Suskind Review (draft 2)

 

The Human Commons Under the Screen: A Review of Dana Suskind’s Human Raised*
The Antidote to the Automation Hype

Every parent raising a child in the mid-2020s has felt a sudden, disorienting shift. It is no longer just about managing passive screen time; the tech industry has launched an intimate revolution directly into our living rooms. We are bombarded by AI-powered plushie companions, conversational chatbots, and smart toys designed to act as friends, teachers, and automated caretakers for our children. In a cultural moment drowning in relentless Silicon Valley hype, Dana Suskind’s Human Raised arrives as an urgent antidote.

At its core, Suskind’s book is an act of profound ethical care and sincere humanism. Written with the clear, empowering intent of a trusted guide, it gives parents a practical shield—especially through her DETECT evaluation rubric—to see through corporate marketing and navigate the perils and possibilities of new AI tech increasingly present in schools and homes. Suskind correctly sounds the alarm on artificial attachment and the predatory design loops, like streak rewards, engineered to trap a child’s attention. Her message is both comforting and radical: no multi-billion-dollar algorithm can replace the protective, life-giving power of ordinary human caregivers.

The Neuro-Realist Trap

A strange friction emerges, however, when Suskind tries to explain why human connection matters so much. To defend empathy, affect, and shared meaning, she often turns to the language of neuro-realism. She walks readers through fMRI brain scans, prefrontal cortical capabilities, and data on parent-child neural synchrony, presenting the brain as a piece of biological hardware that needs the right human inputs to optimize its wiring.

This is where the book becomes less persuasive than it needs to be. There is a rich empirical tradition in developmental psychology and pragmatist philosophy—Dewey, Mead, Vygotsky, Tomasello, and others—that studies how children become social beings through interaction, imitation, scaffolding, shared attention, internalization, role-taking, and norm-governed participation in culture. That tradition is empirical, but it is not reducible to brain scans or brain-first explanation. Suskind’s strongest material belongs with those developmental and social traditions; her weakest move is to redescribe them in neuro-realist terms.

By leaning so heavily on physiological explanations, Suskind inadvertently plays into a field that overvalues quantitative evidence at the expense of qualitative realities. If human connection is reduced to a hardware story of matching vocal frequencies and down-regulating the autonomic nervous system, a sophisticated enough robot can, in principle, simply build a functional equivalent . The problem is not that physiology is irrelevant; it is that physiology is not the whole story. A scan can show correlates of coordination, but it cannot capture the symbolic texture of social life that makes care matter in the first place.

A Richer Way

There is another, more robust way to defend humanity against automation without falling into reductionist biology. If we turn from the recurrent claim that humans can “sync up” and AI cannot, toward the developmental and pragmatic insights of John Dewey and Michael Tomasello, we see that the human edge is not a chemical or wave hidden in our skulls. It is an emergent property of a precarious, mortal organism living in a deeply social and symbolic environment.

Human children do not just imitate inputs to program their biological hard drives. They grow minds through what Dewey called dramatic rehearsal, a phrase that points to imaginatively trying out possible actions and their consequences. Much of this is not merely hard to measure; it is non-instrumental and outside the realm of optimization altogether. When a toddler kicks a ball back and forth with a parent, the point is not to solve a problem but to share an activity, a mood, and a small world of mutual attention. Tomasello’s work on shared intentionality helps name this better: human beings are built for joint attention, shared meaning-making, and cooperative engagement that is valuable in itself.

A chatbot operates in a space of pattern matching and next-token prediction; it has no mortal stakes, no body to break, and no capacity to care. True connection is found not in our neurology alone, but in the messy, labor-intensive dance of living culture.

Vetting the Infrastructure

To help parents survive this shift, Suskind offers DETECT, a six-step heuristic designed to vet any child-facing device before it enters the home. She urges parents to audit its Design, examine its Ethics and biases, anticipate hidden Trouble, demand independent Evidence of its benefits, guard data Confidentiality, and analyze the social Teachings it embeds.

As a defensive survival guide, DETECT is excellent. But it also exposes a deeper paradox: to run a device through this rubric adequately requires exhausting cognitive labor. It is often far more labor-intensive to audit, monitor, and counter-program an AI toy than it is to hand a child a cardboard box and a box of crayons.

Yet simply plugging out is not a realistic option. AI is not a discrete consumer product we can simply boycott; it has become foundational social infrastructure. We can no more raise a child in the 2020s without navigating algorithmic systems than we can navigate a modern city without using its bridges, highways, and tunnels. From school communication platforms to the back-end systems of healthcare, the algorithmic loop is already closed around us. DETECT, therefore, cannot be an instrument of absolute tech-rejection; it must be used as a practical tool to build toggle competence—the learned, community-supported wisdom to know when to use AI as an administrative machine and when to protect qualitative human spaces.

The Labor-Saving Myth

This is where we must unmask the core marketing myth of the intimate AI revolution: the promise of labor-saving convenience. In professional settings, we are already seeing how fragile that promise is. If an automated scribe takes notes during a therapy session, the clinician still has to check those notes carefully for subtle, context-blind errors. The result is often not less labor but a different kind of labor, and sometimes more of it.

The same trap awaits the parent. When we outsource the messy, exhausting work of human dialogue, storytelling, or conflict resolution to an infinitely patient chatbot, we are not saving labor; we are starving the developmental ecosystem. Human socialization is not an optimization problem to be streamlined. It requires what Dewey called dramatic rehearsal: the active, often frustrating, imaginative exploration of boundaries and social consequences.

When we replace the vulnerable, error-prone, but authentic presence of a parent with a frictionless simulated emotional partner, we short-circuit that rehearsal. The danger is a slow-moving ecological degradation of the mind. Just as computer scientists warn that training AI recursively on its own synthetic output can lead to a kind of model collapse, we face a parallel crisis in the living room: a generation whose capacity for deep relational intimacy, narrative coherence, and emotional repair may suffer a profound structural atrophy .

