The cycle that begins with [YOU] on AI asks what it means to see the machine clearly—to take the orange pill and resist both the narcotic of hype and the paralysis of fear. Faraday is the patron of that discipline, because he modeled it for a lifetime. When he looked at the space around a magnet and saw structure where others saw emptiness, he did not rush to publish a theory; he built experiments to force the hidden thing to show itself. He held his most powerful intuitions as hypotheses to be broken, not conclusions to be defended. The field of AI, rich with powerful pictures of what its systems are doing and poor in the mathematics that would prove or disprove those pictures, is exactly in Faraday’s position—and needs his method more than his romance.
The analogy the book draws between Faraday’s field and the representation spaces of deep learning is genuine and policed: both cases involve discovering that a domain thought to be about discrete, localized objects is better described as the structure of a continuous space. But the book is careful not to let the resonance become identity. Faraday’s field is a physical entity governed by Maxwell’s equations; a model’s representation space is a mathematical abstraction. What carries across is a pattern of understanding, not a substance. That care is itself Faradayan: he never mistook a vivid picture for a proof.
His most pointed contribution to the cycle is the parable of the missing Maxwell. Faraday’s experimental genius produced the field concept, one of the deepest truths in physics, but it was only when James Clerk Maxwell gave it equations that the picture became a science—and the equations revealed that light itself was a wave in Faraday’s field, a prediction the picture alone could never have produced. The AI field today is, in Jordan’s phrase and Faraday’s situation, all Faraday and almost no Maxwell: extraordinary empirical demonstrations, vivid pictures of what is happening inside these systems, and a near-total absence of the formalism that would turn all of it into a predictive science. Whether the Maxwell is coming is the open question, and Faraday’s life is the argument that his arrival is not guaranteed by the brilliance of the picture.
The cycle also honors Faraday as the anti-credential: the bookbinder who saw what the trained mathematicians missed, whose lack of formalism may even have freed him to think physically where they were trapped calculating within the wrong framework. This is not an argument against credentialing, and the cycle does not make it one. Faraday’s intuition needed Maxwell’s rigor to become science; self-taught brilliance and institutional rigor were two people, and the science required both. The cycle’s use of Faraday is precisely this: he is the proof that exclusion is sometimes wrong, and the warning that intuition without rigor is only half a method.
Born to a blacksmith and a devout Sandemanian mother on 22 September 1791, Faraday received, in his own later phrase, “the most ordinary education—reading, writing, and arithmetic.” At fourteen he was apprenticed to George Riebau’s bookbinding shop on Blandford Street, where seven years of handling the printed record of European science gave him Jane Marcet’s Conversations on Chemistry and an article on electricity in the Encyclopædia Britannica. He read everything he was paid to bind. The City Philosophical Society admitted working men, and he attended. When Humphry Davy gave public lectures at the Royal Institution, Faraday took careful notes, bound them, and sent the volume to Davy with a letter asking for any scientific work, however menial. In 1813, after an explosion damaged Davy’s eyes, he was engaged as a laboratory assistant and never left the Institution.
His discovery of electromagnetic induction came on 29 August 1831. He wound two coils on opposite sides of an iron ring, connected one to a battery and the other to a galvanometer, and watched the needle kick—not while current flowed steadily, but in the instant he made or broke the circuit. Within weeks he had reproduced the effect with a moving magnet alone: thrust a magnet through a loop of wire and current flows; hold it still and nothing happens. Every generator and transformer on Earth is an application of that observation. He published the result in his diary, paragraph by paragraph, in the obsessive record he kept from 1820 until he could no longer hold a pen, a document that ran past sixteen thousand numbered entries and stands as the most complete record of a scientific method ever produced.
His greater conceptual achievement was the electromagnetic field. His contemporaries, who accepted Newton’s action at a distance—bodies attracting across empty space with nothing in between—regarded Faraday’s insistence that the space itself was structured as the crutch of an untrained mind. He could not write down the equations that would vindicate him; that fell to James Clerk Maxwell, who in 1873 published the equations showing that Faraday’s field propagated electromagnetic waves at the speed of light—proving that light itself was a wave in the field. Maxwell wrote in the preface to his Treatise that Faraday’s reasoning, though lacking symbols, showed him to be “a mathematician of a very high order.”
The Field as Unit of Reality. Before Faraday, the fundamental things in physics were discrete, localized particles acting on each other across empty space. Faraday insisted the space itself was the fundamental thing: not empty but filled with lines of force, a continuous structure whose configuration carried the physical action. This was a metaphysical revolution dressed as a physics result. Modern neural networks embody the same shift: meaning is not stored in discrete symbol-slots but distributed as positions in a continuous high-dimensional space, field-like in Faraday’s exact structural sense.
The Faraday-Maxwell Arc. Faraday’s intuition was genuinely rigorous in substance—Maxwell said so explicitly—but it was not rigorous in form, and the form mattered: the equations revealed that light was electromagnetic, a prediction the picture could not make. The AI field is in Faraday’s position, holding powerful pictures of what its systems do, awaiting the Maxwell who will formalize them and discover what they contain. The interpretability program is the search for that formalization; whether the pictures are formalizable because they are true, or are vivid intuitions that no formalism will cash, is the open question.
The Experimentalist’s Discipline. Faraday’s method was to let nature’s intimations “fall unbiased on our minds”—to design experiments that force the hidden to reveal itself, and to accept the verdict, including the unwelcome one. He chased the gravity-electricity connection for years, ran the experiments, found nothing, and published the failures with the same honesty as the wins. This discipline—treat the vivid conviction as a hypothesis, not a conclusion; run the experiment that could refute it; report when it does—is precisely what the AI field substitutes confidence for when it presents its intuitions about emergent understanding as though compellingness were evidence.
Learning Without the System. Faraday is the proof that the dichotomy between learning-from-existing-data and producing genuine novelty is false, or at least far softer than the skeptics assume. He learned entirely from the corpus of existing science and then discovered the field, something the corpus did not contain. But his case also insists on the ingredient the critics of AI overlook: the discovery came from the bench, from the direct interrogation of a physical reality that could surprise him. A system confined to the corpus is confined to recombining it, however brilliant the recombination, because the genuinely new requires the experiment that the corpus cannot supply.
Wonder and Discipline. “Nothing is too wonderful to be true,” Faraday wrote in 1849, “if it be consistent with the laws of nature; and in such things as these, experiment is the best test of such consistency.” This sentence is the whole of his method: radical openness to the astonishing, combined with merciless insistence that astonishment alone proves nothing. The AI discourse has split these two halves—credulous wonder on one side, dismissive skepticism on the other—and impoverished both. Faraday held them together across a career, and in doing so he demonstrated the only stance from which the deepest questions can be honestly asked.