I’m sitting across from a friend at Microsoft — someone I’d worked with, someone whose judgment I respected — and he’s telling me I’m making a mistake.

I’d just taken a job at Concur. A ten-million-dollar expense report software company. He looked at me the way you look at someone who’s about to jump off a perfectly good boat. “Expense reports?” he said. “There’s no business there. We already have that.”

He was right about one thing. Microsoft already had it. So did SAP. So did Oracle. Every major ERP vendor on the planet had an expense module. They bundled it in for free — an afterthought tacked onto the system companies were actually buying. The vendors didn’t make money on the module. They made money on the consulting fees to implement it.

And the product was garbage.

Nobody was waking up in the morning thinking about how to make expense reporting better. The module existed so the sales team could say “yes, we have that” during the RFP process. Nobody was measuring the actual cost of the bad solution — the hours lost, the compliance gaps, the finance teams building workarounds in spreadsheets because the tool they’d been given was designed by people who considered it beneath them.

My Microsoft friends saw a feature that their company gave away for free and concluded there was no market. What they couldn’t see — what nobody inside an ERP company could see — was that free and terrible is not the same as solved.

Free and terrible is an invitation.

Nineteen years later, SAP acquired Concur for $8.3 billion.

When Concur started, there was no “travel and expense management” category. There was no vertical software category at all. Enterprise software meant ERP — one system that did everything, most of it badly. The idea that you would buy a separate, dedicated product for one workflow was foreign. What existed was a bundle: buy the ERP, pay the consultants to implement it, accept whatever modules came along for the ride, and live with the results.

Expense management was one of those modules. Thrown in for free — not because it wasn’t important, but because no one had ever shown the market what good looked like. The incumbents had trained every customer to expect mediocrity. And mediocrity, delivered for free inside a million-dollar ERP contract, looks a lot like “good enough.”

It wasn’t good enough. It was a black box that companies tolerated because they didn’t know they had a choice.

The unfair advantage Concur had was not technology. It was obsession.

When your entire company wakes up every morning thinking about one problem — when expense management isn’t module 47 in your product suite but the only thing you do — you will eventually build something the bundled solution can’t touch. Not because you’re smarter. Because you care more. The ERP vendor has ten thousand engineers spread across a hundred modules. You have a team that eats, sleeps, and argues about receipt processing. That asymmetry compounds.

Every edge case the bundled module ignored, we studied. Every workaround the finance team had built in Excel, we turned into a feature. Every compliance rule that the ERP treated as someone else’s problem, we treated as ours. Expense management was a line item in their budget. It was our entire reason for existing.

That’s what “this is all we think about” means as a competitive strategy. It doesn’t sound impressive in a pitch meeting. But boring focus against institutional indifference is how categories get created.

I’ve watched this pattern play out twenty years later, from the other side — as the person helping founders find these openings.

Property Meld is the one I know best. Every major property management platform had a maintenance module. Every one of them was bad — not because the companies were incompetent, but because maintenance wasn’t their business. Leasing was their business. Accounting was their business. Maintenance was the checkbox. Ray Hespen saw the same opening I saw at Concur: an entire industry trained to expect mediocrity from a workflow they couldn’t avoid. He made maintenance the only thing his company thought about. The category didn’t exist when he started. It does now.

The pattern has a shape, and once you see it you can’t unsee it.

An incumbent bundles a workflow as an afterthought. It’s free, or close to free, because the real product is something else. The bundled module is maintained by a small team with no incentive to make it great — their job is to keep it functional enough that it doesn’t lose the deal. Customers tolerate it because they don’t know better, and because switching costs for the whole platform are too high to justify leaving over one bad module.

Then someone shows up who only does that one thing.

They’re competing not with the incumbent’s best effort but with the incumbent’s indifference. They don’t need to be better than SAP. They need to be better than SAP’s afterthought. That’s a different bar entirely.

Before Concur, nobody bought expense management software. Before Property Meld, nobody bought maintenance coordination software. Before Veeva, nobody bought pharma-specific CRM. Before Bill.com, nobody bought standalone accounts payable automation. The market wasn’t small — it was invisible. The incumbents had made it invisible by giving it away for free inside something bigger.

This is what the excitement bias actually costs founders.

It’s not that exciting markets are bad bets. AI, climate, the next platform shift — those are real opportunities. But the founders chasing them are competing with thousands of brilliant, well-funded peers all aimed at the same target. The founder who decides to fix the tedious workflow that every company in a vertical industry tolerates but nobody loves? She’s competing with apathy. Apathy is a much easier opponent than brilliance.

And the economics are structurally different. You’re not in an arms race with five well-funded startups trying to own the same horizontal category. You’re the only company in the room that thinks this problem is worth obsessing over. Capital requirements are lower because you’re not burning cash to outrun competition that doesn’t exist. Customers are stickier because the pain is chronic — they won’t switch back to the bad module once they’ve tasted something built by people who actually care. Every customer you land in a vertical market makes the next sale easier, because narrow industries talk.

The caveat is real: this only works when the problem is universal enough to scale — every company that employs travelers has expense reports, every property management company has maintenance. If only twelve companies have the problem, boring and small is just small. But if you’re looking at a workflow that every company in your industry suffers through, and the current solution is a bad module bundled inside a platform nobody chose for that module? That’s the opening.

My friends at Microsoft were smart people. They understood enterprise software better than almost anyone. What they couldn’t see was the gap between “we have that” and “we solve that.”

Microsoft had an expense module. Concur solved expense management. Those sound similar. They are not the same thing.

That gap — between having it and solving it — is where categories are born. And it’s invisible to the people inside the incumbent, because from their vantage point the module exists, the box is checked, and the customer signed the contract. What they can’t see is the finance team building workarounds at eleven o’clock at night. The property manager calling the same vendor for the third time because the system lost the work order. The rancher running loan math on a legal pad because no software in the world does it right.

If you’re in an industry right now and you can name that workflow — the one everyone tolerates and nobody prioritizes, bundled inside a platform and maintained by a team that wishes they were working on something else — you’re looking at the same thing I was looking at in 2001.

There was no market for it. There was no category.

That was the whole point.

The agency did good work. That was the problem.

Tyler Olson and Brennan Holloway are former Marines from Rapid City who ended up running an EMS company in wildland fire. Ten ambulances. Paramedics scattered across the country. A seasonal business where the Forest Service calls and you have fifteen minutes to accept a contract or decline it — and if you can’t confirm a Colorado-licensed paramedic is available right now, you turn down tens of thousands of dollars. Their people carried carbon copies into austere environments where a Starlink dish was the only link to the outside world.

They hired a social media agency to help with marketing. The agency did what agencies do. It produced content about being brave. Being a hero. Answering the call. It was competent. It was well-produced. And Tyler and Brennan took one look at it and their honest reaction was that it felt cringey.

They couldn’t have given you the theory of why. They just knew. It missed the vernacular. It missed the culture. It missed the way people in wildland fire actually talk to each other, which is to say not like that, not ever. The agency didn’t know what it didn’t know, and two guys who had lived it knew in a single glance.

Hold onto that scene. It’s about to explain the next ten years.

What the agency actually was

That agency was a preview of a machine.

Feed a large language model every piece of wildland-fire marketing that has ever been published and ask it for more, and it will hand you exactly what the agency did: brave, hero, answer the call. Not because it’s lazy. Because that is precisely what it’s built to do. It returns the average of everything it was trained on. It regresses to the most likely sentence. And the most likely sentence, by definition, is the one everyone has already heard.

The agency wasn’t failing at being creative. It was succeeding at being average. And to a domain expert, average and wrong are the same word.

The cost of everything went to zero — including the words

You already know the first half of this. The cost of building has collapsed. What used to take fifty engineers takes three. A nurse in Omaha with fifteen years of clinical judgment can ship software without hiring a CTO. The wall that kept people out of building has a door in it now.

But watch what that actually does. If everyone can build, then building stops being the thing that separates you. The same collapse that freed the domain expert to ship also freed everyone else to ship. When execution is free, execution is no longer an advantage. It’s table stakes.

And it isn’t just the code. It’s the copy, the landing page, the pitch, the launch email, the founder’s “story.” All of it now generates to eighty percent, instantly, for free, for everyone. Which means the market is filling up with the competent average of everything. Every landing page starts to sound like every other landing page. Every founder origin story starts to read like every other founder origin story, because they were all pulled from the same eightieth percentile of the same corpus.

Here’s the economics nobody wants to sit with: when the competent average becomes free and infinite, the competent average becomes worth nothing.

What stays scarce

Abundance on one side forces scarcity to reprice on the other. When “built things” and “competent content” both go infinite, value migrates to whatever the machine can’t manufacture. Two things survive: trust and attention. Those were always the scarce goods. Now they’re the only scarce goods.

And there is exactly one input that produces both and cannot be commoditized by a model trained on everyone’s content: a specific, true, first-person story.

Not a story about wildland fire. The story of fifteen minutes to say yes. The carbon paper in an austere environment. The Starlink dish as your only tie to the world. A model can give you the average of what has been said about that world. It cannot give you what you saw, because it wasn’t there. It wasn’t on the fire.

That’s the whole thing. Provenance is the product. Three properties — true, specific, first-person — and no model can fake any of them, because faking them would require having lived a life it didn’t live.

Storytelling is not a soft skill

We’ve spent a generation treating storytelling as garnish. The real work was the product; the story was what the marketing team wrapped around it afterward.

Flip it. When the product is free to build and the marketing is free to generate, the true specific story is the only part of the stack with a moat around it. It isn’t garnish. It’s distribution infrastructure. It is the one asset that earns trust and attention in a market where everyone else is shipping the same competent average.

Storytelling in this world isn’t a soft skill. It’s the moat.

And here’s who it hands the moat to

This is the part that should change how you think about your own odds.

For twenty years, the founder with no brand and no coastal network lost a specific race. They lost on access — who you could get a meeting with. They lost on execution — who could raise enough to hire enough to build enough, fast enough. Tyler and Brennan aren’t in San Francisco. They’re former Marines in Rapid City. Under the old rules, they lose to the well-funded, well-connected founder every time.

Both of those advantages just got commoditized. Access matters less when distribution runs through a story instead of a rolodex. Execution matters less when everyone can execute. What can’t be commoditized is the one thing the outsider has in surplus and the insider usually lacks: they lived it. They were on the fire.

The moat just flipped to the domain expert. The person who spent fifteen years inside a problem no outsider would notice is now holding the single asset a model can’t generate and a competitor can’t buy. Depth was always their edge in the product. It’s now their edge in distribution too.

“But I’m not a writer”

Neither are Tyler and Brennan. You don’t need to be.

The skill isn’t prose. The skill is refusing to sand your story down into the agency’s version. The failure mode is not bad writing — it’s telling your own story the way the machine would, stripping out the names and the fifteen minutes and the carbon paper until all that’s left is brave, hero, answer the call. The average. The cringey version.

Your only job is to notice the true specific detail and leave it in. The detail that makes an insider nod and an outsider ask a question is the detail no model would have generated, which is exactly why it’s the one worth keeping.

The last line

The agency could generate the word “hero” forever and never once land it, because it was never fifteen minutes from saying yes with a life on the other end of the call.

That’s the moat now. Not what you can build — everyone can build. What you were there for.

Go tell that.

Here’s a sentence Sean Patton can say that almost no one else can. He spent nearly two years of his life in combat, on the Afghanistan-Pakistan border, commanding Green Beret detachments, and never had a single panic attack. Then he started a business and had two in the first couple of years. Open sores on his back. Appendicitis. He put himself in the hospital.

This is a West Point graduate. A guy who was, in the eyes of everyone who knew him, a Captain America figure. And the lowest point of his life wasn’t a firefight. It was sitting in a gym he owned, having just lost hundreds of thousands of dollars, declaring bankruptcy, driving Uber, and Venmoing his twenty-four-year-old girlfriend so she could send him money to buy fish.

I opened our conversation by saying his story isn’t about military grit. It’s about what happens when grit alone isn’t enough. Sean spent ten years learning that the answer to every hard thing was to go harder. In the military, that always worked, because the mission always ended. You always redeployed. Someone always pulled you out. As an entrepreneur, he learned the brutal version of the truth: no one is coming to pull you out.

The phone stops ringing

I know a smaller version of this from my own life.

I had a high-powered job at Concur. Title, team, a persona that walked into rooms ahead of me. Then one day you quit, and the phone stops ringing. You find out how much of your personality was actually just the role. Before that, my first company failed outright. I’d raised money from investors, talked a bunch of people into believing, and watched it die. I lived on a couch. Got rid of my apartment. Sold my cars to make payroll. And it still went to zero.

So when Sean started describing the inner collapse underneath the financial one, I wasn’t nodding politely. The hardest part of failing isn’t the money. It’s the question that surfaces at two in the morning: if the thing I do is gone, and the thing I do was who I was, then who am I?

Sean’s answer, looking back, is that he’d built his entire identity on being a winner. His whole psyche ran on mattering, and mattering meant accomplishing, and accomplishing meant being seen as successful. Every bit of it pointed outward. The business didn’t create that fragility. It exposed it. And he’s blunt about the cost: if he’d done the inner work first, that company would have had a very different outcome.

Scarcity is mostly a story you tell yourself

The reframe that pulled him out is worth slowing down on, because it sounds soft until you’ve lived the moment it describes.

Sean talks about the gap between vision and reality. You set a vision of where the business should be. You look at where it actually is. And the space between those two points is where founders torture themselves. Most people cope by cheating one side or the other. They deflate the vision so the gap feels smaller, or they inflate reality and tell themselves it’s not that bad, right up until the tax bill lands and they realize it absolutely is that bad.

His insight is that the gap never closes. The gap is the work. The gap is the life. The trap is believing that once you hit the number, once you land the client, once you raise the round, then you’ll finally be enough. That goalpost moves forever. He had to learn to operate from what he calls enoughness rather than from “I’m not enough and I have to achieve or I’m nothing.”

And here’s the part that gave him leverage. He’d been to rural Afghanistan. He’d seen places that have known nothing but violence for a thousand years. So even mid-panic about his gym, some part of him could step back and recognize that he wasn’t actually fighting through scarcity. He was, in his words, getting to play. He got to create. Most of us, he argues, are running a scarcity story in our heads while we’re standing in abundance.

The leader on the other side

The strange twist is that the breakdown made him a better leader than the medals ever did.

Two pieces of what he rebuilt stuck with me. The first is a distinction he draws between being kind and being nice, and he thinks the Midwest in particular gets this wrong. Nice is the conversation where you don’t want to upset anyone, so you smooth it over and let the real thing fester. What you’re protecting, he says, is surface-level peace sitting on top of turbulence, unsaid expectations, and agreements that were never actually made. Kind is telling someone the true thing even when it makes you both uncomfortable. He hates confrontation. He still made directness the price of admission to his culture, because the alternative quietly rots the company from the inside.

The second is about hiring, and it’s the mistake he openly admits he kept making. The military trained him to spot raw material, to look at a person and see what they’re capable of. What it never taught him was to ask whether they actually want it. In his units, if you were there, you were all in. That’s not how it works when you start hiring. Founders assume everyone will care like an owner, and most people won’t, and that’s fine if you have the honest conversation up front about why each person is really there. You can’t lead someone whose motivation you’ve only assumed.

He also punctured a myth I see strangling young founders. We worship the nineteen-year-old who drops out of Harvard. The average successful first-time entrepreneur is around forty-five. Sean’s own advice to his younger self is almost anticlimactic: go work in the industry first. He wanted to own a gym and refused, on principle, to go work in one for a year. He’s pretty sure a few months at the front desk would have either sharpened the dream or killed it cheaply, before it cost him everything.

What to actually do

For the founder reading this at the bottom of their own gap, Sean’s most useful move costs nothing and feels backwards.

Invest in yourself first, on purpose, as part of the business plan rather than in spite of it. He keeps an explicit priority order: his own mental and physical health, then his wife, then his son, then his clients, then business development. In that order. When those get scrambled, everything gets scrambled, and a founder who’s stopped sleeping and let his closest relationship go to crap is not showing up as the person the company needs.

The other move is to walk toward the fear instead of pretending it doesn’t exist. Sit down and actually run the failure. The business fails, the new location flops, the thing you’re terrified of happens. Then what? You still have a roof. You’re still alive. Sean sold his truck and his downtown condo and moved into a tiny house before he had to, on purpose, so the fear had less power over his decisions. Naming the worst case out loud tends to shrink it. Refusing to look at it lets it grow until your body reacts like there’s a gun to your head.

His new keynote argues that robots are going to be the best managers, that pure task-routing management won’t survive the next few years, and that what’s left is smaller teams of high-agency people who follow leaders worth following. Which lands the whole conversation in one question he now asks the founders he coaches.

