James R. Stewart: Notes

Winning Team, Losing Team

2026-09-11T00:00:00Z

There is a type of football player who looks much better after being traded. Same arm. Same speed. Same brain. Different offensive line.

Suddenly he has another second to throw. The receivers are open. The play calling looks smarter. He stops forcing passes because he doesn’t have to. The team wins, so the coaches trust him more. Because the coaches trust him more, he plays better. A year earlier everyone was debating whether he belonged in the league.

This happens in business too.

I’ve spent much of my career moving between winning teams and losing ones. The main thing I’ve learned is that it’s surprisingly difficult to tell how good you are while you’re on either.

At one startup I worked harder than I have almost anywhere else. It was an early-stage financial technology company serving a difficult customer segment. We needed banks and infrastructure companies to let us operate on their rails. Those companies had their own risk committees, regulators, economics, and reputations to protect. And they did not particularly want our customers.

Partners dropped us. We found replacements. Those replacements dropped us too. Each time, the company mobilized. New meetings. New plans. New integrations. New risk controls. New spreadsheets demonstrating why things were definitely going to be different this time. People worked nights and weekends.

It didn’t matter much. We were playing uphill.

There is an appealing story in Silicon Valley that sufficiently talented people can overcome almost anything. This is useful propaganda if you are trying to convince talented people to join startups. It is also plainly false.

Sometimes you can execute well and still have a bad business. Sometimes the market doesn’t want what you’re selling. Sometimes your suppliers have more leverage than you do. Sometimes the economics are wrong. Sometimes the entire industry has decided that your particular type of customer is radioactive.

Effort matters. But effort is multiplied by circumstances. Ten times zero is still zero.

I didn’t understand this as clearly until I later found myself on the other side.

At a large financial company, I work on crypto products. We have a recognizable brand, millions of customers, capital, lawyers, compliance teams, engineers, and a banking charter. People answer our emails.

This is an underrated corporate asset.

When we started negotiating with a group of infrastructure vendors, several competed for the business. Fees came down. Terms improved. People internally were understandably pleased with the negotiations. They should have been. Good negotiation still matters.

But I remember thinking: of course they gave us a good deal. Look at the jersey.

A small startup might spend months trying to get the same vendor to return an email. We walk into the room carrying customers, distribution, credibility, and future revenue. The negotiator is playing the game. The organization has already moved the ball forty yards down the field.

This is one of the pleasant things about being on a winning team. The world cooperates with you.

Good candidates want to work for you. Partners want your logo on their website. Reporters return calls. Investors assume you know something. Customers are more willing to trust a new product. Even mistakes are interpreted generously.

When a losing company changes strategy, people call it desperation. When a winning company changes strategy, people call it optionality.

The dangerous part is what happens next. You start thinking it’s you.

I saw a more extreme version of this at another company earlier in my career. For a while, we were one of the dominant firms in our corner of crypto. Business came to us. The company grew quickly. Smart people joined. The numbers went up.

Winning creates its own weather.

You become more confident because things are working. Your confidence makes other people more confident in you. Their confidence produces more opportunities, which produces more evidence that your original confidence was justified. It is a nice loop.

Until it reverses.

When the crypto markets collapsed, the company went from winning to fighting for survival remarkably quickly. The personalities did not change overnight. The environment did. And yet behavior changed almost immediately.

People who had been cooperative became territorial. Conversations became sharper. Information became political. Everyone started thinking about who would be blamed, who would survive, and where the exits were. It had a faint Lord of the Flies quality.

This taught me something uncomfortable about company culture. Some cultures are strong because the culture is strong. Others are strong because everyone is getting rich. It is difficult to tell which one you have while the number is going up.

There is another version of the same problem in venture-backed startups.

Sometimes a company raises a huge round, hires famous people, gets glowing press, and starts receiving all the social signals associated with winning before it has actually won anything.

This can be worse than losing. At least losing gives you information. Premature winning can remove it.

People begin optimizing for status inside the company instead of results outside it. Titles matter more. Territory matters more. Meetings multiply. Everyone has a theory of how the organization should work. The company begins behaving like the defending champion while it is still in preseason.

Money makes this easy. A startup with $3 million in the bank has to confront reality fairly often. A startup with $300 million can postpone the meeting.

So I’ve become suspicious of both winning and losing as sources of personal feedback.

If you are losing, you may conclude that you are bad when you are merely playing on a bad team. If you are winning, you may conclude that you are exceptional when you are merely standing in the right place.

The hardest career question may be figuring out what travels with you.

Did the deal happen because you negotiated brilliantly, or because the other side badly wanted your company as a customer? Did the product fail because you made bad decisions, or because the market was structurally hostile? Did everyone listen to you because your judgment was excellent, or because your title was impressive?

You rarely get a clean answer.

This is why changing teams can be so educational. The variables move around. You discover that something you thought was one of your great strengths was mostly institutional leverage. You also discover that something you blamed yourself for was nearly impossible from the beginning.

Both discoveries are useful.

I would still rather be on a winning team. Winning is more fun. It gives you resources. It attracts talented people. It lets you spend time building instead of constantly trying to avoid death.

But winning may be the more dangerous teacher.

Losing constantly tells you that you are not as powerful as you think. Winning constantly tells you that you are. Neither is entirely true.

The trick is to enjoy the wins without confusing the scoreboard for a personality test. Sometimes you made the play. Sometimes the offensive line gave you six seconds.

It helps to know the difference.

Big Company, Small Company

2025-11-10T00:00:00Z

For much of my career, I worked at small startups. Recently, I joined a big company. I didn't expect it to change how I think about work, but it has.

At startups you have a kind of freedom that's almost impossible to replicate anywhere else. It comes from necessity. There isn't a "mobile platform team," a "data pipeline team," a "launch readiness gate," or a "brand team". There's just you, and whoever's sitting next to you, trying to figure out how to get people to use the thing you built. Nobody asks you for permission because there's no one to ask. You decide what matters, you decide when it ships, and you live with whatever happens.

You aren't managing scope, you're managing survival.

