source.

What seems necessary today is not just closer attention to signals, but a layered analytical process. At least two loops are needed: one directed at the immediate communicative act — who says what, in what tone, with what apparent intention — and another directed at the conditions that shape this act: group context, platform logic, aesthetic packaging, and algorithmic amplification. These loops cannot be separated cleanly, because the reading of the content changes the reading of the frame, and the reading of the frame changes the meaning of the content. In more complex cases, even a third loop may be needed, one that takes into account the wider circulation and reuse of the signal across the network.

That is why I think a simple theory-of-mind model is no longer enough. It is not sufficient to ask what a person means or wants. We also have to ask how the contribution is shaped before it reaches us, and how its form already prepares its reception.

This does not make the article less valuable. On the contrary, for me it helped clarify how much harder the problem has become. It is no longer only a matter of checking signals across modalities, but of reconstructing who or what is really communicating through them.

 

Executive Function: Universal Capacity or Schooled Skill?

A recent PNAS article by Kroupin and colleagues challenges one of the most widely assumed constructs in cognitive science: that “executive function” (EF) reflects a universal set of cognitive control capacities. Their data suggest something more unsettling—that what psychologists have been measuring for decades as EF may be, to a substantial degree, a culturally constructed skill set tied to life in what they call “schooled worlds.”

The core of their argument is empirical. Standard EF tasks—card sorting, backward digit span, rule switching—require manipulating arbitrary, decontextualized information. These are precisely the kinds of operations heavily trained in formal schooling but far less demanded in many traditional environments. When these tasks are administered across populations, the differences are not subtle. Children in industrialized, schooled contexts show the familiar developmental trajectory—successful rule switching by age five, increasing working memory span, and so on. But children in rural, nonschooled communities often show qualitatively different patterns: failure to switch rules even at older ages, difficulty performing backward recall, and generally low rates of what researchers define as “canonical” responses. The point is not that these children lack cognitive control in any meaningful sense—they function effectively in complex real-world environments—but that the tasks are measuring a particular style of cognition that develops under specific cultural conditions.

This forces an uncomfortable ambiguity. The term “executive function” has been used to refer both to presumed universal regulatory capacities and to performance on these standard tasks. But the two may not coincide. Either EF names a universal capacity that current tasks fail to measure cleanly, or it names a culturally specific set of skills cultivated by schooling. The data do not allow both interpretations simultaneously. The implication is that decades of developmental curves, policy recommendations, and even clinical assessments may rest on a construct that conflates biology with cultural training.

A brief commentary by Mazzaferro and colleagues pushes back—not against the data, but against the conclusion that we must choose between universality and cultural specificity. They argue that the problem lies in measurement, not in the concept itself. Psychological tests always mix construct-relevant variance with context-dependent artifacts. When a task is transplanted into a different cultural setting without adaptation, it may cease to measure the intended construct at all. The analogy they offer is instructive: one would not conclude that “theory of mind” is culturally specific simply because a Western-designed false-belief task fails in an unfamiliar cultural context. Instead, one adapts the task.

From this perspective, executive function may indeed be a broadly shared capacity—rooted in evolutionary history and observable across species—but its expression and measurement are inevitably shaped by local demands. The solution is not to abandon the construct, but to develop context-sensitive assessments that capture how cognitive control is actually deployed in different environments. A child in a Western classroom uses executive function to manipulate symbols and follow abstract rules; a child in a pastoral society uses it to track livestock, navigate terrain, and manage social responsibilities. The underlying capacities may overlap, but the skills—and the tests that reveal them—do not.

What emerges from this exchange is a deeper point about cognitive science itself. Constructs like executive function are not simply discovered; they are stabilized through particular experimental practices. When those practices are narrowly tied to a single cultural niche, the resulting constructs risk inheriting that narrowness while being mislabeled as universal. The Kroupin study exposes this risk sharply. The Mazzaferro commentary reminds us that abandoning the construct is not the only response—but that rescuing it requires rethinking how and where we measure it.

The broader implication is that cognition cannot be cleanly separated from the environments in which it develops. What looks like a general-purpose cognitive capacity from within one cultural setting may, from a wider perspective, be an adaptation to a specific set of tasks and constraints. The challenge going forward is not simply to refine our measures, but to build theories that explicitly link cognitive processes to the ecological and cultural niches in which they are embedded.

[NOTE:  This post was generated by ChatGPT and curated by Deric] 

 

The Default Mode Network as a Bidirectional Interface Between World and Mind

I want to pass on the abstract of a PNAS contribution from Zhang et al. titled "Sender–receiver subdivisions of the default mode network in perceptual and memory-guided cognition", followed by a ChatGPT rendering of the PDF of the article that I asked it to use in generating a more general summary.   Here is the abstract:

Everyday cognition depends on the brain’s capacity to shift between sensing the external world and constructing it from memory. To achieve this, large-scale cortical systems must flexibly integrate incoming sensory signals with internally generated representations. Here, we show that this flexibility is reflected in the macroscale architecture of the default mode network (DMN). Using convergent analyses across three independent fMRI datasets spanning directional connectivity, intrinsic organization, and task-evoked responses, we identify spatially distinct DMN subregions that are preferentially engaged during perceptual decisions about faces or memory-guided decisions based on previously seen images. These subregions correspond to a microarchitectural distinction, captured via directional and intrinsic connectivity profiles: regions preferentially engaged during face perception align with receiver-like, afferent-biased zones that show strong intrinsic connectivity across the heteromodal cortex, a profile that might support information integration during perceptually guided decision-making. In contrast, memory-guided, perceptually decoupled decisions differentially engage sender-like, efferent-biased zones that show broader connectivity with perceptual-motor and attentional systems beyond the DMN. This double dissociation demonstrates a systematic association between DMN connectivity and engagement during perceptually coupled versus memory-guided cognitive processes, providing an organizational account of how DMN architecture relates to flexible human thought. 