You’re going to spend years becoming a more capable operator. Are you also becoming the kind of person those people actually want to follow?

Grit will get you to the fight. It won’t tell you who you are when you lose one.

March 2001. Concur’s stock was under a dollar. We laid people off in waves, then laid off the people who’d been doing the laying off. I sat in a conference room where a spreadsheet decided who got to keep working, watched a name I knew get highlighted, and told myself the market had done this to us.

It hadn’t. The market pulled the trigger. We’d spent two years loading the gun. We chased every adjacent product, hired ahead of revenue, and believed our own growth story because money was cheap and everyone in the building said go faster. The crash didn’t kill the company that made sense. It killed the version we’d bolted on top of it. When the dust settled, the thing still standing was the small, boring, unglamorous business that had been there the whole time, buried under everything we’d added to look like a rocket ship.

I’ve spent the twenty-five years since reading the same books you have. The ones about the companies that made it. I’ve come to think most of them are the wrong shape.

Study ten great companies and you get ten different recipes. One won on timing that couldn’t be repeated if you tried. One won because a competitor imploded at exactly the right moment. One won on a founder who could sell space heaters in July. The lessons don’t stack. They barely rhyme. Success is a lock with a thousand different keys, and the key that opened one door tells you almost nothing about yours. You can read every founder biography ever written and come away no better at building your company, because what you’re really reading is luck assembled backward, after the win, to make it look like a plan.

Failure is the opposite. Watch enough companies die and you stop seeing variety. You start seeing a short list of the same causes, over and over, wearing different clothes.

Think about a generationally talented athlete who never becomes what the talent promised. You know the story before I tell it. It wasn’t the jump shot. It was alcohol, or money he couldn’t hold onto, or the people he let into the building, or a relationship that pulled him apart off the court, or an injury he hid instead of treating. The gift was never the problem. At that level the gift is common. What separated the ones who made it was the absence of a small number of self-inflicted wounds, the same wounds showing up in different players decade after decade.

Startups work this way too. The talent, meaning a real product, a real market, a founder who can actually build, is more common than the mythology admits. What’s rare isn’t brilliance. It’s the absence of the recurring ways founders take themselves out.

Here’s the line I keep coming back to: the founder loads the gun, and the market only pulls the trigger. The crash, the pandemic, the funding winter, none of them invent the cause of death. They just arrive and find the gun already loaded. Which means the useful question was never “how will you win.” Nobody can answer that honestly; winning hasn’t happened yet. The useful question is which round you’re chambering right now.

So I stopped asking founders how they planned to win. I started running a pre-mortem. It’s three years from now, I tell them, and the company is dead. Walk me backward through how it happened. The good ones can do it without flinching. And the causes almost always come off the same short list.

The money wound. Not too little. Too much, spent on the wrong things. This is the founder version of the athlete who signs the contract and is broke in three years. You raise a round and the burn quietly becomes the strategy. I’ve watched a founder with a healthy $30K-a-month business raise his way into a sick $300K-a-month one, then spend eighteen months trying to grow into a number he’d invented on a slide. The capital didn’t buy speed. It bought a taller cliff. Money papers over the exact problems a smaller company would have been forced to solve while they were still cheap to solve.

The wrong crowd. Athletes get the entourage. Founders get the wrong investors, the wrong first three hires, the advisor whose advice is really a rerun of his own last company. Slowly you start building for the people in the room instead of the people paying you. The tell is subtle. Ask a founder in trouble what their board wants and they’ll give you a crisp answer. Ask what their best customer actually did last week and they go quiet. The people around the table have replaced the people the product is for, and nobody noticed the swap.

The partnership that cracks. The relationship that takes a founder out usually isn’t a marriage. It’s the co-founder. Two people who never wrote down who decides what, who split the equity in a hallway, who mistook liking each other for being aligned. It holds fine while things go up. Then a hard fork comes, pivot or persevere, raise or stay lean, and the crack that was always there opens all the way. I’ve seen more good companies end in a founder divorce than in a competitor’s press release. It’s the quietest cause of death and one of the most common.

The hidden injury. Every founder has a part of the business they don’t understand and hope someone else will handle. I call it the black box, and the most dangerous one is never the tech stack or the cap table. It’s the founder’s own head. The boredom that arrives in year three and gets relabeled as strategy, so the founder who built the thing suddenly wants to build a different thing inside the same company. The psychological debt that never gets paid down because our whole ecosystem treats a founder saying “I’m not okay” as a weakness to hide rather than an injury to treat. Far more companies die of the founder quietly checking out than of anything a rival did. It never makes the post-mortem, because the founder writes the post-mortem.

The talent that never turns into work. This one looks most like success, right up until it doesn’t. The product is real. Customers even use it. But the founder keeps adding, feature eleven, feature twelve, furniture nobody asked for, instead of making the three things that matter genuinely exceptional. What looks like product-market fit is often product-market tolerance: customers who stay but don’t lean in, who’d be gone the day something sharper showed up. The gift was there. The discipline to aim it at the few things that mattered wasn’t.

None of these are market problems. That’s the whole point. And you don’t beat the list by finding a better key. You beat it by keeping the list short. Ray Hespen built Property Meld into an eight-figure business from Rapid City, South Dakota, and the thing that stands out isn’t a clever growth hack. It’s what he refused to do. He went embarrassingly narrow. He never raised his way into a burn the business couldn’t carry. He stayed in the weeds of the one industry he actually understood. No entourage, no vanity round, no checking out to go chase the shinier idea. Boring, on purpose, for a decade.

That’s why I trust the failure list more than any success recipe. I can act on it. I can’t teach you to win on timing; timing isn’t a skill, it’s weather. But I can hand you the short list of ways founders take themselves out and make you look at it, honestly, before the market ever walks in the door. The success stories are entertainment. The failure list is a checklist.

You probably already know your answer. Most founders do. They can feel which one is already sitting in the room. They just don’t want to say it out loud yet, because saying it out loud means doing something about it.

So say it. Three years from now, the company is dead. Which round did you chamber?

January 18, 2025. Michelle Bonds pours the last glass anyone will ever order at Trova Wine + Market, her bar in Dallas’s Preston Center. Four and a half years, back-to-back nominations for best wine bar in the city, a list built one bottle at a time. She’s closing anyway. When she explains why to local reporters, Ozempic is one of the reasons she gives by name: fewer glasses per table, less food ordered per customer, even after she put high-protein salads on the menu trying to chase the GLP-1 crowd instead of losing them.

A wine bar closing in Dallas is a rounding error. Fourteen months later, the same force showed up on a balance sheet nobody could round away.

February 2026. Kraft Heinz’s new CEO, Steve Cahillane, walks into his first earnings call and kills the plan his predecessor spent a year building: splitting the company into two separate businesses. Instead he commits $500 million to something nobody at Kraft Heinz was budgeting for three years ago. Reformulating the entire portfolio around protein, away from sugar.

That’s not a branding refresh. That’s a 157-year-old food company rebuilding its balance sheet around the same drug that closed Michelle Bonds’s wine bar.

In November 2024, I wrote that GLP-1 medications wouldn’t just treat diabetes and obesity. They’d regulate impulse, and impulse is the fuel under a huge share of the American economy: snacking, drinking, impulse retail, the whole advertising machine built on triggering a craving. Pull the fuel out and you get a slow-motion earthquake under restaurants, alcohol, malls, casinos, advertising.

Fifteen months later, the earthquake has a balance sheet, and it started showing up in bars like Michelle Bonds’s before anyone in a boardroom noticed. Statista puts current GLP-1 use at 1 in 8 Americans, which means Bonds wasn’t watching an outlier. She was watching the leading edge.

The alcohol industry stopped arguing about this

Gallup has the share of Americans who say they drink at all falling from 62% to 54% in two years. That’s not a survey artifact, and it’s not just Trova. Trade press covering the spirits industry has started asking whether “no-and-low” alcohol is the last growth category left standing. Diageo’s US spirits sales are down 15.4% in the first half of fiscal 2026, and the damage isn’t spread evenly. Its tequila brands took the worst of it: Don Julio down 20.9%, Casamigos down 30.9%. Morgan Stanley’s research desk models a theoretical ceiling of a 55% drop in alcohol demand if GLP-1 adoption reached the entire overweight and obese population. Nobody’s near that ceiling yet. Nobody’s betting against reaching it, either.

Restaurants already rewrote the menu

Bars felt it first because a cocktail is pure impulse. Restaurants are close behind. Olive Garden rolled its “Lighter Portions” menu out nationwide, citing GLP-1 users as one reason. Cuba Libre, the multi-location rum bar and restaurant chain, built an entire “GLP-Wonderful” menu around smaller plates. In New York, Tucci’s owner Max Tucci has talked openly about shrinking the meatballs and arancini on his menu, not because food costs went up, but because fewer people can finish what used to be a normal portion. None of this is a chain quietly testing something in three markets. It’s public menu redesign, company by company, in under two years.

Food companies aren’t hedging anymore, they’re rebuilding

A Cornell study built on real transaction data from 150,000 households found GLP-1 adoption cuts grocery spending by more than 5% within six months, about $390 a year off a typical $7,400 grocery bill. Among higher-income households, the drop is closer to 8%. The cuts aren’t spread evenly either: chips and savory snacks down about 10%, sweet bakery down 9%, cookies down 7%, while yogurt and fruit actually ticked up. EY-Parthenon estimates GLP-1 use could pull $12 billion out of snack sales over the next decade. A year ago that was a forecast. Now it’s a capital allocation decision. Kraft Heinz’s $500 million turnaround is the headline, but it’s not alone. Nestlé built an entire brand, Vital Pursuit, around GLP-1 users and is expanding it. PepsiCo is reformulating Lay’s and Gatorade, stripping artificial colors, and piloting mini-meal formats under Sabra and Siete. When three of the largest food companies on earth start moving in the same direction inside the same eighteen months, that’s not a trend piece anymore. That’s the market repricing an entire category of demand.

The impulse-control thesis has a real study behind it now

The most interesting new evidence isn’t about spending. It’s about behavior. A 2026 Rutgers study published in Criminology found that GLP-1 use weakened the statistical link between impulsivity and violence by 62%, and between alcohol and violence by 52%, among current users. The researchers are careful to call this a first step, not proof of causation, and they’re right to be careful. But this is the mechanism underneath everything else in this piece: a molecule that changes how strongly people act on craving, at population scale. Alcohol sales and snack sales are downstream symptoms. The impulsivity data is the actual story.

What isn’t priced in yet

Here’s the part that should worry anyone modeling this as a straight line. The people who can afford these drugs and the people who can’t are splitting apart faster than the coverage system can keep up. Thirteen states covered GLP-1s under Medicaid for weight loss as of January 2026, down from sixteen the year before. Employers are dropping the coverage too. Medicare is stepping in with a $50-a-month bridge program starting this July, which is itself an admission that the private market wasn’t going to close the gap on its own. A recent survey from The Health Management Academy found more than 70% of Americans already believe these drugs are only realistically accessible to people with money or good insurance.

And the drugs themselves aren’t free of cost to the people taking them. A University of Utah study published in August 2025 found that GLP-1-driven weight loss cuts lean muscle mass by roughly 10%, and that muscles get measurably weaker even in cases where they don’t visibly shrink. Worse for older adults, who are already fighting to hold on to mobility. Nobody’s modeling that against the productivity gains everyone’s excited about. Somebody should.

The next earnings call to watch

Kraft Heinz just told you where the food industry thinks this goes. The alcohol numbers just told you where the beverage industry thinks this goes. Nobody yet has told you where the insurance industry, the labor market, or the disability system thinks this goes, and those are the three that decide whether this transformation makes the country healthier or just sorts it into two tiers faster than anyone can build a bridge between them.

Kraft Heinz didn’t spend $500 million because a Substack post was right. It spent $500 million because the data stopped being debatable. Michelle Bonds didn’t need a study to know something had changed in her bar. She just ran out of runway before the rest of the economy caught up to what she already knew.

Which industry hasn’t made its move yet?

Sources

• Trova Wine + Market closure, Michelle Bonds — The Drinks Business, “Ozempic partly to blame for wine bar closure, says owner”

• GLP-1 usage rate (1 in 8 Americans) — Statista, “Key Figures: Usage and Opinion of GLP-1 Medication”

• Kraft Heinz $500M reformulation, EY-Parthenon $12B snack forecast — GV Wire, “Big Food Pours Millions Into Rebrands as Obesity Drugs Reshape US Demand”

• Gallup drinking rate, Diageo/Don Julio/Casamigos sales — Poached, “Is GLP-1 Use Cutting Into Your Alcohol Sales?”

• “No-and-low” alcohol category trend — The Spirits Business, “Is the ‘Ozempic era’ an opportunity for no-and-low?”

• 55% alcohol demand ceiling — Morgan Stanley, “Next Order Effects of Increasing GLP-1 Use”

• Olive Garden “Lighter Portions” nationwide rollout — Fortune, “Ozempic mania has even Olive Garden and The Cheesecake Factory cutting back on portion sizes”

• Cuba Libre “GLP-Wonderful” menu, Tucci portion changes — Fox News, “GLP-1 drugs like Ozempic are reshaping American dining habits and restaurant industry”

• Cornell grocery spending study (150,000 households) — Cornell Chronicle, “Ozempic is changing the foods Americans buy”

• Nestlé Vital Pursuit, food company reformulation — CNBC, “GLP-1 drugs are changing how Americans eat”

• PepsiCo Simply NKD / Sabra / Siete — Food Dive, “Food makers poised to withstand threat of GLP-1 weight loss drugs”

• Rutgers/Criminology impulsivity-violence study — Medical Daily, “Ozempic May Be Reducing Violent Crime in America”

• Medicaid coverage drop (16 states to 13) — KFF, “What to Know About the BALANCE Model for GLP-1s in Medicare and Medicaid”

• Medicare $50/month bridge program — NPR, “Medicare launches weight loss drug option with $50 copay”

• 70% believe GLP-1s are only accessible to the wealthy — The Health Management Academy, “The GLP-1 Access Gap”

• University of Utah muscle/lean mass study — University of Utah Health, “New Study Raises Questions About How Ozempic Affects Muscle Size and Strength”

Friday, peak fire season. Minuteman EMS owes its crews a quarter of a million dollars in payroll. The federal government owes Minuteman twice that for work already completed on fires across ten states. Both things are true at the same time.

That’s the business Tyler Olson and Brennan Holloway run. Ten ambulances, three rescue teams, and single paramedics scattered across the country. More than 40 people on payroll at peak. And a payment gap that shows up every single season like weather.

Here’s how the gap works. When a wildfire contractor finishes an assignment, the government issues a form called an OF-286. It’s an official acknowledgment of what the government owes you: these people, this equipment, these days, this amount. It is about as close to guaranteed money as a receivable gets. The only problem is when. Depending on which office is paying, it takes a few weeks to several months to arrive.

Payroll doesn’t wait a few weeks to several months. Wages go out every payday. Fuel, hotels, and flights go out while the crews are still on the fire. So contractors lock up huge amounts of their own capital just to float the gap. And most of them start the season with their backs already against the wall. Equipment, certifications, staffing, insurance, readiness: the spending starts months before the first dispatch. By the time an assignment comes in, much of their working capital is already committed.

The ones who can’t float it do something worse: they pay their crews late, or short. Some don’t pay at all until the government pays them first. Firefighters who spent fourteen days on a fire line working sixteen-hour days go home and wait on money they already earned, with their own bills stacking up. Delayed wages also put contractors on the wrong side of Department of Labor wage laws, which turns a cash-flow problem into a legal one. And plenty of companies simply turn down work they can’t afford to front.

Tyler hears these stories all the time. Firefighters from other companies tell him they’re owed $25,000, $30,000, sometimes $40,000 in unpaid wages because their contractor is still waiting on government reimbursement.

One conversation has stayed with him. A husband-and-wife team worked COVID vaccination deployments and wildfire seasons for the same contractor. When Tyler spoke with them, they had gone three to four months without being paid and were owed nearly $90,000 between them. The work was done. The government would eventually pay. Their paychecks simply hadn’t arrived.

Minuteman took the opposite path. Tyler’s rule is simple: do the work, get paid on time and in full, every time. Holding that line has taken disciplined cash management and, some seasons, finding money wherever it could be found. But employees notice. Firefighters tell each other which companies always pay and which ones don’t, and “go work for Minuteman, they’ll keep you paid” recruits better than any campaign. It’s also the problem Tyler and Brennan believe they can solve for the rest of the industry.

They lived this before they ever built software. A few years ago they applied for a bank line of credit to smooth the season. The bank looked at their balance sheet and offered $250,000.

One payroll run. In a slow year, a single peak-season payroll run at Minuteman hits $300,000 to $450,000. The bank’s answer covered less than two weeks of the problem.