The strange part is that this freedom doesn't necessarily lead to large results. You could have a strong insight, work incredibly hard, and still launch something to six users. If two of them don't like it, there goes 30% of your active user base. Startups are exciting largely because the variance is so high. You could build something that nobody cares about. You could also build something that changes the world. These possibilities sit right next to each other, and you feel that every day.

Big companies are different. You don't own the whole product anymore. You own a part of it - a slice, a pixel. But that pixel will be seen by millions of people. Sometimes millions in a single afternoon. There's something humbling about that. At startups I could change everything, but those changes affected only a few people. At a big company I change almost nothing, but when it moves, it moves the needle.

You trade autonomy for scale.

I used to think the trade wasn't worth it. Now I'm not so sure.

Another difference is the distribution of outcomes. In startups, you can fail spectacularly. I've seen entire companies disappear overnight because one number on a fundraising slide didn't go the right direction. I've also seen products go from zero to ten thousand users in a weekend. It's emotionally violent. You spend a lot of time oscillating between delusion and despair.

At a big company, the highs and lows are capped. You rarely launch something that is catastrophic or transcendent. Most things fall somewhere between "pretty good" and "not so good." That can feel limiting, but it also means your career becomes less of a gamble. Your resume gets better even if the product doesn't change the world.

The same ceiling applies to people.

At startups, you encounter some of the best engineers you'll ever meet. They move effortlessly between architecture, design, and product thinking. They don't ask "what do you want?" They ask "why does it matter?" Some of them think so far ahead that you feel like you're watching someone 4D play chess.

You also encounter the worst engineers you'll ever meet. People who couldn't get a job anywhere else, who are willing to exchange stability for equity that may never exist. Startups compress the talent curve. You get the brilliant and the hopeless. Very little in the middle.

Big companies have a higher floor. Fewer geniuses, fewer disasters. You don't see as many miracles, but you also don't see as many train wrecks.

After enough time in startups, you develop a strange instinct. You start to recognize who is adding value and who isn't. It's not emotional - more like a survival reflex. When you've lived through enough layoffs, you start to see people the way a founder might: if the company had to shrink to five people tomorrow, who would still be here?

That awareness makes you sharper, but also more savage. Startups reward urgency, intensity, and optimization. Big companies reward alignment, collaboration, and patience. Both value creation, but they operate in different timescales. Startups are about combustion. Big companies are about inertia.

If there's anything I've learned from stepping from one world into the other, it's that neither environment has a monopoly on meaning. Startups teach you to build from zero. Big companies teach you to build for scale. If you're lucky, you learn both.

Startups give you the steering wheel.

Big companies give you the highway.

The trick is knowing what you're learning at each stage - and when it's time to trade one for the other.

Low / No Tech

2025-09-08T00:00:00Z

One of my favorite rituals is sitting in the sauna. The heat is quieting. Time slows down. In the corner there’s a small hourglass — a simple sand timer that measures fifteen minutes. No electronics, no beeps, no Bluetooth. Just sand falling through a narrow glass throat. It works as well today as it did centuries ago.

I love that! In a place too hot for anything digital to survive, the hourglass keeps doing its job. It’s honest technology: nothing hidden, nothing networked, nothing to update. Watching the sand run out feels truer than staring at a clock. It’s time you can see, not just measure.

Everything now has an app. Even things that don’t need one. The thermostat is online. The fridge is online. Even the baby monitor has an app. It’s exhausting.

It got me thinking about inventions that never needed any of this. The ones that solved such a fundamental problem they’ve endured for centuries. They look almost the same today as when they were first invented.

The hourglass. Plumbing. The bicycle.

Timers

In the Middle Ages, monks needed a way to pray at fixed times during the day and night. But you couldn’t rely on the sun for that. So in the 14th century, at Salisbury Cathedral in England, they built one of the world’s first mechanical clocks. It struck the hours with a bell. The rhythm of time became something you could measure, not just guess.

Praying Monks → Global Human Coordination

Before that, life was vague. You woke with the sun, worked until dark, and measured in rough stretches: morning, noon, evening. Imagine trying to coordinate anything more precise - like baking bread - without a way to measure minutes.

The hourglass was the personal version of that idea. It didn’t need gears or metal — just gravity. Sailors used them to mark watches at sea. Bakers still use them for bread. I use one sometimes to time my newborn’s feeds. It’s a reminder that the simplest way to track time still works.

Turn it over, and it starts again. No batteries, no software, no patch notes. Just a perfect loop — like breathing.

Plumbing

The Cloaca Maxima in ancient Rome was one of the first great public sewers. Built in the 6th century BC, it was so durable that parts of it still function today. Imagine: stone tunnels carrying waste out of the city for over two thousand years.

Roman Aqueducts → Sprawling Urban Cities

Before systems like this, waste and water mixed freely. People dumped sewage in the street. Wells were contaminated. Epidemics swept through towns again and again. London’s “Great Stink” of 1858 - when the Thames was so polluted the smell shut down Parliament - wasn’t even that long ago.

Plumbing was the unlock. Clean water in. Wastewater out. Cities could suddenly grow without being death traps. Plumbing today is still pipes and pressure. It hasn’t really changed. And it’s invisible, which is probably the best compliment an invention can get.

Bicycles

In Amsterdam, there are more bicycles than people - about ~900,000 bikes for about 800,000 people. At rush hour, you see entire rivers of cyclists moving faster than cars, quietly filling the streets. The bicycle isn’t nostalgia there; it’s the backbone of daily life.

The unlock was personal freedom. Before bicycles, if you couldn’t afford a horse, you walked. Journeys of a few miles were long ordeals. The “safety bicycle” of the 1880s - two equal-sized wheels and a chain changed that. Suddenly anyone could travel 5 or 10x faster than walking, powered only by their own legs. Susan B. Anthony called it the “freedom machine” because it gave women independence in a way nothing else had.

Wheels + Leverage → Women's Suffrage

Even now, a bicycle is one of the most efficient machines ever built. More efficient than walking. More efficient in some ways than a car. Two wheels, a frame, a chain - and yet it’s enough to reshape a city.