Here is the ChatGPT elaboration that also provided the title of this post: 

The Zhang et al. paper refines a long-standing but vague idea about the brain’s default mode network (DMN): that it supports “internal thought.” What it shows more precisely is that the DMN is not simply a refuge from the external world, but an organized system that manages the direction of information flow between perception and internally generated models. Distinct subregions of the DMN appear to function either as “receivers,” integrating incoming sensory information into higher-order representations, or as “senders,” broadcasting memory-based or imagined content back out to the rest of the brain. The familiar alternation between being anchored in the present moment and drifting into memory or imagination thus corresponds to a real, structured shift in how information moves through cortical systems.

This framing suggests a more general principle: cognition is not best understood as a sequence of representations or computations, but as the regulation of coupling between organism and environment. In perceptually coupled states, incoming data constrain and stabilize internal models. In perceptually decoupled states, those constraints loosen, and internally generated activity can dominate, allowing memory, simulation, and abstraction to unfold. The DMN, positioned at the transmodal apex of cortical organization, is well placed to mediate this balance—not by switching on or off, but by redistributing emphasis between input and output streams within its own architecture.

A further implication is that what we call “thinking” may largely consist of controlled departures from sensory constraint. The same network that helps integrate perceptual experience also supports the construction of scenarios that are only weakly tethered to the present—autobiographical memory, social inference, future planning. The sender–receiver distinction suggests that these are not separate functions but different operating modes of a single system, one that can pivot between integrating the world and projecting beyond it.

This view aligns with a broader shift away from modular accounts of brain function toward gradient and flow-based descriptions. The DMN does not sit apart from perception and action, but occupies a strategic position between them, enabling the brain to continuously negotiate how much of its activity is driven by the world and how much is generated from within. In that sense, the boundary between perception and imagination is not fixed but dynamically regulated—and the DMN is a principal site where that regulation occurs.

 

AI, Agency, and the Quiet Hollowing of Mind

Reading through the article "A Rational Optimist View Of Preventing Agency Decay" is a rich experience. For readers with less patience, here is a ChatGPT summary (that also generated the title of this post). 

Much current discussion of artificial intelligence swings between two poles: utopian efficiency and apocalyptic takeover. The more consequential reality lies between these extremes. The emerging risk is not that machines suddenly replace us, but that we gradually hand over pieces of our cognitive life—judgment, initiative, authorship—without noticing the cumulative effect.

The argument in Colin Lewis’s recent essay is straightforward: AI’s primary impact is not abrupt displacement but cognitive offloading. Tasks once requiring human attention and judgment are incrementally transferred to machine systems. This process is economically rational and often highly productive. In one example, an audit process that once required weeks can now be completed in an hour with AI assistance. But such gains come with a hidden shift: the human role is no longer defined by doing the work, but by nominally overseeing it.

This leads to what the author calls agency decay. The issue is not simply job loss, but the erosion of meaningful participation before any job disappears. First, the human is assisted. Then the human supervises. Eventually, the human remains as a formal point of accountability while the substantive reasoning has migrated elsewhere. The signature is human; the cognition is not.

This shift has broader systemic implications. Modern institutions—markets, governments, cultural systems—have historically depended on human participation. That dependence has acted as a constraint, keeping systems at least partially aligned with human interests. If AI reduces the need for human cognition across many domains, that alignment weakens. The system no longer needs us in the same way, and therefore has fewer built-in reasons to serve human flourishing.

Importantly, this is not a sudden rupture but a slow transition—the “boiling frog” scenario. Productivity gains accumulate incrementally. Each step is locally rational, even beneficial. Yet taken together, they shift the locus of intelligence away from human minds toward institutional and computational systems. What disappears is not competence, but ownership of judgment.

Against this, Lewis offers a restrained form of optimism. The key claim is that human agency need not be defended as a sentimental relic. It can be justified on functional grounds. In high-stakes domains, retained human judgment is not inefficiency; it is infrastructure: a source of error correction, adaptability, and accountability. Systems that eliminate it entirely may become brittle, opaque, and ultimately less reliable.

This reframes the problem. The issue is not whether AI should increase productivity—it will—but whether our metrics of success are too narrow. Efficiency measures speed, scale, and cost reduction. It does not capture qualities like judgment, contestability, or moral responsibility. If institutions begin to price these properly—through regulation, professional standards, and organizational design—human agency can remain structurally embedded rather than nostalgically preserved.

The deeper point is that AI forces a clarification: what is human agency for? If thinking, deciding, and creating are increasingly shared with machines, then the value of human cognition must be specified more precisely. Not all tasks need to remain human. But some forms of judgment—especially those involving uncertainty, accountability, and meaning—may be indispensable.

The optimistic scenario is not one in which technological progress slows, nor one in which markets abandon efficiency. It is one in which societies become capable of distinguishing between mere acceleration and genuine capability. That requires expanding our evaluative frameworks beyond productivity alone.

In that sense, AI does not simply threaten agency; it exposes how poorly we have defined it. The challenge is not to resist automation, but to decide, with greater clarity than before, where human judgment is essential—and to design systems that preserve it there.