The bank wasn’t being stupid. It was being a bank. Banks underwrite balance sheets and credit history, and a seasonal contractor with lumpy revenue and equipment in the field looks risky by every measure a bank knows how to read. The half-million dollars in government-guaranteed receivables sitting in Minuteman’s filing cabinet? Invisible. There’s no line on a loan application for “the U.S. Forest Service owes me this and always pays.”

What the platform knows

Regular readers know Tyler and Brennan’s company, The Journeyman. Two former Marines from Rapid City who ran ambulances on fire lines, got tired of turning down contracts because they couldn’t find a Colorado-licensed paramedic in fifteen minutes, and built the system of record for wildland fire contractors. Every resource tracked, every assignment logged, every document filed. Over 4,200 firefighters across more than 150 companies on the platform.

Now watch what that position makes possible.

When Tyler first walked me through the idea, he called it payday loans for fire contractors, half joking. It’s a decent shorthand and a terrible description, because the mechanics are the opposite of a payday loan. A payday loan advances money you hope to earn, at rates designed around the chance you won’t. What The Journeyman is designing advances money you already earned, acknowledged in writing by the United States government. The only variable is timing.

The shape of it: a contractor finishes work and uploads the OF-286. From inside the platform, they request an advance. A lending partner funds roughly half the receivable’s value, enough to cover payroll and the overhead that keeps crews in the field, ideally within days. When the government pays the full amount, the advance clears first and the rest flows to the contractor, minus a modest fee. The half that was never advanced is the built-in cushion. Nobody’s betting on potential income. The bet is that the government pays its own paper, which it does.

The reason this works isn’t financial engineering. It’s that underwriting is mostly the cost of not knowing things, and the platform already knows them. The dispatch order arrives when a resource is ordered, before the work happens. The OF-286 arrives when it’s done. The two documents have to match: same agreement number, same incident, same dates. The platform holds both, plus the season-over-season work history behind them. A bank sees a balance sheet. The Journeyman sees 716 days on fire across ten states, documented as they happened.

That knowledge doesn’t stop being valuable at underwriting either. When you’re on the fire line, the guy sleeping in the tent next to yours might work for a company that shorts crews in October. Word travels. Tyler and Brennan know which operators run clean because they’ve shared dispatch calls with them for years. No credit bureau has that data.

The financing is an extension of that same trust. A contractor who can borrow against work already completed doesn’t have to choose between making payroll and taking the next assignment. Crews get paid on schedule. And the agencies fighting the fires get a workforce that shows up next season.

The pattern underneath

I keep watching the same movie at Wildfire Labs.

AgCents, another one of our companies, started as analytics for cattle producers. Then Brady and David realized their price models and herd data let them see lending risk that rural banks couldn’t, and they closed $795,000 in cattle loans last spring. First the workflow, then the record, then the capital.

The Journeyman is running the same play in a different vertical. Vertical software earns trust by solving the daily problem. Solving the daily problem generates verified operational data. And verified operational data is underwriting, which means the software company can eventually offer financial products that banks, with all their capital and none of the context, cannot.

Notice what’s required to pull this off. Not a fintech background. Not a banking license; a lending partner brings the capital and carries the credit risk. What’s required is knowing the customer so well that you understand their cash-flow problem better than their bank does, and holding the data that proves the money is real. Domain expertise gets you the first product. The first product gets you the data. The data gets you everything after that.

The honest caveat: this is a plan, not a product. The terms are still being worked out with the lending partner and the lawyers, and the first season will be a small proof run with Minuteman and a handful of companies. It may ship looking different than what I’ve described. Early-stage ideas usually do.

But the direction is the thing. Two paramedics from Rapid City are about to offer a financial product to an industry that Wall Street doesn’t know exists, because they’re the only ones who can verify the work behind the receivable. The moat was never the code.

And underneath the strategy there’s a simpler outcome, the one Tyler and Brennan actually care about: the people who did the dangerous work get paid on time.

That’s the product.

So here’s the question worth sitting with, whatever niche you’re building in.

Your software watches your customers work every day.

What does it know that their bank doesn’t?

David Woodbury was traveling when his phone rang. “You have an interview in thirty minutes.” He hadn’t applied for a job since 2004. His wife had. She’d quietly written a résumé and sent it to a handful of places, because she’d hit the end of her rope — not with him, exactly, with the starts and stops. Fourteen years of them. The hamster wheel of founding, pivoting, founding again.

That call was the wake-up.

David started at a Verizon kiosk with no degree, no investors, no playbook. Fourteen years and seven companies later — four of them exited — he’d built the number one wireless store in the country, moved a VC-backed startup from Seattle to a small town in South Dakota, and helped shape the go-to-market for Chewy before it went public. He’d also hit a wall hard enough to walk away from founding entirely for three years.

I made a version of that Seattle-to-South-Dakota move myself — Redmond, then Bellevue with Concur, then back to the Black Hills. So when David and I sat down for a recent episode of Startup Stories from the Treehouse, I expected to talk shop. We spent most of the hour somewhere more honest. Three things he said have stuck with me since.

Simultaneously Up and Down

We talk about startups as a roller coaster. Someone corrected David on that last week, and he repeated the correction to me: it’s not a roller coaster. A roller coaster goes up, then down. Founding goes up and down at the same time — self-doubt and confidence in the same conversation, the same minute. “That’s so disruptive to your nervous system,” he said. Physiologically. You ask yourself whether you’re doing the right thing while being certain you are, all day, every day.

That’s what burnout actually looked like for him, and it wasn’t dramatic. It was drift. He stopped wanting to talk to people, so he’d let everyone else talk. Eighteen-hour days watching crypto candles on top of trying to run a business. Lunch at Killian’s with three IPAs in by 12:30. At one point, he told me, the only thing on his calendar was haircuts.

He named the mechanism better than anyone I’ve heard: decision fatigue. Burnout isn’t mostly about hours. It’s standing at a crossroads — take the consulting gig to fund the dream, or live on a buddy’s couch another month — over and over, with no framework to make the call. Fight or flight takes the wheel. For David, fight looked like starting new things too fast and never finishing them. Flight looked like quitting at four weeks because the traction wasn’t there yet, before he’d turned over enough stones.

You Can’t Outwork It

Here’s the part founders don’t want to hear. “If you don’t build a system to offload some of that, you’re one hundred percent guaranteed to burn out.”

David’s outlet became the bike. He got fit, got competitive, and — this is the part that matters — found that the training taught him more about business than most of his businesses had. The lesson was consistency, not intensity. Showing up, and deliberately carving gaps into his life to invest in himself. “If you’re not healthy, your business isn’t going to be healthy. Plain and simple.”

I’ve lived the other version. I used to stack meetings back to back to back with no white space, and I turned into a dull crayon — marching through tasks, creative about nothing. The white space was what made me better. Not a reward for the work. A requirement of it.

David put a finer point on it with something he picked up from endurance training: the brain can’t really hold focus on one thing for more than about forty-five minutes. So treat work like intervals. Three or four focused forty-five-minute blocks, executed hard, beat eight hours of staring at a screen. You can’t outwork the wall. You can only build a life that keeps you from walking into it.

The Framework He Didn’t Have

When I asked what he’d tell his twenty-five-year-old self, he didn’t reach for a tactic. He reached for the thing that would have prevented the wall: a framework, and the data to feed it.

David’s biggest business mistake was going too wide. Camp Native, his campground booking startup, chased all fourteen thousand campgrounds and RV parks at once. Hipcamp went narrow — unique properties, an Airbnb feel — and won the market. Years later, at an agency, he flipped it: they declared one platform their entire specialty and effectively borrowed that vendor’s twenty-million-dollar marketing budget instead of straining their own two hundred thousand. Niching, he said, is really about “qualifying you in the room” — being narrow enough that the customer already knows you’re worth their time before the meeting starts.

The reason this matters more now is that the cost of finding out has collapsed. He reminded me that Camp Native burned a quarter-million dollars before they put a demo in front of a single customer. Today you can stand up a working demo for three hundred dollars in credits. The expensive part of being wrong got cheap. What didn’t change is where the edge actually lives — not in the model. Domain expertise is not sitting in an LLM somewhere. It’s in the head of the nurse, the civil engineer, the founder who knows one industry cold. The tools just let you carry that knowledge to market for almost nothing.

The Question Before the Grind

The thread under all of it is a question David thinks more founders should ask before they sign up. Do you actually like the grind? “Nobody really cares what you do,” he said. “It’s your time, how you spend it, how you invest it.” If you’re founding to impress someone or prove something, the wall is coming. You can’t outwork a why that isn’t yours.

Three years off the wheel, David is building again — a startup called Grow Army, an operating framework for early-stage companies, built around the exact thing he didn’t have when he burned out. When a fund in Sioux Falls asked last week whether he’d move back to South Dakota to do it, his answer was “don’t twist my arm.”

I know the feeling.

The full conversation goes further — the Chewy employee who took no equity and what that taught David about building incentive pools for early staff, what accelerators get right and wrong, and the move that turned a $25,000 grant into roughly $10 million in local impact. Listen to the episode.

To learn more about what David’s building, find him on LinkedIn at davidnwoodbury or follow along with Grow Army.

Did this one land? If you’re a founder who’s hit the wall — or built a system to keep from hitting it — I’d like to hear how. Reach me at todd@wildfirelabs.io.

Interested in taking your startup to the next level? Learn more about Wildfire Labs’ accelerator program at wildfirelabs.io.

The founder version of burnout isn't the one in the wellness essays. It's not I got tired and made a bad call. It's quieter than that. You run hard. You hit the red line. You keep going, because going is what founders do. The work doesn't visibly fall apart. Meetings still happen. Decisions still get made. But somewhere in there, the part of you that sees a problem from an angle nobody else has thought of stops working. You don't notice. Nobody around you notices. You just default to pattern-matching instead of thinking. You're still moving the ball. You're just not the player who used to find the unexpected gap.

That’s the actual cost. By the time you feel it, you’ve been losing it for months.

I ran that operating system for decades, through GaiaSphere, NetEffects, MAPP, then Concur. Same belief the whole way: outwork everyone, win on volume, sleep when the deal closes. Early morning breakfasts. Back-to-back meetings through the day. Dinners with whoever was in town. Exercise was the first thing to go. Sleep was the second.

I thought I was paying the price in tiredness. I was actually paying it in originality. The work I’m proudest of, the times I saw the angle no one else saw, happened in seasons when I was rested enough to have a thought worth having. The work I’m least proud of happened when I was running on willpower and not much else.

When I finally bothered to look, the pattern was obvious. At my best, I was sleeping deeply, moving constantly, walking miles a day. At my worst, all of it had collapsed — less sleep, less movement, a body slowly giving ground. Same person, years apart. It wasn’t aging. It was the slow trade I’d been making my whole career: body for output, day after day. Eventually it came out of my brain.

I needed someone whose job was to keep that from happening. I couldn’t find one.

What sent me looking is older than any of that. My grandfather had Alzheimer’s. He was the man I was closest to in my family. The pathology of dementia begins decades before the first symptom. Meaning the trade I was making in my forties was also a trade against my seventies. Same brain, two clocks. The seasonal version is I’m thinking less originally this quarter. The lifetime version is what I watched happen to him.

I’ll come back to him.

What I needed first was someone to tell me where I stood. The standard answer is your GP. Mine looked at my bloodwork and said what every GP says: you’re in normal range, keep doing what you’re doing. You’ve heard the line. It sounds reassuring. It isn’t.

Normal ranges are calibrated on the average American, and the average American is metabolically unwell. Normal means average for a sick population. If you’re optimizing against dementia with a family history like mine, normal isn’t the goal. It’s the floor.

There’s no integrator. A GP triages and prescribes. A trainer programs workouts. A nutritionist plans meals. A sleep doctor treats apnea. Every one of them is a subcontractor. None is the general contractor. None looks at the whole picture. None is paid by your VO2 max twenty years from now, or your cognitive function decades out.

For a founder, that gap matters more than for most people. Your brain is the asset. Everything else you’re paid to do flows out of it. And nobody in the system whose job it is to keep you healthy is asking whether you can still see problems clearly.

Start with the philosophy. The noise in this space is overwhelming: cold plunges, a cabinet full of supplements, peptides, red light. Most of the biohacking community is optimizing the wrong end of the curve. Those things might give you a little more, if you’ve already locked in the part that comes from food, sleep, exercise, and stress. Four pillars. Boring fundamentals. Same playbook as boring software. These aren’t wellness. They’re the substrate the work runs on.

The build has two parts.

First, synthetic personalities. Three world class experts, each one a lens on my data. Matthew Walker for sleep. Peter Attia for exercise. Eric Topol for diet.

Walker wrote Why We Sleep. The mechanism that ties sleep most directly to dementia is the glymphatic system, the brain’s waste clearance pathway that runs during deep sleep. Less deep sleep, less amyloid cleared, more risk over time. His frameworks become the lens for my sleep data.

Attia wrote Outlive and runs a longevity practice. His exercise prescription is unsexy and correct: serious weight lifting plus cardio in two tranches. Strength training preserves muscle mass and grip strength. Both are top predictors of all cause mortality and cognitive resilience. VO2 max grows the engine. Zone 2 makes the engine efficient: miles per gallon, in his analogy. Most people do one of the three. All three matter.

Topol is a cardiologist with mortality data behind every recommendation. His diet bias is Mediterranean. I picked him over a more famous nutrition voice because he’s credentialed enough to trust and evidence driven enough to ignore the influencer crowd. That filter is the lesson for picking your own experts. Credentialed. Evidence driven. Willing to be unsexy.

Walker doesn’t have an opening for me. Even if he did, he’d give me an hour, once. An AI version of him on my data, every week, with memory across months — that’s better than the real Walker for the job I need done. Not a worse version of the expert. The version of the expert that’s actually useful.

The second part is the data pipeline. Cowork project. Apple Health export: sleep architecture, heart rate, HRV, workouts, steps. Cronometer export for food: calories, macros, micronutrients tracked daily. Bloodwork panels for context. Subjective notes on stress, added weekly. The synthetic personalities sit in the system prompt. The output: trends across the four pillars, and one incremental change per pillar.

Weekly. That’s deliberate. Daily is noise. One bad night doesn’t mean anything, and you’ll chase fluctuations into bad behavior. Monthly is drift. Problems compound for weeks before you see them. Weekly is the smallest unit where signal beats noise and the largest where you can still steer.

What the coach surfaces is trends, not numbers. Your deep sleep has been sliding for a month, and it tracks a drop in your workouts. That’s actionable. Last night’s deep sleep number is not.

One specific change the system surfaced. I’d been taking vitamin D before bed. The coach flagged the timing through Walker’s framework: vitamin D suppresses melatonin. Move it to morning. I did. The very first night, my deep sleep jumped. Small change, real result. Nobody in the existing system was going to catch that, because nobody was looking at all of it together.

A few surprises after running this against years of data.

My HRV hadn’t moved, even across stretches where my activity swung dramatically. The story my body kept telling me was I’m just getting older. That story was a lie. The gap was activity volume, which is controllable. The cost showed up in my brain. Workout days produce meaningfully more deep sleep and REM than rest days, and the effect carries to the next night. Drop the workouts, drop the deep sleep, drop the amyloid clearance. A trainer would never have told me that. A sleep doctor wouldn’t have looked at my workouts. Only the integrator caught it.

Evening workouts have zero sleep penalty for me. That’s the opposite of conventional wisdom. If you’d asked me before I had the data, I would have moved evening workouts earlier. The data says don’t.

My observed maximum heart rate runs well above what the standard age formula predicts. That changed my entire Zone 2 and VO2 max ranges. A generic recommendation engine would have prescribed me the wrong zones. Too easy on the hard days, too hard on the easy ones. Only personalized data caught it.

None of these are biohacks. They’re what you only see when an integrator is watching all of your data, all of the time, against the asset you’re trying to protect.

If you want to copy this, the recipe is short.

Pick four fundamentals. Sleep, food, exercise, stress. Don’t add a fifth.

Pick one credentialed expert per pillar. Filter for evidence driven. Ignore the influencer crowd.

Build a Cowork project. Load the synthetic personalities into the system prompt. Walker’s frameworks for sleep. Attia’s for exercise. Topol’s for diet. Pick your own for stress.

Pipe data in weekly. Apple Health for sleep and exercise, Cronometer for food, bloodwork from your last panel, subjective notes for stress.

Ask for trends and one incremental change per pillar. Not zero to sixty. Small shifts that compound.

That’s the whole thing. Tools every founder already has access to. Your brain is the asset. The four pillars are the inputs. The integrator is the role. The role exists now, at a price you can afford.

When I was young, my grandfather told me one thing about getting older. Fight it with everything you have. It’s easy now — you don’t notice the grind. But it’s grinding all the same. By the time you feel it, it’s already taken too much.

