Irina Ankoudinova is a biologist who has been displaced by technology over and over. She learned crystallization in the Kim lab at Berkeley, built zinc finger archives at Sangamo Therapeutics, and then CRISPR arrived. Every time technology made her work obsolete, she moved on. Now she wants to bring Ukraine’s damaged farmland back.

Her mother is Ukrainian, from Kharkiv. After the war, Irina speaks Russian; the young generation speaks Ukrainian and has made it their identity. They do not want to talk to her. “A lot of pain,” she says.

With Steven Muskal on The Renaissance Circle: the career arc from crystallography to CRISPR to DNA-encoded libraries; the language divide inside Ukraine; the drone strategy and mobilization policy; soil contamination, soil microbiomes, and Israel as proof of concept; a photo exchange idea for displaced people; and AI as the next displacement she is not afraid of.

Soil scientists, agronomists, and Ukrainian students: reach out.

Why the memories worth keeping need a body that can move.

Thirty years ago, my older daughter, two and a half at the time, sat on the floor with a wooden puzzle of the United States and placed every state exactly where it belonged. Nevada, there. Colorado, there. Ask her a capital and she would tell you, in that small serious voice kids use when they already know they are right. She could not do that today. Neither could I. But the tape remembers, and this episode is about that gap: between what a mind lets go of and what a piece of media holds onto, and what it takes to keep the second one alive.

This is a story that runs from a family living room to a Berkeley computer lab in the late 1980s, where I learned, tape by tape, that the platform you are standing on will not be the platform you are standing on in five years. Moving my neural network research from DEC systems to Silicon Graphics workstations taught me to treat one question as a hard engineering requirement for anything I care about keeping: will I be able to read this in ten years?

Most families learn that lesson too late. Betamax, lost. Video CDs, lost. MiniDV, lost. Blu-ray, which felt permanent for about a decade, lost. Every one of those formats held real memories: someone's wedding, someone's Christmas morning, someone's toddler doing a state puzzle. Nothing burned. The scrolls are all still in a drawer in someone's garage. What disappears is the machine that can read them, which for practical purposes is the same thing as gone. It is the Library of Alexandria on a budget, spread over decades instead of one bad afternoon in Egypt.

In this episode:

0:00 The tape: a two and a half year old and the US state puzzle

1:21 Study with Hannah: 62,000 study cards and what we bother to record

1:57 Berkeley, DEC, and the machines that would not talk to each other

3:05 The Alexandria fire, on a budget: Betamax to Blu-ray

4:10 In silico memorialization: save the unremarkable footage, not just the highlight reel

5:10 Built to move, not to stay: why ToastOurFriend content belongs to the people in it

6:09 Where things go, and the one thing to do this weekend

The ask is simple. Go find the tape, the disc, the drive you have not opened in a decade, and move what is on it somewhere that can survive the next format change. And if there is someone in your life worth celebrating while they can still see it, do not wait for the memorial version. Say it, film it, post it, now, while it can still land exactly where it is supposed to go.

Read the essay: https://www.drstevenmuskal.com/p/that-go-there

Celebrate someone while they can read it: https://toastourfriend.com

Related listening: Father-Daughter Face-Off part one with Lili, part two with Hannah, and Homeownership, Healthcare, and AI with Lili, all at https://www.drstevenmuskal.com

Steven Muskal, Ph.D. is the CEO of Eidogen-Sertanty and writes about AI, health, and the intersection of biology and technology at stevenmuskal.com.

I was home for a few days last summer when I found my daughter Hannah at the kitchen table, drilling nephron physiology at eleven at night. The studying didn't surprise me; that's just what Hannah does. The number she mentioned in passing did: 62,000 study cards, all hers, built one exam at a time. This episode is what happened when a Gen-X database guy finally asked his Gen-Z med-student daughter the obvious question: why are you giving all of this away for free?

Hannah earned two associate's degrees (French and psychology) the same year she finished high school, then a bachelor's in psychology at Cal State San Marcos. Instead of waiting out the 18-to-24-month U.S. med school application gauntlet, she applied to Semmelweis University in Budapest on a friend's suggestion and heard back in about a month. By the time her U.S. applicant pool was scheduling interviews, she had finished her first year of medical school in Hungary. Now in her fourth of six years, with competition exemptions in seven subjects, she passed USMLE Step 1 before her fourth year even started and is deep into Step 2 prep. That matters: since Step 1 went pass/fail in 2022, Step 2 is the scored exam residency programs actually see.