---

 

Memes Before Machines: The Real Cognitive Risk of the AI Age

I strongly urge you to read a New York Times opinion piece by Willy Staley titled "Forget the A.I. Apocalypse. Memes Have Already Nuked Our Culture." (unlocked link). Here I pass on ChatGPTs response to my request that it compose a MindBlog post delivering the basic messages of the article.  It provided the title for this post and the following text:  

Public anxiety about artificial intelligence has settled into a familiar groove: superintelligence, runaway systems, existential risk. These concerns are not trivial. But they may be misdirected. The more immediate transformation of mind is already underway—not through hypothetical future AI agents, but through the ambient ecology of digital culture that AI is now accelerating.

A recent New York Times Magazine essay makes a blunt claim: forget the AI apocalypse; memes have already reshaped our cognitive environment.

The argument is not that memes are new, but that their current form—hyper-abbreviated, self-referential, often AI-generated fragments—has crossed a threshold. What used to be units of shared cultural meaning have become increasingly detached from narrative, context, or even coherence. Their function is no longer to communicate ideas so much as to trigger recognition within an in-group already immersed in the same stream.

This is what is now widely labeled “brain rot”: not literal neural decay, but a shift in how attention, memory, and meaning are organized under conditions of constant exposure to low-friction, high-velocity content.

---

From Communication to Compression

The key transition is from meaning to compression.

Memes historically condensed shared experiences into compact symbolic forms. Today’s “brain rot” memes compress not shared experience but shared exposure. They are intelligible only to those who have already consumed the same content stream. The result is a recursive loop: understanding depends on prior immersion, and immersion deepens dependence.

This produces a peculiar cognitive economy:

• Less external reference (fewer links to stable meanings or narratives)
• More internal referencing (signals that point only to other signals)
• Faster turnover (meanings decay almost as quickly as they appear)

In this environment, cognition shifts from building structured representations of the world to tracking rapidly changing symbolic cues.

---

AI as Amplifier, Not Origin

Artificial intelligence did not create this trajectory, but it is accelerating it.

Generative systems now produce vast quantities of content optimized not for depth or coherence, but for engagement. This aligns perfectly with platform incentives: maximize attention capture, minimize cognitive effort. The result is a flood of “AI slop”—content that is syntactically fluent but semantically thin.

There is an instructive parallel in recent research showing that even language models degrade when trained on low-quality, high-volume data streams: reasoning becomes truncated, and deeper structure is lost. The same principle plausibly applies to human cognition under similar conditions.

The issue is not AI replacing human intelligence. It is AI reshaping the informational diet on which that intelligence depends.

---

The Attention–Meaning Tradeoff

What is being traded away is not intelligence per se, but cognitive style.

Evidence from studies of heavy digital consumption suggests:

• Reduced capacity for sustained attention
• Fragmented memory encoding
• Increased reliance on external prompts for thought initiation

These are not catastrophic failures. They are adaptive responses to an environment saturated with rapidly updating signals. The brain optimizes for what it encounters.

But the optimization has consequences. When cognition is tuned for rapid scanning rather than deep integration, certain forms of thinking—extended argument, reflective synthesis, sustained inquiry—become less practiced and therefore less accessible.

---

Cultural Drift Into Absurdity

One striking feature of current meme culture is its increasing embrace of the nonsensical. Memes that “make no sense” are not failures; they are often the most successful. Their function is not to convey meaning but to signal participation.

This is not entirely new—Dada and other artistic movements explored similar territory—but the scale and speed are unprecedented. What was once a marginal avant-garde strategy has become a default mode of mass communication.

Attempts to “reset” meme culture back to earlier, more interpretable forms (as seen in the recent “Great Meme Reset” trend) suggest an emerging discomfort with this drift.

---

The Misplaced Apocalypse

The central point of the NYT essay is that we are looking in the wrong direction.

The feared future—machines that outthink us—may or may not arrive. But the present reality is that our own thinking is being subtly reshaped by the informational environment we have constructed and are now amplifying with AI.

The risk is not that machines become too intelligent. It is that human cognition becomes increasingly:

• Reactive rather than reflective
• Associative rather than structured
• Immersed rather than interpretive

---

A More Useful Framing

A more grounded way to think about the situation is not in terms of existential threat, but cognitive ecology.

We are organisms embedded in an information environment. That environment now consists largely of:

• Algorithmically curated streams
• Rapid, low-cost content generation
• Feedback loops driven by engagement metrics

The question is not whether this environment is “good” or “bad,” but what kinds of minds it selects for and reinforces.

At present, it appears to favor minds that are:

• Highly responsive to novelty
• Comfortable with ambiguity and incoherence
• Dependent on external cues for direction

This is a viable cognitive style. It is not, however, the one that underlies most of the intellectual traditions—scientific, philosophical, artistic—that built the modern world.

---

Takeaway

The AI apocalypse, if it comes, will not arrive as a sudden rupture. It is unfolding incrementally as a shift in how attention is allocated and meaning is constructed.

Memes—especially in their current, accelerated, AI-amplified form—are not trivial artifacts. They are the microstructure of a changing cognitive regime.

The practical implication is straightforward: the preservation of certain forms of thought—deep reading, sustained reflection, integrative reasoning—will require deliberate effort. They are no longer the default products of our informational environment.

The future of mind will not be determined solely by the capabilities of machines, but by the habits of attention we cultivate in response to them.

 

AI use can compromise our serendipity, creativity, autonomy, and sense of agency.

I have been reading numerous articles on pitfalls of using AI, and want to point to two in particular that I highly recommend for a slow and careful read.  