Some inventions don’t ask to be upgraded - they remind us that the simplest solutions are often the most enduring.

Self Inquiry

2025-09-08T00:00:00Z

Podcasts, practices, and writing I come back to. All with the same thread: radical self-inquiry.

How have I been complicit in creating the conditions I say I don't want? (Video)

What struck me is the lie so many of us carry—that success and happiness can’t coexist. The reminder is simple: I don’t have to choose misery as proof of achievement.

Journaling as clarity (Video)

When I write, the fog lifts. Looking back, I see old struggles with distance, and often I realize they weren’t as permanent or defining as they felt at the time.

Radical Acceptance (Book)

The line that stays with me: “something is wrong with me” is just a trance. Tara Brach shows how noticing and softening that voice is already an act of freedom.

The Big Leap (Book)

We all have ceilings we secretly enforce. Hendricks’ idea—that joy and contribution can live together in a “zone of genius”—makes me notice how often I hold myself back.

Fear-setting (Post)

Writing down fears makes them ordinary. Once they’re on paper, they stop being monsters in the dark. The worst case usually looks manageable, and the cost of inaction becomes clear.

New Dad

2025-07-28T00:00:00Z

On May 24, 2025 at 4:30 am, Taylor and I welcomed our first son, Devereux Benjamin Stewart. We call him Dex. (She had the better last name, so we made her maiden name his first, to let it live on.)

Dex is everything to me.

These early days have come with many surprises, here are some of my observations from my journey into fatherhood:

Meaning: Becoming a dad is easily the most meaningful event of my life - nothing else comes close. Not buying a dog. Not buying a house. Not even getting married. When people say “my pride and joy,” I used to think they were just being sentimental. Now I get it.

What no one tells you is how peaceful babies can be. If your only exposure to infants is on airplanes or in crowded restaurants, you think they’re always screaming. But in reality, there are beautiful, sweet moments - late-night feedings, early-morning stretches, the weight of his small head resting in my arm. These moments aren’t loud or dramatic. But they stick with you. They change you.
Community: One of the strange things about becoming a parent is realizing you're suddenly part of a club - one you didn’t apply for and didn’t even notice before. But now you're in, and the other members find you. A nod from a dad at the park. The subtle smirk from the cashier when your baby’s wailing at checkout. Stories from friends you haven’t heard from in years, now reaching out with advice or commiseration. Parenthood forms its own quiet network.
Struggle: It also makes life harder in the obvious ways. Everything takes longer. Things you used to do without thinking - grabbing coffee, folding laundry, going for a run - now require coordination, negotiation, and sometimes just giving up. You have less time, less energy, and a lot less flexibility. So you have to prioritize like your life depends on it. Because sometimes it does.
Health: I’m what you call a “Huberman Husband.” I followed the whole protocol: early morning sunlight, zero microplastics, daily breathwork, etc. But I've had to relax significantly since Dex arrived. Good luck avoiding plastic when most breast pumps and baby bottles are made from it, even the “glass” ones. Sanitizing often means microwaving plastic, which used to be sacrilege in my household. I've had to let go. You either adapt, or you suffer. Sometimes both :)
Subscriptions: Everything now requires a subscription, and everyone wants your money. The number of paid apps you’re expected to adopt skyrockets, despite any resistance. The baby monitor (we skipped the Nanit because it requires a plan). The sound machine. The nightlight. The sleep tracking app. If it plugs in, it probably has a monthly fee. So we’ve defaulted to buying the “dumb” version of everything. No wifi, no app, no notifications. Just on/off. Simpler is better when you're tired and broke.
Cleanliness: Welcome to hell! Sanitization becomes your new job. Bottles, pump parts, pacifiers. Anything entering Dex's mouth needs sterilizing. Constantly. It’s a draining cycle with no perfect solution. Dishwashers work but run for hours daily. Dedicated bottle washers clutter countertops. Steam bags for microwaves might save time but feel like a health hazard waiting to happen. Hello again, microplastics and cancer!

But still - I’d do it all a hundred times over. Dex is the best thing that’s ever happened to me. He’s not even old enough to crawl, and already I feel more purpose than I ever have. I can’t wait to watch him grow. To teach him to ski, share my love of ribeyes on the grill, and make him grumpy sleeping in tents on remote backpacking trips.

So yes, my life is more complicated, chaotic, and inconvenient. In the end, you realize the inconvenience was actually the point.

I Love LMNT. LMNT Is a Scam.

2024-10-20T00:00:00Z

I love LMNT. It’s my go-to for hydration, especially after a tough workout or a long day in the sun. But let me tell you -- LMNT is also a scam. Or at least, it's overpriced. Let me explain why.

LMNT is an electrolyte drink mix. The benefits of staying hydrated with electrolytes are well-documented: you need more than just water to replenish what you sweat out. Electrolytes like sodium, potassium, and magnesium help your body stay balanced. You feel better, perform better, and recover faster. It's why people reach for sports drinks or coconut water. It's also why LMNT caught my attention in the first place. The packaging is nice, the flavors are bold, and it's effective. It ticks all the boxes.

But what is LMNT, really? It’s salty water. It’s fancy salt in a nice packet. And it’s $1.50 per serving -- which is kind of ridiculous if you think about it. The more I drank it, the more I loved it, but I couldn’t shake the feeling that I was being swindled. How did we get here, paying a premium for what’s essentially sodium chloride and a little potassium and magnesium?

There’s a bit of history. People have been chasing hydration hacks for a long time. Remember Pedialyte? It started as a baby product but found a cult following among athletes and hungover adults. Then came other contenders: Electrolit, Cure, Liquid I.V. -- all with the same promise. Hydration, but better. LMNT is just the latest in a very competitive space. But despite its simplicity, I really do think LMNT is the best.

A TikTok video finally pushed me over the edge. In the video, someone breaks down LMNT for what it is: dressed-up mineral salt. And mineral salt isn’t new. Humans have been consuming it forever. Fleur de sel, Himalayan pink salt -- they sound fancy, but they’re all just salt with a few added minerals. Still, they don’t contain the concentration of electrolytes you find in LMNT.

