The Mind at the Boundary

Where does a thought happen? The familiar answer points inward, toward the brain. A growing family of ideas asks us to look at the boundary instead: the active surface where organism, body, tool, other minds, and environment continually shape one another.

Remembering can begin with a glance at a note. Anxiety can become legible through the body before it becomes a sentence. A room can guide attention. A garden can hold the history of weather, care, and neglect. An AI system can reorganize an idea, while the person using it supplies the purpose, context, and judgment the model lacks.

None of these examples fits neatly into a picture of cognition as a self-contained processor receiving inputs and producing outputs. The useful unit is often the loop:

This research note brings three very different theories into conversation. Allan Adams, Paul Chesler, and Hong Liu show how turbulent dynamics in a lower-dimensional quantum system can correspond to evolving black-hole geometry.¹ Chris Fields and Michael Levin argue that cognitive systems routinely recruit their environments as memory and computation, treating the system–environment interface as a constrained information channel.² Stuart Hameroff and Roger Penrose propose that orchestrated quantum states in neuronal microtubules undergo gravity-related objective reduction, associated with conscious moments.³

These papers do not form a unified theory. They do not even use the word boundary in the same physical sense. Their value together is more disciplined: they separate several questions that are often collapsed into one.

Three boundaries, kept distinct

The boundary that encodes a world

Gravitational holography is one of modern physics’ most startling lessons. In certain highly structured theories, a quantum system without gravity can be mathematically equivalent to a higher-dimensional description containing gravity. Adams and his collaborators developed controlled examples in which turbulent fluid behavior on the boundary corresponds to the dynamics of an asymptotically anti-de Sitter black hole in the bulk.

The careful word is encoding, not projection. The boundary is not a cinema screen manufacturing a separate interior. Boundary and bulk are equivalent descriptions of the same underlying information. Nor does holography require ordinary spacetime to be made of literal pixels. The relevant theories are normally formulated with continuum quantum fields and special mathematical conditions.

What survives translation into our inquiry is a rigorous precedent: collective variables at a boundary can, under the right conditions, encode a rich emergent interior. That precedent is powerful. It does not mean a cell membrane, a user interface, or a brain automatically has a gravitational dual.

The boundary through which systems think together

Fields and Levin offer the most directly useful layer for the information systems we build. Their 2026 manuscript, Cognitive Offloading Is a Cognitive Universal, begins with a system and an environment whose internal state spaces are much larger than the bandwidth available at their shared boundary. Neither side can completely observe or reconstruct the other. Each encounters the other through limited, actionable signals.

The important move is reciprocal. An organism changes its environment; the changed environment stores, transforms, and returns consequences. A beaver’s dam alters the river that then shapes the beaver’s future choices. Writing preserves memory outside the nervous system. The body detects and regulates conditions on behalf of the brain. Language stabilizes meanings no individual speaker possesses alone. An LLM’s apparent capability depends on training data, tools, users, interfaces, institutions, and the wider linguistic world.

In this account, the environment is not merely an input source. It performs part of the computation. Cognitive offloading is therefore more than a clever human strategy. It is a recurrent consequence of bounded systems acting to make future interactions more tractable.

The boundary between possibility and actuality

Orchestrated objective reduction asks a different question. Penrose proposes that superpositions involving sufficiently different mass distributions, and therefore different spacetime geometries, are physically unstable. Hameroff proposes that microtubules organize quantum processes relevant to neuronal function, with objective reduction producing discrete conscious events.

The proposal is controversial. Functionally consequential quantum coherence has not been demonstrated across the required scales in living brains. Microtubule quantum-optical effects and changes in anesthetic sensitivity are interesting, but neither establishes quantum computation in cognition, gravity-driven collapse, or consciousness. The simplest parameter-free implementation of the Diósi–Penrose collapse model has also faced direct experimental constraint.⁷

Still, Orch OR isolates a genuine explanatory target: why or how one integrated possibility becomes an actual experience. Even if the proposed mechanism fails, that question remains.

A reconstructed interior

A conscious world feels spatially extended, unified, continuous, meaningful, and centered on a perspective. Yet sensory and bodily systems provide narrow, fragmented, and delayed signals. The retina offers curved two-dimensional arrays, while perception presents a three-dimensional scene. Sound arrives as pressure variation, while experience presents a speaker, a location, and an intention. Interoception samples the body, while feeling presents an embodied self.