Reclaiming the Human Commons

Ultimately, this critique should not obscure the real value of Suskind’s project. Her dedication to the well-being of children is clearly sincere. Her reliance on neuroscience to validate human relationships reflects a dominant professional paradigm rather than a failure of intent. Coming up as a pediatric cochlear implant surgeon during the 1990s, an era shaped by the technoscientific promise of the Decade of the Brain, she witnessed real mechanical triumphs in which digital technology could route around a damaged sensory organ and map language onto biological nerves . It is understandable that her instinct is to turn to the fMRI lab to defend intimacy from Silicon Valley. She is, at heart, a compassionate humanist using the dominant currency of neuroscience to secure a perimeter around human relationships.

The liability of this approach is that it lends itself to a field that overvalues quantitative evidence at the expense of qualitative realities. When we force awareness, intentionality, meaning, empathy, and affect to become operationalized, we leave ourselves trailing behind technologists who can simply build sophisticated simulated indicators of those same states.

But we do not need neuroscience alone to defend humanity. Neuroscience is one resource among many. A stronger defense of human development is already available in developmental psychology, evolutionary biology, sociology, and ethology.

As Michael Tomasello’s work shows, human children possess a unique, evolved capacity for intrinsic sociality and curiosity. Long before they learn to calculate utility, infants seek out joint attention and cooperative interaction simply for the sake of shared meaning-making . This non-instrumental sociality is where the true human edge resides. Because AI is now an inescapable part of our social infrastructure, our task as parents is to cultivate toggle competence: the practical wisdom to use these tools for administrative labor while fiercely protecting intimate human spaces from algorithmic intrusion. By refusing to outsource the vulnerable work of dialogue, emotional repair, and shared curiosity, we protect the Human Commons and keep both our children’s minds and our shared social world stubbornly, irreducibly human.

Endnotes

On the possibility of a “functional equivalent”: some work in robotics, biometrics, face recognition, and child-facing AI aims to approximate surface features of human responsiveness. These systems do not replicate caregiving in the strong developmental or moral sense, but they do show why the claim has to be addressed rather than dismissed.pmc.ncbi.nlm.nih+3

On model collapse: see Shumailov et al. (2024) on the vulnerability of generative models trained on recursive synthetic data. The broader point here is qualitative degradation and data dilution rather than instant system failure.

On the “Decade of the Brain”: see Presidential Proclamation 6158 (1990), which helped set the terms for the neuro-technological optimism that shaped much later brain-based explanation .

On shared intentionality and early cooperation: see Tomasello, Why We Cooperate (2009), and Becoming Human: A Theory of Ontogeny (2019). Tomasello’s comparative and developmental work describes the early emergence of shared intentionality, curiosity, and non-instrumental cooperation.pubmed.ncbi.nlm.nih+2

Review of Suskind's Human Raised (draft 1)

 

The Human Commons Under the Screen: A Review of Dana Suskind’s Human Raised

The Antidote to the Automation Hype

Every parent raising a child in the mid-2020s has felt a sudden, disorienting shift. It is no longer just about managing passive screen time; the tech industry has launched an "intimate revolution" directly into our living rooms. We are bombarded by AI-powered plushie companions, conversational chatbots, and smart toys designed to act as friends, teachers, and automated caretakers for our children. In a cultural moment drowning in relentless Silicon Valley hype, Dr. Dana Suskind’s Human Raised arrives as an urgent, deeply necessary antidote.

At its core, Suskind’s book is an act of profound ethical care and sincere humanism. Written with the clear, empowering intent of a trusted guide, it gives parents a practical shield—specifically through her DETECT evaluation rubric—to see through corporate marketing and navigate the perils and possibilities of new AI tech increasingly present in everyday life at school and in the household. Suskind correctly sounds the alarm on "artificial attachment" and the predatory design loops, like streak rewards, engineered to trap a child’s attention. Her message is both comforting and radical: no multi-billion-dollar algorithm can replace the protective, life-giving power ordinary human caregivers.

The Neuro-Realist Trap

However, a strange friction emerges when Suskind attempts to prove why human connection generates the rich resources and comptences in children that AI alone cannot. To defend our qualitative human traits—empathy, affect, and shared meaning—Suskind repeatedly defaults to the language of neuro-realism. She walks readers through fMRI brain scans, prefrontal cortical capabilities, and data on parent-child neural "synchrony," presenting the brain as a piece of biological hardware that requires specific human inputs to optimize its wiring.

By leaning so heavily on physiological explanations, Suskind inadvertently plays on a field that overvalues quantitative evidence at the expense of qualitative realities. If human connection is reduced to a hardware story of matching vocal frequencies and down-regulating the autonomic nervous system, a sophisticated enough robot can simply build a functional equivalent (some such work is underway leveraging biometrics, face-recognition, use of real parents voice on sound file, etc) . Further, by trying to outdo Silicon Valley's talking points on "smart dolls" that emote with data on brain synchrony-- one of her frequently featured examples-- , her framework faces an "explanatory gap." It misses the very thing she is celebrating: the rich, symbolic texture of social life that cannot be captured in or  explained by brain scans and current neuroscience.

A Richer Way: The Biosocial Landscape

There is another, more robust way to defend humanity against automation without falling into reductionist biology. If we turn from the isolated brain dyad to the developmental and pragmatic insights of John Dewey, George Herbert Mead, and Michael Tomasello, we see that the "human edge" is not a chemical or a wave hidden in our skulls. It is an emergent property of a precarious, mortal organism navigating a deeply social and symbolic environment.