That’s when I got an idea. If LMNT is just simple ingredients -- salt, potassium, magnesium -- why couldn’t I make it myself? So I set out on a journey to DIY my own electrolyte mix. I bought everything on Amazon. Bags of sodium chloride, potassium chloride, and magnesium citrate. After some trial and error, I finally got the mix right. It was a small eureka moment -- a cheaper, homemade LMNT.

The next challenge was flavoring it. I wanted something more than just salty water. My first thought was freeze-dried and powdered fruit. I tried it, but it was a disaster. The fruit didn’t dissolve well, and the texture was off. Then I discovered citric acid. It’s the same compound that makes citrus fruits taste tangy, and it turns out it’s perfect for electrolyte drinks. After some experimentation with proportions, I got it just right. A simple, citrusy electrolyte drink that tastes… pretty close to LMNT.

I’m not done experimenting. LMNT’s chocolate salt flavor intrigues me, and I think that’ll be my next project. But as much as I love LMNT, I know it's a scam -- at least in the sense that I'm paying a premium for something so simple. The convenience, the branding, the perfectly balanced mix -- it’s all great, but I’ve proven to myself that I can do it at home for a fraction of the cost.

And who knows? Maybe in a few months, I'll have my own lineup of flavors to rival LMNT. For now, I'll keep enjoying my salty water -- both the fancy packets and my own DIY version.

Recipes

2024-10-20T00:00:00Z

Table of Contents

The Best Pancakes
James' Pita Bread

The Best Pancakes

Adapted from @joseybakerbread on Instagram, September 24, 2022.

For the fluffiest pancakes, Baker soaks his flour in milk overnight in the refrigerator, then mixes up the batter in the morning. But don't worry if you forget or haven't planned ahead; these pancakes are great even if you skip that step. Baker gives the option of adding a tablespoon of sourdough starter to the mix, if you've been nursing one along.

Serves 4

1 cup whole wheat flour
3/4 cup milk
1 tablespoon sourdough starter (optional!)
2 tablespoons sugar
2 tablespoons unsalted butter, melted, plus more for buttering the pan and serving
1 large egg
1/2 teaspoon baking soda
1½ teaspoon kosher salt
1½ teaspoon vanilla extract
Chocolate chips, peanut butter, fresh fruit or maple syrup, for serving

Instructions:

In a medium bowl, stir together the flour, milk and sourdough starter, if using. Cover and refrigerate overnight, if desired. Stir in the sugar, butter, egg, baking soda, salt and vanilla and mix until combined; don't worry about overmixing.

Melt some butter on a griddle or in a medium frying pan over medium heat. Ladle some batter onto griddle or skillet, making pancakes of any size you like; the batter will be thick, so you may need to use the back of a spoon to help spread the batter. If you're planning to add chocolate chips (or berries), do it now. Flip pancakes after bubbles rise to surface and bottoms brown, after 2 to 4 minutes.

Cook on the second side until lightly browned and cooked through, about 2 minutes more. Transfer to a plate and top as you like.

Note (from original post):

This recipe omits the overnight soak of the flour, as I have previously instructed in my pancake recipes. I will admit that the overnight soak does in fact yield a slightly better pancake, but I mean wtf we are already dirtying THREE bowls with this recipe, so I didn't dare to ALSO ask you to start this recipe the day before for fear of being labeled a MONSTER. But if you wanna go ALL THE WAY with your pancakes, then please soak your flour and milk overnight and proceed with the rest of the recipe as described here. Or really, if you wanna be THE MASTER OF PANCAKES you should also include some sourdough in your overnight soak. But this might cause you to become known all over the land as the Master of Pancakes, in which case you'd have to quit your day job and just make pancakes for everyone you know for the rest of your days. So proceed with caution and don't say I didn't warn you.

James' Pita Bread

Ingredients

100g discard sourdough starter (about 1/2 cup, optional)
500g bread flour (about 4 cups)
4-5g baking soda (about 1 teaspoon)
15g olive oil (about 1 tablespoon), plus more for cooking if needed
10g salt (about 1 and 3/4 teaspoons)
300ml water (about 1 and 1/4 cups)

Starter and Water Mixture (Optional)

If available, combine the discard sourdough starter and water in a large mixing bowl. Stir well to mix. The starter is optional and is used to enhance the flavor of the pita.

Adding Dry Ingredients

Gradually add the bread flour to the water, stirring continuously. Sprinkle in the baking soda and salt. Drizzle in the 1 tablespoon of olive oil. Mix until all ingredients are combined.

Stretch and Fold

With the dough in the bowl, grab one side of the dough, stretch it upwards and then fold it over itself. Rotate the bowl slightly and repeat about 2-3 times. Let the dough rest for 30 minutes and repeat this process a few times over 2 hours.

First Rise

After the final set of stretch and folds, cover the bowl with a damp cloth or plastic wrap. Allow the dough to rise in a warm place until it almost doubles in size, which should take about 1 to 2 hours.

Shaping the Pitas

Once the dough has risen, gently deflate it. Divide the dough into small balls, around 8-10, depending on your preferred pita size. On a lightly floured surface, use a rolling pin or a tortilla press (if available) to flatten each ball into a round, flat shape, about 1/4 inch thick. Be careful not to press too hard to maintain some airiness in the dough. Cover the dough balls with plastic wrap or a damp towel and let rest for 30 minutes, or up to overnight.

Cooking on Cast Iron Skillet

Preheat a cast iron skillet over medium-high heat. If the pitas start to stick, lightly brush the skillet with olive or avocado oil. Place a rolled-out pita onto the skillet. Cook for 2-4 minutes until it puffs up and the bottom has browned. Flip and cook the other side for an additional 2-4 minutes.

Cooling and Storing

Remove the cooked pita from the skillet and cover it with a clean towel to keep it soft. Repeat with the remaining dough. Serve the pitas warm, or store them in an airtight container after they’ve cooled down.