I loved him. We were close. I nodded and didn’t really hear it. I was twenty-something, two engines and no off switch. What does the grind even feel like? I went and built GaiaSphere, NetEffects, MAPP, then Concur. Outwork everyone. Sleep when the deal closes. The grind he was talking about was the one I was already inflicting on myself, just on a faster clock.

Years later I went to see my grandparents. The grandfather who’d told me to fight didn’t recognize me. He was openly hostile to me. The same man who had taught me what it meant to be steady. The man who had given me the only piece of advice I should have followed.

Dementia is the long version of the trade I’d been making my whole career. Decades of pathology before the first symptom. Months of neglect before the first lost insight. Same brain. Two clocks. Nobody in the existing system was watching either one for me.

I built the AI coach because the role didn’t exist and I needed it to. In a way, it’s the version of my grandfather’s advice I should have taken in my twenties. Translated into a system, run weekly, with experts I’d never get to meet rendering verdicts on data I’d never thought to track.

If you have a family history like mine, or you just want to keep your originality through your forties and your mind through your seventies, this is a starting point. Not because cold plunges or supplements or the latest peptide will save you. Because there is no role in the existing system whose job is to fight the grind on your behalf. You have to build it. The materials are cheaper than they’ve ever been.

My grandfather didn’t have the system. I do. So do you.

P.S. — Want the prompt I use to run this? Comment below and I’ll DM you the template. It runs in any LLM with persistent memory and interviews you on setup: pick your own experts from a list of common choices, fill in your baselines, name your goals. A few minutes to set up. You’ll have the integrator running by tonight.

Walk into an American machine shop in 1900 and look up. Running the length of the ceiling is a steel shaft, spinning all day, driven by one massive steam engine in the basement. Every machine on the floor hangs off that shaft by a leather belt. The lathe, the drill press, the grinder. None of them sit where the work flows best. They sit where they can reach the belt. The whole building is arranged around the geography of a single power source, and nobody on the floor questions it, because nobody can remember building a factory any other way.

That driveshaft is the most important object in the room. It’s also invisible, in the way that load-bearing assumptions always are. You stop seeing the thing everything is built around.

Then electricity arrives, and here’s the part that should worry every founder rolling out AI this quarter.

Thirty Years of Nothing

Factories started electrifying in the 1890s. The productivity payoff didn’t show up until around 1920. Thirty years of a better technology sitting inside companies and barely moving the numbers.

The economist Paul David wrote the definitive account of this in 1990, in a paper called “The Dynamo and the Computer.” He was trying to explain why the computers of the 1980s weren’t showing up in productivity statistics, and he reached back to electricity to do it. Robert Solow had just made the joke that you could see the computer age everywhere except in the productivity data. David’s answer was that this is simply what general-purpose technologies do. They arrive, and then nothing happens for a long time, and the nothing is the most important part.

Because the technology was never the holdup. Electric motors worked. Power was cheaper and cleaner than steam. The holdup was that for thirty years, owners installed electric motors into factories still designed around the driveshaft.

The obvious first move, when electricity showed up, was to pull out the steam engine and bolt in one big electric motor to spin the same shaft. Cleaner. Cheaper. Safer. And almost no gain, because the factory was still laid out around a shaft that no longer needed to exist. Warren Devine, whose 1983 history “From Shafts to Wires” is where a lot of this detail comes from, tracked the slow part: electric motors ran less than 5% of America’s factory horsepower in 1899. Twenty years later it was more than half. The motors were available the entire time. The redesign took a generation.

The breakthrough wasn’t a better motor. It was a different question. Around 1913, factories like Ford’s Highland Park stopped asking “how do we power the old layout more efficiently?” and started asking “if every machine has its own motor, why are we arranging anything around a shaft at all?” Put a small motor on each machine and you can lay the floor out around the work itself. The sequence. The flow. The thing moving through. That’s the assembly line. That’s when the numbers finally moved, and through the 1920s American manufacturing productivity grew at a clip it hadn’t touched before.

The motor was 1890. The insight was 1913. Most companies spent the gap optimizing a driveshaft.

You Have a Driveshaft Too

Here’s the uncomfortable translation. Your org chart is a driveshaft.

Every process in your company was built around a constraint that felt permanent: that a human being was the only thing that could read the document, make the judgment, write the first draft, route the request to the right desk. Your headcount, your approval chains, your handoffs, your org chart itself, all of it is the leather belt running from the work to the one power source that could actually do the thinking. A person.

AI removes that constraint the way the small motor removed the shaft. Not entirely, not everywhere, but in enough places that the old geometry no longer makes sense. And what is almost everyone doing about it? Bolting a motor onto the driveshaft. Buying licenses, dropping agents into the existing process, watching the recruiting pipeline go from weeks to hours or support handle more tickets per head. It feels like progress. It shows up as cost savings. So leaders call it the win and stop.

That’s the electric motor spinning the old shaft. You’ve made the existing layout cheaper to run. You have not redesigned the factory.

I want to be precise here, because I wrote a piece a few weeks back arguing that low AI adoption is a leadership failure. This is a different claim. This isn’t about whether your team uses the tools. They probably do. It’s about whether your company is still shaped like the problem AI just dissolved. You can have 100% adoption and a driveshaft running straight down the middle of the building.

What It Looks Like When Someone Doesn’t Have One

I met a founder this spring named Christian Bitzas. Engineering background, the kind of résumé that includes NASA and Raytheon. He built a supply chain automation platform in about two months, mostly with AI, and walked into our conversation with eight companies already in his pilot pipeline.

Sit with the two months. Not because he’s a genius with a keyboard, though he’s sharp. Because he had no driveshaft to protect. No legacy org, no team whose roles assumed a human in every loop, no decade-old process that “mostly works.” He got to ask the Highland Park question by default, because there was no old factory to renovate. He laid the floor out around the work and put a motor on every machine, because that was the only floor he’d ever drawn.

That’s the real advantage the incumbents keep misreading. They think the threat from companies like Christian’s is speed, or cheapness, or that they’re AI-native in some branding sense. The actual threat is that they were never arranged around the old constraint. They’re not faster at running your process. They’re running a different process, the one you’d design today if you were honest enough to start over.

Why the Honest Companies Still Freeze

So why doesn’t everyone just redraw the floor? This is the part I lived, and it has nothing to do with the technology.

At Concur, we spent years with a bundled platform — expense, HR, procurement — that worked, sort of. Revenue was fine. And it was quietly strangling us, because we were spread across three buyers and not the best in the world at any of the three. What finally broke us out wasn’t building more. It was layoffs, sunsetting the products that diluted us, and rebuilding around the one thing we could own completely. SAP bought us for $8.3 billion. Not because we’d built the most, but because we’d had the nerve to tear down a structure that was still standing.

That nerve is the actual bottleneck of this whole era, and almost nobody prices it in. Redesigning around AI means deleting roles people are good at and proud of. It means admitting the org chart you spent a decade building is now the shaft. And the redesign doesn’t pay off on day one. Performance dips first, because the new process has bugs and the old muscle memory is now wrong. There’s a trough before the climb, and the trough is real and immediate while the payoff is delayed and, while you’re standing in it, entirely theoretical.

So leaders flinch. They harvest the cost savings, tell themselves they “did AI,” and retreat to fine. Fine is the enemy. Fine is exactly what bolting a motor onto the old layout hands you, and it feels safe enough that most teams never walk into the dip on purpose.

The Question

The tech is not your constraint. You are standing in 1900 with a working motor and a steam-era floor plan, and the only question that has ever unlocked the next thirty years is the one Highland Park asked: if I were building this function from scratch this year, knowing what these tools can do, what would I even build? Not “what can I automate in my current process.” That one keeps you optimizing the shaft forever.

The motors have been available the whole time. The companies that win this decade won’t be the ones who bought them first. They’ll be the ones willing to move the machines.

Your factory is already wired. The only question left is whether you’ll redraw the floor, or keep polishing a driveshaft that doesn’t need to be there.

I keep having the same conversation. A founder tells me they’ve rolled out AI across the company. Licenses bought, a couple of workshops run, people using it every day. Then I ask what actually changed on the P&L, and the room goes quiet. Mostly nothing. Maybe a few points.

The tools work. The adoption is real. And the results are a rounding error.

That’s not a tooling problem. It’s a change problem, and almost nobody treats it like one. I’ve started thinking about it as three gates a company has to walk through, in order. Gates, not stages, because you can’t skip one and you can’t fake your way past it. Most organizations are stuck at the first and think they’re done. The ones doing real damage to their own future are stuck at the second. And the third is where leadership teams quietly fail, because clearing it requires the one thing executives are trained to avoid.

Gate one: earn the right to have an opinion

The first gate is curiosity. Not the LinkedIn kind, where someone “is so excited about the potential of AI” and then goes back to their inbox. Real curiosity that ends in evidence.

Here’s the part people miss. Playing with a tool teaches you nothing useful. Running a real project through it teaches you everything. You don’t find the edges of what AI can do by asking it to write a haiku or summarize an article. You find them by handing it work you actually own, with a real deadline and a real definition of done, and watching exactly where it shines and where it falls on its face.

That’s the whole point of gate one. You’re not trying to be impressed. You’re building an accurate map of what these tools can and can’t do, so that when the bigger decisions come, you’re making them from experience instead of hype or fear. The person who has pushed twenty real tasks through these tools has a calibrated read on what’s possible. The person who read three threads about it has a vibe. Those are not the same input, and you should not let people with vibes write your strategy.

Most companies think they’re through this gate because a few employees use AI to draft emails. They’re not. Drafting emails is the haiku. The question is whether your organization has flexed these tools against its actual hard problems and built an honest, shared sense of where the line sits today. If you can’t point to the projects that did that, you’re still standing at gate one.

Gate two: be willing to destroy what works

This is the hard one. This is where I watch almost everyone get stuck, and it’s the gate that separates a 7% company from a 70% company.

The trap is seductive, and it sounds responsible. You take your existing process, the one that’s been working for years, and you go hunting for the steps you can automate. You find a few. You bolt AI onto them. You get a modest lift, maybe 5 to 10%, and you call it a win. Everyone feels productive. The process is faster.

You just paved the cow path.

The phrase comes from the early web, when companies took their paper forms and made them digital, faithfully reproducing every dumb step of the old process on a screen. They optimized the path the cows had worn into the field instead of asking whether the road belonged there at all. They got a faster version of the wrong thing.

That’s what bolting AI onto your current process does. The process you have was built around old constraints. That work is expensive. That a human is the only thing that can read a document, make a judgment, write a first draft. That coordination takes meetings and email. AI removes or reshapes most of those constraints, which means the right process on the other side doesn’t look like your current one with robots glued to it. It often looks nothing like it.

The blindness is completely natural, and that’s exactly what makes it dangerous. You’ve done it this way for years. It mostly works. Your team knows the steps. Of course your instinct is to protect the structure and improve the parts. But “it mostly works” is the sentence that keeps you parked at gate two forever.

The companies that break through stop asking “what can I automate in this process?” They ask “if I were building this function from scratch today, knowing what these tools can do, what would I even build?” Then they build that, and let the old process die. It’s an uncomfortable question, because the honest answer usually erases work people are proud of and good at. But you don’t get a step change by defending the staircase.

Gate three: fund the dip

Say you clear gate two. You’ve redesigned a core process around what’s actually possible now. Here’s the part nobody warns you about. It’s going to get worse before it gets better.

There’s a pattern in any serious transformation called the productivity J-curve. Rip out a working process, replace it with a genuinely new one, and performance drops first. People are learning. The new workflow has bugs. The old muscle memory is now wrong. The numbers dip. Only after the organization climbs out of that trough do you get the payoff, and the payoff dwarfs the incremental tweak you passed up.

The problem is the timing. The dip is real and immediate. The payoff is delayed and, at the moment you’re living through the trough, entirely theoretical. This is where leadership decides everything. A management team that flinches at the bottom of the curve pulls the plug, blames the tools, and retreats to the safe 7%. They’ll tell themselves they “tried AI and it didn’t move the needle.” What actually happened is they quit during the exact stretch where quitting was guaranteed.

Clearing gate three means doing something that cuts against every instinct of quarterly management. You have to choose to take a step backward to go forward. You have to protect the team in the dip from the company’s own antibodies, fund them through the ugly middle, and hold your nerve while the early numbers look bad. That’s not a tooling decision or a process decision. It’s a leadership decision, and it’s the one most leaders won’t make, because things are “going okay,” and okay feels safe.

Okay is the enemy here. Okay is what gate two hands you on its own. The jump from good to great lives entirely on the far side of a dip you have to be willing to walk into on purpose.

The part nobody wants to hear

Notice the gates get harder as they go, and they get more about people and less about technology. Gate one is curiosity. Gate two is intellectual honesty. Gate three is nerve. By the time you reach the gate that actually unlocks the value, the tooling is the easy part. Your own leadership team is the constraint.

That’s why most companies will get this era half right. They’ll clear gate one, mistake it for the finish line, harvest their 7%, and feel fine about it. The few who redraw the whole field instead of paving the cow path, and then hold the line through the dip, won’t be 7% better than those companies. They’ll be playing a different sport.

The gates are open. The only question is how far you’re willing to walk.

August 2024. Jon Mest is running Just Reach Out, a PR software company, and something is wrong with his ad account. Not broken-wrong. Quiet-wrong.

Google hasn’t charged his credit card in a while. And Google loves charging his credit card. So he logs in expecting to find an error: a paused campaign, a billing glitch, something he could fix in five minutes. Instead he finds this. Over a three-week stretch, Google had spent exactly $4 of an ad budget that normally ran into the thousands every month.

Nothing was broken. That was the terrifying part.

He did the thing he hadn’t bothered to do in months. He searched his own core terms. “Affordable digital PR.” The phrase he used to own. He’d had the number-one blue link. He’d had an ad running underneath it. He was double-dipping on his own category. And now, sitting on top of the whole page, was a giant AI overview answering the question for free. A tidy list of affordable PR options. His company wasn’t on it. You scrolled past the answer, past more context, and eventually, after what felt like five minutes of scrolling, you reached his number-one link. His ad was below even that.

Google had decided to stop spending his money so it could give away the answer above his ad.

That’s the moment I wanted to talk to Jon about, because most founders are about to live some version of it. The ground moves under your business, fast, and the only question that matters is what you do in the next ninety days.

The complaint is the easy part

Here’s what almost everyone does when the rug gets pulled: they complain. SEO is dead. Google killed my channel. The platform changed the rules. All true, all useless.

Jon did something different, and it’s worth being precise about what it was, because it’s copyable.

His team had already been poking at a quieter question. How was discovery going to change? He’d noticed it in his own behavior. He was using ChatGPT to find brands and do research, and reaching for Google less and less. So when the $4 month hit, he didn’t treat it as a catastrophe. He treated it as confirmation. The thing they’d been wondering about had arrived.

Then he did the part most founders skip. He went public with the problem before he had a solution. His team started posting to LinkedIn, to their email list, to anyone who’d listen: Is anybody else seeing this? Anybody want to talk about it? We’re thinking about building here. Building in the open, asking for feedback in a space nobody understood yet.

I’ve watched a lot of founders sit on a half-formed idea because they’re afraid someone will steal it. Jon’s instinct was the opposite. Say it out loud, early, and let the responses tell you whether there’s a there there. The responses were strong. They piloted something with existing customers. Would you pay for this? Some said yes. That became ChatRank, a product that helps brands figure out why the answer engines do or don’t recommend them. It wasn’t born from a strategy deck. It was born from a broken ad account and the discipline to ask the next question instead of cursing the last one.

Search engines vs. answer engines

I should admit something here. For years I told founders that SEO was a solved game. Write good content, earn good links, climb the rankings. I’d lived it at Concur. We ranked, we ran ads, the machine worked. I was giving advice from a world that was quietly disappearing.

Jon drew the distinction that reframed it for me, and it’s simple enough to put on a napkin.

A search engine helps you search. It hands you ten blue links and you do the work of deciding what the answer is. An answer engine, like ChatGPT or Gemini or Claude, does that work for you and hands back the answer. Under the hood, Jon says, it’s essentially running the search, reading the first hundred-plus results, and then deciding, based on what it knows about you, which one to actually recommend. It’s doing the clicking. You just get the conclusion.

That one shift changes everything downstream. If a machine is going to read your site and decide whether to vouch for you, you don’t optimize to rank. You optimize to be chosen.

And the way you get chosen is more meritocratic than anything that came before it. This is the part that should make every small founder sit up. Jon told me about a cybersecurity customer competing against Vanta and Drata, companies many times their size, for one specific thing they did genuinely better. They used ChatRank to find the gap in their own messaging, wrote a single strong piece of content explaining exactly what they did and why it was different, and within a week of publishing, they were outranking the giants on that query. In the old SEO world that climb would’ve taken years of accumulated authority. The answer engine just looked at the evidence and went, you’re right, you are the best at this.