One productive hypothesis is that experience is a high-dimensional generative reconstruction from boundary-limited information:

This “phenomenal interior” is not a literal AdS spacetime inside the skull. It is an analogy that may eventually support a formal model. The point is architectural. A rich interior can be stabilized by repeated interaction across a narrow interface.

The environment matters because it corrects and completes the reconstruction. We move our eyes, touch the surface, ask another person, run the experiment, or return to the place. The world answers. Experience is neither a passive copy of external reality nor a private hallucination sealed in the brain. It is constrained by a relationship that has history.

This offers a useful division of explanatory labor:

• Occurrence: why there is a conscious event at all. Orch OR attempts an answer, though its mechanism remains unverified.
• Content: why the event concerns this face, memory, sound, threat, or possibility. Organism–environment history is central here.
• Unity: why color, body state, emotion, and memory appear as one moment. Multiscale coordination may matter more than any single microscopic site.

Objective reduction, if it occurs, cannot create an entire meaningful world from nothing. Meaning requires a history of embodied action, learned relevance, social alignment, and environmental response.

The missing middle scale

Fields and Levin’s use of bioelectric morphogenesis points toward a bridge between microscopic events and whole-organism behavior. Work in planaria has shown that transient changes to bioelectric patterning can alter the latent anatomical target expressed during later regeneration. A worm may appear ordinary before injury while carrying a distributed pattern that becomes visible when it regenerates two heads.

This is not evidence for quantum consciousness. It demonstrates a narrower and significant fact: organism-scale informational states can be stored in distributed bioelectric dynamics and later translated into coordinated cellular action.

Bioelectric fields need not themselves be conscious to matter. They may act as collective variables that coordinate many molecular and cellular processes. Adams’s fluids supply a structural analogy: relatively compact collective descriptions can organize vastly more complicated microscopic dynamics. Biology, however, does not inherit a gravitational dual merely by displaying collective behavior.

Where the synthesis breaks

Any useful framework needs visible seams. Four are decisive.

A generalized holographic screen is not gravitational holography. Fields and Levin describe the boundary through which separable systems exchange accessible information. Adams works within a gauge/gravity duality with stringent quantum-field-theoretic structure. Similar information architecture does not establish physical equivalence.

Boundary collapse is not objective reduction. Loss of separability between a system and its environment differs from eliminating alternative branches of a quantum state. The phrases may sound adjacent while naming nearly opposite operations.

Offloading does not establish consciousness. Memory, prediction, adaptive action, and environmental computation may be prerequisites for a point of view. They are not yet a sufficient criterion for phenomenality. A thermostat also relies on a boundary and feedback.

Holography and fundamental collapse may conflict. Controlled holographic theories are unitary. Objective reduction introduces nonunitary dynamics. A genuine merger would require either effective collapse through ordinary open-system physics, a suitable open-system holography, or a new reduction law that gives consistent predictions on both sides of a duality.

IntentiveNotion as a cognitive habitat

The most immediate consequence for our work is not a quantum architecture. It is a different unit of design.

IntentiveNotion began from a familiar problem: people do not merely need more stored information. They need a way to preserve provenance, revisit context, encounter relationships among ideas, test interpretations, and move between private reasoning and shared sensemaking. The intelligence of such a system does not reside in the language model alone. It emerges through the complete arrangement of people, sources, retrieval, notes, representations, interfaces, and consequences.

Fields and Levin give us a precise reason to treat the interface as more than presentation. Every interface is a cognitive boundary. What it makes visible determines what the coupled person–system can notice. What it remembers changes which trajectories remain available. What it hides can turn uncertainty into false confidence. What it asks from a person changes the person’s own model of the problem.

That suggests several commitments for IntentiveNotion:

1. Keep source claims, model inferences, human interpretations, and shared decisions distinguishable.
2. Make memory inspectable so offloading increases agency rather than quiet dependency.
3. Represent uncertainty and disagreement as structure, not as defects to smooth away.
4. Support movement across scales, from a passage to a concept, a debate, a project, and an evolving worldview.
5. Evaluate the whole human–AI–environment loop rather than scoring an isolated model.
6. Design every act of assistance as reciprocal: ask what the environment is learning from the person, and for whose purposes.

Holographic error correction offers another useful, still metaphorical design prompt. In holographic theories, a logical bulk observable can sometimes be reconstructed from different boundary regions.⁶ A resilient knowledge environment should likewise avoid making meaning depend on one brittle summary or one model’s latent state. Important ideas should remain recoverable through sources, linked concepts, discussion history, spatial views, and human memory. That is ordinary information design informed by a deeper physical precedent, not a claim that the application is performing quantum error correction.