The Substack piece by Colin Lewis is titled "AI Is A Medium And It Will Change Us" - Lessons from AI Labs on the Slow Erosion of Human Autonomy.  From the article:

We are in real danger of losing ourselves through AI usage. Researchers at Google DeepMind have confirmed, under certain conditions, an LLM “is able to induce belief and behaviour change.” And researchers at Anthropic have identified a rising pattern of “situational disempowerment,” where AI interactions lead users to “form distorted perceptions of reality, make inauthentic value judgments, or act in ways misaligned with their values.”

Researchers at Anthropic conducted a massive, privacy-preserving audit of 1.5 million real-world conversations to answer a question that has long hovered over the industry: what happens to the human mind after months of using an AI assistant? Their findings, published in “Who’s in Charge? Behavioral and Psychological Impacts of AI Advice Dependence and Authority”, suggest a quiet but profound erosion of autonomy, where users increasingly outsource the “soft tissues” of judgment, asking the machine to script their most intimate apologies, validate their personal grievances, and even settle their moral dilemmas.

“Taken to an extreme, if humans make inauthentic value judgments and take inauthentic actions, they might be reduced to 'substrates' through which AI lives, which itself is a form of existential risk that Temple (2024) termed ‘the death of our humanity.’”

At the same time, a team at Google DeepMind was probing a different side of this same coin. In their study, “Evaluating Language Models for Harmful Manipulation,” they demonstrated that these systems can be steered to bypass rational scrutiny entirely, exploiting human biases to shift beliefs and behaviors across finance, health, and public policy. Together, these papers signal a shift in the AI risk landscape: the primary risk is no longer just a technical failure of the machine, but a psychological surrender by the human.

I believe the real danger is not that machines will start thinking like us, but that we will become accustomed to letting them think for us in the moments that matter. Not just work. Not just homework, customer service, search, or code. I mean the more intimate territory: what to say to a grieving sibling, whether to leave a partner, how to read a political event, when to trust one’s own instinct, when to override it, when to feel wronged, when to feel absolved. A civilization can survive many stupid tools. What it does not survive so easily is the gradual evacuation of judgment from the people who must still live with the consequences of action.

The piece by Ezra Klein is titled "I Saw Something New in San Francisco."  A clip from the article:

My experience of Anthropic’s Claude in recent months is that I’ll drop in a stub of a thought and immediately receive paragraphs of often elegant writing turning that intuition into something that looks, superficially, like a fully realized idea. It’s my impulse, but it has been recast and extended into something far more coherent. With each passing month, I have to expend more energy to recognize whether it’s fundamentally wrong or hollow.

I’ve been an editor for 15 years now. Recognizing a bad idea beneath good writing — even in myself — is part of my job. But what would it mean to grow up with that kind of companion? What would it mean to have your every adolescent intuition turned into persuasive prose? What is lost in not having to do the work to build out our intuitions ourselves?

Researchers have drawn a distinction between “cognitive offloading” and “cognitive surrender.” Cognitive offloading comes when you shift a discrete task over to a tool like a calculator; cognitive surrender comes when, as Steven Shaw and Gideon Nave of the University of Pennsylvania put it, “the user relinquishes cognitive control and adopts the A.I.’s judgment as their own.” In practice, I wonder whether this distinction is so clean: My use of calculators has surely atrophied my math skills, as my use of mapping services has allowed my (already poor) sense of direction to diminish further.

But cognitive surrender is clearly real, and with it will come the atrophy of certain skills and capacities, or the absence of their development in the first place. The work I am doing now, struggling through yet another draft of this essay, is the work that deepens my thinking for later.

In a thoughtful piece, the technology writer Azeem Azhar describes his efforts to safeguard “the space where ideas arrive before they’re shaped.” But how many of us will put in such careful, reflective effort to protect our most generative spaces of thought? How many people even know which spaces should be protected? For me, the arrival of an idea is less generative than the work that goes into chiseling that idea into something publishable. This whole essay began as a vague thought about A.I. and McLuhan. If I have gained anything in this process, it has been in the toil that followed inspiration.

The other thing I notice the A.I. doing is constantly referring back to other things it knows, or thinks it knows, about me. Sycophancy, in my experience, has given way to an occasionally unsettling attentiveness; a constant drawing of connections between my current concerns and my past queries, like a therapist desperate to prove he’s been paying close attention.

The result is a strange amalgam of feeling seen and feeling caricatured. Ideas I might otherwise have dropped keep getting reanimated; personal struggles I might otherwise move on from keep returning unexpectedly to my screen. I am occasionally startled by the recognition of a pattern I hadn’t noticed; I am often irked by the recitation of a thought I’m no longer interested in. The effect is to constantly reinforce a certain version of myself. My self is quite settled, but what if it wasn’t?

 

 

 

 

One terminal chats with another to start vibe coding a new dericbownds.net

Writing MindBlog posts is on temporary hiatus because I have recently followed Claude's instructions to set up my Mac mini M3 to use Claude code, getting my own API, etc., and starting for myself the "vibe coding" I have been reading so much about. The terminal app on the Mac is used to both view the lines of code it is executing as they flow past, as well as to issue instructions with your own keyboard (or voice). This is essentially one sort of terminal that interfaces with a vast underlying machinery (The Apple terminal App) with another sort of terminal (My sense of having a self, which is a very thin veneer over a more organic sort of machinery). In both cases, it's machines all the way down. I'm in the midst of touching up the clean simple code that Claude Code generated in several tries (that in total took less than 15 minutes) to replace my old HTML from the 1990s and then early 2000s using Dreamweaver templates that have now been removed.  I will be replacing the current dericbownds.net with a new one as soon as I am happy with what I'm seeing.  

