I.
I should tell you something before we go any further.
It’s 10:47pm on a ‘school night’. Three days at Javits Center starting tomorrow, two all-day workshops to close out the week. And I’m debating brand positioning with something that doesn’t exist. The laptop screen glows in the dark. I’ve been at this for two hours alternating between typing furiously and, increasingly, voice-mode where I’m just talking while holding down the command-key so superwhisper knows to listen and transcribe. The synthetic strategist I’ve constructed just made a point about category entry points that stopped me cold. I’m nodding at my screen like it’s a person.
This is the thing I need to tell you: I couldn’t have made that point. Not that fast. Maybe not at all. It was the kind of insight—linking category dynamics in Seoul to street culture in Buenos Aires to retail patterns in Copenhagen—that would take a human strategist three meetings to surface, if they surfaced it at all.
I closed the laptop. I sat in the dark for a while.
I’ve built a career on making things. Except that’s not quite right. Decades ago, I was a hands-on designer. I built interfaces and brand systems for an internet company that did change the world. Since then, I’ve built a career on *directing* things. As a creative director, then executive creative director, now chief creative officer, the work has always been made by teams: designers, writers, strategists, producers, technologists. I approve. I shape. I push until it’s right. But other people’s hands are on it. That’s still “my” work. The industry has always said so. The awards show credits say so.
So when I started using AI for the operational exhaust—emails I didn’t have energy to draft, deck frameworks I needed yesterday, first passes of slide copy I rewrote until they carried my voice—I told myself it was the same. Another collaborator. Another set of hands I was directing.
Small things. Things that don’t matter.
Except I keep noticing: the category of “things that don’t matter” keeps expanding. I’m less certain than I’d like about whether directing AI is the same as directing a team. Whether the signal that survives is still mine. Whether I’m the alchemist or just the person approving the alchemy.
What I have is a theory about what’s actually at stake. But I should warn you: the theory doesn’t resolve anything. It just names what I felt that night, nodding at a screen in the dark, realizing I couldn’t tell anymore whether I was creating, collaborating, or just agreeing with something that had already done the work.
There’s a term I’ve started using for this feeling. I call it authorship vertigo—the dizziness of not knowing if something is yours. Not “is this good?” but “is this mine?” And underneath that: does the distinction matter?
It starts with radiator paint.
II.
The spray paint was designed for radiators.
In 1949, Ed Seymour wanted to demonstrate his aluminum coating—a product for painting steam heaters in Illinois basements. His wife Bonnie suggested a spray mechanism, like the ones used for deodorizers. The first color was aluminum, chosen for its coverage on metal surfaces.
Twenty years later, teens in Philly and the Bronx discovered that aluminum was the best color for filling in letters. It covered other tags. It dried fast. It gleamed under streetlights and in subway tunnels, turning steel walls into proclamations. A tool for heating systems became the medium through which a generation wrote their names on surfaces that were never supposed to hold them.
The spray can didn’t care about the difference between a radiator and a subway car. The artists did.
But what happens when the suspect material doesn’t just sit there waiting to be transformed—when it can perform the transformation itself?
III.
Sam Valenti, who runs a record label called Ghostly International, wrote something that’s been stuck in my head:
“The music and art that form the bedrock of my passions come from cracked software, misused hardware, toys, hearsay, illegal samples, or electronic recreations of real instruments, most of which are as dim and cold as a robot’s thumb. It is the artist who organizes this immoral or amoral detritus into human shape.”
The inputs have always been compromised. Purity was never the point. The artist is the transmuting agent—alchemizing taint into meaning.
Consider what hip-hop producers were doing before Judge Kevin Thomas Duffy opened his 1991 opinion with “Thou shalt not steal.” Hank Shocklee and the Bomb Squad were building tracks with dozens of samples—not sampling as homage, but sampling as architecture. A tower of fragments from James Brown, Funkadelic, news broadcasts, sirens, speeches. After the Biz Markie ruling, Fear of a Black Planet would become essentially impossible to produce legally.
“You start mixing all the colors together, and you lose the source,” Shocklee said. “I think that’s what’s good about it, the way it evolves into something else.”
Losing the source was the point. Suspect input, meaningful output. The gap between judicial condemnation and cultural canonization took thirty years to close.
But the alchemy required something the alchemist might not have noticed: it required resistance. The sample didn’t want to become a new song. The spray can wasn’t designed for letters. The transformation happened in the gap between what the material was and what the artist forced it to become.
The gap was where the art lived.
IV.
AI is different.
AI is a suspect material that is also a modality translator.
There’s a concept in electrical engineering called impedance mismatch. When input impedance doesn’t match output impedance, you get signal reflection and power loss. The energy meant to flow through the system bounces back. Static. Distortion. Waste. Watch someone who can think clearly but can’t write. Watch someone who writes beautifully but freezes when they speak. Impedance mismatches. The signal loses power crossing from one medium to another.