Experimenting with AI

2024-02-24T00:00:00Z

A note from 2026: I wrote this in early 2024, when custom GPTs were new and pasting a credit report into ChatGPT felt cutting-edge. It's badly outdated now. I'm keeping it up as a historical record.

AI's capabilities far exceed funny poems and AI-generated Seinfeld episodes.

My journey into AI

I’ve always been an early adopter of new technology. In 2013, I bought my first Bitcoin from my college dorm room. When it comes to AI, my initial exposure happened by chance – I was listening to an a16z podcast at the gym where the guest spoke of a new wave of “Artificial Intelligence” companies. At the time, they made a convincing argument that AI wasn’t a fading fad, but an enduring trend in its early days. This piqued my curiosity.

I gained practical firsthand AI experience while working as a PM at BlackRock. Our internal Financial Modeling Group developed several ML models for institutional portfolio management and trading use cases. Seeing the models accurately predict trading behaviors was a magical “aha!” moment for me. This was the turning point — I was hooked.

Fast forward to now. I am an early product hire at StellarFi, working to make consumer credit universally accessible. Since my BlackRock days, I haven’t had the opportunity to work directly with AI as part of my day-to-day job. Enter 2022 and the emergence of ChatGPT.

AI and Haikus

When ChatGPT arrived in the fall of ‘22, it immediately captivated the masses, myself included. While many loved the poems and AI-generated Seinfeld episodes, I felt the real story was unfolding in Fintech. Suddenly, a powerful language model wasn't just generating funny haikus, it now had the potential to push fintech in a new and exciting direction. I wasn't interested in parlor tricks; I saw the potential to push the industry in an innovative direction, to truly innovate beyond incremental tweaks. This wasn't just another tool. It was a potential game-changer, a force waiting to be harnessed. And I was ready to dive in.

Beyond jest and verse,
Fintech's potential unearthed,
A new era's birth.

Fintech and AI: A Promising Alliance

Forget the fluff about "exciting use cases" and "potential”. There are real-world applications of AI in fintech today. Some examples:

However great these solutions, I still had an itch I couldn’t scratch. Doing research, I couldn’t find a true application of AI in a consumer-facing fintech product. Specifically with my work at StellarFi, I was interested in credit-building for US consumers. Nothing was out there, and one thought kept burning: could an LLM unlock truly personalized credit report insights and actionable advice? This led me to experiment with language models to uncover their capabilities.

Experimenting with ChatGPT

Inspired by a helpful primer for product people entering AI, I began by inputting my credit report data into ChatGPT. Here's a fictional example of some of the data:

{
  "fullName": "Fictional James",
  "creditScore": 762,
  "recentInquiries": 2,
  "openCreditLines": 5,
  "totalAvailableCredit": 30000
}

Initial Prompt and Output

I got started on crafting a prompt to gather insights from ChatGPT. Here’s the first iteration:

Me:

Output from ChatGPT

This is pretty good! You can see the model is taking my credit data as input and outputting actionable insights for improvement. Even still, there is too much fluff. And some to be desired with its tone.

Refined Prompt and Output

In a number of minutes, I was able to get something workable:

The main challenge was that it required substantial fine-tuning and repeated re-prompting to get the desired output. This iterative approach clearly isn't scalable.

Enter Custom GPTs

Just as I was exploring these possibilities, OpenAI introduced custom GPTs.

Using the new custom GPT creator, I was able to feed the LLM a new system message. With this, I was able to reliably recreate the above results without having to re-prompt and re-tune the model continuously. Pretty cool!

What’s even cooler is that I can even “chat” with my credit report. I can ask it general questions (”How is VantageScore 3.0 calculated?”) and also apply them to my situation personally (”And what might you suggest for me to improve mine?”)

Future Applications: Leveraging GPT-4 via API

While fun, my initial forays were more of a thought exercise than impactful solutions. The real power will come with accessing GPT-4 via API. Pouring through the official API docs, it’s clear the API offers a much higher degree of customization and fine-tuning compared to the standard ChatGPT consumer-facing product.

Some examples:

Tailored responses based on specific financial contexts.
Integration with databases for capture and historical retrieval.
Enhanced performance and cost-savings.

Final thoughts

When I first started experimenting with ChatGPT at the end of last year, I was initially entertained by the amusing poems and stories it produced. But the more I explored customizing the model outputs, the more potential I glimpsed for using this technology to meaningfully help people. This side project applying AI to personal finance and credit reports got me genuinely excited about the future of fintech. I'm still only one guy experimenting with these tools solo, but I can’t help feeling like I’m seeing the seeds of something significant. Who knows where this will lead for me, or what the wider reverberations might be in the future. But I’m eager to keep learning to see where this AI journey takes me.

Principles

2024-01-27T00:00:00Z

In my journey as a product manager and, more fundamentally, as a person, I've come to realize that I hold certain principles. Truths about life, working effectively in the workplace, and building products. Like most others, I hold these subconsciously most of the time. Every so often, I'll find myself in a situation and these principles will re-surface themselves in my mind. This essay is my attempt to articulate these principles, exploring their importance and their personal significance to me.

This list is dynamic and will evolve as I do. It's open to changes as I learn and grow.

How I decide

Seek truth, not compromise (source)

In every team, company, or organization, conflict is a natural part of working together. When faced with conflict, we typically encounter two paths: seeking compromise or seeking truth. Opting for compromise is often the path of least resistance. Jeff Bezos illustrates this: "Take you back in time 10,000 years, and you're in a small village. If you go along to get along, you can survive, you can procreate." He goes on, "If you are the village truth-teller, you might get clubbed to death in the middle of the night." This exemplifies the inherent challenge in truth-telling – it's against our natural instincts for harmony and survival. Yet, embracing truth is crucial in creating a culture of openness and integrity within the teams I lead. It's about facing reality head-on, not shying away from it or reshaping it to fit our desires. I firmly believe that confronting the truth, despite its difficulties, is fundamental to achieving meaningful progress and building genuine, dynamic team relationships.