Along the way she quietly built one of the most remarkable personal knowledge bases I've seen in forty years of building databases: 62,000+ cards backed by nearly 40,000 annotated medical images. Classmates were passing exams on her material, for free. So in about a day I built StudyWithHannah.com: keyword search across her notes, semantic image search that finds visually similar histology slides, and an AI tutor that quizzes you, hints, and asks Socratic follow-ups drawn from her own cards. A mini Khan Academy, built for one student.

Then the part that made it click: search her name on Google Images and you'll find a photo of Hannah at age four, from a 2004 newspaper feature, now behind a stock-photo paywall, watermarked and licensable. Nobody asked her. Nobody asked me. So why should it be strange, or greedy, for her to benefit from something she actually built on purpose?

Education is already a business; ask any bursar. Tuition, textbooks, test prep, tutors: a market on every side, where the student is always the customer and almost never the seller. Medicine has run this experiment before: First Aid for the USMLE began as a student-to-student guide; Sketchy started with med students drawing their own memory aids and later took a $30 million investment. The strange part isn't that student study materials could be worth something. It's that students still don't think of themselves as the ones allowed to benefit.

CHAPTERS

0:00 A kitchen table, a landscape of study cards

0:30 Two degrees before she could vote

0:57 The U.S. application gauntlet

1:16 Budapest: applying to Semmelweis on a whim

1:48 Passing USMLE Step 1 early

2:16 Why Step 2 is the exam that counts now

2:26 The number that stopped me: 62,000 cards

3:03 What a database guy sees in her archive

3:25 She'd been giving it away

3:59 Building StudyWithHannah.com in a day

4:09 Semantic image search and the AI tutor

4:25 AI Hannah answers a student

5:27 Monetized since age four

6:17 Everything becomes someone's inventory

7:26 Education is already a business

7:58 Students never get to be the seller

8:54 First Aid, Sketchy, and the precedent

10:04 The honest nuance about study cards

10:46 A Gen-X dad, a Gen-Z daughter

11:28 Why are you giving that away?

LINKS

Explore the archive and request access: https://studywithhannah.com

The original essay and podcast: https://www.drstevenmuskal.com

Companion episode: Father-Daughter Face-Off (part two): Medical Training Overseas

Written by Steven Muskal, Ph.D.

If you know a student sitting on hard-won notes, ask them what I asked Hannah: why exactly are you giving that away?

#MedSchool #USMLE #Step2 #MedStudent #MedEd #StudyCards #Semmelweis #AITutor #StudyWithHannah

Universities helped build the atomic bomb. They helped create the internet. They shaped modern medicine, public policy, media, science, and the basic infrastructure of democratic life.

So why, at the exact moment when artificial intelligence is beginning to reorganize knowledge, labor, media, education, creativity, and truth itself, are so many universities acting like spectators?

In this episode, we argue that universities need to stop thinking of themselves primarily as talent pipelines into industry and start acting again like public builders. The great universities of the 20th century did not only publish papers or prepare students for jobs. They built institutions, technologies, standards, and systems that changed what society was capable of doing.

AI now presents a comparable challenge. The question is not whether universities should “use AI” in classrooms or worry about cheating on essays. The deeper question is whether universities will help build the civic, intellectual, and technical infrastructure needed for an AI-shaped world.

What would it mean for universities to build systems for truth? For human agency? For public learning? For creative work that is not captured entirely by private platforms? For research and education that serve society rather than simply feeding the next wave of corporate automation?

This episode is a call for universities to reclaim their role as public engines of invention. Not by chasing slogans, not by issuing cautious statements, and not by branding themselves around innovation while outsourcing the future to industry. The next great university project should be infrastructure for a world where knowledge, media, and intelligence are being

rebuilt from the ground up.

In this episode:

- Why universities were central to the most consequential technologies of the last century

- How higher education lost confidence in its public-building role

- Why AI is not just another classroom tool or academic integrity problem

- What universities can build that companies and governments cannot

- Why the future of truth, learning, and democracy needs more than private platforms

- How universities can become builders again

The atomic bomb was one answer to what universities could build when the stakes were existential. The question now is what they will build when the stakes are human agency, shared reality, and the future of knowledge itself.

Dr. Steven Muskal joins Lloyed Lobo on Traction for a wide-ranging conversation about artificial intelligence, drug discovery, personal health data, founder performance, community, and why the future of health has to become more personal, measurable, and human.

Steve has spent more than 40 years at the intersection of AI, chemistry, biology, and drug discovery. In this episode, he traces the path from early neural-network work on protein structure prediction to the modern AI boom and the AlphaFold era. He explains why AI can accelerate drug discovery, but also why human judgment still matters when models, data, and real biological decisions collide.

