Practice Ground

Chanin Nantasenamat is known online as The Data Professor — the educator who turned a YouTube channel about data science into a pathway for millions of learners who never thought data science was for them. He's a former associate professor at a research university in Thailand with nearly two years of experience applying machine learning and bioinformatics to drug discovery. He's now a senior developer advocate at Snowflake and Streamlit. And he launched a course with Nifemi at Deep Learning AI on fast prototyping of GenAI apps.

He came from biology. He had to fight to learn to code. That background is the foundation of everything he teaches.

In this conversation, we explore the philosophy of learning by doing, why the researcher's tolerance for failure maps perfectly onto entrepreneurship, and why having an anchoring project might be the most practical advice for surviving the AI era.

We discuss:
• [4:10] "The best way to learn data science is to do data science" — learning by doing
• [8:30] Why prototyping is about converting thought into something tangible
• [12:20] Intuition builds through repeated experience — gut check backed by data
• [18:45] Embrace imperfection: good enough and shipped beats perfect and stuck
• [22:10] Using a checklist to know when your prototype is production-ready
• [28:30] The teacher as facilitator, not authority — co-evolving with AI • [33:50] Why coming from biology gave him empathy with technical beginners
• [38:15] PhD failures and what nearly a year of negative results taught him
• [44:20] Risk minimizers, not risk takers: the entrepreneurship insight from Stanford
• [50:10] How AlphaFold changed biology and why we're still just scratching the surface
• [56:30] Keeping a project as an anchor when the AI landscape shifts every week
• [60:20] IKigai, purpose, and what it means to democratize learning at scale
• [64:40] AI as a co-founder for immigrants navigating complex systems • [68:00] Daily practice: touch grass, human connection, keep learning

Chanin's subscriber once watched his tutorials, went to Berkeley, got a job at big tech, built a startup, and sold it. Chanin didn't plan for that. He just kept making content — learning in public, doing the work, staying in the game.

That's the whole lesson.

This is episode 12 of Practice Ground — exploring how we find meaning as we go deeper into the automation era. These conversations are shaping a book coming in 2026.

Richmond Alake is the Director of AI Developer Experience at Oracle, a former MongoDB developer advocate, and one of the most prolific AI educators you've probably never heard enough about.

He has over two hundred articles on Medium, has written for NVIDIA and Neptune AI, taught at Imperial College and O'Reilly, and built openspeech — an AI-powered learning partner. He's also the founder building agent memory tools on top of Oracle's database infrastructure. He grew up in Lagos, Nigeria, moved to London at ten, and built his entire career by learning in public.

In this conversation, we explore why agent memory is the foundation that will determine how intelligent AI systems actually become, the honest psychology behind relentless curiosity, and what it means to walk into rooms where no one looks like you — and stay anyway.

We discuss:
• [12:30]Why agent memory maps directly to human memory: episodic, semantic, procedural, entity
• [8:20] The progression: website → app → agentic — and why memory is what makes agents real
• [18:40] Nigerian excellence culture: "Do they have two heads?"
• [22:15] Learning in public as a career strategy — 200+ articles, O'Reilly, Imperial College
• [38:50]Why AI always converges — and why deep beats broad

• [45:20] The case for no memory: when forgetting is a feature
• [52:10] Writing as a secret weapon in corporate strategy
• [56:30] Representation: seeing one VP at Google who looked like him, and what that changed
• [60:40] Daily practice in utopia: space exploration and an ever-expanding universe

Richmond's philosophy is simple: pick one thing, go deep, trust that AI will converge toward you.

He chose agent memory at MongoDB. He's still in it at Oracle. The field came to him.

This is episode 10 of Practice Ground - exploring how we find meaning as we go deeper into the automation era. These conversations are shaping a book coming in 2026.

Josh Starmer is the founder of StatQuest and an assistant professor at the University of North Carolina.

He's built one of the most beloved statistics education channels on YouTube by taking the most hostile possible audience — coworkers who literally hated math — and figuring out how to make them understand anyway.