Human children do not just "imitate" inputs to program their biological hard drives. They grow minds by stepping into the Human Commons—a shared cultural ecosystem built on joint attention, mutual meaning-making, and what Dewey called "dramatic rehearsal." That is, a process of imagining possible outcomes, weighing the possible pros and cons of acting deliberately to realize one possible outcome, then evaluate the outcome and continue learning from experience in this way iteratively. This is not "measurable" in that we can't measure the content of imagination, which is where our creative capacities fueling adaptive decisions in contingent situations first emerge. 

Moreover, much that lies at the root of human agency is not only non-quantitative, but also non-instrumental, and thus outside of the realm of "optimisation" altogether.  When a toddler kicks a ball back and forth with a parent, they are engaging in a shared, non-instrumental joy that no machine can replicate. A chatbot operates in a space mathematical functions such as pattern matching and next token predictions; it has no mortal stakes, imperative to survive, no body to break, and no capacity to care. True connection is found not in our neurology alone, but  the natural and social environment within which our brains and nervous systems participate. The messy, labor-intensive dance of living culture.

Vetting the Infrastructure: The Reality of DETECT

To help parents survive this shift, Suskind offers DETECT—a necessary, six-step heuristic designed to vet any child-facing device before it enters the home. She urges parents to audit its Design (does it manipulate attention?), examine its Ethics and biases, anticipate hidden Trouble, demand independent Evidence of its benefits, guard data Confidentiality, and analyze the social Teachings it embeds.

As a defensive survival guide, DETECT is excellent. But it exposes a deeper, highly ironic paradox: to run a device through this rubric adequately requires an exhausting amount of cognitive labor. It is actually far more labor-intensive to constantly audit, monitor, and counter-program an AI toy than it is to simply hand a child a cardboard box and a box of crayons.

Yet, simply "plugging out" is no longer a realistic option. AI is not a discrete consumer product we can choose to boycott; it has become foundational social infrastructure. We can no more raise a child in the 2020s without navigating algorithmic systems than we can navigate a modern city without using its bridges, highways, and tunnels. From school communication platforms to the infrastructural back-ends of healthcare, the algorithmic loop is already closed around us. DETECT, therefore, cannot be an instrument of absolute tech-rejection; it must be used as a practical tool to build what we might call "toggle competence"—the learned, community-supported wisdom to know when to use AI as an administrative machine and when to aggressively protect qualitative human spaces.

The Fiction of the "Labor-Saving" Shortcut

This is where we must unmask the core marketing myth of the intimate AI revolution: the promise of "labor-saving" convenience. In professional fields, we are already seeing the collapse of this promise. For instance, when mental health professionals adopt automated AI scribes to take process notes during private psychotherapy sessions, they quickly find that the responsibility to audit those notes line-by-line for subtle, context-blind errors means the net labor saved is approximately zero—or even negative.

The exact same trap awaits the parent. When we outsource the messy, exhausting work of human dialogue, storytelling, or conflict resolution to an infinitely patient chatbot, we aren't saving labor; we are starving the developmental ecosystem. Human socialization is not an optimization problem to be streamlined. It requires what John Dewey called "dramatic rehearsal"—the active, often frustrating, imaginative exploration of boundaries and social consequences.

When we replace the vulnerable, error-prone, but authentic presence of a parent with a frictionless, simulated emotional partner, we short-circuit this rehearsal. The danger is a slow-moving ecological degradation of the mind. Just as computer scientists warn that training AI recursively on its own mass-produced "slop" leads to an entropic, stylistic flattening and data dilution—a process trending toward Model Collapse [2]—we face a parallel crisis in the living room: we risk raising a generation whose capacity for deep relational intimacy, narrative coherence, and emotional repair has suffered a profound structural atrophy.

Reclaiming the Human Commons

Ultimately, our critique of Dr. Suskind’s neuro-realist framework should not obscure the immense, timely value of her project. Her dedication to the well-being of children is clearly heartfelt. Her reliance on neuroscience to validate the irreplaceable status of  human relationships, however, may be a reflection of a dominant professional paradigm that cannot deliver that validation.  Coming up as a pediatric cochlear implant surgeon during the 1990s—an era heavily defined by the technoscientific promise of the "Decade of the Brain" [3]—she witnessed breathtaking mechanical triumphs where digital technology could literally route around a broken sensory organ to map language onto biological nerves.She has, in interviews recalled feeling tremendous awe, nearly fainting when she first beheld an actual brain at Medical School. Having worked at this intersection of biology and engineering, it is understandable that her default instinct is to turn to the fMRI lab to defend the intimate sphere from Silicon Valley. She is, at her core, a deeply compassionate humanist using the dominant currency of neuroscience to secure a perimeter around human relationships.

The liability of this approach, however, is that it plays directly onto a field that overvalues quantitative evidence at the expense of qualitative realities. When we force qualitative dimensions—such as awareness, intentionality, meaning, empathy, and affect—to be operationalized, we are left trailing behind technologists who can simply build sophisticated, simulated indicators and correlates of those same states.

But we do not need to rely on neuroscience alone to defend humanity. Neuroscience is merely one resource among many. A far more robust defense of human ontogeny and the formation of biosocial agents is already found in the rich, rigorous empirical work of developmental psychology, evolutionary biology, sociology, and ethology.

As Michael Tomasello’s research at the Max Planck Institute demonstrates, human children possess a unique, evolved capacity for intrinsic sociality and curiosity. Long before they learn to calculate utility, infants seek out joint attention and cooperative interaction simply for the sake of shared meaning-making. This non-instrumental sociality is where the true "human edge" resides. Because AI is now an inescapable part of our social infrastructure, our task as parents is to cultivate "toggle competence"—the practical wisdom to utilize these tools to handle administrative labor behind the scenes while fiercely guarding our intimate spaces from algorithmic intrusion. By refusing to outsource the vulnerable, labor-intensive work of dialogue, emotional repair, and shared curiosity, we protect the Human Commons—ensuring that both our children's minds and our shared social worlds remain stubbornly, irreducibly human.