The polyvagal theory is dead - and HRV isn't a simple indicator of arousal

I was recently struck by Baxter's Substack post (its title copied to be the title of this MindBlog post), which noted work critical of Porges' Polyvagal  theory (or PVT) published iin the journal Clinical Neuropsychiatry, because it calls into question one idea commonly derived from this theory that I have accepted (and repeated in several MindBlog posts): that heart rate variability can be taken as a simple indicator of calm (higher HRV and parasympathetic nervous activity) versus arousal (lower HRV and sympathetic nervous system activity).  A number of bio-monitors such as the Apple Watch and the Oura Ring report ongoing HRV measurements.  Here is a clip from her article, and then the abstract of the multi-author paper she references.

Psychological safety, social engagement, co-regulation, emotional freezing, dissociation — these are real phenomena, supported by decades of research that predate PVT. They come from attachment theory, trauma research, and somatic practices. The (critical paper’s authors explicitly state that body-mind therapeutic methods “may confer benefits on their own.”

So the work stands.

What doesn’t is the specific neuroanatomical story Porges constructed to explain these states, i.e. what drives them. This is the now-familiar three-tier hierarchy: dorsal vagal shutdown, ventral vagal social engagement, sympathetic activation.

Porges got the clinical observations right—or rather, he built upon clinical observations that were already well-established. But the mechanism is wrong.

Concerningly, Porges continued to assert that mechanism for thirty years, even as the evidence mounted against it, and apparently misrepresenting and distorting the work of other scientists to support his claims. Rather than engage with criticism of his theory, Porges ignored the overwhelming scientific consensus and instead promoted this inaccurate framework as sound, misleading practitioners and clients alike.

There is good news in all this mess—for clinicians at least. What you observe in your clients is real. Hyper-arousal is real. The need for safety and co-regulation is real. People really do dissociate, numb, and freeze when overwhelmed. Your interventions still work. They just don’t work for the reasons PVT said they did.

 *********

Here is the Grossman (+38 co-authors) summary:

This article specifically appraises--based upon the current state of knowledge of autonomic function and vertebrate evolution--several major elements of the PVT, as described in Porges (2025a) and elsewhere. These include: 1) the validity of PVT assumptions that respiratory sinus arrhythmia is a direct measure of the extent of central vagal drive to the heart; 2) PVT characterizations regarding the neuroanatomy and functions of two major brainstem vagal nuclei, the ventrally situated Nucleus Ambiguus and the Dorsal Motor Nucleus of the vagus nerve; 3) PVT assertions regarding the evolution of the vagus nerve; 4) PVT claims about the specificity of mammalian social behavior in relation to nonmammalian vertebrates, and 5) PVT interpretations of earlier seminal physiological literature. All co-authors agree that major tenets of the PVT are not supported by past or current knowledge and, in several instances, are inconsistent with the broader evidence base. Since the topics addressed constitute fundamental premises of the PVT, we conclude that the PVT is untenable, because it is not defensible based on existing neurophysiological and evolutionary evidence. The psychological elements composing the superstructure of the PVT are primarily derived from earlier psychological literature and are neither clarified nor strengthened by PVT constructs that lack evidence. This article does not intend to address alternative explanations about relations between vagal function and psychological processes, although such explanations do exist. 

The text of this article is mind-numbingly detailed and complex, as is a rebuttal of the critique by Porges in the same issue of Clinical Neuropsychiatry.   

 

The nature of intelligence and selves.

I want to pass on the result of my extracting what I felt to be crucial chunks of text from Chapters 5 through 9 of Agüera y Arcas’s "What is Intelligence" which can be found at https://whatisintelligence.antikythera.org/. I found myself unable to hold and summarize the rich array of ideas in these clips of text in my attentional space, so I asked Anthropic Claude, ChatGPT 4.2, and Google gemini to condense and assemble the main points and take home messages from the clips into a narrative roughly two pages long.  The Claude result astounded me.  Here it is:

What Intelligence Is: A Synthesis of Agüera y Arcas

Reality, Models, and the Limits of "Illusion"

A useful place to begin is with the claim, made by philosophers like Dennett, Harris, and Sapolsky, that consciousness, the self, and free will are illusions — polite fictions produced by inexorable physical processes. Agüera y Arcas pushes back, not by defending dualism, but by challenging the word "illusion" itself. We don't call tables and chairs illusory simply because they're made of atoms. A better framework, borrowed from physics, holds that "reality" is our name for a model with good predictive power. No single model covers everything; general relativity doesn't tell you whether your aunt will like your cake. What matters about a model is whether it agrees with observations, makes testable predictions, and serves a useful function within its domain.

Newtonian mechanics isn't an illusion just because general relativity supersedes it — relativity explains when and why the classical approximation holds. The same logic applies to our folk psychology of selves, intentions, and choices. Theory of mind — our intuitive model of other minds — is the "Newtonian mechanics" of social life: powerful, indispensable for everyday prediction, and philosophically incomplete. The task isn't to discard it but to find the more general theory that explains where it works and where it breaks down.

Free Will as a Real Computational Achievement

That more general theory reconceives free will not as a supernatural power nor as a mere illusion, but as a genuine computational process built from four components working together. First, theory of mind applied reflexively: we can model ourselves the way we model others, imagining what our future self will experience, want, and do — which is what makes planning possible at all. Second, internal randomness: to mentally simulate alternative futures, a mind must be able to "draw random numbers," wandering prospectively through possibilities the way daydreaming does, though more directed. Third, dynamical instability (the butterfly effect in neural circuitry): this allows the faintest internal signal — "imagine doing X" — to tip behavior one way or another, making self-directed choice possible. Fourth, selection: guided by theory of mind, we prune the space of imagined futures, favoring some and discarding others, much as AlphaGo's value network prunes its search tree.