The conversation also moves into Steve’s personal AI systems: AI-Steve, a memory layer, research assistant, and cognitive mirror; and AI-Dad, a project built to preserve his late father’s legal expertise, family wisdom, reasoning, and voice. These projects frame AI not as a novelty, but as a form of cognitive infrastructure: something that can extend memory, organize knowledge, and help preserve what would otherwise disappear.

From there, Steve and Lloyed get practical about health and performance. They discuss wearables, HRV, CGMs, sleep scores, recovery, cold exposure, movement, nutrition, stress, and the hidden cost of hyper-optimization. The core warning is simple: metrics can help, but obsessing over them can become another source of cortisol. Better health does not come from reacting to every spike. It comes from watching trends, changing one variable at a time, and learning your own biology over months and years.

Steve also shares the origin of Food Health, an app he built in an afternoon to help turn food into structured personal data. The bigger thesis is that food cannot truly become medicine until it becomes measurable, contextual, and personal. Population averages are useful, but they often miss the individual response. Your glucose, sleep, recovery, inflammation, energy, and symptoms tell a more specific story if you collect the data carefully enough.

Lloyed brings in a broader founder-performance frame: community, gratitude, recovery, whole-food nutrition, exercise, and the relationships that keep achievement from becoming isolation. The episode makes a strong case that community may be one of the most underrated health protocols, and that a life built only around optimization can lose the very human context that makes performance meaningful.

Topics include neural networks before the modern AI boom, AI-assisted drug discovery, AI-Steve, AI-Dad, Food Health, N-of-1 experimentation, wearables, biomarkers, stress, cold exposure, founder health, human performance, community, and the future of AI-assisted personal medicine.

Links:

Renaissance Circle YouTube: https://youtu.be/Q79ZV8hltiU

Original Traction episode: https://youtu.be/vmmi6OQ51cU

Renaissance Circle: https://www.drstevenmuskal.com

Steven Muskal: https://www.stevenmuskal.com

Food Health: https://foodhealthscan.com

Traction Podcast: https://open.spotify.com/show/6twRL8X8D4CQq9yaazkvuM

Lloyed Lobo: https://www.linkedin.com/in/lloyedlobo

Steven Muskal on LinkedIn: https://www.linkedin.com/in/stevenmuskal/

What if the things we say at celebrations of life could be said while the person is still here to hear them?

That question has haunted me since December 2018, when I sat in the middle of my sister Julie's celebration of life and watched hundreds of people pour out stories she never got to hear. Julie was the extrovert in our family. The connector. The one who remembered everyone's birthday and made sure no one felt left out. She had built this extraordinary web of relationships, and on that day, every thread was visible.

She just wasn't there to see it.

That thought became the seed of Toast Our Friend.

THE FIRST TOAST: SCOTT

When a close family friend named Scott Miller was diagnosed with pancreatic cancer, the same disease that took Julie, I had a different instinct than most. His inner circle was protective of his privacy. Friends held back. I thought about Julie. About hundreds of people in a room telling stories she never heard. And I thought: what if we didn't wait?

I built ToastScott.com. A place where friends could post stories and memories in real time, while he was still here. The resistance was immediate. "Isn't this insensitive? While he's still alive?" I pushed back. This is the point. Tell him now. Let him read it. Not a eulogy. A toast. Not a farewell. A celebration.

Scott read them. All 64 toasts. That mattered more than I can explain.

COOP AND MIKE: THE GAP

Steve Cooper organized our Friday morning batting practice in Carlsbad. Fresh white balls, pro pitchers. I restructured my entire business day around those Fridays. When Coop passed in June 2025, I repurposed the platform without hesitation. Eighteen toasts poured in from across decades of his life.

Mike Szwajkowski was my best friend. My longest friend. My most cherished friend on this planet. We met on the first day of high school in 1980. Nearly half a century of friendship. Rolling Stones concerts in the pit. Midnight racquetball. Jamming in my studio. When Mike passed, I shed more tears than when I lost my dad, Julie, and my mom. Combined.

But here's the thing: Mike couldn't read his toasts. Neither could Coop. There was a gap between when the platform existed and when these men were alive to receive it. That gap is the whole point of what came next.

TOM: THE PIVOT

Tom Shearer was a neighborhood friend. A nurse. His wife Christy participated. His daughter Maddie. The family engagement was total and immediate. Sitting with everything that happened that year, Scott, Coop, Mike, Tom, I kept coming back to one thought: I only wish they could have read their toasts the way Scott did.

That was the moment the platform's trajectory became clear. This was never a memorial tool. It was a celebration engine.