There are no solutions, only tradeoffs (Thomas Sowell)

Every decision I've made at work looked like a solution at the time. It wasn't. It was a trade. Ship sooner and you take on risk. Add a feature and the product gets harder to use. Hire the senior person and you lose the scrappy one. So I've stopped asking "what's the right answer?" and started asking "what are we giving up, and are we okay with that?" When nobody in the room can name the cost of a decision, it usually means nobody has looked for it yet.

If it's not a "Hell yeah!", it's a no (source)

The easiest framework for decision-making. If you are ever met with a high-stakes decision — be it personal, professional, or otherwise — ask yourself: is this a "HELL YEAH!"? If there's any hesitation that stops it from being a resounding yes, then it's a no. This approach is like the Net Promoter Score (NPS) used for products, where only those consistently rated 9 or above are truly considered exceptional. By applying this 'Hell Yeah!' framework, I ensure that my efforts are dedicated to pursuits that genuinely matter and resonate with me. It helps me maintain a laser focus, leading to a more meaningful and satisfying life. This is my personal NPS framework for life.

How I work

Fall in love with the process (James Clear)

James Clear says you don't rise to the level of your goals; you fall to the level of your systems. Nick Saban built a dynasty on the same idea. His teams were taught to ignore the scoreboard and win the next play. Goals are good for picking a direction. They're bad for getting you there, because everyone in the race has the same goal. What separates people is what they do on an ordinary Tuesday. The Zen version is older and shorter: chop wood, carry water. I keep relearning it.

Calculation is miscalculation (The Inner Game of Tennis)

Tim Gallwey noticed that his tennis students played worse the harder they thought. He described two selves. Self 1 is the voice in your head that calculates, judges, and corrects. Self 2 is the part of you that already knows how to hit the ball. Every time Self 1 steps in, it gets in the way. Calculation is miscalculation. I see it most in big moments, like a hard conversation or a pitch. The more I try to compute my way through, the worse it goes. The work happens beforehand. In the moment, the job is to trust it.

How I grow

What got you here won't get you there (Marshall Goldsmith)

The habits that make you good at one job often get in the way of the next. At startups I was rewarded for doing everything myself. At a big company that same instinct makes you the bottleneck. The hard part is that old habits don't feel like weaknesses. They feel like the reason you got here. Every time I change roles, I ask which of my strengths just turned into a liability.

Chase the person ten years ahead (Matthew McConaughey)

When Matthew McConaughey won his Oscar, he said his hero is himself, ten years from now. He picked that hero at fifteen. Every decade, the hero moves another ten years out. He'll never catch him, and that's the point. I like this better than most advice about role models. It turns an ordinary day into a question: is this setting up the person I want to be in ten years? Usually the answer is in small things. What I read. Who I spend time with. Whether I go for the run.

How I live

This is water (David Foster Wallace)

Two young fish swim past an older fish, who says, "Morning, boys. How's the water?" A little later one turns to the other: "What the hell is water?" The most important parts of your life are the ones you've stopped noticing: your assumptions, your habits, the people you see every day. Wallace's point is that paying attention is a choice you have to keep making, usually when you're tired and stuck in a checkout line. It's the hardest principle on this list, because it never stays done.

Build cathedrals, don't lay bricks (source)

As a product leader, my role is to set a compelling product vision, embodying the principle of "Build cathedrals, not just lay bricks." It's crucial for me to ensure that every team member sees beyond the day-to-day tasks to the grand vision we're constructing together. I strive to make everyone aware that they are not just laying bricks, but are integral in building an awe-inspiring catheral. My responsibility is to keep this vision clear and present in all our efforts, ensuring that each action contributes to our magnificent end goal. This approach not only enhances our work but also fuels motivation and a sense of purpose within the team.

Man in the Arena (source)

This quote from Theodore Roosevelt powerfully captures the mindset I aspire to in work and life. It's about setting audacious goals, taking risks despite critics, and finding fulfillment in the journey - win or lose. It stands as the guiding principle I strive to embody in my life — so much so, that it's permanently embedded in the source code of my website. To me, "the man in the arena" represents courage - the courage to create, build publicly, and push limits. It's less about outcomes than the grit required to attempt great things. This quote drives me to take that first step and pour my energy into efforts that deeply matter, whether or not they succeed. If I had to choose one guiding principle, this would be it.

To be continued…

Good Eggs

2024-01-21T00:00:00Z

I love good eggs. Seriously, my love for eggs is fast becoming a signature trait. Have you ever delved into the myriad health benefits of eggs? These oval wonders aren’t just delicious, but they’re brimming with high-quality protein, vitamins, and minerals. Recent studies suggest that eggs might actually improve heart health. And let's not overlook their role as a delicious, bountiful source of Choline, a nutrient that turbocharges the brain.

My obsession for eggs is no casual affair. I consume roughly 20 eggs a week — that's just shy of two dozen, tallying up to about 80 a month, or just over a grand total of 1,000 eggs per year! So, yes, I take my eggs very seriously. This egg-centric lifestyle has propelled me on a quest, a quest in pursuit of the perfect egg. And what, you might ask, qualifies the perfect egg? Allow me to introduce "The Egg Quotient," a slightly satirical formula for gauging eggcellence.

Let’s dive in.

Formula:

Variables Explained:

Yolk Radiance Factor (YRF): Measures the color intensity of the egg yolk on a scale from 1 (pale yellow) to 10 (vibrant orange). The more vibrant, the better. Typically, more vibrant egg yolks indicate a higher concentration of Omega-3s and other important nutrients.
Shell Sturdiness Score (SSS): Ranks the robustness of the eggshell on a scale from 1 (fragile as a soap bubble) to 10 (hard as a rock). Healthier chickens lay sturdier shells.
Egg Expenditure Index (EEI): Reflects the cost of the egg. It starts at 1 (wallet-friendly) and goes up to 10 (might require a small loan). The cheaper the egg, the better for the index.

With this rigorous scientific framework in mind, I embarked on an eggspedition to crack the code of the ultimate egg.

Trader Joe’s Organic Cage Free

Despite the touted organic certification, these eggs came up short. Mildly yellow/orange yolk color and average shell sturdiness. Decent price at $3.50 per dozen!