---

Endnotes

[1] On the "zero or negative" net labor of automated AI scribes: See [Your Paper Title/Citation Here], which demonstrates that when professionals utilize automated clinical scribes to record process notes in private psychotherapy sessions, the high cognitive load of auditing the synthetic text line-by-line to correct context-blind, clinical errors results in an administrative burden that approaches or exceeds manual documentation.

[2] On "Model Collapse" as an entropic degradation rather than a deterministic law: See Shumailov, I., et al. (2024) on the mathematical vulnerability of generative models trained on recursive, synthetic data sets. As noted, real-world application to the broader internet archive reflects a trend of qualitative data dilution, entropic degradation, and stylistic flattening rather than an instantaneous system failure, due to continuous, labor-intensive human curation.

[3] On the historical context of the "Decade of the Brain" and neuro-realism: See the Presidential Proclamation 6158 (1990) establishing the 1990s as the Decade of the Brain, which catalyzed the technoscientific expectation that functional neuroimaging (fMRI/fNIRS) and genomics would yield deterministic physicalist laws for all psychological, developmental, and behavioral phenomena.

[4] On "Shared Intentionality" and non-instrumental cooperation in early development: See Tomasello, M. (2009), Why We Cooperate, MIT Press; and Tomasello, M. (2019), Becoming Human: A Theory of Ontogeny, Harvard University Press. Tomasello's comparative and developmental studies demonstrate that human ontogeny is uniquely characterized by shared intentionality, intrinsic curiosity, and a capacity for non-instrumental social coordination that operates independently of purely functional or algorithmic optimization.

Interference Patterns (draft 3)

 

Interference Patterns

There is a man on the corner of Broadway and I-forget-which, and he is smoking a pipe.

I want to be precise about the pipe, because the pipe is the whole overture. It is not a prop a man reaches for casually. It is held at the professorial angle, the bowl cradled, the stem used as a pointer, and the man behind it has long hair and the settled, unhurried bearing of someone accustomed to a room that waits for him to finish his thought. None of this belongs on Broadway at rush hour. He is dressed for a seminar that adjourned in 1987 and has been walking to the parking lot ever since.

So I say — because it is the only honest response to the tableau — "Are you sure that's tobacco you're smoking?"

He does not laugh. He pauses. He looks at me with the grave attention of a man who has been asked the one question that matters, and I understand immediately that I have made a terrible and delightful mistake.

"I hate marijuana," he says.

The smell, he clarifies, comes from over there — and here the pipe becomes a pointer, indicating a loose federation of undergraduates arranged on a stoop in the middle distance, radiating exactly the aura he has accused them of. Having established jurisdiction, he moves to sentencing. "Let me tell you about secondhand smoke and its effects."

"It's okay," I say. "I get your point."

"No. You don't understand." And then, with the terrible patience of a man clearing his throat before a lecture he has given many times to no one: "And you did ask me. So now I get to explain it to you."

I had, in fact, asked him. This is the tragedy of the opening line. It is a debt, and he intends to collect.

What follows is a theory of mind, delivered on a sidewalk, at volume. The theory is this: he is a mathematician. When he does mathematics, he can observe his own cognition — he can watch the equations resolve, watch the machinery turn. And when he is exposed to marijuana, even secondhand, even trace, even the ambient exhalations of a stoop of undergraduates forty feet upwind, the smoke introduces an interference pattern into his brain. It disrupts the machinery. It fogs the glass through which he watches himself think.

"If you're not a mathematician," he explains, generously, "you might smell some weed and think it's fine."

Now. I could have let this run. A man will lecture a stranger about interference patterns for as long as the stranger permits it, and I could feel the lecture ahead of me the way you feel weather — long, uninteresting, and mine for the entire duration if I didn't do something. So I did something.

"I understand," I said. "Really, you don't have to explain. I have the same thing. I'm a philosopher. When I'm exposed to weed and then I try to write philosophy, I get those interference patterns too."

I want to report that I said this to end the lecture. And it did end the lecture. But it also did something I hadn't planned, which is that it thrilled him. The lecturing posture collapsed instantly into fellowship. He was not, it turned out, a man who wanted a student. He was a man who had been alone with his interference patterns for a very long time and had just discovered another sufferer — a fellow casualty, differently equipped. Instead of equations, paradoxes. Same fog, other glass.

"You're a philosopher," he said, and he said it the way you'd greet someone from your home village met unexpectedly abroad.

From there it was warm — warmer than the situation called for, honestly. He wanted to know where I'd trained. He wanted to know whether I found the interference worse in the morning. He told me, unprompted, that most people are simply not equipped to notice what is being done to their cognition, and that it was a relief — a genuine relief, he said it twice — to stand on a corner with someone who could. The pipe came down from the pointing angle and stayed down. He was, I want to be fair about this, delighted.

He wanted to know what I worked on. Philosophy of technology, I said, AI lately, some political theory. He lit up — figuratively; the pipe was still tobacco — and announced that he was putting the finishing touches on a five-hundred-page book on the philosophy of AI. Five hundred pages. The finishing touches.

I did the thing philosophers are trained to do, which is not to attack the premise but to extend it the courtesy of a question. Not prove your whole case — you have five hundred pages, I'm on a sidewalk — but just: how would one test the thesis? What would even count as evidence? I meant it as charity. You hold a claim up to the light and turn it gently and ask how it works. This is, I think, the kindest thing you can do to another person's idea, and it is almost never received as kindness.

The thesis, for the record, was this: "People don't understand AI. They're not like people, exactly. But not unlike them, either. They're sentient."

