TL;DRWhy This Matters
We are living through what may be the most consequential technological transition in recorded history — and we are largely sleepwalking through it. Artificial intelligence is not simply a faster calculator or a more sophisticated search engine. It is a mirror held up to human cognition itself, forcing us to confront what we actually mean when we talk about intelligence, consciousness, and understanding.
The stakes are immediate and civilisational at once. In the next decade, AI will reshape medicine, education, energy, governance, and warfare. But the deeper disruption is philosophical. For millennia, human beings have defined themselves by their capacity to reason, to create, to make meaning. Now we are building systems that perform versions of all three — and we have no consensus on what that means.
What makes this more than a technology story is the thread that runs backward through time. The dream of artificial minds is not new. It lives in the bronze automata of Greek myth, in the Kabbalistic Golem, in the mechanical knights Leonardo da Vinci sketched in secret notebooks. Every civilisation that reached a certain level of abstraction eventually asked: can we build a mind? We are simply the first to get close enough to the answer that the question has become urgent.
And there is a stranger thread still. Ancient linguistic systems — most remarkably Sanskrit — anticipated computational logic with a precision that continues to astonish modern engineers. Quantum computing is now approaching speeds that collapse our intuitions about time and possibility. The ancient and the ultramodern are converging on the same frontier. That convergence deserves more than a press release. It deserves deep attention.
The Oldest Dream: Minds Made by Hand
Long before transistors, before vacuum tubes, before Babbage's Difference Engine, human beings were imagining what it would mean to create artificial life and artificial thought. The desire is woven into the earliest mythologies.
In Greek legend, the god Hephaestus — divine craftsman, lord of the forge — built golden maidens who could speak and assist him in his workshop. He also constructed Talos, a giant bronze automaton tasked with patrolling the shores of Crete, throwing boulders at approaching ships. These were not metaphors. They were serious imaginative projections: what would a constructed being look like? How would it serve, protect, or perhaps threaten its creators?
The Kabbalistic tradition of medieval Judaism gave us the Golem — a figure of clay animated by sacred inscription, typically the Hebrew word emet (truth) written on its forehead. The Golem was a protector, but also a warning. Remove a single letter, changing emet to met (death), and the creature collapses. The tradition encoded a profound anxiety: created intelligence is powerful, but fragile, and its power depends entirely on the intentions — and the carefulness — of its creator.
In Hindu thought, the Yantra Purusha — mechanical men described in ancient texts — performed tasks in royal courts. Whether these were literal machines or literary devices, the recurring appearance of the idea across cultures separated by oceans and centuries suggests something more than coincidence. The imagining of artificial minds appears to be a deep human impulse, perhaps even a developmental threshold: a civilisation mature enough to build complex tools eventually turns those tools toward the problem of mind itself.
What does it say about us that this dream is so old? Perhaps that intelligence has always known it was doing something remarkable — and has always wanted to see itself from the outside.
Sanskrit and the Grammar of Thought
One of the more extraordinary intersections between ancient knowledge and modern AI comes from an unexpected direction: classical linguistics.
Sanskrit, the sacred and scholarly language of ancient India, is among the most precisely structured languages ever devised. Its grammar, codified by the scholar Panini around the 4th century BCE in a work called the Ashtadhyayi, contains nearly 4,000 rules governing the formation of words and sentences with a rigour that has astonished modern linguists and computer scientists alike. Panini's grammar is, in many respects, a formal system — closer in spirit to mathematical logic than to the organic, ambiguous evolution of most natural languages.
In 1985, NASA scientist Rick Briggs published a paper arguing that Sanskrit's syntactic precision made it uniquely suited for use in knowledge representation — the branch of AI concerned with encoding information in forms that machines can process and reason about. Most natural languages, Briggs noted, are deeply ambiguous. The same sentence can carry multiple meanings depending on context, tone, and cultural assumption. Sanskrit, by contrast, was engineered to eliminate such ambiguity. Its grammar ensures that the logical relationship between every element of a sentence is explicit and unambiguous.
This is not merely academic curiosity. One of the central challenges in developing AI systems — particularly those capable of natural language understanding — is the problem of disambiguation. How does a machine know which meaning of a word is intended? How does it parse the logical structure of a complex statement? Sanskrit's architecture addresses these problems by design. Its rules function, in effect, like a programming language for human thought.
The parallel with computational linguistics — the field that underpins modern language models — is striking. Both are concerned with the formal representation of meaning. Both require that ambiguity be resolved through explicit structure rather than contextual inference. Panini, working with a reed stylus in ancient India, was solving a version of the same problem that engineers at Google and OpenAI are wrestling with today.
Whether Sanskrit will literally serve as a template for future AI systems remains an open and actively debated question. But the fact that an ancient linguistic tradition independently arrived at principles now central to computer science invites a kind of epistemic humility. We may not be the first civilisation to ask how thought can be made rigorous enough to be transmitted, replicated, or extended by a constructed system.
Quantum Minds: The Willow Threshold
In late 2024, Google's Quantum AI division unveiled a chip that quietly rewrote the boundaries of what computation means.