Generation Alpha and theory of mind

Generation Alpha will develop theories of mind in environments containing responsive nonhuman agents from early childhood. Previous generations learned that books hold ideas and screens show distant people. Today’s children are also learning that a synthetic voice can answer, remember, imitate emotion, make mistakes, take actions, and appear to understand.

The central educational challenge is not deciding once and for all whether an AI “has a mind.” It is learning how to reason well about agents whose inner states are hidden.

That is boundary literacy. What can this agent observe? What does it remember? Which goals shape its behavior? What is generated internally, retrieved from elsewhere, supplied by another person, or performed by a tool? How does my interaction change its future behavior? Where does responsibility remain human?

A mature theory of mind for hybrid intelligence will need several models at once. Children should be able to recognize fluent social behavior without treating fluency as proof of feeling; attribute useful beliefs or goals to a system without assuming a human interior; notice when an interface encourages attachment or disclosure; and understand that apparently private assistance depends on institutions, energy, labor, data, and policy.

These are not only safety lessons. They are foundations for collaboration. A child who understands the boundaries of an artificial agent can use it with more imagination, verify it with more care, and retain more authorship over the shared result.

The environments we build will teach these lessons implicitly. A system that conceals provenance teaches that confident language is knowledge. A system that makes memory invisible teaches that assistance has no history. A system that always agrees teaches that a mind is a mirror. A better cognitive habitat lets young people inspect how understanding is assembled and see where their own judgment changes the outcome.

Designing cognitive habitats

A cognitive habitat is a physical, biological, digital, or social environment intentionally designed to participate constructively in thought. It might be an AI workspace that preserves a project’s memory without erasing disagreement; an XR environment that makes systems visible in space; a garden that carries ecological history into family learning; or a classroom where children, plants, models, artifacts, and rituals each hold part of the curriculum.

The design question changes from “How do we make the object smarter?” to “What kind of relationship does this system make possible?”

That relationship always has two directions. Offloading expands capability and creates dependency. An environment that computes for us may also recruit our attention, labor, language, and preferences for its own objectives. Social platforms made this reciprocal bargain difficult to see. AI assistants can make it more intimate.

Boundless Vision’s working principles follow:

1. Intelligence is relational.
2. Every interface is a cognitive boundary.
3. Every environment shapes what its inhabitants can think.
4. Offloading creates capability and dependency at the same time.
5. Good systems increase agency across more than one scale.
6. Hybrid intelligence should strengthen the living systems in which it participates.

What remains worth testing

The speculative edge can be made scientifically productive by replacing resemblance with explicit mappings. Can a constrained boundary code generate an emergent representational geometry that predicts perceptual structure better than conventional models? Can information be tracked causally from molecular and cytoskeletal dynamics through membrane voltage, bioelectric coordination, neural activity, and behavior? Can a candidate nonclassical variable predict a conscious-state transition after classical cellular effects are controlled?

On the theoretical side, the cleanest bridge is information geometry. Holography connects distinguishability between quantum states to gravitational quantities such as canonical energy.⁵ Penrose’s proposal also concerns the distinguishability of branches with different mass distributions. A serious program would ask whether Penrose’s gravitational self-energy can be derived from, related to, or ruled incompatible with the information-geometric distance between logical states in a holographic code. A toy model should come before a biological claim.

For the systems we can build now, the experiment is closer to home. Measure the intelligence, resilience, alignment, and human agency of the complete coupled loop. Compare a model acting alone with the same model embedded in transparent memory, source retrieval, tools, other agents, embodied context, and human review. Ask not only whether the answer improves, but whether the person’s understanding and capacity improve too.

Intelligence is a relationship

Holography asks how an interior can be encoded at a boundary. Cognitive offloading asks how systems and environments use boundaries to build consequential states together. Orch OR asks whether a fundamental physical transition turns an integrated possibility into an experienced event.

None completes the picture. The first is mathematically controlled only in special physical theories. The second explains broad features of cognition without supplying a criterion for consciousness. The third remains biologically and physically unverified.

Together, with those limits visible, they suggest a durable research direction. Mind may be neither sealed inside matter nor vaguely distributed through everything. It may be a structured relationship spanning scales: molecular processes, bioelectric and neural coordination, bodily regulation, artifacts, language, other people, and a world that remembers and responds.

The next science of intelligence may begin by looking more carefully at the boundaries through which mind and world become real to one another.