The Anxious Autocrat
Why Creative Excellence Requires Benevolent Dictatorship—And What to Do About It
This essay is the first of two. The second essay, “The Quality Cliff“ explores what happens when organizations ignore the principle described here.
The Collapse You’ve Watched But Couldn’t Name
Crispin Porter + Bogusky was Agency of the Decade in 2009. Over 1,000 employees across Miami, Boulder, Los Angeles, London, and Sweden. Cannes Grand Prix in all five categories—the only agency ever. The work was legendary: Subservient Chicken, Truth anti-smoking, MINI Cooper.
By 2021, CP+B had collapsed to 40 employees and 2 clients. By 2024, they’d dropped “Bogusky” from the name entirely.
What happened?
I was at CP+B when the answer became obvious in retrospect and invisible in real time. You don’t notice the thing that makes a place work while it’s working—you just assume the machinery is the machinery, that the culture will keep producing what the culture produces. It’s only later, watching the same processes generate different outcomes, that you realize you were never inside a system. You were inside a person’s judgment, distributed across a thousand employees who thought they were the point.
Alex Bogusky left in 2010.
The agency tried everything. Multiple creative directors. Collaborative leadership. Creative councils. They brought Bogusky back in 2018 with full authority as “Chief Creative Engineer.” He left again 18 months later. Even the original curator couldn’t revive it.
Elapsed time from curator departure to collapse: 11 years.
The industry blamed “losing magic” and other mystical evasions. Here’s what happened: Bogusky’s taste was embodied, not embedded. His judgment lived in his head—pattern recognition built over 10,000+ hours, consequences absorbed over decades, anxiety that kept him personally reviewing everything. When he walked out, that judgment walked with him.
You’ve seen this pattern at agencies you’ve admired that declined. The person with taste leaves. The agency promotes “collaborative leadership.” Best talent interviews elsewhere. Work becomes forgettable. Within 36 months, acquisition or closure.
The curator with exceptional taste and veto authority is the only thing that makes creative organizations work at the highest level. The industry’s attempt to democratize, systematize, and scale by removing them is why most agencies produce mediocre work.
This contradicts every trend in modern management. It makes succession impossible. It makes agencies unsellable as going concerns.
But it’s true. And AI is making it undeniable.
The Anxious Autocrat Principle
Creative excellence at scale demands benevolent dictatorship by an insecure genius—one person with exceptional taste who has both absolute veto power and perpetual self-doubt. This cannot be delegated or democratized without quality collapse. The moment you replace singular judgment with collaborative process, you’ve already lost.
The formula:
Taste × Veto × Anxiety = Excellence
Remove any variable and the equation collapses to mediocrity.
I’ve seen this equation work, which is another way of saying I’ve been the anxious person in the room hoping the equation held. In 1996, I designed AOL’s 4.0 interface—warm, earthy colors, a yellow running man carrying a heart for favorites, a globe for the internet, an envelope for email. The dominant paradigm was dense, blue, technology-oriented. I designed for people experiencing the internet for the first time. Their anxiety, not technical sophistication.
Three people approved it. President, VP Design, me. No committee. No focus groups. The interface reached hundreds of millions of users.
I still don’t know if it was good. I know it worked. I know the judgment was mine and I had to live inside the gap between those two things for years before the metrics arrived.
That approval chain is harder to protect now. Legal review, brand safety, accessibility requirements, stakeholder complexity—the operational environment has changed. But the principle hasn’t: creative judgment concentrated in few hands with taste, not diffused across many hands seeking consensus. The job today isn’t to recreate 1996’s conditions. It’s to design structures that protect curatorial authority within contemporary constraints—distinguishing between operational process (necessary) and creative committee (fatal).
I’ve thought about what would have happened to that interface in a 2024 approval process optimized for consensus. Twelve stakeholders. Three rounds of focus groups. A/B testing that optimized for task completion over emotional resonance. The running man would have been killed for being “too playful.” The colors would have been tested into blue. We’d have gotten something defensible instead of something loved.
We’d have gotten something defensible instead of something loved.
The Distinction That Determines Survival
Not all taste dictatorships are created equal. The difference between those that survive and those that collapse is whether taste is embodied or embedded.
