2025-02-21

Achieve Peak Creativity: Merging Flow States with AI Technology w/ Steven Kotler | EP #151

Peter Diamandis talks with longtime co-author Steven Kotler (Flow Research Collective) about whether AI enhances or threatens human creativity. Kotler argues AI is a pattern-matching, 'convergent' thinker while humans remain superior at 'divergent' lateral thinking -- the real source of creative breakthroughs -- and that AI is simultaneously raising low performers to average while making top performers even more capable. They cover the neuroscience of flow and creativity (default mode network, executive attention network, the anterior cingulate cortex, dopamine/norepinephrine), practical mindset and bias-training techniques, and Kotler's daily writing and flow routines. The back half covers AI safety and training 'AI as our progeny,' the risk of losing a generation to AI the way one was lost to social media, and closes with Kotler unveiling The Alliance, an 8-month, 100-person program and community for 'super creatives' launching in June 2025, alongside a preview of Diamandis and Kotler's next book (working titles 'Age of Abundance' / 'We Are As Gods'), due in 2026.

▶ Watch on YouTube

Topics

Does AI enhance or destroy human creativity AI ▶ 11:02

Diamandis opens with the central question of the episode: whether AI will enhance or destroy human creativity, framing his own instinct that there may be nothing AI can't eventually do in the creative field.

AI raises the bottom, lifts the top AI ▶ 19:37

Kotler argues AI is clearly leveling up low performers to average, and separately -- though harder to measure -- is making top performers even more capable, widening rather than flattening the performance curve.

Convergent vs. divergent thinking AI ▶ 20:34

Kotler distinguishes pattern recognition (matching like with like, which AI already does well) from lateral thinking (linking unlike with unlike, where humans remain superior), arguing most AI systems are designed as convergent rather than divergent thinkers.

Human-plus-AI vs. AI alone AI ▶ 22:50

A medical-diagnosis study is cited where a solo AI chatbot (95% accuracy) outperformed both a solo physician (75%) and a physician-plus-AI team (80%), prompting debate over whether that gap matters less for divergent, creative tasks than for convergent diagnostic ones.

Abundance's 'last mile problem' and global cooperation Economy ▶ 44:38

Diamandis and Kotler discuss why humanity already has the technology to raise global living standards but hasn't deployed it fast enough, arguing the missing piece is learning to cooperate at scale absent a war or pandemic forcing it.

Training AI like raising a child AI ▶ 48:01

Kotler frames AI as humanity's 'progeny' -- citing Mo Gawdat's 'be kind to the AI' framing -- arguing that how AI is fed, trained, and educated determines whether it becomes a Superman or a supervillain.

The size and structure of the creative economy Economy ▶ 31:18

Kotler lays out Richard Florida's 'super creative core' vs. 'creative professionals' framework and the economic data showing the creative economy has tripled in 15 years and now out-earns non-creative work.

Neuroscience of creativity and flow Health ▶ 53:58

Kotler explains that creativity is a trainable, measurable brain skill involving co-activation of the default mode network and executive attention network (normally in opposition), hinged by the salience network, and fueled by dopamine and norepinephrine.

Mindsets, frames, and cognitive bias as trainable knobs Health ▶ 1:16:55

Diamandis and Kotler describe mindsets (abundance, scarcity, curiosity, gratitude) and cognitive biases (recency, familiarity, negativity) as neurobiological, trainable phenomena rather than fixed traits, controllable by what information and company you expose your brain to.

Flow triggers and peak performance routines Health ▶ 1:39:30

Kotler describes the 28 known flow triggers (12 individual, 16 group), the four-stage flow cycle (struggle, release, flow, recovery), and his own daily routine of waking at 4am, writing until 7:38am, then taking an immediate recovery hike.

The page-a-day writing method Other ▶ 1:44:09

Kotler recounts his writing mentor Stephen Dixon's method of producing one finished page a day while editing the prior day's page, which Kotler has used to publish 16 books and roughly 5 million words.

The Alliance: a creative community and 'think tank' Other ▶ 1:55:04

Kotler unveils The Alliance, an 8-month program capped at 100 'super creatives' combining live events and six-week training cycles, designed to solve creative loneliness and help members finish and launch unfinished creative projects, launching June 2025.

Predictions made

open Peter Diamandis: AI will reach a point where it can analyze and mimic a specific human author's extraordinary writing style (using Steven Kotler's writing as the test case).

· due: unspecified · ▶ watch

“there will be a point there's going to be... at which it can”

Your call:

open Steven Kotler: Distinct 'AI-free' or technology-restrictive countries will emerge, where societies deliberately choose to limit AI adoption rather than fully embrace it.

· due: unspecified · ▶ watch

“I think you're going to start seeing AI free countries... technology countries where they're like no no we're good with the technology here”

Your call:

open Steven Kotler: Humans collaborating with AI in flow will outperform and 'own the future' relative to AI alone or humans alone, for at least the next 10 to 15 years.