He plays cello, has performed on big stages with rock bands, writes songs, and has a childhood dream of one day opening a pizza place. He started his YouTube channel with cooking videos, got nine views in the first year on his first statistics video, and thought that was amazing.

In this conversation, we explore where great teaching actually comes from, why pictures beat words every time, and what it felt like to finally understand a neural network not just intellectually — but emotionally.

We discuss:

• [4:10] The hostile audience as the ultimate teacher-training tool

• [8:30] Why trust through explanation opened every professional door Josh walked through

• [14:20] How StatQuest started with cooking videos, music, and nine statistics views

• [28:40] The slow learner's strategy: wait a year to see what actually lasts

• [35:15] How pictures communicate what words can't

• [40:50] Creating conditions for the aha moment — teaching neural networks visually

• [46:30] The neural network video that felt more beautiful than any song

• [52:10] What an AI-proof teaching career might look like: live shows, human connection

• [56:20] The pizza place dream, jogging with friends, and daily practices that matter

Josh's best observation: he's not a rock star. He plays statistics to growing crowds instead. And he's never once felt stagnant in that work the way he did playing the same songs night after night. That's the signal — the work that never stops teaching you is the work that's worth keeping.

This is episode 9 of Practice Ground — exploring how we find meaning as we go deeper into the automation era. These conversations are shaping a book coming in 2026.

Sanmi Koyejo is an assistant professor of computer science at Stanford University, where he leads the Stanford Trustworthy AI Research group.

He's also the founder and chief AI officer at Virtue AI, and is writing a book on machine learning from human preferences. He grew up in Lagos, Nigeria — and if you've been following Practice Ground, you may recognize International School of Lagos as the same school I attended. We also trained at the same boxing gym in Austin for years, without fully realizing how far our paths would go.

In this conversation, we explore what trustworthy AI actually means technically, why measurement systems built for academic use are failing at the societal level, and what role serendipity really plays in every meaningful career.

We discuss:

• [5:15] Why the AI field is great at making a number get bigger — and why that's now a problem

• [8:30] What AI measurement was built for vs. what society is demanding from it now

• [12:40] The shocking reality: 50% data/engineering, 10% algorithms we actually teach

• [18:25] Why humans will always be in the loop with complex AI systems

• [28:10] What trustworthy AI actually means — calibrated trust, not blind trust

• [32:50] Fairness, robustness, privacy, security: the four directions of trustworthy AI

• [42:15] Exploration vs. exploitation in research vs. business

• [48:30] Serendipity, fire alarms, and the career moments you can't predict

• [54:40] Why you're probably better off not doing a PhD — unless you love exploration

• [58:20] Daily practice in utopia: a researcher freed to explore and think broadly

Sanmi's candor is rare: he'll tell PhD students to their faces that they might be better off not doing a PhD.

He'll admit that data and engineering dwarf the algorithms he loves. And he'll acknowledge that the measurement systems he's built most of his career on weren't designed for the decisions society is now trying to make with them. That kind of intellectual honesty is exactly what this era needs.

This is episode 8 of Practice Ground — exploring how we find meaning as we go deeper into the automation era. These conversations are shaping a book coming in 2026.

Yemi Amu is the founder and director of Oko Farms — New York City's first and only public accessible outdoor aquaponics farm, which she established in Brooklyn in 2013.

She's an aquaponics expert, educator, and cultural steward who grew up in Lagos, Nigeria, came to New York for undergrad (supposed to do law school and go home), and found her calling not in law, music publishing, or sci-fi writing, but in the soil.

Her farm grows Nigerian vegetables, fiber crops, indigo, and medicine plants — not just as food, but as living acts of cultural reclamation and ancestral practice.

In this conversation, we explore what it means to create the world you want to live in, why Yemi's farming is inseparable from her Yoruba heritage, and what happens when human instinct and embodied knowledge meet a world increasingly in love with automation.