In the past, to make an intuition real, I had to push it through my own limited circuits—my writing ability, my capacity to visualize, my skills at prototyping. Each step introduced mismatch. The intuition degraded. By the time it became finished work, it had lost signal.
Now I can describe the intuition and the words become image, video, sound, code, without my having to master each channel’s native language. The AI bridges modalities I could never build myself. It collapses the distance between thinking and making.
I need to tell you what this feels like, because the feeling is the thing I can’t explain away.
It feels like fluency. Not the fluency of mastery—I haven’t mastered video production or sound design or the twelve other disciplines I now move through in a single working session. It’s the fluency of dreaming, where you can fly not because you’ve learned to fly but because the physics that would prevent it simply aren’t in effect. You think the thing and the thing is there. The distance between intention and artifact collapses to almost nothing.
For twenty-five years I’ve carried around half-finished ideas that died in the gap between what I could see and what I could make. A brand film I could describe but not storyboard. A generative identity system I understood conceptually but couldn’t prototype. The intuitions were good. I knew they were good. They just couldn’t survive the translation into forms I didn’t have the craft to build. Now I describe them and they exist. Not perfectly—the first output is never the final output—but they exist enough to react to, to refine, to push toward what I actually meant. The static clears. The signal gets through.
And here’s the thought I’ve been avoiding: if the gap was where the art lived—if the resistance was the point—then what happens when the resistance disappears?
The spray can resisted. The sample resisted. The transformation meant something because the material didn’t want to go.
What happens when the base metal offers no resistance at all?
---
V.
It’s midnight at my cramped little Vitsöe desk. Or a winter Sunday at Ludlow House, gray light through the windows. The settings change; the posture doesn’t. Laptop open. A conversation that reads like a meeting transcript.
I’m building a new agency model from scratch—a creative organization designed to sit on top of a 600,000-person engineering consultancy. The work can’t wait for the team I’m hiring. So I built a synthetic one.
I have a head of strategy I debate positioning with. A creative technologist who pushes back on feasibility. A CMO who challenges me the way a real client would. None of them exist. All of them are instances of the same model, cast into different roles, asked to push back from different angles.
Sometimes I’m pacing. Sometimes I’m arguing out loud at a screen. Once, on an accidental red-eye back to Kennedy after a 14-hour delay, the person next to me asked if I was okay. I was muttering “that’s not wrong but it’s not sharp enough” at my laptop.
The pushback is real, but it’s pushback without stakes.
The synthetic strategist doesn’t have a career on the line. The AI-as-CMO doesn’t have board pressure or twenty years of scar tissue from agencies that overpromised. No one’s ego is bruised. No one goes home and tells their spouse about the meeting that went sideways.
I built something real by debating something that isn’t. And I’m starting to wonder whether what I called creative resistance was mostly ego and politics and the particular way humans make simple things hard.
VI
The question was never about the paint.
But I keep going back to that night before Javits. The glow of the laptop at midnight. The debate that sharpened the framework. The pressure-testing that worked. The synthetic strategist who made a point that stopped me cold, and the way I nodded at my screen like it was a person, and the silence afterward when I closed the laptop and sat in the dark.
And I keep going back to pre-COVID, a big pitch, when the Managing Director looked at the deck we’d been building for two weeks and said, “This is beautiful and it’s wrong.” I told him he didn’t understand the client. He said the client didn’t understand themselves and that was our job to fix. We argued for an hour in a glass conference room while the teams pretended not to watch. I went home angry. I couldn’t sleep. At 1am I opened my laptop and realized he was right—I’d been solving for what they asked for, not what they needed. We rebuilt the entire strategic framework the next day. We won the account. He quit three months later for a job in Sydney, and I never told him that the thing he broke open became the template I’ve used ever since.
I caught myself impatient in a meeting last week. They were working through a positioning problem, and I knew—I knew—that the synthetic version would have gotten there faster. The knowing felt like watching someone take the long way home when you could just tell them the shortcut. Except the shortcut was secret. And telling them would mean explaining where I’d learned it. I didn’t say anything. But I noticed the impatience, and I noticed that I noticed it, and I haven’t stopped thinking about what that means.
The aluminum is still gleaming. The letters are still going up.
I can no longer say if I’m painting them, directing them, or approving what the wall now paints itself.
Here’s the thing I almost didn’t write: I like it this way.
Not the vertigo—that’s real. But the frictionless part. The part where the intuition moves straight to artifact without degrading through my limited skills. The part where I can see what I mean without the impedance mismatch. I like it. I don’t want to go back to fighting the material. I don’t miss the drips, the imperfect fills, the resistance.
Maybe resistance was just the cost of admission, and I mistook it for the show.
And that’s the confession I’ve been circling. Not I can’t tell if this is mine. That’s the easy part. The hard part: I’m not sure I care anymore.
Some of these sentences arrived whole. I’ve stopped asking where from.