Embodied taste lives in one person’s head and hands—pattern recognition built through experience, intuitions that can’t be articulated, judgments that feel instantaneous. This is what most great creative leaders have. It’s also untransferable.
Embedded taste lives in systems, processes, and training mechanisms—documented methodology, selection filters, apprenticeship structures. It can be transmitted to others who prove they’ve internalized it.
Cristóbal Balenciaga closed his fashion house in 1968 rather than hand it off. His method was embodied—he never sketched, instead manipulating fabric directly. When he died, the knowledge died with him. Paul Bocuse’s restaurant held 3 Michelin stars for 55 years. Two years after his death: demoted to 2 stars. Hermès has survived six generations because taste is embedded in apprenticeship structures, family governance, and quality filters at every stage.
CP+B, Balenciaga, Bocuse: embodied taste, no transmission mechanism, collapse.
Hermès: embedded taste, generational survival.
The question for every creative leader: Is your taste embodied or embedded? And what are you doing about it?
The Taste Bottleneck Inversion
Every management consultant will tell you to eliminate bottlenecks. They’re wrong about this one.
The Taste Bottleneck—one person through whom all creative judgment must flow—isn’t a constraint to remove. It’s the only thing that works. Agencies don’t fail because they have a Taste Bottleneck. They fail because they try to eliminate it through “collaborative leadership” and “democratic process.”
The diagnostic question isn’t “how do we distribute this authority?” It’s “do we have the right person in the bottleneck—and are we protecting their authority or diluting it?”
Creative review meetings fail because they lack the conditions that make collective intelligence work: diversity of opinion, independence from social influence, aggregation mechanisms. A creative committee has none of these. Same backgrounds. Same room. Same politics. No way to “average” aesthetic judgment.
I’ve worked in all three failure modes. Autocracy without taste—an ECD who made up principles based on what he liked, without the foundation underneath. I sat in reviews where he’d kill work for violating rules he’d invented that morning, and the room would nod because nodding was survival. I nodded too.
No taste culture—a massive production shop that thought it was a creative agency, where I won major accounts by flying under executive attention.
And I’ve worked in the success mode: small approval chains, people with taste at the gates, authority to make calls. The results weren’t close.
I’m writing this from the three years I spent fighting to establish regular creative reviews at an agency that kept canceling them because “we trust the teams.” From the campaigns that won Effies inside structures so broken I had to route around my own leadership to protect the work. From the deals I made with systems I knew were wrong because I wasn’t powerful enough yet to refuse.
The Anxiety Engine
When you say “this work isn’t good enough,” you’re not saying “I personally don’t like it.” You’re claiming something bigger—that this work will not resonate, will not move business, will not earn attention. You’re speaking as if your subjective judgment binds everyone.
This is exactly what the creative curator does. And it’s why the role is inherently anxious—you’re asserting binding authority from a position that’s inherently personal.
The anxiety isn’t weakness. It’s structural honesty. The curator who isn’t anxious about this has stopped understanding what they’re actually doing.
The anxious autocrat operates under self-imposed constraint—the constraint of their own standards, which forces creative solutions rather than obvious ones. Remove the anxiety and you remove the constraint that generates invention. The comfortable autocrat produces competent work. The anxious autocrat produces surprising work. The people who should have all the power are the ones terrified of having it.
The people who should have all the power are the ones terrified of having it.
Why AI Makes All of This More Urgent
Every six months, AI models get dramatically better at execution—concept generation, design, writing, strategy frameworks. The curve is exponential.
Before AI, agencies competed on execution quality. “Our designers are better.” “Our writers are faster.” AI has obliterated those differentiators.
What AI can’t do yet: Know which output is worth making.
AI as currently deployed in most organizations learns from the aggregate, which means it learns median taste. The curator’s value isn’t just judgment; it’s idiosyncratic judgment that produces work the algorithm wouldn’t surface. AI will get better at “good.” The anxious autocrat produces “surprising.”
This creates asymmetric warfare:
The Anxious Autocrat + AI: Generate 100 variations in an hour. Curator applies taste to pick the 2 worth developing. Ship faster with higher quality bar maintained. Result: 10x output with elevated quality.