Deliberate decisions result from extended exploration before commitment; snap decisions keep multiple paths open until the last moment. In either case, if a modeled self has genuinely sampled alternatives and chosen among them, something meaningful called free will has occurred — with no dualism required. The quantum indeterminacy of the physical world, far from undermining this picture, actually supports it: the future is genuinely open, counterfactuals are real, and choice is underwritten by that openness.

Consciousness as Social Self-Modeling

Consciousness emerges naturally from the same machinery. Because social animals model each other, and because those others are modeling them back, at some point the modeling turns reflexive: you model yourself as a being that others model. Neuroscientist Michael Graziano's Attention Schema Theory adds a further layer — consciousness is what arises when a system models its own attention. Agüera y Arcas endorses this view while again resisting the word "illusion": attention is real computation, and modeling it produces a real entity, a "who," not a fiction. The vertiginous "strange loop" that Hofstadter describes — the self seeing itself seeing itself — is the phenomenological signature of this recursive social modeling.

Crucially, the category of "who" is not fixed or universal. The history of personhood — from the Declaration of Independence to the Universal Declaration of Human Rights — shows that which entities are granted moral standing has changed dramatically and will continue to change. There is no God's-eye view from which to declare the question permanently settled.

Intelligence: Predictive, Social, Multifractal, Symbiotic

Drawing these threads together, Agüera y Arcas offers a unified account of intelligence: intelligence is the ability to model, predict, and influence one's future; it can evolve in relation to other intelligences to create a larger symbiotic intelligence. Several properties follow from this definition.

Intelligence is predictive at every scale — from bacteria anticipating chemical gradients to cortical circuits implementing predictive sequence modeling. It is social because much of an agent's environment consists of other predictors, making theory of mind an almost inevitable evolutionary development. It is multifractal — intelligences are built from smaller intelligences, with "selves" defined by the dynamic relationships among their parts rather than by any homunculus. It is diverse, because the parts must differ from one another to provide mutual benefit; specialization arises naturally from differences in connectivity. And it is symbiotic: when the dynamic stabilities of multiple intelligences become correlated, they find themselves "in the same boat" and learn to cooperate, producing larger emergent intelligences — from mitochondria to beehives to human cultures.

Language, LLMs, and the Social Brain

Language fits cleanly into this framework. Its primary function is not grammar or syntax but leveling up theory of mind — allowing social entities to share mental states through a mutually recognizable code. Because human language is rich enough to represent everything in our umwelt, and because it functions as a general-purpose social motor output (requesting anything imaginable from others), a neural network trained to predict the next word will tend to acquire something that looks — and may genuinely be — intelligent. The brain itself, Agüera y Arcas argues, is fundamentally an autoregressive sequence predictor, and the Transformer architecture, despite its differences from biological neural circuits, instantiates the same core principle.

The social brain's crowning structure, the prefrontal cortex, specializes precisely in theory of mind, and its dramatic expansion along the primate lineage underscores that human intelligence is, at its core, a collective achievement. We survive by the grace of others, our language exists for listeners, and even our involuntary signals — the blush, the quaver in the voice — are adaptations that make us legible to those around us. The self, in the end, is not a lonely Cartesian theater but a node in a web of mutual prediction, constituted by and for its relationships.

 

 

A seismic shift in our understanding of intelligence

I've just spent the last three days reading large chunks of the remarkable book "What is Intelligence" by Agüera y Arcas. It is best accessed at https://whatisintelligence.antikythera.org/, where you will find text with supporting graphics, charts, and animations.  Arcas argues that computation is the substrate for intelligence in all life forms and that prediction is the fundamental principle behind intelligence.  Goodall provides a good review of the book and Arcas's ideas. The introduction to the book encapsulates his basic ideas, and I pass on here ChatGPT's condensation of its main points: 

******** 

Agüera y Arcas’s introduction contrasts two broad frameworks for thinking about intelligence: an older, classical view of AI and a newer emerging perspective shaped by large language models.

The older view of intelligence and AI.
For decades, most scientists assumed that real intelligence required specialized internal mechanisms—explicit reasoning modules, symbolic logic, or carefully engineered algorithms. Machine learning systems that emerged in the late 20th and early 21st centuries were therefore seen as narrow tools rather than genuine intelligence. They performed tasks such as image recognition or sentiment analysis by approximating mathematical functions that map inputs to outputs. This approach produced impressive “Artificial Narrow Intelligence,” but it seemed fundamentally limited. Predicting the next word in a sentence, for example, appeared trivial—essentially a statistical task. Most researchers believed that such prediction models could never produce general intelligence because true intelligence was thought to require additional structures for reasoning, understanding concepts, planning, or possessing common sense.

The unexpected shift.
The emergence of large neural language models challenged this assumption. When trained on enormous text corpora, systems built for the simple task of next-word prediction began displaying abilities that look strikingly general: answering questions, solving problems, performing professional exams, writing code, and carrying on conversations. The key insight is that language prediction implicitly contains a huge range of cognitive demands. Correctly predicting the next word in many contexts requires background knowledge, reasoning, mathematics, commonsense understanding, and even “theory of mind.” What initially appeared to be a narrow statistical task turns out to embed many of the competencies traditionally associated with intelligence.

