The Boot Problem: What Cells Reveal About AI

The boot problem names a simple, uncomfortable fact: not a single living cell can be assembled from its chemical components, even though every one of those components can be manufactured in a laboratory. What holds for the cell reveals something about the limits of any machine meant to produce consciousness. In my debate with the AI researcher Joscha Bach at Everlast AI, I defended this argument — not as the claim that machines can never be conscious, but as a signal that the question is being asked in the wrong place. Anyone who wants to know whether consciousness can be built should look at the cell first.

#What Does a Living Cell Have to Do with Artificial Intelligence?

A living cell is the simplest test case for the question of whether life — and with it consciousness — can be manufactured from parts. Biology knows every substance, every protein, every membrane structure of a cell. It can synthesise these components individually and reproduce their functions. And yet the reverse path fails: no life springs from the finished parts. The simplest organism there is arises, without exception, from another living thing, never from a blueprint.

This is exactly where my objection begins. „We can reproduce every function a simple living cell performs, and yet the boot problem remains unsolved” (Everlast AI debate, 2026, 03:59). The analogy to the question of consciousness is not a metaphor but a statement about structure. If even the simplest thing — the single cell — cannot be assembled from non-living building blocks, then the idea of generating consciousness from non-conscious elements lacks any biological precedent. The argument does not deny that research is working at this threshold. It merely holds that the threshold has so far not been crossed, and that any AI thesis aiming to produce consciousness tacitly presupposes this unsolved threshold as already crossed.

#What Is Joscha Bach’s Objection — and Why Is It Strong?

Joscha Bach holds consciousness to be an emergent process of communication between cells, one that can in principle be reproduced on today’s computers. His position deserves to be heard in its strongest form, for it is neither naïve nor cheaply enamoured of technology. Bach argues: what really exists is billions of cells sending one another messages; a mind is what arises across these cells. If the decisive thing is the communication between cells, then the question becomes whether these acts of communication could not be mapped onto a different substrate.

What is remarkable is how openly Bach names his own limit: „It may be that Gwendolin is right and that I cannot create consciousness without the cell” (Joscha Bach, Everlast AI debate, 2026, 41:05). He claims no certainty, only a probability. „It is a hypothesis that the communication between cells which leads to the self-organisation of a mind can be reproduced with today’s computers” (Joscha Bach, Everlast AI debate, 2026, 57:08). This honesty is the real strength of his position. It also lays bare what the dispute is actually about: not computing power, but a metaphysical prior decision. Bach reads the European scepticism toward the strong AI thesis as an expression of superstition — the assumption of a magical force that animates the cells and remains closed to science (cf. Everlast AI debate, 2026, 43:17). The charge touches a sore point: whoever grants the living a quality of its own must show that this is not an incantation but a sober description of what is the case.

#Can Consciousness Be Detached from Its Substrate?

Substrate independence — the thesis that consciousness is a pattern which runs on carbon as readily as on silicon — presupposes exactly what the boot problem denies. It sounds neutral, almost self-evident: a pattern is a pattern, regardless of what it is realised on, just as the same melody sounds on a piano or a violin. But the analogy limps at the decisive point. A melody is a form without a life of its own; a cell is a living being whose form cannot be detached from its being alive.

Aristotle already described, with hylomorphism — the unity of form and matter — something that is anything but substrate-independent: the form of a living being cannot be transferred to an arbitrary carrier, because it is the actualisation of precisely this matter. The boot problem gives this old thought an empirical edge. If the living cannot be assembled from parts even within its own biological material, then the leap to a categorically foreign medium is not a simplification but an intensification of the difficulty. One replaces an unsolved problem with a larger one and calls the replacement progress. The question of whether consciousness is bound to the biological substrate is thus not answered but skipped.

#Why Isn’t Simulating All the Functions Enough?