Yolk Radiance Factor (YRF)	Shell Sturdiness Score (SSS)	Egg Expenditure Index (EEI)	Egg Quotient (EQ)
3	5	3	2.67

Trader Joe’s Pasture Raised

A step up from the TJ’s Cage Free eggs! Decent shell sturdiness, however not the deep orange color I’m after. These are my default when I can’t get eggceptional eggs.

Yolk Radiance Factor (YRF)	Shell Sturdiness Score (SSS)	Egg Expenditure Index (EEI)	Egg Quotient (EQ)
5	7	3	4.00

Trader Joe’s Organic Pasture Raised

I don’t understand how they manage to charge $5.00 for these. Mediocre color and WORSE sturdiness than their non-organic counterpart. An egg hoax!

Yolk Radiance Factor (YRF)	Shell Sturdiness Score (SSS)	Egg Expenditure Index (EEI)	Egg Quotient (EQ)
5	6	5	2.20

Naturally Omega Cage Free

Certifiable hot garbage. If these eggs truly contain more Omega-3s, it is nowhere evident in the abysmal shell sturdiness and dull yellow color. A true disappointment - skip at all costs.

Yolk Radiance Factor (YRF)	Shell Sturdiness Score (SSS)	Egg Expenditure Index (EEI)	Egg Quotient (EQ)
1	1	6	0.33

Happy Egg Co. — Free Range

I am torn by these Happy Egg Co. eggs. Sublime shell sturdiness indicates a healthy mama hen, however, I was disappointed to find a run-of-the-mill yellow/orange color. Close, but no cigar.

Yolk Radiance Factor (YRF)	Shell Sturdiness Score (SSS)	Egg Expenditure Index (EEI)	Egg Quotient (EQ)
5	8	7	1.86

Happy Egg Co. — Heritage Free Range

A beautiful tragedy. These eggs are truly eggstraordinary — from the beautiful blue and brown speckled shells to the gorgeous deep orange yolks. I would eat these eggs any day. The one problem — they cost $10.00 per dozen. A special occasion egg.

Yolk Radiance Factor (YRF)	Shell Sturdiness Score (SSS)	Egg Expenditure Index (EEI)	Egg Quotient (EQ)
10	10	10	2.00

Vital Farms — Pasture Raised

Eggregious price ($9.48!) for subpar eggs. Strip away the fancy packaging and you will find a dozen sad excuses for eggs. A surefire disappointment.

Yolk Radiance Factor (YRF)	Shell Sturdiness Score (SSS)	Egg Expenditure Index (EEI)	Egg Quotient (EQ)
5	4	10	0.90

Whole Foods — Pasture Raised

Indeed, the egg odyssey has reached its zenith. At the modest sum of $4.99, what a joyous triumph it is to lay one's hands on these paragons of poultry perfection. Each carton, a treasure trove, revealing orbs that are the epitome of ovular opulence. As the shells crack, they give way to yolks of such resplendent golden hue, akin to the first rays of dawn breaking over an idillic farmstead. To savor these eggs is not simply to dine; it is to revel in a feast that transcends the ordinary, each mouthful a melody of flavors that sings to the soul. Indeed, for a price that doesn't demand one to break the proverbial bank, the sheer pleasure these exemplary eggs provide is an unequivocal triumph of taste.

Yolk Radiance Factor (YRF)	Shell Sturdiness Score (SSS)	Egg Expenditure Index (EEI)	Egg Quotient (EQ)
10	10	3	6.67

The Final Tally

Rank	Name	Final Rating
1	Whole Foods — Pasture Raised	6.67
2	Trader Joe’s Pasture Raised	4.00
3	Trader Joe’s Organic Cage Free	2.67
4	Trader Joe’s Organic Pasture Raised	2.20
5	Happy Egg Co. — Heritage Free Range	2.00
6	Happy Egg Co. — Free Range	1.86
7	Vital Farms — Pasture Raised	0.90
8	Naturally Omega Cage Free	0.33

Advice for Aspiring PMs

2023-05-12T00:00:00Z

Quick advice if you want to break into product management.

Decide what type of PM you want to be

If you're considering a career in product management, the first step is not to look outward, but inward. Consider the type of PM you wish to become:

Consumer/Growth PM: This path is for those interested by the nuances of human behavior and psychology. It's about delving into why people do what they do and leveraging that understanding to drive user engagement and growth. Ask yourself: Do you find joy in decoding user data to inform product decisions? Are you the kind of person who thrives on experimenting with different features to see what sticks?
Platform PM: Here, the focus shifts to the backbone of the product – the systems and infrastructure. It suits those who find beauty in the technical side of things, who enjoy collaborating with engineers and love the challenge of managing complex systems.
Technical PM: This role is a blend of eng and PM, ideal for those with a technical background or a strong interest in tech. It's about diving deep into APIs, understanding technical documentation, and working closely with developers.
Enterprise PM: If you're drawn to building tools that streamline workflows and enjoy the intricacies of stakeholder management, this might be your calling. It's a role that combines creativity in problem-solving with the discipline of project management.

These categories are not rigid. The best PMs often draw from multiple areas, blending skills to fit the unique challenges they face.

Startups vs. Big Tech

When considering startups versus big tech, think of it as a choice between two very different games:

Startups: Think of it as joining a small, agile team where everyone wears multiple hats. It’s fast-paced, high-impact, and your work’s influence is clear. But it comes with risks and uncertainties.
Big Tech: It’s like becoming part of a well-oiled machine. You’ll likely specialize, gain structured learning, and access vast resources. It’s more predictable and offers a clear career trajectory.

Both paths have their perks. Choose based on what excites you more.

How to break into your first PM role?

