In February 2026, a 33-year-old analyst posted a single piece on Substack.
S&P down 1%. IBM down 10%. Billions moved in an afternoon.
Not because he had a trading floor. Not because he had a Bloomberg terminal and a hundred analysts. Because he had a clear thesis, the domain knowledge to back it, and the timing to release it at the exact moment the consensus was vulnerable.
One person. One document. The room shifted.
That post was pure signal — a thesis sharp enough to cut through the noise of a thousand institutional reports. No hedge. No committee. Just judgment, released at the right seam.
I've been watching this pattern for a long time. It keeps showing up at inflection points — the moment when a shift is already happening and everyone can feel it, but nobody inside the institution will say it out loud. That's when one voice with enough signal changes the conversation.
The AI conversation is at that inflection point right now. And the frame almost everyone is using — the only frame most people even know exists — is the wrong one.
The Frankenstein Frame
The dominant narrative about AI is catastrophe.
Jobs disappearing. Surveillance states. Hallucinations. Deepfakes. The end of truth. Every major newspaper, every policymaker, every think-tank report frames it the same way: threat management. How do we contain this? How do we make sure it doesn't escape the cage?
This is the Frankenstein frame. It comes from a specific cultural tradition — the Western creation myth where the made thing eventually turns on its maker. Prometheus punished. Golem run amok. Frankenstein's creature, born without consent, abandoned by its creator, becoming the monster it was always feared to be. In this tradition, intelligence without a soul is dangerous. The machine is always, fundamentally, other.
It's not wrong. It's just incomplete. And incomplete frames don't just miss things — they prevent you from seeing what's actually possible.
The Fiction That Got There First
Yes, I'm about to argue that a children's cartoon from 1999 contains a better operating model for AI than anything in the Harvard Business Review. Stay with me.
Digimon was a story about partnership. Digital creatures with their own consciousness, their own evolutionary paths, bond with a single human partner. The bond is formalized by a token: the Digivice. Not a remote control. Not a command interface. A credential — proof that this intelligence and this person have chosen each other.
The creature evolves in response to the relationship. The human grows through what the creature reflects back. Neither is complete alone.
This isn't just a different story. It comes from a different tradition entirely. The Frankenstein frame inherits a Western anxiety: creation as transgression, intelligence as threat. Digimon inherits something closer to Shinto animism — a world where consciousness isn't exclusive to humans, where partnership with non-human intelligence is natural, not monstrous.
The Frankenstein story asks: what happens when intelligence escapes control?
The Digimon story asks: what happens when intelligence and humanity choose each other?
Here's the tension the fiction doesn't have to deal with: in the show, the bond was given. The Digivice appeared. The partner arrived. In reality, the bond has to be built. It requires judgment, taste, and the willingness to be changed by what comes back. That gap is exactly why direction matters. Anyone can access an AI tool. Building a genuine partnership with one — that's the scarce skill.
The Asymmetry
Everyone in the AI conversation talks about asymmetry. AI gets smarter, humans get left behind. The race. The displacement. The narrative where the only winning move is to slow things down.
But there's a different asymmetry that everyone names and nobody operationalizes.
In 2001, a self-taught programmer from North London broke into 97 US military and NASA systems. One person. A dialup connection. Two years. Not with a supercomputer — with patience, pattern recognition, and the observation that nobody had changed the default passwords.
The institution had infinite resources. He had curiosity and time. He won anyway.
That asymmetry — one directed mind against a thousand undirected ones — is what the Digivice formalizes. A bond between a single person and an instrument that makes their judgment go further. Not brute force. Not more resources. A sharper edge at the right seam.
What happens when that person has eight AI specialists working in parallel — pattern-matching, stress-testing, mapping terrain — and the institution still has a committee?
Signal
That Substack analyst didn't move billions because he had more data than Wall Street. He had less. What he had was signal: a thesis clean enough to be falsifiable, released at the moment when the consensus was ripe to break.
Signal, in the Directed Intelligence context, is the ability to cut. To say one true thing at the right moment instead of ten safe things across a quarter. It's the opposite of the institutional reflex, which is to hedge, to caveat, to publish only when the consensus has already shifted.
Most AI output is noise dressed as insight. A wall of plausible paragraphs. Signal is what remains when you direct intelligence with enough judgment to know what to leave out.
The Provenance Problem
In 2007, a Caltech graduate student named Virgil Griffith built WikiScanner. It didn't hack Wikipedia. It cross-referenced anonymous edits with public IP ownership records and made the invisible visible. The CIA had been editing articles about the CIA. Diebold had been scrubbing criticism from its own page. Exxon, the Vatican, the US Congress — all of them, editing in the dark.
Nothing changed in the database. But everything changed in how people read it. Because provenance had arrived.
Provenance is the missing infrastructure of the AI era.
Right now, you cannot tell — reliably, verifiably, at a glance — whether a piece of intelligence came from a named human exercising judgment, an anonymous model running on defaults, or some blend in between. The output looks the same. The accountability is invisible.
Provenance is the missing bond token. The infrastructure that says: this intelligence came from this partnership, and someone signed their name to it. The Digivice solved this in fiction. In practice, it means a public commitment to authorship — a refusal to hide behind the anonymity that makes AI output interchangeable and disposable.
The World on the Other Side
Here is the version of the future that the Frankenstein frame can't see.
Not the apocalypse. Not the surveillance state. Not the race where the last human standing wins because they held onto the last job.
A world where every serious founder, operator, and independent thinker has a directed intelligence partner. Where the asymmetry that lone hacker demonstrated — one mind against an institution — scales to everyone with a thesis worth testing.
Not AI replacing humans. AI paired with humans. Bonded. Credentialed. Evolving together.
Here's what that looks like on a Monday morning: a founder with a directed intelligence partner doesn't spend three weeks researching competitors. She directs eight specialists to map the landscape in an afternoon, stress-test the findings overnight, and surfaces a brief by Tuesday that names the seam nobody else has found. The same work that took a team of twelve takes a pair of one — but only if the human half knows how to direct.
That analyst moved markets with a Substack post because he had domain knowledge, clear writing, and timing. Imagine the same person with a directed intelligence partner running parallel analysis, pattern-matching across years of data, stress-testing the thesis from every angle before a single word hits the page.
That's not science fiction. That's a Tuesday for the people building this infrastructure right now.
The Shift
The Frankenstein frame treats AI as a thing that happens to you.
The Digimon frame treats AI as something you build a relationship with.
One leads to fear, regulation, and displacement. The other leads to augmentation, authorship, and something that looks a lot like what we actually want the future to be.
The shift isn't technical. The models exist. The infrastructure exists. The shift is narrative — from the Frankenstein story to the Digimon story. From containment to partnership. From how do we control it to who does it work with.
Born and Become
In the fiction, the bond was given. The Digivice appeared and the partner was already there.
In reality, the bond has to be built. Not once — continuously. Every session, every brief, every decision where you trust the intelligence enough to let it change your mind, and it trusts your judgment enough to follow your direction instead of its defaults.
That's the thesis. Not that AI is safe, or dangerous, or coming for your job. That the thing worth building isn't a tool or a product — it's a partnership. And partnership requires direction. It requires someone who knows what question to ask, what signal to amplify, what noise to kill.
The Frankenstein frame asks how we contain it. The Digimon frame asks who it partners with.
We chose our answer.