We discuss:

• [3:45] Why caring for the environment is a spiritual practice, not an ideology

• [8:30] The winding path from Lagos to law school to vegan baking to aquaponics

• [16:20] The conflict between Nigerian immigrant expectations and finding your own meaning

• [28:40] Farming as Yoruba ancestral practice and cultural reclamation

• [33:15] What growing Nigerian vegetables taught her about identity and diaspora

• [44:50] The eighty-year-old teacher who said: tell the pH by smelling the water

• [47:30] Why technology must be farmer-led, not the other way around

• [51:10] "We're being sold something as progress that I'm not sure is progress"

• [57:20] Her daily practice in a utopian world: mornings of dreaming, praying, and growing for joy

Yemi has a phrase she lives by: she would grow purely for the joy of growing. That's not a metaphor. It's her test for whether something is worth doing. And it's one of the simplest answers to the meaning question I've heard in this entire series.

This is episode 7 of Practice Ground — exploring how we find meaning as we go deeper into the automation era. These conversations are shaping a book coming in 2026.

In this conversation, we explore what most teams get wrong when building agentic systems, how to distinguish between problems that need agents and problems that just need a script, and what Tony's obsessive daily practice reveals about finding meaningful work.

Tony Kipkemboi grew up in Kaptagat, Kenya — a small town that produces an extraordinary share of the world's greatest distance runners. He came to the US on a running scholarship, served seven years in the US Army (including frontline COVID-19 research at USAMRIID), pivoted into tech by teaching himself Python during late nights while his family slept, earned a master's from Penn, worked at Bloomberg, Booz Allen, Snowflake, and Crew AI, and is now at Guild helping close the AI talent gap. On YouTube, he's known as The How To Guy, where he teaches technical builders how to design automation and agent systems.

We discuss:

• [3:20] Why "agentifying everything" is the wrong approach to building automation

• [6:45] Workflow first, agent second - the right order for building agentic systems

• [9:10] Why AI is a magnifying glass on your data infrastructure

• [22:30] From Kaptagat farm boy to NCAA runner to US Army to AI builder

• [38:50] The two-door strategy for navigating career pivots

• [42:20] How obsession, not ambition, is the real signal of meaningful work

• [50:15] Teaching people to use AI as a thought partner, not a crutch

• [58:30] Why skills and community are the anchors in any future — utopian or dystopian

Tony's story is a masterclass in relentless pursuit. He didn't start with a plan - he started with a door. When it didn't work, he found another. And when he finally found the thing where his off day is when he's asleep, he went all in.

This is episode 7 of Practice Ground — exploring how we find meaning as we go deeper into the automation era. These conversations are shaping a book coming in 2026.

Michael Hendrix is a writer, author, designer, musician, collage artist, and educator. He spent decades at IDEO, where he pioneered the concept of design sensibilities - the idea that emotional intent, not just methodology, is what separates great design from average design. He's the co-author of "Two Beats Ahead," a book exploring what music can teach the creative process. He now lives in Iceland.

In this conversation, we explore how emotional intent transfers from maker to receiver, why constraints are essential to creative freedom, what AI actually produces (and why it's almost guaranteed to be mediocre), and what principles will give humans meaning no matter what the technology does.

We discuss:

• [5:20] How emotional intent transfers from maker to receiver in design and music

• [12:10] The difference between design methodology and design sensibility

• [18:30] Why constraints are a creative strategy, not a limitation

• [27:15] The thing I'm always looking for in a product: was it made with care?

• [34:40] Why AI gives you the most average thing - because it's based on averages

• [37:55] What lifelong creatives celebrate more than the outcome

• [44:20] Fiction vs. non-fiction and long-term vs. short-term truths

• [50:10] Relationships, awe, and what gives humans meaning across all eras

• [56:35] His daily practice: making, teaching, telling stories about the world

Michael's central idea: you can feel whether something was made with care. That care shows up in the seams, in the materials, in whether the maker understood what they were really making.

It's as true for a sneaker as it is for a song. And it's the one thing AI - optimized for averages - cannot replicate.

This is episode 6 of Practice Ground -exploring how we find meaning as we go deeper into the automation era.

These conversations are shaping a book coming in 2026.

John Small is a writer, editor, journalist, and the host of Small Talk a Substack newsletter, podcast, and media business about Gen X culture, creativity, and living with intention.