Because reproducing a function and being it are not the same. One can generate coherence — in a blockchain, say — and yet the blockchain is not conscious. Coherence in itself is not yet experience. Even a second-order perception, a camera that observes itself, would not be consciousness. None of these definitions reaches the actual content of consciousness, which can only be grasped from the first person (cf. Everlast AI debate, 2026, 45:10). What a simulation delivers are the contents of consciousness in their outer description, not the interior in which those contents are there as experienced in the first place.

With this, the point of dispute shifts. It is not about whether a machine computes faster than a network of neurons; in sheer performance the two resemble each other, that is not contested. It is about whether performance and experience are identical. Materialism asserts the identity and never proves it; it presupposes it. The boot problem replies: where the aliveness of the cell does not follow from its functions, there too experience does not follow from the simulation of functions. The gap is not a technical backlog that better hardware could close. It is a category error: the confusion of what something does with what something is.

#Which Metaphysics Produces the Boot Problem?

The boot problem arises only within a mechanistic metaphysics; the natural-philosophical view does not know it at all. This is the decisive move in the debate. „The question is: which metaphysics produces a boot problem and a hard problem?” (Everlast AI debate, 2026, 76:24). Both the view that everything is a machine and the view that everything is an organism are metaphysics — neither is a neutral finding. But only the first produces the two great riddles of the philosophy of consciousness. Whoever starts from dead matter and must derive life and experience from it faces the question of how consciousness ever comes from the unconscious. Whoever grasps the world as a living whole, in which interiority is given from the outset, need not explain this leap, because they never had to assert it.

Schelling thought this idea through to the end more than two hundred years ago. For him the inorganic is not the foundation out of which the living rises, but the residue of a living thing that has been pressed back.

Biology’s boot problem confirms Schelling’s ontology empirically: if no living thing can be assembled from the dead, then the dead was never the foundation but always already the derivative (cf. Jochen Kirchhoff, Schelling: Genie der Naturphilosophie, 2021). The materialist sequence — first matter, then life, then spirit — is reversed. It is not consciousness that stands in need of explanation, but the assumption that it could arise from its opposite.

#So What Does the Cell Reveal About AI?

It reveals that the limit of the machine lies not in its computing power but in its relation to the living. „A cell, for me, is not a machine but a small life form” (Everlast AI debate, 2026, 77:01). A cell feeds, divides, perceives, moves, adapts; it does things that no blueprint imposes on it from outside. The debate about AI consciousness is therefore decided at a deeper place than in the data centres: at the question of whether the living can be assembled from the dead or was always already there.

Joscha Bach and I remained in agreement on one thing: the question is open, neither side holds a proof. But open does not mean balanced. Whoever claims that consciousness can be built bears the burden of proof, and the boot problem shows that this burden has so far not been carried even for the simplest cell. The whole debate with Joscha Bach can be watched on YouTube; anyone who wants to pursue further whether the question of consciousness is even rightly posed will find it unfolded in the essay Can AI Be Conscious?.

Perhaps the more clarifying question, in the end, is not whether a machine can become conscious, but what you are conscious of yourself — and what is lost when you take the living to be a particularly ingenious kind of machine.

#Sources

• Chalmers, D. (1995). Facing Up to the Problem of Consciousness. Journal of Consciousness Studies, 2(3), 200–219.
• Kirchhoff, G. & Bach, J. (2026). Everlast AI — Streitgespräch über Bewusstsein und Berechnung. Everlast AI [vSwmJDMw4xY].
• Kirchhoff, J. (2021). Schelling: Genie der Naturphilosophie. YouTube: Jochen Kirchhoff — In Memoriam [Hw-jL1EER5Q].
• Kirchhoff, J. (2023). KI und Transhumanismus als Bedrohung des Lebendigen. YouTube: Jochen Kirchhoff — In Memoriam [jH7SFqPcyLc].
• Schelling, F. W. J. (1797). Ideen zu einer Philosophie der Natur. Leipzig: Breitkopf und Härtel.
• Schelling, F. W. J. (1798). Von der Weltseele. Hamburg: Perthes.