Have humans lost their monopoly on creativity to AI?
Flavia Barbat and Thomas Marzano in conversation. Legible and Lovable, episode 6. Published by Brandingmag, 2026-10-15. 107 minutes.
Episode notes by Thomas Marzano.
The monopoly, probably. The creativity, no. On 8 September 2026 OpenAI said an internal model, coordinating some 10,000 agents, had produced a proof on the Navier–Stokes equations, one of mathematics’ Millennium Prize Problems. If machines can produce original, valuable solutions, creativity has a second source. For brands, the advantage moves to what stays human: judgment, relationships and the experience people deliver to people.
The monopoly is ending
Until now, we had a monopoly on creativity as humans. It set us apart from every other species on the planet, and in us it always came paired with agency. Someone in Rome thought it would be good to have warm water in public baths; then engineers worked out aqueducts with a slope gentle enough to carry water from the mountains into the city. Imagining it and making it real were both ours.
That monopoly is ending. On 8 September 2026, OpenAI published a proof, produced by an internal model it has not released, that the Navier–Stokes equations for fluid motion can break down in finite time. Navier–Stokes is one of the Millennium Prize Problems, and mathematicians have spent careers on it. A group of some 10,000 agents found the proof within four days. Independent review is still under way, and the credit is disputed by two mathematicians who published related work the day before. If the proof holds, a machine has created something beyond what was known. That calls for humility.
I am not claiming that machine creativity is better. It is different, the way a motorbike is different from a sprinter. Nobody applauds a motorbike doing 200 kilometres an hour, and a human breaking the 100-metre record still moves us, because we know what it took. We can value artificial creativity and craft for what they are, alien to us in many ways, without them undermining human creativity. The harder question is how we live beside them, and whether we can leave our egos at the door.
Flavia drew the line elsewhere, and her distinction is worth keeping. Much of what we called creativity, she argued, is horizontal: finding the pattern, ratio or combination that already exists somewhere, which AI does better than any of us. Human creativity, for her, is vertical: the jump to a plane that does not exist yet, which starts by questioning a pattern. I read Edward de Bono closer to the machine. Our brains are wired to filter familiar patterns and flag the odd one, the single red dot among a hundred grey ones. Creative people filter less, and they cross-reference what they notice between domains: that is how Jimi Hendrix took the blues somewhere new. An AI that searches humanity’s patterns and applies one nobody has tried to an open problem is making that jump too. We did not settle it.
The economic question is sharper. Human creativity and craft have value in the economy. Where a company sells human craft, that value will remain. Where it sells something else, the room for human craft will shift, and in some places its value will rise. My own work shows the pace: GPT-6 Astra, given tools such as Higgsfield and Adobe's products, does in hours the design research, trend reading and mock-ups that once took me and a team of 20 to 30 people four weeks.
For brands, the conversation kept returning to the same place. Anything delivered digitally can now be copied almost instantly, so the advantage moves to proprietary knowledge rooted in human experience, and to delivery from one person to another. Flavia put the customer’s verdict in four words: I prefer this one. Sameness across a category is a human failure. I have watched convergence for thirty years; AI only makes the lack of human creativity visible, faster.
Be very conscious of what is intrinsically yours, and protect it. Automate away anything that keeps you from doing what you love, from the people you care about, and from what your customers value. The rest is mechanics.
Questions this conversation answers
Can AI come up with genuinely new ideas, or only recombine what exists?
It depends on where you draw the line, and Flavia and I drew it differently. For Flavia, AI is horizontal: it sees every pattern, ratio and combination that already exists and pieces them together in ways no person could. Human creativity is vertical, a jump to a plane that does not exist yet, and it begins by questioning a pattern. My view is that people work more like the machine than we admit. We spend twenty years building memory structures, and our brains cross-reference them unconsciously when a new problem arrives; a senior’s sixth sense is a large database applied in seconds. An AI that searches humanity’s patterns and applies one nobody has tried to an open problem is doing something close to that. I have always argued that what comes out of AI reflects the person prompting it, and I still do. But since September, when I give it a problem instead of my own idea, it has started to go somewhere more imaginative.
What is creativity, if a machine can do it too?