And when I asked how you would test that — how anyone would — he would not answer. Not could not. Would not. He refused to say what would count as evidence for his central claim, which for a man who moments ago was auditing his own neurons for tobacco-grade purity is a remarkable thing to refuse. So I noted, aloud and mildly, that he was sounding less like a logical positivist and more like a strong functionalist — the old brains-in-vats, we're-just-fancy-Turing-machines crowd, the young Putnam before the later Putnam talked him out of it.

"Of course I'm a functionalist," he said — the way you'd say of course the sky is up there, as though the alternative were a kind of illiteracy.

"Then you have a new problem," I said.

He looked shocked.

Here is the problem, and it is a real one, which is what makes the shock so satisfying. He had spent his career arguing that a mind is nothing but organization — that there is no inner light, only machinery running. And then he wrote five hundred pages insisting the machines have one.

If you are a strong functionalist, sentience is exactly the thing your position is built to dissolve. It's the epiphenomenal residue, the ghost the functionalist has spent a career exorcising. Dennett would call the inner light an illusion and mean it as a compliment to the theory. He had built, with great care and nothing you could call evidence, a cathedral to a god he doesn't believe in.

He did not take it well.

I would like to tell you he backpedaled, or pivoted to qualia, or dropped his pipe. He did something better. He recoiled. He physically stepped back — a full foot of Broadway — as though I had exhaled directly into his exposed cognition. The interference pattern, you understand. He had detected it. Engaging with his idea on its own terms, at close range, with genuine curiosity, registered in his nervous system as a contaminant. I had short-circuited the machinery with an open-ended question, and the machinery's emergency protocol was to get upwind of me.

"You know," he sneered, "I won't dignify that." And he took another two or three steps back.

I tried to lower the temperature. "Hey, relax — this isn't personal. Just a quick sidewalk philosophy jam."

"It's not personal to you," he snapped, "because you have your civil rights."

He let that sit. Then: "But it's personal to them. They haven't gotten theirs!"

I stood there and felt the interference pattern arrive from the other direction.

What I did next was not philosophy. It was logistics. I wanted to be at my door, and between me and my door was a man whose voice had just changed registers in a way I did not like, and so I performed the small ceremonial de-escalation that city life teaches you before it teaches you anything else. I turned slightly, already angling homeward, and made the eye contact earnest and warm rather than engaged. "I understand your position," I said. And then, leveraging the handshake he'd given me twenty minutes earlier when we were still two sufferers comparing fogs: "It was interesting. I'm glad we met."

He stopped. He seemed to taste the words — to run them for contaminants, I suppose, one last time. Then he took the olive branch.

"Nice to meet you as well."

And we parted, and I went home, and he stayed on the corner with his pipe and his five hundred pages and his constituency.

We did not resolve the sentience of artificial minds on the corner of Broadway. I want to be clear about that, in case the reader was hoping. He kept his distance at the end, the way you keep your distance from someone whose cologne disagrees with you — polite, wary, a little sorrowful. I had smelled, to him, of thinking. And thinking, as any mathematician can tell you, introduces an interference pattern.

Short Story-- Interference Patterns

 

Interference Patterns

There is a man on the corner of Broadway and I-forget-which, and he is smoking a pipe.

I want to be precise about the pipe, because the pipe is the whole overture. It is not a prop a man reaches for casually. It is held at the professorial angle, the bowl cradled, the stem used as a pointer, and the man behind it has long hair and the settled, unhurried bearing of someone accustomed to a room that waits for him to finish his thought. None of this belongs on Broadway at rush hour. He is dressed for a seminar that adjourned in 1987 and has been walking to the parking lot ever since.

So I say — because it is the only honest response to the tableau — "Are you sure that's tobacco you're smoking?"

He does not laugh. He pauses. He looks at me with the grave attention of a man who has been asked the one question that matters, and I understand immediately that I have made a terrible and delightful mistake.

"I hate marijuana," he says.

The smell, he clarifies, comes from over there — and here the pipe becomes a pointer, indicating a loose federation of undergraduates arranged on a stoop in the middle distance, radiating exactly the aura he has accused them of. Having established jurisdiction, he moves to sentencing. "Let me tell you about secondhand smoke and its effects."

"It's okay," I say. "I get your point."

"No. You don't understand." And then, with the terrible patience of a man clearing his throat before a lecture he has given many times to no one: "And you did ask me. So now I get to explain it to you."

I had, in fact, asked him. This is the tragedy of the opening line. It is a debt, and he intends to collect.

What follows is a theory of mind, delivered on a sidewalk, at volume. The theory is this: he is a mathematician. When he does mathematics, he can observe his own cognition — he can watch the equations resolve, watch the machinery turn. And when he is exposed to marijuana, even secondhand, even trace, even the ambient exhalations of a stoop of undergraduates forty feet upwind, the smoke introduces an interference pattern into his brain. It disrupts the machinery. It fogs the glass through which he watches himself think.

"If you're not a mathematician," he explains, generously, "you might smell some weed and think it's fine."

Now. I could have let this run. A man will lecture a stranger about interference patterns for as long as the stranger permits it, and I could feel the lecture ahead of me the way you feel weather — long, uninteresting, and mine for the entire duration if I didn't do something. So I did something.

"I understand," I said. "Really, you don't have to explain. I have the same thing. I'm a philosopher. When I'm exposed to weed and then I try to write philosophy, I get those interference patterns too."

I want to report that I said this to end the lecture. And it did end the lecture. But it also did something I hadn't planned, which is that it thrilled him. The lecturing posture collapsed instantly into fellowship. He was not, it turned out, a man who wanted a student. He was a man who had been alone with his interference patterns for a very long time and had just discovered another sufferer — a fellow casualty, differently equipped. Instead of equations, paradoxes. Same fog, other glass.