Breaking into your first PM role can seem challenging, but there are multiple paths:

(Easiest) Join a company that aligns with your values and interests. Often, it's easier to transition to a PM role internally where you're already a known quantity.
1. “If you're offered a seat on a rocket ship, don't ask what seat!” - Sheryl Sandberg
(Harder) Consider roles adjacent to PM, like product operations or product marketing, especially in startup environments. These can be stepping stones to your desired role.
(Hardest) The more challenging but rewarding route is getting into a big tech rotational program. Yes, the competition is fierce, but it's a solid foundation for a career in PM.
1. FB RPMs
2. Uber APM program
3. Google APM program

Interviewing

PM interviews are hard! They’re a mix of interpersonal skills (do I like this person?), product skills (does this person have a good, intuitive product sense?), and technical/analytical ability (is this person data-driven and uses data to solve problems?).

A big reason the PM interview is so challenging is that the questions asked are very broad. The focus and goal of many of these interviews is to assess how well you can break down complex problems into component pieces and think on your feet. Because of this, you must learn some basic product management frameworks to help you organize your thoughts and approach.

The best resource I have found is Stellar Peers: https://stellarpeers.com/. There are paid features to the site, but you should be able to get by by reading the basic frameworks.

In addition to Stellar Peers, a newer resource that's become genuinely top-tier is Ben Erez's work on PM interviews. His two viral posts in Lenny's Newsletter are, in my opinion, the current gold standard for breaking down big-tech style interviews. They're not generic tips; they give you an actual system you can follow, plus templates you can use to structure answers in a way that consistently impresses interviewers. His method really does work if you put in the reps.

Ben also teaches a Maven course (https://maven.com/ben-erez/pm-interview) that’s probably the most comprehensive PM interview course out there. It’s pricey, but it’s the best version of structured, modern PM interview prep I’ve seen.

One question you will almost surely get is to name your favorite product. What do you like about it? What would you improve? Preparing for this question is a good starting point to help you understand PM frameworks and what types of things interviewers are looking for.

Other useful websites:

Exponent
Igotanoffer
levels.fyi

The classic book that many recommend is Cracking the PM Interview, however I find this material outdated and not relevant to how PM interviews are conducted (in my experience). If you are looking for a book, I’d recommend The Product Manager Interview by Lewis C. Lin. This is good if you want to interview in Big Tech.

New Products

2023-02-23T00:00:00Z

A note from 2026: I wrote this in early 2023, before AI reshaped how products get built. My process looks very different today. I'm keeping this up as a historical record of how I worked in the pre-AI era.

Thoughts on launching new products. Here's a look at my product development process, from the first idea to the final product.

Plan

James' discovery process

Define initial product vision, goals, & user personas
1. What are we setting out to do and why?
2. What might a long-term product mission look like?
3. Who are we aiming to help?
Speak to users - early and often. Ask questions.
1. Resources I love for product discovery:
  1. The Mom Test, Rob Fitzpatrick
  2. 5 Whys, Taiichi Ohno
  3. The Lean Startup, Eric Reis
Inventory current pain points. Assess the severity, impact, and frequency of pain points. Prioritize ruthlessly.
Form themes and hypotheses around themes.

Partner

Partner with business stakeholders
1. Business goal alignment: how might the introduction of a new product help or hinder existing company goals? Are we performing due diligence and communicating the product vision effectively to leadership for buy-in?
2. Establish a Steering Committee: comprised of project leads, plus company leadership. The purpose of the steering co. is to provide regular check-ins with leadership and relevant business leads for progress updates, brainstorming, and product feedback. Steering co. meets on a regular cadence — usually monthly.
Partner with design. Build low-fidelity prototypes. Test with users.
1. Important to not put engineering “hands on keyboards” before features are validated with users. Fail fast!
Partner with engineering.
1. Assess feasibility: ensure features are technically viable to create and do not cause system slowness or otherwise negative user experience.
2. Discuss resourcing: understand current engineering priorities and agree upon appropriate team staffing.

Execute

Prioritize features - shape the MVP.
1. Seed the backlog foundational user stories
2. Partner with design on high-fidelity prototypes
3. For a given user journey - bucket features into “must haves”, “performance”, and “delighter” features (kano model).
Build!
1. Agile ceremonies to consider:
  1. Backlog refinement: to ensure a healthy product backlog of epics, user stories, and bugs — ranked in priority order. Owned by product team in collaboration with tech lead.
  2. Sprint planning: deciding the proper mix of stories to include in each sprint iteration. Depends on story complexity and team capacity. Co-owned by product and tech leads, with participation by engineering and design.
  3. Daily Standups: (don’t have to be daily) - Check in on what you accomplished yesterday, what you plan to do today, and any blockers. Meeting is led by tech lead, with participation from product, engineering, design, and QA.
  4. Sprint review: engineering’s chance to brag about the cool stuff they shipped! Product, engineering, design, QA, and the occasional relevant business stakeholder.
  5. Retrospective: an opportunity to reflect on what is going well, what isn’t going well, and potential next steps. Participation from product, engineering, design, and QA.

Iterate, Launch, Iterate

More user testing :)
1. Usability testing (users are now in the driver's seat) to uncover feature gaps and potential bugs.
2. Run experiments: A/B testing with treatment groups vs. control groups

Experiment design

Dogfooding and UAT testing.
1. Partner with QA on regression suite and acceptance testing.
2. Document bugs and assess impact.
Launch 🚀
1. Partnerships & Collaboration:
  1. Marketing: does marketing have all the market, user, and product context to go to market? Are they set up for success with the proper tooling for growth?
    
    Marketing channels are a function of product maturity
  2. Support: is support aware of the launch and potential day-one bugs? Is there sufficient internal or external documentation?
2. Launch strategy:
  1. Hypothesis-driven micro launches. Start small and expand via learnings!
    
    Micro launches enable risk reduction
  2. Rollback and contingency planning
  3. Ensure event logging/metrics tracking are implemented
  4. Perform user outreach. Collect feedback. Synthesize into learnings.
    1. Survey, email, 1:1 sessions
Prioritize future product iterations
1. Refine user stories and define acceptance criteria
2. User interviews (try for weekly)
3. Prioritize backlog based on potential reach, impact, confidence, and effort (RICE method)
4. Determine the right mix of bugs vs. enhancements to include each sprint (painkillers vs. vitamins). Weigh against existing tech debt initiatives.