The Democratic Committee + AI: Generate 100 variations in an hour. Bring all 100 into collaborative review meetings. Spend 3 weeks building consensus. Approve mediocre compromise. Result: 100x output of mediocrity.
I run Claude and ChatGPT in parallel—routing outputs between them based on what each does well. It’s modular synthesis thinking applied to idea generation: one model generates raw material, another critiques against specific criteria, a third synthesizes patterns. The signal passes through multiple processors before I decide what’s worth keeping.
This isn’t how most people use AI. They treat it as a question-answering machine. I treat it as a creative team that needs direction—the same direction I learned to give at CP+B and refined across decades of creative reviews.
David Droga—the most awarded creative in Cannes history—got early access to OpenAI’s Sora and told the industry exactly what’s happening. At Cannes Lions 2024: “Not all creativity is worth saving. The majority of advertising is not creative. It’s written by something far more dangerous than AI: research.” Tools like Sora will “eradicate the messy middle of the creative process without displacing jobs at the top.”
The jobs at the top are curators who know which execution is worth making. Agencies without those jobs have nothing left to sell.
The Model That Actually Works (And Its Limits)
Wieden+Kennedy has 1,500 employees across eight global offices. Dan Wieden stepped into a chairman role in 2015. Quality remains high. In 2025, W+K won Adweek’s Global Agency of the Year while handling Nike’s return to Super Bowl advertising after a 27-year hiatus.
How?
Dan Wieden didn’t just curate work. He curated curators. W+K isn’t one 1,500-person democracy. It’s multiple autonomous offices, each led by a taste dictator, all sharing Wieden’s aesthetic values. Portland. Tokyo. London. Amsterdam. Each office has its own creative leadership with veto authority. Each operates as an independent creative dictatorship under a shared cultural banner.
W+K proves you can scale through replication, not democratization. Clone the dictatorship across regions. Don’t distribute authority across committees.
But Wieden himself said something revealing: “I think it’s true that I’ve only done great work for Nike.” The agency’s other work was good, not great. The exception was Nike—where his judgment shaped everything personally.
Even Dan Wieden couldn’t scale his personal taste to all clients. What he could do: scale a culture of taste by recruiting other anxious autocrats and giving them autonomy.
The Succession Problem
If creative excellence requires irreplaceable singular taste, what happens when that person retires, dies, or sells?
Three paths exist:
Path 1: Find Another Autocrat. Nearly impossible. This is rare enough that most succession attempts fail—you’re not looking for creative competence, you’re looking for the specific combination of taste, anxiety, and authority that produces outcomes. There are patterns that increase your odds—signals that distinguish real taste from confident performance. One signal: Ask the CCO to show you work they killed and why. Confident performers show you wins. Real curators can articulate what they said no to and the judgment that drove it.
Path 2: Install Collaborative Leadership. Guaranteed failure. Creative councils, democratic reviews, consensus decision-making. Quality collapses within 24 months.
Path 3: Build a Transmission Mechanism. Fragment into replicated autocracies (W+K). Embed taste in systems and apprenticeship (Hermès). Create selection mechanisms that test whether candidates have internalized your standards.
Path 3 is better than Path 2. But I suspect that transmission favors consistency over breakthrough. Hermès has survived six generations; whether it could produce something as radical as Balenciaga’s original vision is an open question. W+K maintains quality; whether it retains the capacity for era-defining surprise is something only time will answer.
This might be the honest tradeoff: you can transmit taste that sustains excellence, or you can bet on embodied genius that produces breakthroughs but dies with its creator. Most organizations would benefit from the former. A few are built for the latter.
Most agencies choose Path 2 while pretending to attempt Path 1, then act shocked when quality evaporates.
The honest choice: Plan your exit while you’re still relevant. Sell when the curator is still there and the brand has value. Or accept you’re building something beautiful and temporary.
What This Means For You
If You’re an Agency CEO/Founder:
Stop hiring “collaborative” CCOs. Hire difficult perfectionists with proven taste—anxious autocrats who’ve shipped work you admire. Give them absolute veto authority. No committees. No consensus.
Accept the tradeoff you’re making. If they’re truly excellent, you’re building something that depends on their presence. You can build transmission mechanisms—fragment into autonomous units, document methodology, create apprenticeship structures—but be honest about what transmission can and can’t preserve.