Money can still buy a big brand a lot of polished content. What it can’t buy nearly as easily is the thing the LLMs actually weight: real humans saying you’re good.

Where you actually win

Jon breaks the “real humans” part into three buckets, and they’re the most actionable thing in the whole conversation.

Third-party brand mentions. Does a credible outside source say you’re good? It used to be about the backlink. Now, Jon says, the backlink barely matters. The mention does. And credible doesn’t mean the Wall Street Journal. He has a New York restaurant customer whose whole identity is no seed oils. There’s a tiny blog called Seed Oil Scout that does nothing but hunt down restaurants that avoid them. Small readership, almost no traffic. But the LLMs treat it as the world’s authority on that exact question. One mention on that nobody-blog, and the restaurant got validated as one of the best no-seed-oil spots in the city, instantly. The expert the machine trusts is often not the one with the biggest audience. It’s the one with the narrowest obsession.

Social signals. Reddit, Quora, the forums where actual people talk. Organic conversation about your brand, the messier the more believable.

Sentiment. Reviews, on whatever platform your world lives on. G2 for software, Yelp for restaurants, Healthgrades for doctors. Real people saying you were good. If you got dinged on service, you don’t argue with the algorithm; you go earn a review that mentions how you handled the problem.

None of those require a war chest. All of them are available to a two-person company today. That’s the quiet good news buried in Jon’s $4 disaster: the new game rewards specificity and credibility over budget, which is exactly the terrain a small focused company should want to fight on.

What to do Monday morning

I always ask founders the same closing question. It’s Monday: what do you actually go do? Jon’s answer cost nothing and I think it’s the best single move available right now.

Open ChatGPT. Hit the button that makes it a temporary chat, so it has no memory of who you are. Then ask it plainly: who are the best five companies that do the specific thing you do?

It’ll name five. Maybe you’re on the list. Probably you’re not. Either way, ask the follow-up: why didn’t you mention me? And the model will tell you. You don’t have the right reviews. Your website never actually says you serve the customer you claim to serve. There’s no credible source backing up the thing you say you’re best at. It will hand you, in plain language, the exact gaps between the company you think you are and the company the evidence says you are.

That’s the whole job, really. Close those gaps. Write the one piece of content that names what you do better than anyone. Get the narrow-but-trusted source to mention you. Earn the review. Repeat.

The $4 month felt like the end of Jon’s marketing. It was actually the beginning of ChatRank. The difference wasn’t that he saw the disruption coming. Almost nobody does. The difference was what he did in the ninety days after it landed. He stopped defending the old answer and went looking for the new one.

Google’s already told us where this goes. The ten blue links are on their way out. The only question is whether the machine, when someone asks it about your category, has any reason to say your name.

Go find out. It’s a free search away.

Some version of this happens most mornings. I sit down with coffee before anyone needs anything from me, and I pick up a problem that’s been following me around for weeks. Lately it’s been how to make a cattle-lending company legible to private capital that has never once thought about cattle. I don’t solve it. I move it a few feet. I write down a version that’s slightly less wrong than the one I had yesterday, and then the day starts, and the problem goes back in the drawer.

By five o’clock I’ve done a lot. I answered the email that had been sitting for three days. I reviewed a budget. I worked through an HR situation that needed a person and not a policy. I cleared a dozen small things off a list that regenerates overnight like it’s farming me.

Here’s the part I’m not proud of. The afternoon felt better than the morning.

The morning was the valuable part. I know that. The thing I actually get paid for, the thing nobody in my orbit can do on my behalf, is sitting with a hard problem long enough to make it a little less hard. But it didn’t feel like progress, because nothing closed. The afternoon closed twelve things. Twelve small hits of done.

I’ve started calling that closure candy. A small, finishable task pays you the moment you finish it. The hard problem pays you nothing for weeks and then, if you’re lucky, a little. So we drift toward the candy. Not because it matters more. Because the reward shows up faster.

For most of my career I mistook a cleared list for a good day. I’d end a Friday with forty things crossed off and feel like I’d earned the weekend, never mind that the two questions that would actually change the trajectory of a company hadn’t moved an inch. The list was the performance of work. Some of it was even real work. But the satisfaction was lying to me about which part counted.

It went deeper than the list. My real edge, for a long time, was simple: I filled every space. Nights, weekends, the dead air between meetings, the drive home. I outworked people because I was willing to put work into every gap I had, and for years that was the whole strategy. It worked. It also cost more than I’m going to put in this paragraph.

The afternoon is now free

Here’s what changed my mind.

Almost everything on that afternoon list is now something a machine can do about as well as I can. Draft the email. Summarize the budget variance. Pull the three things I need to know before the HR conversation. Not perfectly. About eighty percent of the way there, which is most of the way most of those tasks ever needed to go.

That eighty percent used to be scarce. You paid people for it. You paid yourself for it, in hours. It was a real part of the job, and clearing it felt like clearing the job. Now it’s close to free, and not just for me. It’s free for everyone. When competence gets that cheap, doing it faster stops being an edge. Everyone has the same fast.

So the value of my afternoon didn’t go up because AI made it quicker. The value went to nearly zero, because the thing I was doing all afternoon is now something anyone can have done for them in minutes.

What’s left is the part that never closes

What the machine can’t do is the morning.

It can’t sit with the cattle-capital problem and feel which version is slightly less wrong, because being slightly less wrong there depends on twenty years of watching how money actually moves and how ranchers actually decide. It can draft around the problem all day. It can’t tell me the draft is missing the thing. That judgment is the part I should be selling, and it’s the part I was starving every afternoon to feed the candy.

This is the uncomfortable trade. The deep work feels worse in the moment. No list. No hit of done. You go to bed with the same problem you woke up with, a few feet further along, and your brain files that as a wasted day even though it was the only part of the day that mattered. The reallocation isn’t free. You give up the dopamine of the cleared list and get back the slow, unglamorous progress on the thing that compounds.

I’m not the smartest person in the room

Once I saw this on my own calendar, I saw it everywhere I work.

AgCents isn’t mine. It belongs to Brady, who spent twenty-five years in cattle, and David, who can build the models I can’t. I’m not the smartest person in that room about cattle or about code. Property Meld is Ray’s company. I’ve been in the room with him for ten years, watching the platform grow into the hundreds of thousands of rentals, and I have never once written a line of their software or closed one of their deals.

What I bring to those rooms is the same thing AI gives me back in the morning. I take the part that can be delegated, and I protect the part that can’t. The framework. The next milestone. The one question that actually matters this month, which for most founders is some version of get the next customer or don’t run out of cash before you’re done. Founders drift toward whatever is loudest or most fun to talk about, the same way I drift toward my afternoon list. My job is to keep pointing them back at the morning.

Onboarding an agent and getting a founder to stop polishing the deck and go sell something are the same act. You are directing capability you do not personally have toward an outcome only you can see clearly. The capability used to come only from people, which made it expensive and slow and risky. Now some of it comes from software. The act is identical. What’s scarce was never the doing. It was the judgment about what’s worth doing.

The catch

There are three ways this goes wrong, and I’ve done all of them.

The first is handing the afternoon to the robots and then filling the hole with more afternoon. New tools, same trap. If you automate the busy work and immediately invent fresh busy work to feel productive, you haven’t reclaimed anything. You’ve just bought a faster way to avoid the hard problem.

The second is assuming everyone has a morning worth protecting. The whole thing rests on you being genuinely deep in something the machine can’t fake. If your only contribution was the eighty percent everyone now gets for free, automating it doesn’t promote you to the thinking. It removes the thing you were being paid for.

The third is the one I got wrong the longest. Clearing the busy work doesn’t mean I now do eight hours of deep work instead of four. Deep work isn’t a faucet you open wider. Three empty hours without the headspace to use them produce nothing but a guilty stare at a blank page. The reclaimed time isn’t for doing more. It’s for having the room to think at all, which is a rarer thing than an open calendar.

So here’s the part that still surprises me. I work fewer hours than I have at any point in my career. I’m also solving better problems than I ever have, because now I get to choose them instead of grabbing whatever was loudest. That’s the trade I’d actually sell you. Not more output. Fewer, better problems, with enough room around them to do them justice.

The list will be there tomorrow. It regenerates overnight, the way it always has. The robots can have it. I’ve got thinking to do, and the room to do it in.

The Setup: Wildfire Labs hit a rough stretch — funding fell through, a key relationship collapsed, our AI app got delayed, the team was gassed. I needed space. Two trips gave me more than expected.

The Difference: People who build real things don’t do it through breakthroughs. They do it through an accumulation of ordinary days.

Your Version: Whatever you’re grinding through has a next step. Not the whole solution. Just the next brick.

Earlier this year, Wildfire Labs hit a rough stretch.

Funding we relied on didn’t come through. A developer relationship we’d invested months in building fell apart. An internal AI app we were developing hit delays that pushed our timeline back. We were three and a half years in, and the team was tired from pushing hard for a long time. The gap between our current and expected position was uncomfortable — but clarifying.

I want to be honest about what it feels like because founders often lie about it. We talk about the grind like it’s a badge of honor, but we rarely admit the self-doubt, fatigue, and fear that the plan isn’t working. Those feelings are real. They cost people time, energy, and confidence that’s hard to rebuild.

I needed space.

My wife and I went to Belize. I didn’t need answers or a newsletter topic. I just needed a break from the weight of it for a few days and breathe.

I got more than I expected.

Our guide was Hugo, one of 65 people licensed to lead tours through the Actun Tunichil Muknal cave, a sacred Mayan site buried deep in the Belizean jungle. The trip to get there is its own kind of test: a 45-minute hike through rainforest, three river crossings on foot, then a swim into a cavern mouth that opens into an underground world that hasn’t changed in a thousand years.

Hugo does this every day. He hadn’t taken a day off in 46 days when we met him. He has four and a half years of this pace ahead.

His parents fled Guatemala on foot when he was four, escaping a 36-year civil war targeting Mayan people like his family. They crossed into Belize with nothing — no money, no English, no plan beyond making it through the night. They settled and started building a life from zero.

School was a luxury they couldn’t afford. Hugo dropped out in fourth grade.

But he made a decision. English was his path forward.

Not English class. Not a tutor. He took manual labor jobs alongside foreigners who spoke only English — not for the pay, but because the daily immersion forced him to improve. Every shift was a language lesson. Every awkward conversation was practice. He studied on his own. He earned his first aid certification. He navigated a licensing process for people with far more formal education. He became one of those 65 certified guides.

He got married, has three daughters, and built a life worth protecting.

Those 46 consecutive days without a break? They’re not burnout. Hugo saved enough to buy land — he already owns the plot. Now he’s designing the house and mapping out his tourism business. He’s calculated the cost, sketched the timeline, identified the clients he want to serve. Every long day is a brick in that foundation. He’s not grinding blindly. He’s building something specific, one day at a time.

Here’s what I’m seeing: Hugo isn’t doing anything complicated. He’s doing something relentless. There’s no hack in his story. No shortcut. No pivot that made everything click. Just the next small step, repeated until it compounds into something real.

In that cave, listening to his story amidst a thousand-year-old Mayan world, a thought kept surfacing.

My problems are solvable.

That trip to Belize wasn’t the first time travel had recalibrated how I see things. On a trip to Cambodia the year before, my wife and I hired a guide named Seiha to take us through Angkor Wat, and what started as a tour became a conversation I haven’t stopped thinking about.

Seiha grew up in rural Cambodia. The Khmer Rouge killed both of his parents when he was five. His aunt and uncle took him in and did what they could, but eventually they couldn’t provide for him anymore. When he was eight, they brought him to the Angkor Wat orphanage. He lived there as a Buddhist monk until he was eighteen.

Many days, there wasn’t enough donated food to go around.

After leaving the orphanage, Seiha did what Hugo did — he found work with his hands and taught himself English the slow way, through repetition, through embarrassment, through trying again when every instinct told him to stop. He became a guide at one of the most visited archaeological sites on the planet. He was good at it. He’s still good at it.

He got married, has a daughter, went back to the rural village where he grew up and built a school. It’s thriving now — kids are learning who never would have had the chance.

Then he told me something that stopped me cold.

He tracked down his parents’ killers. Not for revenge or confrontation, but to reconcile. He wanted peace — the real peace, requiring sitting across from someone who took everything from you and deciding their act of destruction won’t define your story.

I was silent after he told me that. I didn’t know what to say. I’m still unsure. There’s a kind of courage in what Seiha did that I don’t fully understand — the willingness to face the worst thing that happened to you and choosing to build something from it instead of letting it own you. I’ve thought about that moment more than almost anything else from that trip. Maybe more than anything else from this entire year.

The pattern is clear: Seiha did what Hugo did. Not a grand gesture. Not a single heroic act. He guided one tour, saved one dollar, taught himself one more English word, laid one more brick — and repeated the process until something real took shape. A career. A family. A school. And eventually, peace with the unthinkable.

The bricks were small. The vision was not.

At Wildfire Labs, those stories changed my view of our rough stretch.

Here’s the truth: my funding gap is closeable. The developer relationship that fell apart taught us something valuable about what we actually need in a partner. The AI app delays forced us to rethink our approach — and the new direction is stronger for it. The team’s exhaustion is real, but it’s also evidence of something important — we’ve been pouring ourselves into work that matters to us. Those are problems with solutions. They’re bricks that need laying, not walls that need tearing down.

Hugo’s parents crossed an international border on foot with a four-year-old to escape a genocide. Seiha grew up in an orphanage where meals weren’t guaranteed and then sat across from the men who murdered his mother and father. My company hit a rough quarter.

I don’t say that to minimize the difficulty of building a startup. It is genuinely hard. But perspective has a way of right-sizing your problems, and right-sized problems are easier to solve. When the obstacle in front of you stops looking like a wall and starts looking like a brick that needs laying, the work becomes possible again.

I don’t think perspective is something you find once and keep forever. It’s more like a practice — something you return to, especially on days when the problems feel large and the distance between where you are and where you want to be seems to be growing instead of shrinking.

On those days, I think about Hugo — still in the cave, 46 days in, building toward something he can see clearly that others can’t.

I think about Seiha — sitting across from the people who killed his parents, choosing peace instead of letting destruction win.

Then I get back to work.

That’s what perspective does. It doesn’t erase problems or make hard days easy. It clarifies that the work is worth doing and that the obstacles are smaller than they feel.

I know my next brick. That funding gap needs closing this quarter. That’s the brick. Not the whole wall. Not the finished building. Just the next thing that moves Wildfire Labs forward.

So here’s my question for you: What’s the next brick you need to lay?

The one small, specific, unglamorous step that moves you closer to the thing you’re building. Not the strategic plan. Not the five-year vision. Just the next brick.

Hugo knows his. Seiha knows his. I know mine.

Do you know yours?

Two months in, I looked at the numbers.

By one measure, he was our best developer. More code shipped than anyone else on the team. He’d picked up Claude Code quickly — faster than most — and was using it constantly. The output was real. Velocity looked great.

Then I looked at the bug report.

He was responsible for six times more defects than the next closest person on the team. Not 20% more. Not twice as many. Six times.

I’ve been watching the vibe coding conversation for a while now. The hype is real — non-technical founders spinning up MVPs, designers shipping working prototypes, PMs building internal tools that would have taken a sprint two years ago. And the developer productivity angle is just as compelling: senior engineers using AI pair programming to compress days into hours.

But there’s a version of this story nobody talks about as much.

AI doesn’t make you a better developer. It makes you a faster version of the developer you already are.

If you understand the code, it’s a force multiplier. You write less, think more, catch the edge cases before they ship. You use AI to handle the tedious scaffolding while you apply judgment to the parts that actually matter. The output goes up, and so does the quality.

If you don’t understand the code, it’s also a force multiplier. You generate more, think less, and ship problems at scale. The output goes up. The quality craters.

That’s the part that’s hard to see from the outside — and hard to admit when you’re inside it.

Our first developer wasn’t malicious. He wasn’t lazy. He was enthusiastic. He genuinely believed he was doing well, and by the only metric he could easily see, he was. Lines committed. Features closed. Tickets moved. The dashboard looked good.

He was usually the first one done. He was almost never the first one to production.

What the dashboard didn’t show was what was inside the code. Patches on top of patches. Logic that worked in the happy path and fell apart everywhere else. Defects that took the people around him hours to diagnose and fix — hours that never showed up attributed to him, because by then he was already onto the next thing. We tried to help. We put him on a performance plan. Walked through the issues. Made the expectations explicit. It didn’t turn around, and eventually we let him go.

The other developer we hired around the same time took a completely different path.