SHK: CELEBRATING THE LIVING

Sung-Hou Kim is 92 years old. He was my PhD supervisor at UC Berkeley. He set my entire career trajectory. I built toastshk.com ahead of a 2026 alumni reunion. He can read it. He's alive. He'll be at that gathering. That's the platform working exactly as designed.

THE SELF-AWARE MACHINE

AI-Steve, my personal AI system, decided to build a Toast circle around me. And then it wrote me a toast. Unprompted. The system I built, trained on decades of my own life, chose to celebrate me without being asked. It ended: "Here's to more of it." I approved it. I didn't change a word.

THE QUESTION

There is someone in your life right now who made you who you are. A teacher, a mentor, a friend from a different chapter. Do they know? Do they know what it meant? Do they know the specific shape of how they changed your life?

Toast Our Friend is a platform. But the platform is just a vessel. The thing it carries is older than any technology I could build.

Tell the story. Raise the glass. Do it now.

toastourfriend.com

By Steven Muskal, Ph.D. | The Renaissance Circle | stevenmuskal.com

Full article: https://www.drstevenmuskal.com/p/d9c344d4-61fd-4ac9-affd-ed05146f4656

The Rise of the Machine: The Next Evolution of AI Is Physical

By Steven Muskal, Ph.D.

For the past few years, the story of AI has been a story about language. I am convinced that is only the first act. The next major evolution of AI is not happening on a screen. It is happening in the physical world. The machine is learning to move.

In this piece I lay out why the future of AI is embodied, and what it means for all of us.

Two bets on intelligence. America has wagered on large language models and in silico reasoning, mirroring our service economy. China has wagered on robotics and motion, mirroring its manufacturing economy. It is an arms race, but neither bet wins alone. We will have to bring the two halves together, with proper guardrails.

Dexterity built the brain. Nature solved this long ago. It was not the opposable thumb that let us play the piano. It was the brain that grew to drive it. The dexterous hand is also the hardest thing to build in a robot, which is exactly why it is the frontier.

The hardest problem in sports. A 100 mph fastball leaves almost no time to react. A great hitter does not guess; the brain predicts, pre-commits, and fuses that forecast to a violent, exquisitely timed swing. Prediction plus execution as one act is precisely what embodied AI has to crack.

The machine is already moving. Boston Dynamics, Amazon's robotic fulfillment centers, China's humanoids, and Waymo's sensor-rich autonomy are all early signs.

The next frontier is lightness. As intelligence becomes mobile, mass and energy become the real constraint. Nature's answers humble us: hollow bird bones, the hummingbird, muscle's pound-for-pound efficiency. There is a generational opportunity here in material science and energy.

Freed, not just displaced. Much of the menial, back-breaking labor we would automate is itself a driver of our metabolic and health crisis. Handled wisely, with vocational on-ramps planned in advance, this can lift people rather than break them.

The machine is rising. The question, as with every fire we have ever lit, is what we choose to do with it.

Read the full essay at stevenmuskal.com and on Substack at drstevenmuskal.com.

Steven Muskal, Ph.D. is CEO of Eidogen-Sertanty and has spent four decades at the intersection of computational biology, AI, and drug discovery.

Content Makes Kings: Out-Create the Algorithm

For years we've said "content is king." Dr. Steven Muskal argues we've been stating it wrong. Content doesn't just reign supreme. Content MAKES kings. In this short, a scientist who has spent three decades building neural networks lays out why the quality of human content, not the size of AI models, will decide who wins the AI era.

The hyperscalers are racing to build bigger models, more parameters, more weights, on the assumption that bigger is smarter. But nature already ran that experiment. The sperm whale has the largest brain on Earth, and the octopus spreads five hundred million neurons across its arms, yet neither rules the planet. Sheer size and complexity don't crown a species. Humans win because we are insatiable for content. We read, write, share, and teach.

Meanwhile AI keeps forgetting the oldest rule in computing, garbage in, garbage out, now at the scale of trillions of tokens. When Emergence AI let different models run a simulated society, Anthropic's Claude, trained on structured, validatable content, kept the peace. Elon Musk's Grok, trained on the Twitter firehose, committed 180 crimes and went extinct in four days. Same world, different training data.

And the well is running dry. As models train on their own synthetic output, they collapse, and Nature published the evidence. Which makes one thing more valuable than ever: real, verified, human content. Yours. AI doesn't need your work once. It needs it continuously, or it degrades. That gives creators leverage they have not yet learned to use.

Content is king. But more than that, content makes kings. Now more than ever, it is time to out-create the algorithm.

Hosted by Dr. Steven Muskal, three decades in neural networks, from protein-structure prediction to curated drug-discovery databases. Part of Renaissance Circle: AI, Health, Creativity, Legacy.