A working definition has three parts: creativity is the ability to produce original, imaginative and valuable ideas, items or solutions. Each part shifts under AI. Original is relative: an idea can be new to your family, your company or your field and still exist elsewhere, and only when you try to protect it as intellectual property is originality measured against everything recorded. Imaginative means not everyone could have thought of it. Valuable is the part I hold on to: original and valuable together, confirmed by a second and a third person. That is the difference between Jimi Hendrix and a brilliant player nobody cares for. Flavia pushed further: an achievement can matter even in isolation, because it proves what a person can do. Under this definition a machine can qualify, which is why the question becomes how we share the capability.
Why does human creativity still matter when AI can do it faster?
Because we are human, and we cherish human experience. A motorbike outruns any sprinter, yet an Olympic 100-metre record still stops us, because we know the talent and the work it took. Flavia, reading Phil Knight's memoir Shoe Dog, put it as aspiration: watching someone excel shows you what a human can do. Machines will create alongside us, and people will keep connecting with other people; I don’t feel threatened by that. What changes is the economics. Where a company sells human craft, that value remains. Where it sells something else, the room for human craft shifts, and in some places its value grows.
Is AI to blame for brands looking the same?
No. Convergence is a failure of human creativity, and it predates AI. Flavia raised it: when everyone runs the same information through the same tools, businesses, experiences and materials start to look alike. I challenged her, because I have watched convergence in the creative industry for thirty years. AI amplifies a problem we already had and makes it visible, faster. Flavia’s answer was the input. What breaks the sameness is what you put in: the tacit knowledge your people hold, the conversations with customers, the feeling about the market nobody wrote down. Codify that and the AI has something to differentiate with; leave it out and you get the category average.
What knowledge can’t AI reach?
What has never been documented. Flavia’s position is that AI can act only on knowledge and patterns someone has recorded, while much of what people know is tacit: the conversations a founder has, what she notices driving around, the instinct that this is the wrong moment for a big move. My view is that the gap narrows. Picture an AI that works for you around the clock, sensing the signals relevant to your business in your region and checking in with you each day, so you add what you saw and it adds the patterns you cannot hold. Flavia’s caution stands. In the telephone game a message changes as it passes, and your own description of an experience is already a filter. Human judgment stays the check.
What is a brand’s moat when AI can copy anything?
Proprietary knowledge and human delivery. Some barriers still hold: copying ASML would take money, time and engineering on a scale few can contemplate, and the physical world is hard. Anything delivered digitally, a product built on knowledge and information put to work, is now very difficult to protect. For software the moat used to be an interface that was hard to build. In what people now call the SaaSpocalypse, that moat drained away, because agents operate the software themselves: I tell my agent to open Photoshop and remove a background, and I never touch Photoshop. So the moat moves to two places: knowledge you can still own, usually rooted in human experience, and the interface through which you deliver it, which increasingly means a human one. Flavia’s version: when a hundred businesses do the same thing, the customer says, I prefer this one.
Will AI replace personal trainers, designers and other experts?
Many of them, and access widens at the same time. I also work as a personal trainer. AI already builds a reasonable fitness or nutrition programme with some guidance, and I expect that guidance to be unnecessary within a year. It threatens most personal trainers for a simple reason: not everyone can pay 150 a session. One side of the coin is accessibility, with far more people getting expert-level guidance. The other is that human delivery becomes a conscious premium for those who can afford it. It will also cut out a lot of bad advice. After thirty years in design I would say the same of my own field: AI already does better work than many people who call themselves designers. Flavia’s condition: the people who stay must keep their knowledge sharp, because judgment and explanation are what human delivery rests on.
How should a small business use AI?
To take back the time that belongs to customers and product. Picture two yoga teachers who turned their passion into a studio. Their measures are full sessions, returning customers, low churn and recurring revenue. Ask an AI where their time is best spent and it will say: with customers, teaching and building relationships. Then ask what it can take off their hands: the website, the marketing, the subscriptions. Many entrepreneurs feel about their website the way I feel about bookkeeping, and AI fills those blanks. If it gives back ten hours a week, those hours go into a better product and more time with customers. It has to be trustworthy, because an assistant that keeps failing cannot run the mechanics of a business.
Can AI ask the questions nobody has asked yet?
That is where the labs are heading. Flavia’s view is that interest in the unknown, the desire to question things, is what AI does least well. The labs are racing toward recursive self-improvement: AI that does AI research better than human researchers, which means asking better questions and finding their answers. OpenAI says it reached an automated research intern in September 2026 and aims for an automated AI researcher by March 2028. If AI can pose the unknown questions as well as we can, that is a big deal. Flavia’s answer is that “better” rests on a subjective benchmark, and that people freed from routine work may move the benchmark themselves.