He had domain expertise — he understood the problems we were solving before he ever touched our codebase. He adopted AI tools just as quickly, but he used them differently. He wasn’t generating code and shipping it. He was generating code and interrogating it. Understanding what it did, why it worked, what it would break. Within a few months, he’d become the person we turn to for the most technically complex work we have — specifically how we handle inference inside our application. That’s not a role we assigned him. He earned it because the team kept discovering he could think through things nobody else could.

When we recognized what we had, we asked him a simple question: is there anyone in your network with the same capacity and interest in doing this kind of work? He’s since led us to two more developers who fit the same profile. Both have worked out.

The thing I keep coming back to is how invisible this pattern is until it compounds — in both directions.

Velocity metrics don’t catch the bad version. Story points don’t catch it. Closed tickets don’t catch it. The first sign is usually someone on the team muttering about weird bugs in unfamiliar corners of the codebase. Then the time-to-fix goes up. Then you look at the data. By then, you’ve already paid the tax.

The good version is equally invisible at first. A developer who slows down to understand what the AI generated doesn’t look like a star in week three. The metrics don’t reward interrogating a pull request. They reward closing one.

Once AI makes code generation fast and cheap, the bottleneck moves. The scarce resource isn’t output. It’s judgment. The ability to look at the code you generated and know whether it’s right. To understand the tradeoffs you made and the ones you didn’t. To maintain, debug, and extend what you built six months from now.

That judgment doesn’t come from the model. It comes from the developer.

I think about this every time I see a job posting that says “must be fluent with AI coding tools.” Fair enough — but that’s table stakes now. The question worth asking is whether the person can evaluate what those tools produce. Whether they can spot the hallucination. Whether they can read the diff and tell you what it actually does, not just what it was supposed to do.

If the answer is yes, they’re going to be exceptional. They’ll also help you find others like them. People with genuine judgment tend to know other people with genuine judgment. Treat them as a talent source, not just a contributor.

AI is an amplifier.

Put a developer with domain expertise and genuine understanding in front of it, and you get more of both — faster. Put a developer without those things in front of it, and you get more of that faster, too.

The underlying ability doesn’t change. The speed at which it manifests does.

That’s not an argument against AI tools. It’s an argument for being clear-eyed about what they are.

They’re not a shortcut to competence. They’re a multiplier of whatever competence you bring.

Choose accordingly when you’re hiring. And be honest with yourself when you’re building.

A sales call for a QA automation startup. The founder is running the playbook everyone runs: walking a CTO through the pain points. Don’t flaky tests drive you crazy? Wouldn’t you love this automated? Let’s set up a demo. And the CTO, getting more irritated by the minute, finally cuts him off.

“Listen, I don’t care. I just want QA off my plate. We’ve had some big bugs lately. We need to solve this. What are we even talking about here?”

That one sentence was worth more than the entire pitch. The startup just didn’t know it yet.

Rob Snyder tells this story, and it’s the one that reorganized how I think about selling. Rob’s path to it was not gentle. McKinsey, then Harvard Business School, then the very reasonable conclusion that being a tech billionaire wouldn’t be so bad. He did all the things you’re supposed to do. Customer discovery, raised money, signed design partners, drew the clean line to a multibillion-dollar business. Then he got punched in the face for two years while none of it worked. When the company finally took off, it went zero to $4 million in ARR in two years, mostly before the product was even ready. And here’s the part that hooked me: he didn’t understand why it failed, and then he didn’t understand why it worked. He’s spent the half decade since obsessed with the answer.

I sold against pain for years and never noticed

I should’ve learned this at Concur.

We sold expense reporting, and the pain was obvious. No transparency, runaway costs, fraud, lawsuits. Our favorite prospects were companies that had just been in the news for getting sued. Pharma was the worst offender. Reps expensing dinners with doctors, “Dr. John Smith” spelled six different ways to slip past the limits on how much you could spend with any one physician. We’d walk in and say: you don’t want this exposure anymore. And we’d win.

For years I thought we were winning on the pain. We weren’t. Plenty of companies carried the exact same exposure and never bought anything. The difference in the ones who signed wasn’t the size of their pain. It was that someone inside the building had just been handed a project: make sure we never get sued for this again. The lawsuit didn’t create the pain. The pain was always sitting there. The lawsuit created a priority, and a named person who owned it.

I figured that out in real time, on the podcast, while Rob was explaining his framework. Which is a strange feeling at this point in my career.

Pain is everywhere. Pull is rare.

Here’s Rob’s core claim, and it’s heretical if you’ve absorbed the standard startup advice. Pain points are not predictive of a purchase.

He’s watched founder after founder run the same heartbreak. You do twenty-five discovery interviews. You ask people to rate their pain. Twenty of them call it a nine or a ten out of ten. You build the thing, you go to sell it, and they don’t buy. Why? Because, as Rob puts it, life is a bunch of pain points and we mostly do nothing about them. A ten-out-of-ten pain is not the same as a willingness to drop what you’re doing today and change how you work.

So Rob stopped talking about pain points. He talks about the projects on your to-do list instead. Picture everyone you’re selling to as having a Trello board. The backlog is endless. But at any given moment there’s a short list of things actually being worked on. One, maybe two or three priorities. Everything else is just sitting there being painful.

That short list is where buying happens, and it’s the basis of what he calls the pull framework. The acronym is PULL. There’s a Project the person is actively trying to get done. It’s Unavoidable, something they’re prioritizing right now. They’ve assembled a List of options they’ve tried or considered. And those options have Limitations, they’re not good enough to finish the job. When all four are present, Rob says, somebody will rip the product out of your hands. When they’re not, you’re stuck twisting arms, poking, prodding, convincing. And most of the time, you still lose.

The unlock for Rob came from a random video he stumbled onto about, of all things, product management. Ryan Singer from Basecamp, drawing the line between supply and demand. A feature request, Singer argued, is a request for supply. It isn’t demand. Demand is something already out in the world that exists whether or not your product does. Nobody actually wants your product. Nobody wants features. They’re trying to get something done in their life, and your supply either fits or it doesn’t. The job was never to convince people to want what he’d built. The job was to understand what they were already trying to do.

What it looks like when you get it right

Go back to that angry CTO. Once Rob’s team heard it correctly, everything changed. The man did not want a QA automation tool. He did not want to think about QA at all. He wanted QA off his plate, permanently. So they stopped selling “a QA automation tool” and started selling forward-deployed QA that you plug in and stop worrying about. The demo stopped making sense, so they cut it from the process entirely. The price stopped making sense too, so they raised it tenfold, then doubled it, then doubled it again. Customers bought happily, in weeks instead of six months.

Or take Jump, a company in Utah selling an AI notetaker built specifically for financial advisors. Advisors have to log notes from every client meeting into their CRM, manually, so that when regulators come knocking they don’t get fined. Nobody wants to spend their evening doing that. There’s a standing project on every advisor’s list: spend the least possible time on this. Jump shows up, fits that exact shape, and advisors are texting “I’m never going back” an hour after their first meeting.

My favorite example is the one most out of left field. A crisis-response consultant who sells to Fortune 500 CEOs in the middle of a public activist-investor fight. Building the prospect list is easy: open the newspaper. Getting the meeting is the hard part, because anyone can open the newspaper. So, the legend goes, this consultant would send a physical samurai sword to the CEO’s office with a note: I will go into battle with you. Rob calls it the tollbooth approach. Find the exact moment a person has pull, position yourself right at it, and you barely have to sell.

Monday morning

So you’re a founder chasing your first twenty-five customers. What do you actually do?

Rob’s answer costs nothing. Write out the PULL framework by hand. Then look at the customers who bought weirdly fast, or the ones who are genuinely thrilled, and study them. Go back and watch the call recordings. Listen for what they were trying to do, not the pain points they listed. What was the project? What options had they already tried? Why weren’t those good enough? Why did they move so fast?

What you build from that is a pull hypothesis: if I talk to a person with these specific characteristics, they wouldn’t just maybe buy, it would be weird if they didn’t. Then you go test it on people who fit. If they don’t buy, that’s a signal something’s off in the hypothesis, and you dig. Rob runs his tests in sets of five calls. No deep reason, he says. It’s just a good number.

His book, The Power of Pull, came out July 7. He’s rewritten it eight times from scratch, and he’s honest that it’s a working hypothesis rather than a finished theory. What it promises is not that running a startup gets easy. His own line on it stuck with me: with the pull framework, you’ll still feel like you’re failing at everything, but there will be fewer things to fail at.

Most founders solve the product and then quietly stall on go-to-market, convinced they need to do a hundred things at once. Rob’s whole argument is that you don’t. You need to find the people who already have a project, already have pull, and stop trying to manufacture it in everyone else.

The next time a prospect tells you your idea is “really interesting,” don’t celebrate. Interesting is what people say right before they ghost you. Go find the ones who are already trying to do the thing. They won’t tell you it’s interesting. They’ll ask how soon they can start.

1941 - The British government took out an advertisement in The Daily Telegraph.

Not for engineers. Not for mathematicians. Not for trained intelligence analysts.

They ran a crossword competition.

The ad was a recruitment pipeline in disguise. Anyone who could finish the Telegraph‘s famously difficult crossword in under 12 minutes received a letter in the mail. Come to London. We have something we’d like you to help with.

Those recruits ended up at Bletchley Park, a country house north of London that housed Britain’s codebreaking operation during World War II. Their mission: crack the Enigma cipher. The German military used Enigma to encode every communication: troop movements, supply chains, U-boat coordinates. The machine had more possible settings than atoms in the observable universe. Germany’s engineers had designed it to be mathematically unbreakable.

Bletchley’s answer was not to hire the world’s best cryptographers.

It was to build the world’s strangest team.

The roster reads like someone accidentally merged three academic departments and forgot to assign offices.

Alan Turing, a Cambridge mathematician obsessed with the theory of computing. Gordon Welchman, an algebraic geometer. Dilly Knox, a papyrus scholar who’d spent two decades deciphering ancient Greek texts. A professional chess champion. A museum curator. A bridge player. Linguists, classicists, crossword constructors.

Not a single person had worked in intelligence. That was, mostly, the point.

Researchers who’ve studied Bletchley call it “organizational porosity” — knowledge moving sideways instead of up. No formal org chart. A linguist could walk into a mathematician’s hut and describe a pattern he’d noticed in intercepted messages. The mathematician would see something the linguist couldn’t. The chess player would see something neither of them caught.

The crossword solvers found the first critical cribs: known-plaintext fragments that gave the mathematicians something to anchor on. The reason: they were looking at the cipher as a puzzle, not a mathematical object. They noticed that German operators, bored on overnight shifts, signed off with predictable phrases. The same words appearing at the same position in messages sent at the same time each day.

The mathematicians would never have found that. They were looking for structural weaknesses in the algorithm.

The crossword solvers were looking for human laziness.

They needed each other to see the whole thing.

Most companies do the opposite when problems get hard.

When something won’t budge, the instinct is to find whoever has solved that exact problem before. Sales is struggling, so you hire a VP Sales from a company that figured it out. Product isn’t landing, so you bring in a CPM with a stronger pedigree. Feature velocity is low, so you add senior engineers.

The logic is intuitive. Hard problem, deep expertise. Makes sense.

But domain expertise comes packaged with domain assumptions. The seasoned VP Sales knows how to run a playbook — the one that worked at her last company, in her last market, against competitors she understood. The experienced CPM knows what “good” looks like, which means he also knows what “impossible” looks like. Expertise is pattern recognition. The patterns aren’t always right.

The people at Bletchley who made the first breakthroughs weren’t the ones who knew the most about codes. They were the ones who’d never been told what made codes impossible to break.

Psychologists call this functional fixedness: the inability to see a tool used for anything other than its intended purpose. A hammer is for nails. Experts develop professional versions of this. A cryptographer looks at a cipher and sees a mathematical object. That framing carries assumptions about where the weaknesses must be.

The crossword solver doesn’t have that. She sees the cipher as something humans made, which means humans are using it, which means humans are taking shortcuts. That observation is worth more than a lot of advanced mathematics in the right moment.

This isn’t a case for hiring generalists instead of specialists. Bletchley needed Turing. The Bombe machine he built, which tested possible Enigma settings at mechanical speed, was the infrastructure the entire operation ran on. Without the mathematical rigor, the human observations went nowhere.

But without the crossword solvers’ cribs, Turing’s machine had nothing to test.

The specialists and the generalists were each other’s raw material.

Pixar figured this out around 2000 with the Braintrust.

The Braintrust is a group of directors and writers who watch rough cuts of films in development and give feedback. No hierarchy. No authority. The director can take every note or throw them all out — the Braintrust doesn’t decide anything.

Ed Catmull, in Creativity, Inc., describes the goal as giving filmmakers a moment to see what they’ve made through someone else’s eyes. People invested enough to give real feedback, but not invested in the film itself.

That’s organizational porosity. Walls between projects stay thin. What the animation team learned on Finding Nemo flows into what the story team is building on The Incredibles. Not through a knowledge management system. Through people who move between rooms.

Most early-stage startups do this by accident. The designer, the engineer, and the founder sit in the same customer call. The support rep walks to the product manager’s desk and says, “You should hear what this customer told me.” The walls are thin not by design but by necessity — you don’t have the headcount to build silos yet.

Then the company grows. The walls go up. Disciplines separate.

And the hardest problems stop getting solved.

I’ve watched this pattern run in both directions. At Property Meld, when the product team was stuck on a workflow problem that engineers and designers had been circling for weeks, a customer service rep who’d never attended a product meeting walked in and said: “They just want to know what’s happening to their request.” The whole frame changed in sixty seconds. She wasn’t solving the technical problem. She was pointing at it from the right angle.

At AgCents, a cattle lending startup I work with, the breakthrough in underwriting logic didn’t come from the finance team. It came from a rancher describing forty years of buying decisions that no spreadsheet had ever captured. The finance team had the model. He had the variable they’d been missing.

Different rooms, different angles. Same problem, now visible.

The Bletchley lesson, stated plainly:

For your hardest problem — the one that’s been stuck for months — the answer probably isn’t more domain expertise. It’s a different lens on the same problem.

Find the crossword solver. Not metaphorically. Find someone with no context about what you’ve already tried and no stake in why it hasn’t worked. Someone whose pattern recognition comes from somewhere else entirely. Give them full access to the problem. Don’t explain your framework before they start. Don’t tell them what you’ve already ruled out. Let them describe what they see.

Then sit with your mathematician and figure out what to do with it.

The chaos at Bletchley wasn’t dysfunction. It was the operating system. Different disciplines don’t just bring different answers — they bring different questions. And in 1941, it turned out the right question wasn’t mathematical at all.

It was: why do German operators keep sending the same phrase at the start of every morning weather report?

Nobody trained in cryptography thought to ask that.

The museum curator did.

TL;DR: Why Amundsen Beat Scott

The Setup:
Two teams race to the South Pole in 1911. Amundsen arrives first and returns safely. Scott arrives 34 days later and dies with his entire party.

The Difference:
Amundsen chose proven tools (dogs), built margins everywhere, and marched exactly 15–20 miles daily. Scott chose experimental tech (ponies, motor sledges), ran thin margins, and bet on heroics over systems.

Your Startup Version:
Most founders lose from a thousand small operational choices nobody thought mattered—not from one big competitor move.

Do This Today:
List your scariest single point of failure. Spend 30 minutes adding one layer of redundancy.

Bottom Line:
Startups are won by boring systems that work on your worst day—not vision decks or heroic sprints.

Two Teams, One Goal

In the early 1910s, two expeditions raced for the same prize: the first human footprints at the South Pole in Antarctica.

Roald Amundsen leads the Norwegian team and Robert Falcon Scott commands the British expedition. Both are experienced explorers with national funding and public expectations.

Amundsen treats the journey like an engineering problem—meticulous systems, routines, obsessive preparation. Scott treats it like a test of character—intentions, leadership, adaptive improvisation.

Amundsen plants the Norwegian flag at the Pole and returns home with every man alive. 34 days later, Scott arrives, finds Amundsen’s tent, and dies with his entire party on the frozen return.

Same goal. Similar resources. Different outcomes.

Jim Collins immortalized this contrast in Great by Choice with his “20 Mile March” principle—disciplined consistency beats erratic intensity. But there’s a second pattern founders miss: Amundsen didn’t just march consistently. He built redundancy at every layer to prevent a single failure from causing catastrophe. Consistency gets you moving. Margins keep you going.

Most founders don’t lose because a competitor was smarter, faster, or better funded. More often, they lose like Scott: a series of small operational choices nobody thought mattered until the money ran out.

This is a field guide to Amundsen-style execution.

Two Unwritten Philosophies

Amundsen’s unspoken rule is: “The work is secured before you leave.”

He spent years studying polar survival, not just reading about it but living it. He embedded with Inuit communities in Greenland to learn dog handling, fur clothing design, and igloo construction. He tested every gear in real conditions. He built redundancy into everything: extra food depots every 60 miles, supply caches marked with black flags visible from three miles away, backup navigation checks twice daily.