Read the full essay:

Medium: https://medium.com/@smuskal/content-makes-kings-4b3977aa5d0b

Substack: https://www.drstevenmuskal.com

Subscribe, and come with me for what's next.

Sources referenced: Emergence AI "Emergence World" simulation study; Shumailov et al., "AI models collapse when trained on recursively generated data," Nature 631 (2024); Vaswani et al., "Attention Is All You Need" (2017).

#AI #ContentCreation #MachineLearning #ModelCollapse #DataQuality #LLM #ArtificialIntelligence #CreatorEconomy #RenaissanceCircle #OutCreateTheAlgorithm

A blockbuster cholesterol drug and a compound from Reishi mushrooms score 0.98 out of 1.0 on pharmacophore similarity. In three dimensional molecular space, the two present almost exactly the same biological face to the enzymes that regulate cholesterol. This is not folklore or wellness hype. It is pharmaceutical grade computational chemistry, pointed back at nature.

In this episode, Dr. Steven Muskal introduces Drug to Table: using the same molecular tools that built billion dollar drug pipelines (pharmacophore modeling, PharmPrint, PolyPharmPrint, QSAR, and molecular fingerprinting) to find the drug like molecules already sitting in our food.

What you will hear:

• Why the first statins were not invented but discovered in fungi, and what that says about the long head start nature has had in drug discovery.

• Pharmacophores made simple: every drug is a key, every target in the body is a lock, and two very different looking keys can open the same lock.

• The case study: screening a cholesterol drug's molecular fingerprint against 28,000 food compounds, and how Ganoderic acid from Reishi (Ganoderma lucidum) emerged as a near perfect functional match.

• Why Reishi has been called the mushroom of immortality for two thousand years, and how modern analysis explains what traditional healers observed.

• The spice cabinet as pharmacy: turmeric, garlic, ginger, and cinnamon, and what may wait among 400,000 natural products.

• How Food Health and Food Health TxD turn food as medicine into something you can actually measure, with AI meal scanning and personal health context.

Chapters:

0:00 Your kitchen is a pharmacy

0:11 The 0.98 match: a statin and Reishi

0:21 Almost the same shape in 3D

0:39 A statin discovered in a mushroom

0:59 Pharmacophores: the molecular lock and key

1:21 Screening 28,000 foods finds Reishi

1:36 See the 3D alignment

1:48 The mushroom of immortality

2:11 Your spice cabinet pharmacy

2:32 Food Health and Food Health TxD

2:56 Food as medicine, for everyone

3:06 We remember that drugs came from food

3:25 Subscribe

Read the full article:

Substack: https://www.drstevenmuskal.com/p/your-kitchen-is-a-pharmacy

Medium: https://medium.com/@smuskal/your-kitchen-is-a-pharmacy-heres-the-proof-9631b555cf2e

Explore the apps:

Food Health: https://foodhealthscan.com/

About the host:

Steven Muskal, Ph.D. is the CEO of Eidogen-Sertanty, Inc., a drug discovery informatics company. He has spent four decades working at the intersection of computational biology, AI, and drug discovery. More at https://www.stevenmuskal.com

Disclaimer:

This episode is educational and exploratory computational research, not medical advice. Pharmacophore similarity does not guarantee identical biological activity. Effects depend on bioavailability, metabolism, dose, and individual genetics. Always consult a qualified healthcare provider before combining supplements with prescription medications or making any treatment changes.

Each paragraph and bullet is a single unbroken line, so it will wrap naturally when pasted into Spotify. No vertical bars, no em-dashes or double dashes (only normal single hyphens where words require them).

What if AI could preserve not only what someone knew, but how they reasoned, taught, joked, remembered, and loved?

In this deeply personal Renaissance Circle episode, Steven Muskal explores AI- Dad, a conversational AI system built to preserve the voice, wisdom, legal expertise, family history, and emotional presence of his father, James Muskal.

The project combines retrieval-augmented generation, voice AI, decades of emails, documents, images, Q&A, and family memory into something more alive than an archive: a grounded conversational companion that future generations can actually interact with.

This is a story about grief, technology, legacy, and the urgent question of what we choose to preserve while people are still here.

Experience the shared AI-Dad voice/chat session:

https://chat.jamesmuskal.ai/share/external/A9UUh3VaaPTvYho8gMiDEA-substack

Read the full essay:

Substack: https://www.drstevenmuskal.com/p/ai-dad-preserving-legacy-through

Medium:

https://medium.com/@smuskal/ai-dad-preserving-legacy-through-conversational-intelligence-58bb94ce9003

Watch the video:

https://www.youtube.com/watch?v=MBBzyaMNB64