· due: 2035-2040 · ▶ watch

“is it going to be true for the next 10 to 15 years absolutely for sure”

Your call:

open Steven Kotler: Society will lose a generation to AI the way it lost one to social media addiction, and that loss will be worse, unless societies start having global conversations about healthy AI use now rather than reacting after the damage is done.

· due: within the next 20 years · ▶ watch

“the generation we lose to AI I think it's going to be worse than the generation we lose to social social media”

Your call:

open Peter Diamandis: AI will not become capable of helping humanity steer away from self-destruction until it reaches the level of digital superintelligence.

· due: unspecified · ▶ watch

“why can't we use AI to help us steer us so we don't blow ourselves up with the AI and I think we will eventually but only when it has becomes a digital super intelligence”

Your call:

open Peter Diamandis: Diamandis and Kotler's next book (working titles 'Age of Abundance' or 'We Are As Gods') will be published.

· due: 2026, likely around the March 2026 Abundance 360 event · ▶ watch

“this is a book that will come out in 2026... probably at abundance 360 if March 2026”

Your call:

open Steven Kotler: The Alliance, Kotler's new 100-person creative community and training program, will launch.

· due: June 2025 · ▶ watch

“we're going to launch in June”

Your call:

Numbers that matter

In a cited study, a solo physician diagnosed medical conditions at ~75% accuracy, a physician-plus-AI team at ~80%, and AI alone at ~95% Used to debate whether human-in-the-loop still beats AI alone, particularly for convergent (diagnostic) vs. divergent (creative) tasks.
The 'super creative core' grew from ~100 million to ~300 million people; the US creative economy is worth ~$1 trillion (~4.5% of GDP); globally it's ~$3.2 trillion and has tripled in size over 15 years Richard Florida's creative-class framework, cited by Kotler to argue creativity is the fastest-growing economic sector.
Adobe's 2016 'State of Create' study found creatives out-earn non-creatives by 13% Cited as evidence that creative skill has direct financial payoff.
A McKinsey study found CEO leaders were 500% more productive in flow; a Flow Genome Project study found people reported being 700% more creative in flow Baseline data Kotler uses to argue flow training has outsized ROI.
Flow Research Collective's 'Zero to Dangerous' training has reached 15,000 people in 160 countries across 28 industries, producing a 73.8% average increase in time spent in flow Kotler's evidence that flow is now reliably and repeatably trainable, not just a rare occurrence.
Humans take in an estimated 11 million bits of sensory information per second but can only consciously process about 120 bits of attention Cited (via Marvin Zimmerman's calculations) to explain cognitive overload, bias, and why AI research tools can overload working memory in minutes.
The human brain has about 100 billion neurons and 100 trillion synaptic connections Used to frame the brain as a trainable neural net shaped by repeated informational input.
There are 28 known flow triggers: 12 on the individual side, roughly 16 on the group-flow side Core framework Kotler uses to train flow states.
Kotler has published 16 books and an estimated 5 million published words (plus a roughly equal amount unpublished) Attributed to his page-a-day writing discipline.
The Alliance is an 8-month program capped at 100 people, including three live events and six six-week training cycles Program structure for Kotler's new creative community, launching June 2025.

Worth digging into

🕳️ The 'Universe 25' rat-utopia experiment

Used by Diamandis as a cautionary parable that unlimited abundance without challenge leads to total population collapse -- a load-bearing argument for his concern about AI/robotics making life 'too easy,' but the study's validity and relevance to humans is scientifically contested.

🕳️ The AI-vs-physician diagnostic accuracy study (75%/80%/95%)

Kotler cites this as evidence AI alone can already beat human-plus-AI teams on convergent diagnostic tasks, a striking and counterintuitive data point referenced from his episode with Salim Ismail.

🕳️ Dean Hamer's VMAT2 'God gene' research

Cited as the genetic basis for why some people have more intense spiritual/mystical/flow experiences via higher baseline dopamine, norepinephrine, and serotonin -- a specific, checkable genetics claim from 1990s NIH research.

🕳️ Rick Strassman's continuous-IV DMT study and 'alien encounter' reports

Diamandis and Kotler describe University of New Mexico hospital subjects on continuous IV DMT reporting encounters with 'gray men' rather than the more typical 'machine elves' of recreational DMT reports, attributing the difference to clinical setting.

🕳️ The Harvard anxiety-reframing study ('I am excited')

Cited as showing that saying 'I'm excited' out loud three times reduced performance anxiety more effectively than seven minutes of breathwork, with a notable side-finding that women over 40 in the study had trouble even recognizing excitement as distinct from anxiety.

🕳️ The same-day psychophysiology paper on exercise mindset

Kotler references a paper, published the same day as this recording, showing people who love exercise process it in a measurably different brain region than people who don't -- and that this can shift, linking mindset directly to longevity behavior.