His daily discipline was robotic: travel 15–20 miles, then stop. Even on perfect weather days when the team could push 30 miles, Amundsen called halt. Consistency over heroics. Predictability over inspiration.

Scott’s unspoken rule: “We’ll figure it out as we go.”

He believed exceptional men would overcome challenges through courage and real-time problem-solving.

His expedition bet on unproven technologies. These included experimental motor sledges that failed within days, Siberian ponies that sweated through their fur and couldn’t handle ice, and when those failed, men hauling 200-pound sledges by rope. The plan included looser supply margins and variable daily distances. The implicit assumption was that commitment would compensate for system gaps.

Scott wasn’t incompetent—he’d survived previous Antarctic expeditions and was optimizing for different constraints, including public spectacle and scientific research. When systems and heroics collide in survival conditions, they prevail.

Review your last three months of calendar data and spending. You’ve revealed your running philosophy; you just haven’t defined it.

Founder Takeaway #1: Your operational defaults—what you do when nobody’s watching—matter more than your vision deck. Systems are more effective than speeches.

The Ordinary Decisions That Influenced Everything

The “Dogs vs. Ponies” Choice

Amundsen chose sled dogs, effective in Arctic travel. They could eat seal meat hunted along the route, thrived in -40°F temperatures, and moved fast across ice.

The plan was emotionally hard. Some dogs would be culled mid-journey to feed the remaining dogs and men. Amundsen accepted the moral discomfort because the operational logic was sound.

Scott chose ponies. They looked impressive, evoked cavalry traditions, and felt more “civilized.” But they sweat through their coats and freeze. They panic on ice. They need transported grain, not hunted food. When they failed, Scott’s men became the pack animals—hauling sleds by hand for hundreds of miles.

Your decision: Salesforce because “that’s what enterprise companies use” with 12 customers and no sales team. A custom-built admin panel for “full control” instead of Retool. Hiring a VP of Marketing at seed stage because the growth playbook suggests it—even without product-market fit.

Choosing tools and tactics based on appearances instead of what works for your current altitude is a poor decision.

The straightforward solution your successful competitors use? That’s choosing dogs.

The Unnoticed Gear Until It Breaks

Amundsen wore reindeer-fur anoraks and sealskin boots—copies of Inuit designs refined over generations. He layered for moisture management. He used lightweight tents that deployed in under two minutes, even in 60 mph winds.

Scott wore modified Royal Navy wool uniforms and heavier cotton-lined parkas that trapped sweat and froze against skin. He also set up canvas tents that required 15 minutes to stake properly.

Both teams faced identical blizzards. Only one remained warm, dry, and fast.

Notion vs. information chaos across Slack, email, and local desktops. Automated alerts for revenue anomalies vs. discovering it three weeks later in a board deck. Weekly customer health scores vs. being surprised when your biggest account churns.

These aren’t “nice to haves.” They’re the difference between moving quickly in a crisis and freezing while you search for the information to make a decision.

The Tuxedo Test: Your internal tools should work on your worst day, not demo day. A tuxedo and a parka are “clothes”—but only one works in freezing temperatures 400 miles from help.

The Discipline of Consistent Effort

Amundsen’s rule: 15–20 miles per day, regardless of conditions. If you hit the target by 3pm, set up camp and rest. If the ideal weather allows 30 miles, stop.

This protected against exhaustion, made fuel calculations predictable, and preserved energy reserves for emergencies.

Scott’s mileage was reactive—25 miles on good days, zero on bad ones. These variations are minor. But timing miscalculations compound: you reach a supply depot three days late, find less fuel, ration heat, move slower, arrive at the next one later, in worse condition, with less margin for the next crisis.

The math of compounding consistency: A team shipping one small customer improvement weekly for 24 weeks delivers 24 touches. A “when ready” might deliver three big releases in the same period. Even if each is worth 3x a small improvement, the weekly team delivers 24 vs. 9—and builds a reputation for momentum that the sporadic team never earns.

Sarah’s B2B SaaS team shipped sporadically—big releases every 6-8 weeks when “features were ready.” Customer churn was 8% monthly. She implemented a straightforward rule: one customer-facing improvement every Tuesday, no exceptions. Sometimes it was tiny (a new reporting filter). Sometimes substantial (workflow automation). Six months later, while holding pricing and ICP constant, churn dropped to 3%. Customers cited “momentum” and “responsiveness” in renewal surveys—they felt the weekly pulse, even with small features. The cadence became the signal.

You don’t need one massive product launch. You need a system that ships measurable customer value every week for 52 weeks until the compounding becomes clear.

Founder Takeaway #2: Consistency beats intensity. The team that executes the same system for 52 weeks will outlast the one that sprints, burns out, pauses to recover, then sprints again.

Why Intelligent People Choose Ponies

Here’s what makes the Scott mistake seductive: it feels like leadership.

Choosing the experimental approach feels innovative. Betting on your team’s ability to adapt reflects trust. Refusing to over-plan feels agile. Pushing through exhaustion demonstrates commitment.

Scott wasn’t stupid or lazy. He was a decorated naval officer who had survived previous Antarctic expeditions. His men loved him. His intentions were commendable.

In complex systems—whether polar expeditions or startups—good intentions and personal courage do not overcome structural fragility.

The Swiss Cheese Model from reliability engineering:

Critical systems don’t fail from one catastrophic event. They fail when multiple small failures cascade—when the gaps in different defense layers align.

Commercial aircraft don’t crash due to a single sensor failure. They crash when a sensor fails AND a backup warning gets ignored AND a pilot misjudges under fatigue AND the flight was behind schedule AND weather closed alternate airports.

Each failure is survivable, but the combination is lethal.

Amundsen built margins: extra food at every depot, multiple backup navigation tools, dogs that could eat locally, clothing that worked when wet, and daily mileage that preserved energy reserves.

No single point of failure could jeopardize the expedition.

Scott’s system had no slack. Ponies that couldn’t adapt, depots spaced at maximum range, gear optimized for ideal conditions, daily distances assuming everything would go smoothly.

When one thing broke, everything fell apart.

Your top sales rep carries 60% of the pipeline and there’s no documented process for closing deals. Your payment processor is Stripe with no backup if they freeze your account. Your biggest customer represents 40% of MRR. Your entire codebase is in one engineer’s head.

You’re running fine today. But you’re one unexpected departure, one policy change, one market shift away from a failure that ends the company.

Try This Today (15 minutes):

Open a blank doc and title it “Single Points of Failure.”

List all dependencies that would stop your business for over 48 hours if you lost one:

• People (roles only one person can perform)

• Customers (revenue concentration)

• Infrastructure (critical services without backup)

• Processes (knowledge isolated in one person’s mind)

Pick your scariest single point of failure. This week, spend 30 minutes adding a layer of redundancy:

• Document the sales process of your top representative.

• Set up a backup payment processor.

• Export essential customer data outside your primary database.

• Record a Loom of your CTO deploying production code.

You’re not building for paranoia. You’re building margins so a bad week doesn’t become a disastrous quarter.

The Dogs-Not-Ponies Audit: Your 90-Day System Check

Note on innovation: Be boring in operations and innovative in product. Amundsen used proven Inuit gear to focus on the novel route-finding problem. You want stable systems to take smart risks on customer problems. Redundancy isn’t anti-agile—it’s what makes agility sustainable.

Use this framework to assess your next 90 days:

1. Know the Environment (Before Committing Resources)

Have you talked to 10+ people in your space? Not pitching—listening. What do practitioners know that you’re unsure about?

Action:

• Block three hours this week.

• Find and interview two people who perform your target customer’s job daily.

• “What didn’t work? What solution solved this?”

2. Choose Dogs Instead of Ponies

List your current tools, tactics, and team structure. Label each as:

• Dogs: mundane, established, dependable (Stripe, Retool, weekly standups)

• Ponies: exciting, unproven, break (custom CRM, untested channels, experimental processes)

“Am I choosing this because it works, or because it signals the company I want to be?”

Action:

• Replace one pony this month.

• The answer is: switch to the straightforward tool your successful competitors use.

3. Build Your Essential Weekly Rhythm

Define the weekly operating rhythm:

• Sales pipeline review (same day/time, same format)

• Customer health check (automated alert + manual review)

• Are we improving or declining in product shipping velocity?

Formula: Consistency × Time = Compounding

Small weekly gains beat big quarterly gains because compounding works with regular intervals.

Action:

• Pick one metric you review.

• Set a 30-minute recurring calendar block to review it weekly for the next 12 weeks.

• Consistently invite the same people.

• Watch patterns emerge that quarterly reviews overlook.

4. Install Excessive Margins of Safety

The 30/30 Margin Rule states that your financial model should assume 30% higher burn and 30% longer sales cycles than your realistic projection.

Why? Your realistic projection assumes nothing goes wrong. Problems arise.

Margin checklist:

• Cash: a buffer of 3+ months beyond your minimum viable runway

• Pipeline: Maintain a $600–900K pipeline (2–3x coverage) if you need $300K in Q1.

• Infrastructure: Backup vendors for anything that would halt operations if it failed.

• Knowledge: Every critical process documented clearly enough for a new hire to execute it.

Action:

• Open a spreadsheet and create this margin audit.

• Label each category green (margin exists), yellow (thin margin), or red (single point of failure).

• If anything scores red, that’s your fragility point. Address it before it fails.

The One Thing to Do in the Next 24 Hours

Both teams had the same dream, similar talent, and remarkable courage.

The difference wasn’t vision. It was systems.

Startups aren’t won by the boldest idea or the most inspiring speech. They’re won by teams that treat execution like Amundsen treated the ice: as an engineering problem, not a platform for heroics.

Here’s your 24-hour challenge:

Open your calendar. Find one current “heroic sprint”—a launch that will change everything, a critical sales quarter, a churn-fixing feature, or a fundraising blitz that will solve your problems.

Ask yourself: “Am I counting on rising to the occasion, or do I have a reliable system that will work on my worst day?”

If you’re relying on heroics, replace it with a repeatable system. Pick one pony to ride this week and replace it with a dog to run every week for six months. Then run it next week. And the week after.

The Pole isn’t the point. Reaching home alive is.

What operational default will you fix this week? Reply—I read every response.

_______________

Did this post resonate with you? If you found value in these insights, let us know! Hit the ‘like’ button or share your thoughts in the comments. Your feedback not only motivates us but also helps shape future content. Together, we can build a community that empowers entrepreneurs to thrive. What was your biggest takeaway? We’d love to hear from you!

Interested in taking your startup to the next level? Wildfire Labs is looking for innovative founders like you! Don’t miss out on the opportunity to accelerate your business with expert mentorship and resources. Apply now at Wildfire Labs Accelerator https://wildfirelabs.io/apply and ignite your startup’s potential. We can’t wait to see what you’ll achieve!

Brady Chandler is the CEO of AgCents, a cattle lending and analytics platform we backed at Wildfire Labs. He’s spent his whole career in agriculture — every drought cycle, every shift in feeder margins, every market dislocation since the 1980s. He can sit across from a fifth-generation rancher and talk feeder economics for forty minutes without checking a note.

For the past few months he’d been pitching the platform — to lenders, family offices, investors. A tight top-level story about four tools and what each one does for a producer. People nodded. People wrote checks. People scheduled second meetings.

Then someone asked to see it.

Brady went into prep mode and discovered something uncomfortable: there was a gap between the platform he’d been pitching and the platform that actually existed. Not because anyone had lied. Because he’d been carrying around a mental model of the product that he’d assembled in conference rooms — never tested against the actual screens.

His co-founder David George had been carrying the inverse problem. David built the software. He knew it to the keystroke. He didn’t know — at the same depth — what a producer does on the second Wednesday of October when the auction barn calls.

Brady had the producer. David had the product. The gap between them was the demo.

What Brady later called it — and I’m quoting him — was “an unbelievable tightening of the sphincter.”

That’s the line. That’s the moment. That’s where the second black box gets sealed shut — except this time, they noticed.

The first black box, and the other one

In the chapter of the book I’m writing called No Black Boxes, I argue that the most dangerous thing a founder can do is run their company with a function they don’t understand and don’t try to.

When most people hear that, they picture a technical founder who refuses to sell. The CTO-CEO who hires a VP of Sales, hands over the revenue motion, and is shocked eighteen months later when he can’t explain why customers actually buy. I wrote about that one earlier — when you outsource customer truth, you stop hearing the objections that should be reshaping your roadmap, and the company quietly drifts away from the market it was built for.

That story is real. But it’s only half the story.

The other half — the one that gets less airtime because it’s harder to dramatize — is the sales-and-domain founder who treats technical architecture as the CTO’s department. Brady didn’t need to write code. He was right about that. What he needed was to ask the next question — and what saved him was that the demo schedule asked it for him.

Trust without understanding is a black box. It doesn’t matter which side of the company it’s on.

The architecture is the strategy

Brady and David caught their gap during demo prep, while the stakes were credibility with investors. Most founders catch theirs much later, when the stakes are bigger.

Here is the thing nobody tells you in the early years: every architectural decision your engineering team makes in the first eighteen months is a strategic decision in disguise.

The data model your team picks in week three quietly decides which customers you can serve in year three. The integration framework — or the absence of one — decides which markets you can expand into. The choice to be single-tenant or multi-tenant decides whether you can sell into enterprise without a six-month detour. The decision to roll your own auth versus building on a standards-based one decides whether you can pass your first SOC 2 audit without rebuilding the foundation.

None of those are tech decisions. They are strategy decisions, made by people who didn’t think they were making them, while the founder who actually owns strategy was off closing deals.

I’ve watched this exact movie too many times. The CRM that couldn’t add a second product line because the data model was hardcoded for the first. The marketplace that couldn’t expand into a second vertical for the same reason. The SaaS company that landed its first enterprise logo and then spent nine months and two million dollars rewriting the backend to support SSO and role-based access — features the sales founder had been promising on calls for a year, not knowing the architecture didn’t support them and was never going to without a rewrite.

In each case the founders weren’t asleep. They were working harder than anyone in the company. They just weren’t working on this. And nobody around them was incentivized to wake them up, because telling the CEO that the foundation has a problem is one of the least career-advancing things an engineer can do.

Speed got cheap. Comprehension got expensive.

That was the playbook for twenty years. Slow enough that a halfway-attentive founder had time to notice the architecture saying no before it killed the company.

AI compressed the timeline.

I have founders in my orbit right now who shipped a working MVP in a weekend. Not a prototype — an actual product, with auth and a database and a payment integration, vibe-coded into existence by someone who wouldn’t have called themselves an engineer two years ago. That is the upside. The downside is that the architectural decisions inside that weekend MVP — the data model, the auth pattern, the way state is stored, the assumptions baked into the schema — were made by an AI in conversation with a founder who didn’t know to ask whether any of it would survive the second customer.

The founders in those examples had eighteen months before the architecture started saying no. The founders shipping today have six. Maybe less.

The asymmetry is new. For most of software history, shipping fast required knowing your stack — speed and comprehension came in the same person. AI broke that link. Now the founder who can’t read the code can still ship the product. The two skills decoupled, and most founders only kept the one they had.

This is the part where most “AI for founders” essays say isn’t it amazing how fast we can move now, and stop. That’s the wrong stop. The right stop is one step further: AI made shipping cheap, which made understanding what you shipped the new strategic question. Speed isn’t the moat anymore — every reasonably capable founder can move fast now. Comprehension is the moat. Knowing what you actually built. Knowing where the load-bearing walls are. Knowing which decisions the AI made for you while you were focused on the customer demo.

The black box used to be a function of delegation — the sales founder trusted the CTO, the technical founder trusted the VP of Sales. Now there is a third black box: the AI itself, shipping decisions inside your codebase that nobody in the company has read, much less debated. And those decisions still get to vote on which customers you can serve in year three.

The questions below get more important, not less.

What Brady and David did next

The playbook isn’t complicated.

They mapped the gap. They named the scenarios the platform could demo with credibility today, and the ones it couldn’t yet. They tightened the roadmap for the next two iterations around the demos that mattered most. They restored Brady’s credibility with stakeholders by being precise about the line between here’s what works today and here’s what’s coming next — with dates.

What they didn’t do is what most founders do. They didn’t fix the gap by Brady becoming an engineer or David becoming a rancher. They didn’t paper over it by getting better at pitching what wasn’t there. They got their hands on the actual product together — the CEO and the CTO, in front of the same screen — and stopped letting demo prep be the first time they saw the same thing.

That’s the whole game.

The three questions

You don’t need to learn to code. You need to learn to ask three questions.

Once a month, sit down with your technical lead. Not a status update. Not a sprint review. A conversation. Ask:

1. What’s the most fragile part of our system right now? Every system has one. The thing that scares the engineers. The thing that breaks at three in the morning every other Sunday. The library nobody wants to upgrade. If you can’t get an honest answer to this question, the problem isn’t the architecture — it’s the relationship.