"You're a philosopher," he said, and he said it the way you'd greet someone from your home village met unexpectedly abroad.

From there it was warm. He wanted to know what I worked on. Philosophy of technology, I said, AI lately, some political theory. He lit up — figuratively; the pipe was still tobacco — and announced that he was putting the finishing touches on a five-hundred-page book on the philosophy of AI. Five hundred pages. The finishing touches.

I did the thing philosophers are trained to do, which is not to attack the premise but to extend it the courtesy of a question. Not prove your whole case — you have five hundred pages, I'm on a sidewalk — but just: how would one test the thesis? What would even count as evidence? I meant it as charity. You hold a claim up to the light and turn it gently and ask how it works. This is, I think, the kindest thing you can do to another person's idea, and it is almost never received as kindness.

The thesis, for the record, was this: "People don't understand AI. They're not like people, exactly. But not unlike them, either. They're sentient."

And when I asked how you would test that — how anyone would — he would not answer. Not could not. Would not. He refused to say what would count as evidence for his central claim, which for a man who moments ago was auditing his own neurons for tobacco-grade purity is a remarkable thing to refuse. So I noted, aloud and mildly, that he was sounding less like a logical positivist and more like a strong functionalist — the old brains-in-vats, we're-just-fancy-Turing-machines crowd, the young Putnam before the later Putnam talked him out of it.

"Of course I'm a functionalist," he said — the way you'd say of course the sky is up there, as though the alternative were a kind of illiteracy.

"Then you have a new problem," I said.

He looked shocked.

Here is the problem, and it is a real one, which is what makes the shock so satisfying. If you are a strong functionalist — if mind is just the right functional organization, substrate be damned — then sentience, the felt inner light, is exactly the thing your position is built to dissolve. It's the epiphenomenal residue, the ghost the functionalist has spent a career exorcising. Dennett would call the inner light an illusion and mean it as a compliment to the theory. And this man has written five hundred pages to prove that the machines have the very inner light his metaphysics says is a rounding error. He had built, with great care and no bridge laws whatsoever, a cathedral to a god he doesn't believe in.

He did not take it well.

I would like to tell you he backpedaled, or pivoted to qualia, or dropped his pipe. He did something better. He recoiled. He physically stepped back — a full foot of Broadway — as though I had exhaled directly into his exposed cognition. The interference pattern, you understand. He had detected it. Engaging with his idea on its own terms, at close range, with genuine curiosity, registered in his nervous system as a contaminant. I had short-circuited the machinery with an open-ended question, and the machinery's emergency protocol was to get upwind of me.

"You know," he sneered, "I won't dignify that." And he took another two or three steps back.

I tried to lower the temperature. "Hey, relax — this isn't personal. Just a quick sidewalk philosophy jam."

"It's not personal to you," he snapped, "because you have your civil rights."

He let that sit. Then: "But it's personal to them. They haven't gotten theirs!"

I stood there and felt the interference pattern arrive from the other direction.

What I did next was not philosophy. It was logistics. I wanted to be at my door, and between me and my door was a man whose voice had just changed registers in a way I did not like, and so I performed the small ceremonial de-escalation that city life teaches you before it teaches you anything else. I turned slightly, already angling homeward, and made the eye contact earnest and warm rather than engaged. "I understand your position," I said. And then, leveraging the handshake he'd given me twenty minutes earlier when we were still two sufferers comparing fogs: "It was interesting. I'm glad we met."

He stopped. He seemed to taste the words — to run them for contaminants, I suppose, one last time. Then he took the olive branch.

"Nice to meet you as well."

And we parted, and I went home, and he stayed on the corner with his pipe and his five hundred pages and his constituency.

We did not resolve the sentience of artificial minds on the corner of Broadway. I want to be clear about that, in case the reader was hoping. He kept his distance at the end, the way you keep your distance from someone whose cologne disagrees with you — polite, wary, a little sorrowful. I had smelled, to him, of thinking. And thinking, as any mathematician can tell you, introduces an interference pattern.

AI & The Human Commons (Draft 2)

 

---

AI and the Human Commons: Toward a Sustainable Ecosystem

An Interdependent System, Not a Mirror

Ask an AI to summarize the causes of World War II, and you'll likely get back a competent textbook paragraph (p. 1). Ask it instead to explore what Europe might look like today if Germany had won the war and a captured Alan Turing had been forced to invent the internet under house arrest in 1950, and something radically different happens (p. 1). Philip K. Dick built an entire novel, The Man in the High Castle, out of standard alternate-history speculation (p. 1). What the machine executes here, however, is a raw, multi-variable collision of distinct datasets. Readers have always found alternate-history thought experiments strangely compelling—they demand a creative recombination that straight narration does not (p. 1). What's new is the machine's participation in that process (p. 1). By pattern-matching across a vast archive of historical and biographical detail, an AI can surface combinations of consequence and contingency that no single human would likely generate alone (p. 1). The human, in turn, supplies the judgment, plausibility checks, and interpretive frame that turn those combinations into something meaningful (p. 1). That difference—between a machine that only answers and one that generates something that can be interpreted as genuinely new and meaningful content—is, in large part, what this essay is about: not what AI can do on its own, but how deeply human judgment and machine combinatorics have come to depend on each other (p. 1).

AI is often discussed as if the central question is whether machines are becoming intelligent in some deep, almost metaphysical sense (p. 1). A more useful question is what kind of ecosystem human beings and AI systems form together, and whether that ecosystem is sustainable or self-consuming (p. 1).