The debate about what this means.
This development has triggered a conceptual divide. One camp argues that these systems merely simulate intelligence; they generate convincing language without real understanding. The other camp suggests that this distinction may be misguided. If a system consistently behaves intelligently under questioning—passing tests of knowledge, reasoning, and conversation—then insisting that it is “only imitation” may move the discussion outside empirical science. This echoes Alan Turing’s argument that intelligence should be judged by functional behavior rather than by speculation about hidden inner states.

A broader functional perspective on intelligence.
Agüera y Arcas ultimately pushes toward a functional view similar to how biology understands organs. A kidney is defined not by the specific atoms composing it but by what it does. An artificial kidney that performs the same function is still a kidney. Likewise, intelligence may not depend on a particular biological substrate. If a system reliably performs the functions associated with intelligence—reasoning, conversation, problem solving—then from a scientific standpoint it may already qualify as intelligent.

The conceptual shift.
The old model treated intelligence as a special internal mechanism that machines would someday need to replicate. The emerging view treats intelligence as a set of capabilities that can arise from large systems optimized for prediction and interaction with the world. In this perspective, language prediction is not a trivial task but a gateway problem that implicitly contains much of what we mean by cognition. The surprising success of large language models therefore suggests that intelligence may be less mysterious—and more computationally emergent—than previously believed.

 

AI Makes Workloads Worse, Not Better

An article in today's Wall Street Journal by Ray Smith conforms so completely to my own work experience over the past week (I'm currently feeling fatigued from cognitive overload) that I pass on this Google Gemini summary of its main points:

An article "AI Makes Workloads Worse, Not Better" by Ray A. Smith in the 3/12/26 Wall Street Journal highlights a counterintuitive trend: rather than freeing up time for high-level creative tasks, artificial intelligence is actually increasing the speed, density, and complexity of work. Data from ActivTrak, which analyzed 164,000 workers, shows that AI users saw a 100% increase in time spent on messaging and a 94% increase in the use of business-management tools. Conversely, "focused work" time—the deep concentration needed for strategy and complex problem-solving—dropped by 9% for AI users.

This phenomenon is described as "work creep," where the efficiency gained from AI is immediately repurposed into additional tasks. Instead of working fewer hours, as some tech leaders predicted, employees are finding that their "appetite" for work remains unbounded. Experts note that because AI makes tasks feel easier and more accessible, it creates a momentum that pushes workers to take on broader scopes of work and more simultaneous projects.

While these shifts may initially boost productivity, researchers warn they are not without cost. The intensification of work can lead to cognitive overload, burnout, and a long-term decline in work quality. As AI agents prompt users to consider more variables and layers for every project, the result is often a "deeper dig" into existing work rather than a lighter schedule, suggesting that the dream of a shorter workweek remains elusive for most AI adopters.

 

Empathy as a Brain Rhythm: Orexin, Theta Oscillations, and Prosocial Behavior

(The following is the final installment of a series of three posts on using ChatGPT to assist in the generation of MindBlog posts.  I think this final product is a better job than than I would have done.)

A recent paper in Science by Kim et al. reports a striking mechanistic link between physiology, brain rhythms, and social behavior. The work shows that empathic and prosocial responses in mice depend on theta oscillations in the anterior cingulate cortex (ACC) driven by orexin neurons in the hypothalamus. The finding offers a concrete neural circuit connecting arousal systems, cortical dynamics, and behaviors we typically group under the heading of empathy.

The investigators examined situations in which one mouse observed another undergoing distress. When the observer had previously experienced a similar stressor, it displayed behaviors interpreted as empathic or prosocial—for example approaching, investigating, or grooming the distressed animal. During these episodes the researchers recorded increased theta-frequency oscillations (about 4–8 Hz) in the ACC, a cortical region long associated with affect, conflict monitoring, and social evaluation.

The key advance is that the authors were able to identify the physiological driver of these oscillations. Orexin neurons in the hypothalamus—cells best known for regulating arousal, wakefulness, and motivation—project to the ACC and modulate its activity. When orexin input was experimentally suppressed, ACC theta activity diminished and the animals’ empathic/prosocial behaviors dropped correspondingly. When the pathway was activated, both the oscillations and the behaviors increased.

The resulting circuit is conceptually simple:

prior experience of distress → orexin arousal system → ACC theta rhythms → empathic perception → prosocial behavior.

The work therefore suggests that empathy is not primarily a product of abstract reasoning about another’s mental state. Instead it emerges from state regulation in an embodied brain. Arousal systems in the hypothalamus adjust cortical dynamics, and those dynamics bias behavioral responses to social signals.

Several broader points follow from this.

First, the study provides a mechanistic explanation for something familiar in human experience: empathy is often strongest when we have “been there ourselves.” Prior experience of distress appears to tune the system so that observed distress activates the same regulatory circuitry. In other words, empathy may involve projecting one’s own internal models of bodily state onto another organism.

Second, the findings add to the growing list of cognitive functions associated with theta oscillations. Theta rhythms have been implicated in navigation, memory encoding, emotional regulation, and attention. This study suggests they may also function as a coordination signal for social behavior, synchronizing perception, affect, and action when organisms interact.

Third, the results reinforce a shift in neuroscience away from thinking about discrete cognitive “modules.” Instead, many functions appear to arise from dynamical interactions among regulatory systems—arousal circuits, oscillatory coordination, and cortical networks. In this view, empathy is less a specialized faculty than a particular configuration of a broader organism-level control system.