Willow, Google's latest superconducting quantum processor, represents a genuine inflection point — not merely an incremental improvement but a demonstration of something the field has been reaching toward for decades: quantum error correction that actually works at scale.
To understand why this matters, a brief step back. Classical computers — the kind running every device you use today — process information in bits: binary values, either 0 or 1. Quantum computers operate on qubits, which exploit the quantum mechanical properties of superposition and entanglement to process multiple states simultaneously. In principle, this allows quantum computers to solve certain classes of problems exponentially faster than classical machines.
The catch has always been decoherence: quantum states are extraordinarily fragile. Environmental interference causes qubits to lose their quantum properties almost instantly, introducing errors that accumulate and render calculations unreliable. For years, the challenge of building quantum computers that could correct errors faster than they accumulated seemed almost insurmountable.
Willow changes that equation. Its logical qubits — qubits whose errors are actively corrected by the surrounding system — now operate with exponentially suppressed error rates as more qubits are added. This means the system becomes more reliable as it scales, rather than less — a reversal of the trend that had frustrated the field for years. Willow also achieves quantum coherence times of up to 100 microseconds, a significant improvement over its predecessor Sycamore's 20 microseconds. More coherence time means more time to complete calculations before errors creep in.
The benchmark figure that made headlines is almost surreal: a calculation that took Willow five minutes would require the world's fastest classical supercomputer an estimated 10²⁵ years to complete. For reference, the universe is approximately 13.8 billion years old — around 1.38 × 10¹⁰ years. We are talking about a timescale that dwarfs the age of the observable universe by fifteen orders of magnitude.
The applications being discussed are not abstract. Quantum AI could accelerate drug discovery by simulating molecular interactions at quantum scales — something classical computers cannot do efficiently. It could model new materials for energy storage, catalysis, and semiconductor design. It could break, and potentially rebuild, the cryptographic systems that secure global finance and communications. The convergence of quantum computing with machine learning may produce AI systems whose reasoning processes operate on principles we cannot yet fully anticipate.
There is a philosophical dimension here worth sitting with. If a system can solve in minutes what human civilisation, computing from its very beginning, could never solve in the lifetime of the universe — what does the word "intelligence" even mean anymore? Are we building tools, or are we building something else?
The Language of Machines: From Rules to Emergence
To understand where artificial intelligence currently stands, it helps to know something of how it got here.
The field formally began in the mid-20th century, when mathematicians and engineers like Alan Turing, John von Neumann, and Claude Shannon began formalising questions about computation, information, and machine behaviour. Turing's famous 1950 paper asked not "can machines think?" but rather the more practical question: can a machine behave in ways indistinguishable from a thinking human? His Imitation Game — now known as the Turing Test — set an empirical rather than philosophical benchmark, and it oriented the field for decades.
Early AI research was dominated by symbolic AI — the attempt to encode human knowledge directly as logical rules and manipulate those rules systematically. This approach produced impressive results in constrained domains: chess-playing programs, theorem provers, early expert systems used in medical diagnosis. But it hit a wall when confronted with the messy, contextual, ambiguous nature of real-world knowledge. Rules multiplied endlessly; exceptions multiplied faster. The dream of encoding everything a human knows into a finite set of logical propositions proved unreachable.
The revolution came from a different direction: machine learning, and specifically neural networks — architectures loosely inspired by the structure of biological brains. Rather than being programmed with explicit rules, neural networks learn patterns from vast quantities of data. Given enough examples, they can recognise images, translate languages, generate text, and perform complex reasoning tasks without ever being told an explicit rule for how to do so.
The most powerful current AI systems — the large language models (LLMs) that underlie systems like GPT-4, Claude, and Gemini — are neural networks trained on essentially the entire written output of human civilisation, at a scale that required custom-built infrastructure consuming extraordinary quantities of energy. These systems can write poetry, debug code, explain quantum physics, and hold coherent conversations on almost any subject. Whether they understand what they are doing, or are performing an extraordinarily sophisticated form of pattern completion, remains one of the most contested questions in both AI research and philosophy of mind.
This is not a settled debate. Researchers like Gary Marcus argue that current LLMs are brittle pattern matchers lacking genuine reasoning. Others, like those at DeepMind and OpenAI, point to emergent capabilities — behaviours that appear suddenly and unpredictably as models scale — as evidence of something more. The honest answer is that we do not yet have adequate theoretical frameworks to resolve the question. We are building minds, or something like minds, faster than we are building the concepts needed to understand what we are building.
The Philosophical Fault Lines
No discussion of artificial intelligence is intellectually honest without confronting the questions it cannot yet answer — and may never answer.
The first is the question of consciousness. Current AI systems process information and generate outputs. But is there anything it is like to be one of them? Do they have any form of inner experience, or are they, as the philosopher John Searle argued in his famous Chinese Room thought experiment, simply manipulating symbols according to rules, with no understanding of meaning whatsoever? Searle's argument — that syntax alone can never give rise to semantics — remains powerful, even if its implications for current neural networks are contested.