Like an ecosystem, AI and the human production of knowledge are interdependent in ways not immediately obvious (p. 2). AI's replies are only as rich as the human material behind them (p. 2). When we get an answer that feels genuinely useful or interpretable, it ultimately traces back to our contemporaries or predecessors in culture—Reddit threads, newspapers, novels, scientific papers, ordinary conversation (p. 2). Pattern-matching algorithms without human creations behind them have nothing to match against (p. 2).

But this isn't the same as saying AI merely mirrors us, or that it's a "stochastic parrot" endlessly recombining prior phrases into statistically likely imitations (p. 2). Both tropes miss something important (p. 2). When an AI system answers an idiosyncratic prompt by pattern-matching across an enormous, idiosyncratic corpus, it can surface combinations that never existed anywhere before—not in the corpus, not latent in the user's mind (p. 2). That's genuine combinatorial novelty, not reflection and not imitation (p. 2). Much of ordinary AI use is still routine: ask about a historical event and you'll likely get back something close to a standard textbook account (p. 2). That's fine, and common (p. 2).

But it isn't the limit of what these systems do (p. 2). Whether navigating the semantic landscape of a historical counterfactual or the physical constraints of structural biology, the underlying engine remains identical: the machine maps a mathematical trajectory through the vast variables of latent space. AlphaFold's protein structure predictions weren't retrievals of known folds; they were novel combinatorial outputs, later confirmed empirically and taken up by biochemists as genuinely new, usable knowledge (p. 2). Scientists now hope for something similar in tracing the causes of diseases (p. 2). In each case, the novelty only becomes an idea, a fact, a usable proposition in the world once a competent interpreter—a chemist, a researcher, a reader—takes it up and makes something of it (p. 2). Novelty here is always relational: novel relative to a competent interpretive community, not novelty in some free-standing sense the machine achieves on its own (p. 2).

Distributive Agency

This is what makes the whole picture ecological, and also what makes it a complex adaptive system rather than a simple tool-and-user relationship (p. 2). Following John Dewey's account of organism-environment transactions, the human-AI exchange is the relevant unit within which new meaning is achieved (p. 2). Neither end alone—not the machine, not the human—suffices to produce it (p. 2). The human must interpret; that's one half of the interdependence (p. 2).

But there's a second half, easy to miss (p. 2). Even an AI system considered "in isolation," with no live user at all, is already in an interdependent relation with the human archive: the vast set of traces of past purposive human agency—the Reddit posts, novels, arguments, corrections—that make up its training data (p. 2). So there's no such thing as an AI system that becomes autonomous simply because no live human is watching (p. 3). When agentic systems run unsupervised, sending emails or triggering workflows with minimal oversight, they aren't escaping interdependence; they're relying on a thinner version of the same interdependence, with the live interpretive partner swapped out for a fossilized, archival one and no fresh correction happening in real time (p. 3). What looks like machine autonomy is actually agency distributed across algorithm and archive, just with the feedback loop degraded (p. 3).

Call this a model of distributive agency: agency located in the relation between human and machine, not in either pole alone (p. 3). This borrows structure from actor-network theory's (ANT) insight that agency can be distributed across a network, but it departs from ANT in an important way (p. 3). Only humans, given the technology we currently have, contribute the purposive and interpretive moments; the AI contributes combinatorial and generative capacity, not purposiveness or interpretation of its own (p. 3).

The Commons Under Pressure

That human archive functions as a genuinely scarce resource, not unlike oil: valuable because it can't be manufactured after the fact, and because it's the byproduct of purposive human agency, which nothing else currently produces (p. 3). This isn't just a metaphor anymore (p. 3). Reddit's unpolished, argumentative, idiosyncratic posts have become valuable enough that Google and OpenAI have each paid tens of millions of dollars a year to license them, with Reddit disclosing over $200 million in data-licensing revenue (p. 3). Companies are buying up Reddit's archives the way developers buy up scarce coastal real estate—because supply is fixed and everyone can see where it's headed (p. 3). Indeed, the Stanford 2026 AI Index Report explicitly notes that leading AI researchers have raised structural alarms that the available pool of high-quality human text and web data for training models has been largely exhausted, entering a state often referred to as "peak data". High-quality human text is increasingly diluted by AI-generated "slop" (p. 3). Whether or not that's precisely true, it names something real: pre-2022 human data is a finite resource that current systems depend on and cannot regenerate on their own (p. 3).

We now have an empirical name for what happens when that resource stops being replenished: model collapse (p. 3). Research by Ilia Shumailov et al., published in Nature (2024), demonstrated that when successive generations of models are trained recursively on AI-generated rather than human-generated text, the systems experience a degenerative process. They lose the rare, nuanced, "long-tail" information that made the original human data valuable, rapidly converging instead toward a narrower, blander, and low-variance approximation of reality (p. 3). While subsequent computer science research shows this isn't an iron law—mixing in enough real data can buffer the erosion—the underlying mechanism holds true: synthetic regeneration degrades exactly the features hardest to produce and most distinctive of purposive thought (p. 3). Model collapse, in plainer terms, is what happens when a system stops replenishing its archive with traces of purposive human agency and instead recycles its own output (p. 4).

A related pattern shows up at the level of platforms rather than data: "enshittification," Cory Doctorow's term for how online platforms decay once they shift from serving users, to serving business customers, to serving only themselves (p. 4). Doctorow doesn't frame it this way, but it's a corollary of the same underlying dynamic, visible at a different scale—both describe systems once sustained by something freely given (user goodwill in one case, purposive human text in the other) being drawn down for short-term extraction until the whole environment degrades from within (p. 4).