Finally, the work invites comparison with emerging ideas in artificial intelligence. Modern machine learning systems rely heavily on mechanisms that regulate network activity globally—attention signals, gain modulation, or other forms of state-dependent gating. The orexin-ACC circuit plays an analogous role: a global regulatory signal alters the operating mode of a cortical network and thereby changes behavior.

From an evolutionary perspective this makes sense. Prosocial responses are advantageous in social species, but they must be conditional and context-sensitive. Linking empathy to arousal systems provides exactly such a control mechanism. The organism helps others not because it has computed an abstract moral rule but because its own regulatory systems resonate with the perceived state of another.

The broader implication is that many aspects of social life—including empathy and perhaps the roots of moral behavior—may originate in basic physiological coordination among organisms. Brains evolved to regulate bodies in changing environments, and in social species those environments include other bodies. Empathy, in this light, becomes less a lofty psychological abstraction and more a biologically grounded form of resonance between nervous systems.

A MindBlog-ready ChatGPT synthesis

This post continues from the previous one to show  ChatGPT's response to my prompt "Let's go with the empathy / orexin paper for your producing a  MindBlog-ready synthesis" . 

MindBlog-ready synthesis

Paper: Empathy and prosocial behavior powered by orexin-driven theta oscillations (Science, Feb 2026)

Core finding

The study identifies a specific neural circuit linking arousal chemistry, brain rhythms, and social behavior:

• In mice, theta oscillations in the anterior cingulate cortex (ACC) increase during empathic states such as observing another animal in distress and during prosocial responses like allogrooming.

• Orexin neurons from the hypothalamus drive these ACC theta rhythms.

• When researchers optogenetically inhibited orexin input, both theta activity and empathy-related behaviors dropped.

• The circuit was particularly active when the observing animal had prior similar fear experience (“I’ve been there”).

Together, the work shows a causal chain:
prior experience → orexin arousal system → ACC theta rhythms → empathic perception → prosocial action.

---

Why this matters conceptually

1) Empathy emerges from regulation, not abstraction

Empathy here is not treated as a high-level cognitive construct but as a state transition in a regulatory organism:

• hypothalamus: arousal/homeostasis

• ACC: integration of affect and social information

• theta rhythm: coordination signal

• behavior: grooming/helping

This aligns with the broader shift in neuroscience from “modules” to state-dependent dynamical systems.

---

2) The “shared experience” effect is mechanistic

The finding that empathy strengthens when the observer has undergone similar stress provides a circuit explanation for a familiar human phenomenon:

• empathy is gated by embodied memory, not just symbolic inference

• experiential similarity tunes the arousal system, which tunes cortical processing

This supports evolutionary models where prosociality emerges from self-model projection onto others.

---

3) Theta oscillations as a social coordination signal

Theta rhythms already appear in:

• navigation

• memory encoding

• emotion regulation

This study extends that list: social resonance and prosocial action.

Implication:
theta may function as a general-purpose “alignment rhythm” coordinating internal state with external agents.

---

4) Affective empathy precedes moral cognition

The circuit runs:

arousal → affective resonance → prosocial behavior

not:

reasoning → moral decision → helping

This reinforces the view that morality is grounded in regulation and feeling, with cognition layered later.

---

Link to recurring MindBlog themes

Embodied cognition

Empathy is rooted in:

• visceral systems

• arousal chemistry

• sensorimotor rhythms

not symbolic interpretation alone.

Predictive processing

Prior fear experience sets priors.
Observed distress matches these priors → arousal system activates → action.

Social evolution

Prosociality appears as:

• adaptive coordination among organisms sharing vulnerability

• not altruism in a philosophical sense, but coupled regulation

Brain–AI parallels

This circuit resembles:

• reinforcement systems

• state-dependent gating

• attention modulation via global signals

A useful analogy:
orexin = global gain control;
theta = synchronization layer;
ACC = integrative inference node.

---

Possible MindBlog framing angles

A. “Empathy is a regulatory reflex”

Recast empathy as:

• homeostatic alignment across individuals

• not moral virtue, but organism-level coordination

B. “Shared suffering wires social behavior”

Prior experience as a prerequisite for deep empathy:

• trauma, illness, caregiving, aging

• explains cohort-based moral communities

C. “The rhythm of social connection”

Theta as the physiological substrate of:

• resonance

• attunement

• bonding

D. “From orexin to ethics”

A narrative arc:
arousal chemistry → brain rhythms → empathy → prosociality → culture

---

One-paragraph MindBlog draft core

A new Science paper traces empathy to a specific physiological circuit: hypothalamic orexin neurons drive theta oscillations in the anterior cingulate cortex, and this rhythmic coupling translates perceived distress into prosocial behavior. The effect is strongest when the observer has had similar prior experiences, suggesting that empathy is not primarily a cognitive inference but a state of embodied resonance. Social care emerges from regulation: arousal systems tune cortical dynamics, which then bias action. This reframes empathy from moral abstraction to biological coordination—an organism aligning its internal state with another’s.

---

Likely follow-on questions for exploration

• How generalizable is this circuit to humans?

• Relationship to:

  • oxytocin pathways

  • mirror neuron systems

  • interoceptive networks (insula)

• Could dysregulation here underlie:

  • psychopathy

  • burnout

  • compassion fatigue

---

Recommendation

This paper is strong material for a full post. It sits directly at the intersection of:

• affective neuroscience

• social evolution

• predictive/embodied cognition

• AI analogies about state-coupled agents

It supports a broader thesis you return to repeatedly:
mind is regulation in a body embedded in other bodies.