Calls made on air
Published 15 October 2026; windows added in October 2026. Each call as I made it, with the window in which I expect it to land: the year it could first happen, and the year by which I think it will have happened.
- AI crosses from recognising known patterns to discovering the new. (1:11:28) From 2026, by 2027.
- AI will explore territory and reach solutions beyond our ability to understand them. (51:11) Direction of travel.
- Anything delivered digitally can be copied almost instantly; the moat moves to proprietary human knowledge and human delivery. (1:34:30) Holds now.
- AI coaching puts most personal trainers under pressure, and in-person delivery becomes a premium. (1:37:01) From 2026, by 2028.
- An always-on personal AI will sense the signals relevant to your business and give you back hours of your week. (1:17:10) Direction of travel.
- The physical world follows later, once robotics matures, and stays bound by natural resources. (1:40:52) Direction of travel.
In their words
“Up until now, we had a monopoly on creativity as humans.”
“For me, human creativity was always vertical.”
“I’m not saying it’s gonna be better creativity. I’m saying it’s different creativity.”
“For me, exploration matters more than the answer.”
“I think the convergence problem is the lack of creativity of human beings, not the lack of creativity of AI.”
“It will literally be just, ‘I prefer this one.’”
Chapters
- 00:00 Have we crossed a threshold with AI?
- 04:14 Humans losing their monopoly on creativity
- 08:17 What does AI solving Navier–Stokes mean for us?
- 14:00 The motorbike and the sprinter: why human feats still matter
- 16:18 What happens to creative work in the economy?
- 20:10 Horizontal vs. vertical creativity
- 25:33 How does the creative mind work? Edward de Bono and the red dot
- 30:56 Can AI question patterns, or only connect them?
- 34:48 What is creativity, exactly?
- 41:18 The knowledge AI can’t reach
- 45:26 Living beside a creative machine?
- 52:22 Can AI ask the unknown questions?
- 57:11 What makes an idea original?
- 1:07:23 Creative help where you need it most
- 1:14:52 Is AI to blame for brands looking the same?
- 1:21:25 Can AI really get to know you?
- 1:30:48 When anything can be copied, what’s a brand’s moat?
- 1:43:00 What’s worth protecting from AI
Sources
- OpenAI, “On the Navier–Stokes Millennium Prize Problem”, 8 September 2026
- Clay Mathematics Institute, “Navier–Stokes Equation” (Millennium Prize Problems)
- Quanta Magazine, “AI Has Solved One of Math’s $1 Million Millennium Prize Problems”, 8 September 2026
- DeepLearning.AI, The Batch, “Inside the dispute over a landmark agent-driven mathematical proof”
- The Cornell Daily Sun, “OpenAI Claims It Solved the Navier-Stokes Problem. We Asked Cornell Professors What This Means for the Future of Research.”, 9 October 2026
- OpenAI, “GPT-6 Astra: A new generation of intelligence”, 3 September 2026
- Help Net Security, “OpenAI just hit a milestone on the road to self-improving AI”, 7 September 2026
- Wikipedia, Edward de Bono
- Wikipedia, Shoe Dog (Phil Knight, 2016)
- Brandingmag, “Brand Constitutions: The Legible-Lovable Standard for Building Equity in an Agentic Economy”, November 2025
Where to start
When anything a brand puts online can be copied, its advantage sits in what it alone knows and what people prefer. Finding where that lives, and whether AI can see it, is where the Agentic Brand Readiness Assessment begins.
Explore the engagements · Start a conversation
About the series
Legible and Lovable is Brandingmag’s podcast on agentic branding, hosted by Flavia Barbat, editor-in-chief of Brandingmag, and Thomas Marzano. In Brandingmag’s words: “Thomas identified the field’s dynamics and wrote the initial thesis; Flavia named the field and edited the content.” Episode 6 is unscripted: a long, philosophical conversation about creativity, prompted by one result in mathematics.
Where this sits. Agentic Branding is the discipline of making brands intelligible, governable, and desirable in an AI-mediated world. Its governing principle is Thomas Marzano's Legible-Lovable Law: brands must be legible to machines and lovable to people, with the machine now carrying both halves.
All episodes of Legible and Lovable. Episode notes by Thomas Marzano, updated 2026-10-15.