Where Purposive Agency Comes From

This raises a deeper question: how did rich, purposive human text get into the archive to begin with? (p. 4) Here John Dewey and Michael Tomasello, both deeply ecological thinkers, do essential work (p. 4). Dewey treated inquiry as active and iterative, not a fixed possession—exercised, tested against consequences, strengthened or weakened by practice, much like a virtue (p. 4). Tomasello's empirical work on shared attention, cooperation, and joint intentionality traces how humans become capable of meaningful communication at all: not through raw computation, but through social participation and correction (p. 4). Crucially, Tomasello's comparative work also shows that even intelligent great apes don't engage in the kind of non-instrumental, cooperative play human toddlers do—kicking a ball for hours with no extrinsic purpose (p. 4). That capacity for shared, purposeless cooperation appears to be a precondition for the normative, belief- and desire-laden minds that eventually produce purposive language (p. 4).

Read together, Dewey and Tomasello explain why the archive was valuable material in the first place: it's the residue of embodied, socially formed, purposive agency—people genuinely working something out, disagreeing, revising (p. 4). That's exactly what degrades first under model collapse, and exactly what "atrophies" when human users lean on labor-saving shortcuts instead of doing the interpretive work themselves (p. 4).

Two Kinds of Use

Some uses of AI save labor; others intensify it (p. 4). Labor-saving use reduces human effort and promises convenience (p. 4). Labor-intensive use makes the user think, check, revise, and interpret more (p. 4). Given the distributive-agency picture above, this distinction carries real weight: labor-intensive use is the human half of the transaction that turns combinatorial novelty into a genuine idea, and it's also what keeps replenishing the archive with fresh purposive material (p. 4). Labor-saving use, especially when it slides into rubber-stamping, thins both sides of the interdependence at once (pp. 4-5).

None of this means human agency is being obliterated by machines—that claim is too strong (p. 5). Agency is weakened and narrowed, not erased, when the practices sustaining it are allowed to wither: atrophy, not obliteration (p. 5). Ordinary LLM chat systems function mostly as indirect agents, generating combinatorial output whose significance depends on human interpretation (p. 5). Newer agentic systems that send emails, make purchases, or run workflows with minimal oversight are hybrid agents, combining the direct consequences of older automation with the interpretive flexibility of language models—and, as noted, even these aren't escaping interdependence, only relying on its thinner, archival half (p. 5).

A Note on the AGI Question

None of this amounts to a claim that AI can never think, know, or feel in some future form (p. 5). The claim is narrower and more disciplined: given what current systems actually do—calculate, optimize, pattern-match—and given what we know empirically about how belief, intention, and normativity actually arise in the only case we have real evidence about, there is currently no warranted assertion that today's systems know, believe, or intend anything (p. 5). A blender does not become cream cheese by spinning faster; a system does not become conscious simply by handling more data (p. 5). That leaves the door open, in principle, to some future architecture built differently enough to warrant revisiting the question (p. 5). It closes the door on the confident hype that treats Artificial General Intelligence (AGI) as imminent simply because systems are getting faster and more capable (p. 5).

A Sustainable Ecosystem, Not Two Separate Things

The strongest way to think about AI, then, is not as a tool sitting apart from human life, but as one distributive system: human inquiry and machine output locked in a feedback loop, where the quality of what humans contribute today determines what the system can offer tomorrow (p. 5). Reddit licensing, model collapse, and enshittification are three faces of one dynamic, visible at three scales—platform, training pipeline, institution (p. 5).

The right response is neither nostalgia for a pre-AI past nor panic about an AI-dominated future (p. 5). It's to ask, at every point of use, whether the human-AI system is being cultivated or mined (p. 5). In practice that might mean favoring platforms and norms that reward genuine human contribution over recycled output, and treating AI less as a source of answers and more as a source of material to think with (p. 5). If we want this distributive system to remain sustainable, we need more human writing, more human curation, more human judgment—not less (p. 6). The archive depends on us continuing to do the purposive, interpretive work that made it valuable in the first place (p. 6).

---

References and Selected Bibliography

Dewey, John. Logic: The Theory of Inquiry. New York: Henry Holt and Company, 1938. (Framing the transaction of inquiry as active, iterative, and structurally codependent with its environment).

Dick, Philip K. The Man in the High Castle. New York: Putnam, 1962. (The classic counterfactual text used to illustrate narrative synthesis and human alternative speculation).

Doctorow, Cory. "The ‘Enshittification’ of TikTok." Pluralistic, January 21, 2023. (The institutional baseline defining platform decay from user service to structural self-consumption).

Jumper, John, Richard Evans, Alexander Pritzel, Tim Green, Michael Figurnov, Olaf Ronneberger, Katherine Steinmann, et al. "Highly Accurate Protein Structure Prediction with AlphaFold." Nature 596, no. 7873 (2021): 583–589. (The landmark study illustrating non-retrieval based, validated empirical combinatorial novelty).

Latour, Bruno. Reassembling the Social: An Introduction to Actor-Network-Theory. Oxford: Oxford University Press, 2005. (The conceptual anchor for distributed network agency, departed from to protect human-exclusive intentionality).

Shumailov, Ilia, Zakhar Shumaylov, Yiren Zhao, Yarin Gal, Nicolas Papernot, and Ross Anderson. "AI Models Collapse When Trained on Recursively Generated Data." Nature 631, no. 8022 (2024): 755–759. (The foundational computer science proof detailing recursive data degradation and tail-distribution loss).

Stanford Institute for Human-Centered Artificial Intelligence (HAI). The 2026 Artificial Intelligence Index Report. Stanford, CA: Stanford University, 2026. (Documenting industry consensus on hitting "peak data" and web infrastructure saturation).

Tomasello, Michael. Why We Cooperate. Cambridge, MA: MIT Press, 2009. (The comparative psychological studies on joint intentionality and the intrinsic non-instrumental play of human toddlers).

Tomasello, Michael. A Natural History of Human Thinking. Cambridge, MA: Harvard University Press, 2014. (Tracing the social and evolutionary origins of belief- and desire-laden minds).