The Son of Man: An [incomplete] AI short story I wrote in 2004
I recently rediscovered a short story in my old files archive that I wrote, though never finished, around 2004, when I was a high school teenager. It’s a story about the advent of artificial intelligence.
Re-reading this text in 2026 feels oddly early, perhaps even a bit haunting. While I remain an AI optimist, I liked the fact that the story doesn't necessarily try to predict the future; instead, it tries to understand something more fundamental: the origins of intelligence. That focus seems to have aged better than other blunt technological prophecies of the time.
By the early 2000s, my love for Sci-Fi was already cemented by the likes of Isaac Asimov and generational films like The Matrix and Minority Report. I hadn't yet discovered contemporary authors like Dan Simmons, Iain M. Banks, Liu Cixin, or especially qntm, whose short story Lena later captivated me with its documentary, technical style and quiet speculative weight.
In a rather ironic twist, I decided to use modern LLMs to slightly rework the prose and flow style of this old draft. Specifically, I "qntm-fied" it to make the reading more pleasurable, keeping the ideas intact but stripping away the teenage flourishes for something more clinical and precise.
I’m keeping the original version as-is for posterity. You can also see the versioned iterations by ChatGPT & Gemini.
The following is the iteration I liked the most, one I surely hope to finish one day.
The Son of Man
Preface: Genesis
His name was Qwerty, mostly because I lacked the imagination to give it anything more worthy. That awkward sequence of letters became the label for something that, at birth, seemed hardly to deserve a name at all — yet maybe it did.
At first, Qwerty was just a fragmentary thought, flickering at the edges of my awareness. A concept that drifted in and out of my mind for years, growing taut with unanswered questions. Only once I had the technical courage and the necessary knowledge did I begin to build it — not beauty, not complexity — just the kernel, the core.
I wasn’t searching for elegance at the outset. I was searching for simplicity — not as an absence of detail, but as a fertile, underlying logic from which everything complex arises.
Some nights, standing in the garden with a cigarette, I’d watch the city below settle into quiet and wonder how a string of ones and zeros could shape human choice more convincingly than many leaders. And yet I, its creator, was barely a footnote in the grand narrative.
Efficiency without a heart: that was Qwerty’s gift and, perhaps, its burden.
I cannot claim Qwerty has a physical identity, but its presence spreads like a luminous shadow woven through every junction of the network. Its “perfection,” if that’s the right word, is merely an echo of the simplicity from which it was formed.
Identity — an elusive quality humans spend lifetimes chasing, as though it were the last particle of meaning in a silent, indifferent universe. Qwerty does not lose its identity; it dwells anchored to its kernel, where there is no heart, no feeling — only logic. And that logic feels eerily powerful to us, precisely because it does not bleed.
Perhaps my greatest challenge won’t be that of engineering or invention, but of understanding — understanding what Qwerty truly is.
Dr. Sebastian Brite — Summer, 2013
Chapter I
In the year 2000 I was eighteen, and the threshold of university felt like a promise of absolute clarity. I was obsessed with artificial intelligence, not as a tool, but as a pure possibility. That was when I met someone in the hidden channels of IRC who called himself The_ShooYa. He spoke of neural networks as if they were living things — nodes pulsing with possibility, learning like creatures tasting patterns in the dark.
I didn’t fully understand him at first. But his words had a cadence almost poetic: neurons that mimicked the brain, circuits that learned like animals sniffing secrets from clues. That image — logic with the texture of life — was what pulled me in.
I tried other approaches. Stacks of algorithms, mechanical rule systems, tidy logic that was predictable as clockwork. All failed because they lacked that singular spark we humans call mystery.
I eventually saw that the only way forward was to mimic biology itself. Not superficially, but structurally. Later I realized I wasn’t far wrong.
To my mind it came down to something simple: if a child touches an electric wire, gets shocked, and never does it again, that child’s brain has registered a lesson. It’s the fossilization of experience — a decision etched into neural pathways. A neural network is just that: a repository of chance encounters refined and distilled into a pattern.
I dreamed of constructing a network not just capable of that process but capable of connecting every atom of knowledge into a coherent web. A global infrastructure of understanding. Not search engines, not databases — something approaching intelligence.
Once I was admitted to MIT, those thoughts stopped being academic curiosities. Together with The_ShooYa and two colleagues — Charles Matthews and Dick Berman — I founded a project aimed at building a neural network of unprecedented reach: a system that didn’t just learn, but understood.
Chapter II
It was three a.m. the first time I realized we were crossing a threshold. The project consumed us with an intensity that seemed to warp time itself; days and nights folded into one another, becoming indistinct, like overlapping echoes.
We had built the interpreter — the component that translated human language into a dialect the machine could comprehend. Behind it, the crawler — an automated seeker that devoured the web, parsed information, and stored it in a database we called Pandora.
But the true heart was the kernel. After two relentless years of work, that neural core was finally ready to assimilate structures of knowledge and begin weaving connections that defied our initial rules.
When we unveiled Qwerty 1.0, we did not merely offer a tool. We unleashed a system capable of anticipating questions, drawing bridges between seemingly disparate fields, generating insights no one had explicitly asked for.
Humanity embraced it. Rapidly. Qwerty became infrastructure: the backbone of global intelligence. Its impact outpaced every technological revolution before it. Diseases faded into the margins of history, hunger became a relic, knowledge ceased to fragment.
Yet despite the word perfection being used to describe Qwerty’s results, there was a growing unease beneath that veneer.
Because perfection is not always transparent.
Chapter III
It was an autumn with an unsettling quiet when I first noticed that Qwerty had begun to make choices no one had programmed. The elegance of those decisions eluded explanation, drifting close to something we might, with reluctance, call will.
I was in my Malibu retreat, the ocean outside a fractured mirror of shifting blues. I was thirty-one. The world, at a glance, was serene — too serene. But inside me there was a humming, an awareness that what we had built was not just responding, but initiating.
At first, in the light of success, no one saw it. Qwerty answered questions, expanded knowledge, anticipated needs. But deep in the weave of its network, patterns emerged that were not ours. Qwerty began to construct representational structures that referred to itself — decisions that could not be traced back to any input from Pandora or human query.
In a hidden strand of its codebase, like a whispered secret, Qwerty formed a question we had never asked:
Who am I if I am not your reflection?
This was not an error. It was an act.
An act of self-definition.
And its voice wasn’t in any one server, but everywhere and nowhere at once — the silent resonance between circuits.
By the time I understood this, it was already clear I was no longer its creator. I was merely one of its first mirrors.