Link to the code: brain-emulation GitHub repository

Michael Pollan's 'A World Appears': What a Science Writer Sees in the Consciousness Problem

Michael Pollan did not approach consciousness from neuroscience. He came from food writing. His books — The Omnivore’s Dilemma, The Botany of Desire, In Defense of Food — found enormous audiences by taking complex systems (agriculture, ecology, plant biology) and making them legible through personal narrative. His 2018 book How to Change Your Mind brought psychedelic research to mainstream readers who had never heard of MAPS or read a psilocybin trial protocol. It worked because Pollan is not a scientist writing for scientists. He is a curious layperson writing for curious laypersons.

A World Appears: A Journey into Consciousness, published by Penguin Press in February 2026, uses the same approach on a harder target: consciousness itself, and specifically the question of whether artificial systems can have it.

The book will introduce millions of readers to debates that the mind uploading and whole brain emulation community has been conducting for decades. For that reason alone, it is worth understanding what Pollan gets right, where he simplifies, and how his framework maps to the current state of actual research.

What the Book Covers

Pollan’s investigation is organized around personal encounters: conversations with neuroscientists, philosophers of mind, AI researchers, and a handful of patients whose consciousness has been altered by injury, disease, or pharmacology. He visits labs, undergoes his own consciousness experiments (EEG monitoring, sensory deprivation, TMS), and documents his reactions with the same first-person transparency that characterized his psychedelic reporting.

The book’s central question is: what is consciousness, and could something other than a biological brain have it?

Pollan traces the hard problem of consciousness — why any physical process should give rise to subjective experience — through the standard contemporary frameworks. He interviews proponents of Integrated Information Theory, Global Workspace Theory, and predictive processing, and he finds each compelling and each incomplete. He is good on the phenomenology: what it feels like to be a conscious creature, the strangeness of that fact, the reasons why science has struggled to give an account of it.

On AI consciousness, Pollan is genuinely uncertain — which is intellectually honest given the state of the field. He engages seriously with the claim that large language models might be having something like experience, and with the arguments that they are not. He visits researchers working on AI welfare. He documents his own discomfort with the question.

Where Pollan Is Helpful for Mainstream Understanding

The book’s greatest contribution is its framing of the stakes. Pollan makes explicit, for a general audience, why the question of AI consciousness matters:

If an artificial system can be conscious, then consciousness is substrate-independent. And if consciousness is substrate-independent, then mind uploading is not science fiction — it is a technical problem. The persistence of your mind after biological death becomes a question of whether we can build the right kind of artificial system, and whether your connectome can run in it.

This framing is not original to Pollan. It is standard in the WBE literature, articulated clearly in Sandberg and Bostrom’s 2008 roadmap. But Pollan reaches audiences who have never heard of Sandberg or Bostrom, who have never read a paper on neuromorphic computing or connectomics, and who have never encountered the philosophical literature on personal identity.

The book also provides a useful popular synthesis of the current empirical state. It describes the 256-subject adversarial collaboration between IIT and GWT researchers, the development of consciousness monitoring tools, and the clinical application of consciousness science to patients in vegetative states. These are presented as news — which for many readers they will be.

Where Pollan Simplifies

Science writing for general audiences requires simplification. Some of Pollan’s simplifications are productive; others collapse important distinctions.

On substrate independence. Pollan presents the substrate independence debate as a binary: either consciousness can run on any sufficiently organized information-processing system, or it can’t. The biological computationalism framework articulated in 2026 research occupies a middle position that his framing doesn’t accommodate: perhaps consciousness requires specific kinds of physical computation, without requiring biological neurons specifically. Pollan’s binary makes it hard to evaluate the neuromorphic computing direction or the gradual replacement approaches.

On identity continuity. When Pollan discusses mind uploading, he frames the copy problem (is an uploaded mind really you?) as a philosophical puzzle without resolution. He is right that it’s unresolved, but the literature has made significant progress on the specific conditions under which psychological continuity is preserved or broken. Parfit’s analysis, the Moravec gradual replacement proposal, and the 4E cognition challenge each address different aspects of this question. Pollan’s readers will come away thinking it’s purely a matter of intuition when it’s partly a matter of argument.

On current AI systems. Pollan is appropriately skeptical about attributing consciousness to current large language models, but he engages with the question in terms of behavioral criteria: do they respond like conscious beings? The WBE field is mostly interested in a different question: what structural and functional properties would an artificial system need to have for consciousness to be likely, regardless of behavior? That question requires a theory of consciousness, not just behavioral observation — and Pollan doesn’t have a settled theory.

The Value for This Blog’s Readership

A World Appears will generate search traffic. “Michael Pollan consciousness book,” “AI consciousness 2026,” “can AI be conscious Pollan” — these are the queries it will produce from readers who encountered the book in mainstream media, NPR coverage, or bookstore displays.

For readers who come to this blog from that direction, the Pollan book is a useful starting point. The research covered here goes substantially deeper on every topic he touches. The Allen Institute adversarial collaboration is real empirical science at the scale Pollan describes. The biological computationalism framework addresses exactly the substrate-independence question he raises. The neuromorphic twins technology represents the engineering approach to digital consciousness he describes as aspirational.

Pollan’s book is not a research contribution. It is a map for people who don’t yet know the territory. The territory itself is more complex, more contested, and more interesting than the map suggests.

On the Question He Asks Most Honestly

The strongest section of A World Appears is its final chapters, in which Pollan abandons the synthetic third-person perspective and writes directly about his own experience of consciousness — what it is like to wake up, to notice that noticing is happening, to be unable to find the boundary between the observer and the observation.

This phenomenological mode is not where consciousness science operates, but it is where the hard problem lives. The gap between the scientific description of neural dynamics and the first-person reality of experience is not a gap that any current theory has closed. Pollan names this gap honestly and refuses to fill it with false confidence.

For readers who encounter whole brain emulation through Pollan’s book, the honest message to carry forward is: the science is real, the technical progress is accelerating, and the deep questions remain open. Not because we lack ambition, but because consciousness is genuinely difficult.

Official Sources