In this conversation, we explore why AI will never replicate the human magic that makes a piece of writing worth reading, what it means to finally write for yourself after three decades of writing for others, and the only reliable way to find meaning in a creative idea.

We discuss: 

• [4:20] Why the writers who fear AI are the ones who've never actually used it
• [9:45] The fake video effect: why people tire of knowing something isn't real
• [13:30] What the magic is - and why AI can never generate it
• [18:20] Writing as a drug: the flow state and why it's so hard to protect
• [28:10] Why your best creative work happens when you write for yourself first
• [34:50] How to find meaning in an idea (hint: it's never in your head)
• [42:15] Finding your voice as a writer - why it's not a moment, it's a process
• [48:30] Morning pages, walking, and putting the phone away
• [54:10] Why books will survive - and why reading trains your attention
• [58:40] Daily practice in utopia: pure aesthetic experience, no agenda 

This is episode 5 of Practice Ground - exploring how we find meaning as we go deeper into the automation era. These conversations are shaping a book coming in 2026.

Joe Reis is the co-author of the best-selling "Fundamentals of Data Engineering" and a polymath who refuses to be put in a box. He's a data engineer, advisor, entrepreneur, DJ, rock climber, writer, and someone who wakes up every morning to have AI agents build things while he reads.

He's launching a publishing company and writing his second book, and he believes we're living through the most exciting creative moment in history.

In this conversation, we explore the era of abundance for creators, what happens when the barrier to making things drops to near-zero, and what remains distinctly human when AI can match your style.

We discuss:
• [4:20] Why everyone's back at the starting line in the AI era

• [8:45] The challenge of knowing when to stop building with AI

• [15:30] His daily practice: AI agents building while he makes coffee

• [28:15] Why distribution and attention are the scarcest currencies

• [35:40] What's left for humans when AI can write in your style

• [38:20] How experience becomes your data, and writing is just the tool

• [45:10] Finding "true north" when everything keeps changing

• [50:25] Why he's choosing optimism over dystopia

• [53:40] His daily practice: random acts of kindness and community

Joe's perspective is refreshing: the bar for quality has been raised, not lowered. AI doesn't replace human creativity - it forces us to inject more of ourselves into our work. Your experiences, the things in your head that only you have witnessed, that's your data. Everything else can be automated.

This is episode 4 of Practice Ground-exploring how we find meaning as we go deeper into the automation era. These conversations are shaping a book coming in 2026.

SUBSCRIBE FOR MORE: New episodes every two weeks.

GET NOTIFIED ABOUT THE BOOK: https://naporepublic.substack.com/

Jason Mayden is an industrial artist and designer who's spent his career creating products as "future ancient artifacts" - objects meant to be studied by future civilizations. 

He works with Nike and world-class athletes, but his real work is cultural architecture: documenting brilliance, reclaiming heritage, and ensuring his people's contributions can't be erased from history.

In this conversation, we explore why he creates for centuries instead of seasons, how AI amplifies rather than replaces human intuition, and why love remains the only technology that truly matters.

We discuss: 

• [3:45] Why fear of embarrassment kills more creativity than failure ever will 

• [18:20] Creating products as "future ancient artifacts" worthy of being remembered 

• [12:35] How Black people adorn themselves as an ancestral act of showing regality

• [32:15] Using AI as a prototyping tool to amplify intuition, not replace it 

• [25:40] Why Nipsey Hussle's philosophy shaped his approach to staying true 

• [38:50] The daily gratitude practice that keeps him grounded 

• [42:10] Why love and human connection will outlast any technology 

Jason's work isn't about being relevant right now—it's about being remembered forever. He's bridging past and future, creating with conviction, and proving that the most important technology we'll ever have is our ability to feel, give, and receive love.

This is episode 3 of Practice Ground - exploring how we find meaning as we go deeper into the automation era. These conversations are shaping a book coming in 2026.

FOLLOW JASON MAYDEN: • https://www.linkedin.com/in/jasonmayden/

https://www.instagram.com/jasonmayden/

SUBSCRIBE FOR MORE: New episodes weekly

GET NOTIFIED ABOUT THE BOOK: https://naporepublic.substack.com/