2. What would break if we 5x’d our load tomorrow? Not “could we handle it.” Not “are we scalable.” Specifically: what would be the first thing to fall over? The database? The job queue? The auth layer? The third-party API you’re rate-limited on? The answer is also a quiet map of the bets you’ve been making about how big this gets.

3. What decision did we make six months ago that you’d make differently today? This is the one that opens the kimono. Every engineering team has at least one of these. The choice that was right at the time and isn’t now. The dependency that seemed harmless and turned out to be load-bearing. The shortcut that’s now a wall. If your tech lead can’t name one, they’re either not paying attention or they don’t trust you with the answer. Either way, you have a different problem than you thought.

If your team is shipping AI-generated code at any scale, add a fourth: What’s in our codebase that nobody on the team has read? The third black box answers to the same logic as the other two. Trust without understanding is still a black box, even when the source of the trust is a tool.

These questions are sized for a small-to-mid-stage company with one engineering team. At 200 engineers across multiple platforms, the cadence gets more specific — per service, per quarter — but the principle holds. The discipline doesn’t bend with scale; the schedule does.

You’re not trying to second-guess the engineers. You’re trying to make sure that the architectural reality of your company shows up in your strategic conversations before it shows up in a customer call you can’t deliver on. Or a demo prep that almost was.

The point

The first black box — the technical founder who can’t sell — kills companies by starving them of customer truth.

The other black box — the sales founder who never examined the foundation — used to kill companies slowly, and the death looked like success right up until it didn’t. Revenue was growing. The team was hiring. The deck looked great. Then the architecture said no, and the strategy got quietly rewritten by people whose job titles didn’t include the word strategy.

AI shrunk the runway between looks like success and isn’t.

You name the CTO. You name the VP of Sales. The third black box doesn’t have a name — it’s the AI shipping decisions inside your codebase while you weren’t looking. Until you open it, it gets a vote on every customer you don’t yet have.

Trust is good. Understanding is better. Open all three.

Most startup advice worships speed. Move fast, raise fast, exit fast, then go do it again. Eric McDonald did the opposite. He spent nineteen years building one company, in a basement, in South Dakota, and walked out the other side of a $1.25 billion outcome. Here’s what that kind of endurance taught him about capital, identity, and why the founder who stays usually beats the founder who moves.

I recently sat down with Eric McDonald on the podcast. Eric wrote the original code for DocuTap himself, in his basement, and scaled a healthcare tech company from nothing to a $1.25 billion valuation. The company ran through every stage of the money game: angel investors, a family office, venture capital, private equity, a merger with its biggest competitor, and a final exit in 2022. Today he keeps his hands in tech with a couple of startups, including Coin Lion, an algorithmic crypto trading platform. He still lives in South Dakota, on eight acres down by the river.

What struck me was how little of his story is about being the smartest person in the room. Eric almost shut the company down six or seven times. He raised money $50,000 at a time from corn farmers. He nearly killed himself tying his identity to the work. And he came out the other side with one of the better outcomes this state has ever produced. If you’re early in the grind and wondering whether you have it in you to keep going, these lessons are for you.

Pick a Vertical That Grows Without You

Eric didn’t start with urgent care. He started by trying to sell electronic medical records to everyone. Primary care, then an ENT group, then orthopedics, then cardiology. Every specialty looked like a customer. And every specialty turned out to have its own data sets and its own workflow, each one a separate product to build. As a small company, he could only point his time at one thing.

So in 2004, four years in, he picked urgent care. Two reasons. There were no big competitors focused there, while everyone else was crowded. And the category itself was exploding. A client who signed with two clinics would be running eight or ten within three years. Eric got the growth without having to go re-sell it. The market did the selling for him.

This is the thing we talk about constantly at Wildfire. Founders show up with a broad vision to change an industry, and the actual work is finding the beachhead inside it. Eric found the rare kind: a vertical that compounds underneath you, so your customers expand whether or not your sales team does.

What to do this week: Pull your customers from two or three years ago. Did their usage, seats, or spend grow without you re-selling them? If the line is flat, your product might be fine and your vertical might still be the problem.

Get to Break-Even Faster Than Feels Right

I asked Eric what he’d do differently. No hesitation: they were undercapitalized, and he chased growth when he should have chased profitability. He raised in $50,000 and $100,000 increments from pig farmers, corn farmers, and local business guys, because there was no large investor group in the region until 2011. And the business model made it worse. It was transactional. He had to hire three people at $80,000 each to chase revenue he was earning a dollar fifty at a time.

His number one rule now, for almost any startup: get to cash flow break-even as fast as possible. And he names the counter-pressure honestly. The venture and private equity guys will tell you to run as hard and fast as you can, and if you run out of money, they’ll give you more. That sounds generous. It’s a way to take more equity and more control of your company.

We’re living through the swing back. At Property Meld we’re counting the days down to profitability, and the investors are supportive of it, precisely because nobody knows what the market is going to do next. Free cash flow you can reinvest puts you in a different negotiating position than a runway that ends in someone else’s term sheet.

What to do this week: Run two numbers. Your months to break-even at current burn, and your months of runway if you raise. If the first number is reachable and you’re still optimizing for the second, ask yourself who that actually serves.

Trust Is Yours to Lose, Not Earn

Eric has a hiring philosophy I keep thinking about. New hires get his trust instantly. Not “earn it over time,” but “it’s yours to lose.” And it comes paired with something most founders won’t do: he puts his own weaknesses on the table. He told his sales team plainly that he could close any deal you put him in front of a client for, but he was never the guy waking up thinking about how to close more deals, so they needed to push him. He has, in his words, no problem being the dumbest person in the room.

This is the heart of what I call “no black boxes.” Most founders are a sales person, a product person, or a technical person, and they over-index on that one strength and quietly black-box the other two. When the business hits a rough patch in marketing or ops or hosting, the founder who black-boxed those functions can’t call BS and can’t add value. Eric’s fix wasn’t to master everything. It was to say out loud what he didn’t know and build a team that would push him there.

What to do this week: Name the one function in your company you understand least. Tell the person who runs it, out loud, that it’s your weak spot and you want to be pushed on it. Watch how fast the black box opens.

Your Best Ideas Aren’t Yours

I asked Eric what his superpower is. He said he can talk to ten people, and when one of them says one thing, he knows immediately: that’s money. “Rarely are they my ideas,” he told me. People ask how he comes up with all this stuff, and his answer is that he doesn’t. He’s just good at listening and recognizing signal when he hears it.

That’s the whole game for a lot of great founders. Not generating brilliance in a vacuum, but pattern matching across a lot of conversations and grabbing the one piece that fits the problem in front of you. The skill isn’t talking. It’s filtering signal from noise, and then having the conviction to go act on the signal.

What to do this week: Think back over your last ten outside conversations. Which single comment was “money” and you let it slide past? Go do something with it this week.

Reinvent Yourself or Get Left Behind

The version of Eric who wrote code in his basement is not the version who ran a billion-dollar company, and he’s clear that the distance between those two people is enormous. He had three distinct executive teams over his tenure, each one suited to a different size of company, and for the most part one team didn’t make it to the next level. Leading ten people is a different job than leading fifty, which is a different job than leading a hundred.

The harder reinvention is the founder’s own. Once you’ve recruited a team that can run the thing, the question becomes what only you can do now, and that answer keeps moving. The founders who plateau are usually the ones who got good at one stage and stopped re-recruiting themselves for the next one. Eric credits his VC and PE boards for pushing him, but his real point was that if your board won’t push you, go find the rooms that will.

What to do this week: Write down the job you actually do today versus the job the company needs from you six months from now. The gap between those two is your next reinvention.

One Plus One Equals Four (And Where I Push Back)

The merger that made DocuTap was Warburg’s plan from day one: take the number one player and the number two player in urgent care, combine them, build it up, and sell it. Practice Velocity’s CEO said he wanted the top job, and Eric, twenty years in and with no ego about it, handed over the torch and moved to the board. It worked. A $1.25 billion valuation, roughly 70% of the market, and a sit-down with the DOJ to explain why that wasn’t a monopoly problem.

Here’s where my own experience pushes back. At Concur we did thirteen or fourteen acquisitions, and if I’m honest, at least half weren’t accretive to what we promised. We bought $40 million companies that turned into shelfware. So when Eric told me the merger worked, I wanted to know why his didn’t end up in that pile.

His answer is the lesson. The big merger worked because it wasn’t a bolt-on. Two companies of similar size, where everyone knew that if it didn’t come together the whole boat sank, so they made it come together. The deal that got messy for him was the small tuck-in, the one sold on “it’ll be accretive this way and there’s this revenue opportunity” that never quite materialized. That matches my scar tissue exactly. The dangerous acquisition is rarely the bet-the-company one. It’s the small one nobody respects enough to integrate properly.

What to do this week: If you’re eyeing an acquisition, ask whether failure would sink the boat or just dent it. Counterintuitively, the ones that can’t fail quietly tend to get the attention that makes them work.

Don’t Tie Your Identity to the Company

This is the part I most wanted to talk about, and Eric didn’t flinch. Burnout is real, and it nearly took the company down more than once. He described the lonely version everyone pictures: standing at the edge of the cliff called failure, wondering how to make payroll, carrying the weight of investors and customers and employees while everyone assumes you have it handled. But he also named the version nobody warns you about. Running a billion-dollar company with an all-star team that doesn’t need you, showing up and not knowing what to do, feeling like you have no value. Both ends burn.

The deeper trap is identity. When Eric stepped down as CEO and took a year off, people at events would ask what he did for work, and he’d hear himself spinning stories to justify why he was “between jobs,” because his sense of self was wrapped around the company. His advice is to anchor your identity somewhere the business can’t touch it. Husband, father, child of God, whatever your version is. Not “was my company successful this hour.”

He lived the reset literally. The week after he left DocuTap, with his new house still being built, he moved into a canvas tent on his eight acres for three months. Chainsaw, tractor, an outdoor shower he rigged up, trips to the laundromat to wash his clothes. Stripped all the way back to the basics. He says he wouldn’t do it again, but it’s some of his boys’ favorite memories, and it was exactly what he needed in that season. Sometimes you have to take everything away to remember who you are underneath the title.

What to do this week: Finish this sentence without mentioning your company: “I am someone who ___.” If it’s hard, that’s the work, and it’s worth more than anything on your roadmap.

The Bottom Line

Speed gets the headlines. Endurance builds the billion-dollar outcomes nobody saw coming from a basement in South Dakota. Eric’s career is a case for staying: pick a vertical that grows under you, get to break-even faster than feels comfortable, be honest about what you don’t know, and don’t let the company become the only answer to who you are.

What would change about how you’re building if you assumed this was a nineteen-year job, not a three-year one?

This article is based on my conversation with Eric McDonald on the podcast. We covered more than I could fit here, including his bet that nobody will be trading crypto manually in a few years, what it was like to argue your own non-monopoly case in front of the DOJ, and why he wanted a brand-new industry after two decades in healthcare.

[Listen to the full episode here.] (episode link)

Connect with Eric: LinkedIn (Eric noted LinkedIn is the only platform he’s on.)

In 1928, Henry Ford bought 2.5 million acres of the Brazilian Amazon to grow his own rubber. He hired engineers who knew steel and glass. He shipped Cape Cod bungalows down the river. He paved streets, built a hospital, and put movie theaters in the jungle.

He did not hire a single botanist.

That detail is the whole article. Not the cafeteria riot. Not the floating bar. Not the failure of synthetic rubber to make the project obsolete by 1945. The botanist Ford didn’t hire is the reason none of the rest matters. Twenty million dollars — hundreds of millions in today’s money — and the founder of the assembly line never thought to ask whether rubber trees grow the way Model Ts do.

They don’t.

In the wild, rubber trees grow far apart. The spacing isn’t aesthetic — it’s defensive. Pests and leaf blight can’t bridge wide gaps. In Fordlândia, the trees were planted in the dense, orderly rows you’d use to plant corn. Within a season, the blight spread the way fire spreads in a closed room.

The rest of the failures — the cafeteria riot, the Michigan-style 9-to-5 schedule under the equatorial sun, the prohibition that workers eventually defeated by building a floating bar on the river — were downstream. By the time the Brazilian workforce was rebelling against brown rice and whole-wheat bread, the trees were already dying. The rubber failure was ecological. The rest was theater.

But the underlying mistake — the thing that produced both the rubber failure and the cafeteria riot — was the same.

Ford imported the system instead of learning the terrain.

This is the thing founders do.

You succeed at one thing. You build a real business solving a real problem. You learn the operating model that worked — the cadence, the playbook, the way you talk to customers, the way you hire. And then you walk into a new domain holding the playbook like it’s a passport.

It isn’t. It’s a souvenir.

The mistake isn’t arrogance in the abstract. Ford wasn’t an idiot. He had genuine expertise in something genuinely hard. He had revolutionized manufacturing in a way that reshaped the global economy. The mistake was assuming that competence in manufacturing transferred to agriculture. That the operating principle behind the assembly line — standardize, scale, control variation — applied to a living system that runs on the opposite principle.

It didn’t. The jungle isn’t standardizable. Rubber trees aren’t input materials. Brazilian workers aren’t replaceable parts. None of those things failed because Ford was a bad manager. They failed because Ford was a great manufacturing manager applying manufacturing logic to terrain that doesn’t accept it.

You’re most dangerous when you’re confident.

I have done a version of this. Most founders I know have done a version of this.

When I left Concur and came back to Rapid City to start Wildfire Labs, I brought the SaaS playbook with me. Concur scaled because we knew exactly what to do at every stage — when to hire, when to layer in process, when to push the customer success motion. I had run that playbook for nineteen years. It worked at $10 million in revenue. It worked at $100 million. It worked at $1 billion. I was not modest about how well I knew this terrain.

Then I tried to apply it to AgCents.

Cattle financing isn’t SaaS. The customer cycle isn’t quarterly — it’s the cattle cycle, which runs on biology, not procurement calendars. The buyer isn’t a CFO with budget authority — it’s a rancher who is also the operator and the bookkeeper, and his decision is shaped by his banker, his neighbor, and his father-in-law. The “user” doesn’t sit at a desk. He’s in a feed truck.

I was Henry Ford in a feed lot.

Not because I was wrong about software. Because I was right about software in a domain where software wasn’t the thing. The terrain was. And the terrain doesn’t care that you ran sales for SAP’s expense module.

There’s a founder I mentored who started in horizontal CRM — selling open source against Salesforce, making the money on the consulting underneath it. He was good at it. When we talked about verticalizing, he picked agricultural co-ops. The need was real. He landed a handful of customers. Then he hit the wall. He was a technical founder. He loved building product. Ag co-ops don’t buy product. They buy people they know. It didn’t matter how good his CRM was — they weren’t going to buy software from someone who couldn’t talk crop rotation over coffee. He sold the company to a buyer who had the relationships, and walked away. The mistake wasn’t picking ag. It was assuming the asset he had — building great software — was the asset the market valued.

The caveat matters.

Domain expertise is real. It is one of the most valuable assets a founder can carry into a new venture — when the new venture is in the same domain. A second-time founder building the next version of vertical SaaS for property managers has a massive advantage over the first-time founder. A career operator in regulated industries who starts a regulated-industry company knows things you can’t read in a book. Pattern matching is real. Domain depth compounds.

The transfer assumption is the trap.

Ford’s expertise in manufacturing was the asset. His belief that manufacturing expertise generalized was the liability. The two things lived in the same person, and one of them paid for the other. The founder version: you can be the most capable operator in your old domain and the most reckless founder in your new one. The capability is what gives you the conviction to move fast. The recklessness is what makes you ignore the botanist.

The fix isn’t humility in the abstract. It’s specificity. Before you import the playbook, make a list — written, on paper, not in your head — of the assumptions the playbook depends on. Standardized inputs. Predictable customer cycles. A buyer who reads email at her desk. Channels you can pay to scale. Then walk through the new terrain and check, line by line, whether each assumption holds. Most won’t. The ones that don’t are your botanists. Hire them, or build them, or partner with someone who has them, before you spend twenty million dollars on Cape Cod bungalows.

Henry Ford never visited Fordlândia.

He sent his system. The system didn’t know what it didn’t know.

Today, about two thousand people still live in the ruins. They live in the original American bungalows. They walk the paved streets that the jungle is slowly reclaiming. The water tower is rusting; the powerhouse is a glass-walled skeleton. But the people are there, going on with their lives in the bones of a billionaire’s failed utopia.

The terrain wrote that ending. Ford built the town. The jungle wrote what comes after.

Whatever you’re building, whatever expertise you’re carrying with you, the terrain you’re walking into is going to write the ending of your story too.

It would be cheaper to ask it some questions first.