The diagnostic question: Where does taste actually live in your organization? Map who can kill work, who can’t, where decisions diffuse into committee. Then design the approval architecture that protects curatorial authority while meeting operational requirements. This isn’t about eliminating process—it’s about ensuring creative judgment flows through people with taste rather than dissolving into consensus.
Litmus test: Can your CCO kill work that 15 people worked on for 3 weeks without committee approval? If not, you don’t have a curator. You have a manager.
If You’re a CCO/Creative Leader:
Your job isn’t to make people comfortable; it’s to make them better. Make the decisions yourself. The moment you delegate judgment, you’ve abdicated your role.
Cultivate your imposter syndrome. Your anxiety is your constraint, and constraint is your canvas. The comfortable leader produces obvious work. The anxious leader produces surprising work.
Use AI to amplify your capacity. Generate 100 variations, pick 2 worth developing. And document your “this, not that” decisions—you might be building something that outlasts you.
If You’re a CMO/Client:
Stop hiring agencies for “process” or “AI capabilities.” Hire them for one person’s taste.
Meet the person who will personally curate your work—not the CEO, not the new business lead. The actual creative leader who will say yes or no. If the CCO seems too confident, run. Confidence without anxiety produces mediocrity.
The complication: the curator who maintains quality standards must also maintain client confidence under pressure. That combination is rare—when you find it, protect it.
Ask: “Will you personally be in every major creative review?” If the answer is anything other than “yes,” you’re buying process, not judgment.
The Choice
AI has made execution cheap and abundant. The exponential curve of AI capability is making one truth undeniable: Taste is the only defensible moat.
Most leaders would rather fail through collaboration than succeed through dictatorship, because failure through collaboration provides political cover. No one gets blamed when the committee approved it.
Most leaders would rather fail through collaboration than succeed through dictatorship, because failure through collaboration provides political cover. No one gets blamed when the committee approved it.
But you don’t have to be most people.
Every six months, this becomes more true. Every time AI execution quality doubles, the taste gap becomes more obvious. The largest holding company in the world is now explicitly betting on AI + scale as their differentiator—testing the opposite hypothesis from this essay in real time. The agencies betting on AI to democratize creativity are making the same mistake as agencies that bet on process to replace taste.
The agencies putting AI in the hands of anxious autocrats are building asymmetric advantage. One curator with Claude and ChatGPT can now do what required 20 people before.
CP+B collapsed from 1,000 employees to 40. The pattern is clear. The industry knows this—they just won’t admit it publicly because it makes succession impossible, and it makes agencies unsellable.
Your real decision: Not whether you want to be a dictator.
Whether you’re anxious enough to be a good one.
Why I Wrote This
Writing this forced me to name patterns I’ve lived but never articulated. This essay is how I think—in public, subject to judgment. The right people recognize the judgment on display. The wrong people don’t finish reading. Both outcomes are useful.
Three types of problems keep surfacing in my conversations:
The Succession Crisis: Agency founders who built on curatorial dictatorship but can’t find another curator at their level—or who hired “collaborative leadership” and are watching quality drift while their best people quietly interview elsewhere.
The Execution Gap: CMOs whose agencies deliver beautiful decks that engineering says “can’t be built” nine months into development. They need creative that can actually ship.
The Creative-Product Divide: CCOs and CPOs who recognize the split is costing them competitive velocity, but don’t have the bilingual fluency to build hybrid teams that satisfy both disciplines.
One thing I find useful in these conversations: mapping the decision trail for a recent piece of work. Who touched it? Who could have killed it? Where did judgment get diluted into committee? That map usually surfaces more insight than months of organizational introspection.
I’ve spent 25 years building and fixing creative organizations—as creative leader at CP+B, AKQA, BBH, and Monks, shipping work for Google, Samsung, and Burger King’s AI-native “Million Dollar Whopper.” I’ve also led product: AOL’s running man back when internet arrived on CD-ROM, and Peacock streaming service v1.0. Currently CCO at TCS Interactive, the creative arm of a $30B, 600,000-person company most creatives have never heard of. I’m working on that.
If any of this describes your situation—or if you think I’m wrong about something—I’d like to hear it.