Skip to content
EO StudioEO Studio

Why Your Best Ideas Come When You Stop Trying | Marcus du Sautoy

AI just cracked a mathematical conjecture that had stayed open for decades. It found the counterexample — and still couldn't tell us why it was true. Marcus du Sautoy is a professor of mathematics at Oxford and the Simonyi Professor for the Public Understanding of Science. He spends his life on the line between what machines can do and what only people can. In this conversation he gives you a working definition of creativity — three specific types, in order of difficulty — and then shows exactly which two AI has already taken and which one it still can't reach. Along the way: why AlphaGo's Move 37 counts as genuine machine creativity and deserves the credit over its programmers, and why the human edge turns out to be something nobody puts on a résumé. We're lazy. AI isn't. That's the whole advantage. What you'll learn: - The three-type framework for classifying any creative act, and which two AI already does well - Why finding a counterexample and proving why are different jobs, and which one is still yours - The local maximum test for spotting when your whole field is stuck on a false summit - How to tell genuine machine creativity apart from the programmer's creativity - The laziness heuristic: use your refusal to do the donkey work as a shortcut detector - The one behavior from an AI that would signal real intention, not statistics 00:00 Intro 01:35 The Three Levels of Creativity 02:41 Why mathematics is closer to fiction than to science 04:37 Type 1: Exploratory creativity 05:15 Type 2: Combinational creativity 05:44 Type 3: Transformational creativity 06:40 How Creative is AI in 2026, Actually? 06:58 AI cracked a decades-old conjecture 07:47 AI is a digital telescope, not an artificial brain 08:20 Move 37: the move every expert called a mistake 09:31 The local maximum problem 10:09 Why the credit belongs to the AI, not the coder 10:52 The one signal that would mean there's a ghost in the machine 11:25 What Only Humans Can Do 12:35 Gauss as a schoolboy, and the birth of the algorithm 14:07 Why AI never bothers to find the shortcut EO is a global media brand for builders. We tell the defining stories of founders shaping the future: people who see what others don’t and build what they believe in. Subscribe to EO: https://www.youtube.com/@eoglobal EO Magazine: https://www.eomag.io Instagram: https://www.instagram.com/eostudio.official/ X: https://x.com/eostudi0 LinkedIn: https://www.linkedin.com/company/eo-studio EO Studio: https://eo.team/ Business inquiries: partner@eoeoeo.net Build what you believe in.

Marcus du Sautoyguest
Jul 28, 202615mWatch on YouTube ↗

At a glance

WHAT IT’S REALLY ABOUT

Creativity’s three levels, AI’s limits, and humans’ shortcut advantage

  1. Creativity can be usefully split into exploratory, combinational, and the rarest form—transformational creativity that breaks prior conventions.
  2. Mathematics resembles fiction in that it invents rule-bound worlds that need not match physical reality, making it inherently creative rather than purely descriptive science.
  3. AI excels at exploratory and combinational creativity and can function like a “digital telescope,” revealing patterns and counterexamples that humans miss.
  4. AlphaGo’s Move 37 illustrates machine-generated novelty that can reshape human practice, suggesting AI can sometimes reach transformational-like breakthroughs.
  5. Humans retain an edge in intention and in inventing elegant shortcuts (algorithmic leaps), and du Sautoy argues AI should be treated as augmented intelligence and a collaborator, not a competitor.

IDEAS WORTH REMEMBERING

5 ideas

Defining creativity is prerequisite to judging AI’s creative capacity.

Du Sautoy argues debates about “can AI be creative?” go nowhere without specifying which kind of creativity—rule-exploration, cross-domain combination, or convention-breaking transformation—you mean.

Transformational creativity is rare and hardest to mechanize.

Because many AI systems learn from past styles, they naturally extend existing rule-sets; truly breaking those conventions in a way that opens new territory is the toughest test.

Mathematics is closer to fiction than most people assume.

Unlike empirical sciences that must match observed reality, mathematics can explore internally consistent “possible worlds” (e.g., non-Euclidean geometries), paralleling how novelists define rules and investigate consequences.

AI’s most reliable power today is pattern-revelation, not understanding.

He frames AI as a tool like Galileo’s telescope: it can surface structures (including counterexamples to long-held conjectures), but often does not provide the explanatory “why” behind them.

AlphaGo’s Move 37 shows AI can expand the human search space.

The move looked like a mistake to experts but proved decisive, pushing human Go strategy beyond what du Sautoy calls a “local maximum” to a higher peak that humans couldn’t see through their conceptual fog.

WORDS WORTH SAVING

5 quotes

We really can regard AI in a similar way. It's almost like a digital telescope. It's allowing us to see into the digital world, see patterns emerging.

Marcus du Sautoy

The most difficult and the rarest form of creativity is transformational creativity. That's where something seems to come out of nowhere.

Marcus du Sautoy

We thought we climbed a mountain peak and we, we knew the top place to play this game. But what the AI revealed is, okay, that might be a high peak, but it's only what we mathematicians call a local maximum, that there's actually a, a much higher peak if you go down the valley and up the mountain just across the valley.

Marcus du Sautoy

I'm not saying that it won't get to that stage, but at the moment, it will suddenly write a novel because it wants to tell you what it's like to be an AI and that, that intention to express itself will be, I think, um, our first indication that, you know, oh, maybe there is a ghost in the machine.

Marcus du Sautoy

One of my central messages is to remember that artificial intelligence is not a competitor. It's a collaborator.

Marcus du Sautoy

Three levels of creativity (exploratory, combinational, transformational)Mathematics as creative world-building vs empirical scienceAI as augmented intelligence / “digital telescope”AI finding counterexamples in mathematicsAlphaGo Move 37 and changing expert conventionsLocal maximums vs higher peaks metaphorHuman shortcut-thinking (Gauss summation) and collaboration with AI

High quality AI-generated summary created from speaker-labeled transcript.

Get more out of YouTube videos.

High quality summaries for YouTube videos. Accurate transcripts to search & find moments. Powered by ChatGPT & Claude AI.