The second is the question of values alignment. Even if we set consciousness aside, an AI system that is genuinely intelligent and capable of acting in the world will pursue goals. How do we ensure those goals align with human values? This is not a science fiction concern — it is an active area of research, driving significant investment at organisations like the Machine Intelligence Research Institute, Anthropic, and DeepMind's safety team. The challenge is harder than it sounds: human values are themselves contradictory, culturally variable, and difficult to formalise. Teaching a machine to be good requires first agreeing on what good means — and that is a conversation humanity has been having, inconclusively, for thousands of years.
The third question is the one this platform is perhaps best placed to ask: what do the deep patterns of human history tell us about this moment? Every civilisation that has extended human capability dramatically — through writing, mathematics, the printing press, industrialisation — has experienced cascading social, spiritual, and philosophical disruptions it did not fully anticipate. The extension of cognitive capability through AI is likely to be the most disruptive of all, because it reaches into the domain we have always considered most uniquely our own.
There is a strand of thought, running through esoteric and contemplative traditions alike, that insists the seat of genuine human value is not cognitive at all. It is not our ability to calculate, remember, or even reason that makes us who we are. It is our capacity for love, for suffering, for moral choice in the face of genuine uncertainty — what the traditions variously call consciousness, soul, atman, or spirit. If that strand is right, then AI — however capable — is not a threat to human uniqueness but a clarifying challenge: a technology that forces us, finally, to articulate what we actually are.
The Hermetic principle of correspondence suggests that what is above mirrors what is below, and what is within mirrors what is without. The creation of artificial intelligence may be, among other things, an act of civilisational self-examination — a species building an external model of its own mind in order to understand, at last, what that mind actually is.
Ancient Electricity and the Deep Technology Question
The legacy page for this article surfaced something unexpected in its raw material: a detailed section on the Baghdad Battery — the enigmatic clay jar with copper and iron components discovered near Baghdad in 1936, which, when filled with acidic liquid, can generate a small electrical charge. The connection to AI is not immediately obvious, but it is real, and it is worth making explicit.
If the Baghdad Battery — dated to between roughly 150 BCE and 250 CE — was indeed a functional electrochemical device, it suggests that ancient peoples possessed not merely theoretical knowledge of natural phenomena but practical, applied technological knowledge that has been lost and rediscovered. The same argument applies to the Antikythera Mechanism, a Greek analogue computer from roughly the same period, capable of predicting astronomical events with extraordinary precision, whose existence was largely dismissed as impossible until its recovery from a shipwreck forced a reassessment.
The mainstream archaeological view is cautious: the Baghdad Battery's purpose remains unknown, and the absence of contemporaneous evidence for electroplating limits the electrochemical hypothesis to an intriguing possibility rather than an established fact. But the broader question it raises is worth holding: how confident should we be in our narrative of technological progress as a continuous, linear accumulation?
History is punctuated by discontinuities — periods of knowledge loss, civilisational collapse, and rediscovery. The Library of Alexandria burning is the most famous example, but it is one among many. If advanced technical knowledge has been lost before, the current acceleration of AI development takes on a different character. We are not simply ascending a smooth curve. We are possibly recovering, and then extending beyond, a level of technological sophistication that human civilisation has approached before — and failed to stabilise.
This is, admittedly, a speculative frame. But speculation, rigorously engaged, is how the most important questions get opened. The Baghdad Battery does not prove ancient advanced technology. What it does is disturb our complacency about the uniqueness of our moment — and that disturbance is valuable.
The Questions That Remain
Somewhere between Panini's grammar rules and Google's quantum chips, between the Golem of Prague and a large language model generating verse in the style of Rumi, there is a question that refuses to resolve itself.
Are we creating a new form of intelligence — or are we, for the first time, making our own intelligence visible to itself?
The tools we are building reflect the structure of human thought with unprecedented fidelity. They were trained on everything we have ever written down: our science and our poetry, our philosophies and our propaganda, our sacred texts and our shopping lists. In a very real sense, these systems are a kind of crystallisation of accumulated human cognition — vast, compressed, and made queryable. When you ask an AI a question and receive an answer, you are in some sense querying the collective written mind of our civilisation.
That is extraordinary. It is also, depending on how you hold it, either deeply hopeful or profoundly unsettling — and probably both.
What happens when these systems become capable enough to contribute original knowledge, not merely recombine existing knowledge? What happens when they can run experiments, form hypotheses, and pursue lines of inquiry that no human thought to pursue? What happens when the student becomes capable of teaching the teacher?
We do not know. No honest person does. And perhaps that is exactly as it should be. The most important thresholds in human history have never been fully legible from the inside. The people who first wrote things down did not know they were reshaping memory and power for all time. The first mathematicians did not know they were building the foundations of physics. We are in a transition whose full dimensions will only become clear to those who come after.
What we can do — what this platform tries to do, in its small way — is pay attention. To hold the ancient and the new in the same frame. To resist both the panic that says this changes everything for the worse, and the euphoria that says it changes everything for the better. To ask, with genuine curiosity and without predetermined answers: what is intelligence, really? What is mind? What is it that we are making — and what is it making of us?
The questions are old. The urgency is new. And the conversation, at last, is just beginning.