Not a lab in a kitchen. There was no lab. There were no customers either. There were three of us, Shevy, Thanh, and me, with renderings of a lab that didn’t exist yet and a very long list of things we hadn’t figured out. Eventually, that turned into a full production lab.

Here’s the strange part about the in-between. At no point did the work feel dramatically harder than the month before. But every few months, one of us would look up and ask the same question: what was I even doing six months ago? The days back then felt full. The days now felt full. What counted as a full day had quietly doubled, and neither of us could point to the week it happened.

It took me a long time to realize that feeling has a name.

The oldest program there is

Legend has it that Milo of Croton, a six-time Olympic wrestling champion in ancient Greece, built his strength by lifting a newborn calf onto his shoulders every day. The calf grew a little each day, so the weight grew a little each day, and Milo grew with it. Four years later he was carrying a bull.

Lifters call this progressive overload: do slightly more than last time, recover, repeat. Every serious training program is built on it. And the legend only works because Milo did both halves. He showed up every day, and the weight kept going up. Skip months and the calf outgrows you. Carry the same weight forever and you stop growing at all.

Founders talk a lot about consistency. What we talk about less is that consistency is only half the program.

Start with things that don’t scale

There is one place the gym version misleads you, and it’s right at the start. A new lifter begins with the empty bar and works up slowly. A new founder should not.

The best early-stage advice is still Paul Graham’s: do things that don’t scale. His essay is a catalog of founders starting heavy on purpose. The Airbnb founders going door to door in New York to recruit hosts and fix up their listings. The Collison brothers setting up Stripe on a founder’s laptop on the spot instead of emailing them a link. Our version was less glamorous: a kitchen, some renderings, and brute-force figuring out what a production lab would even need to be. This kind of work is heavy for what it produces, it’s uncomfortable, and it breaks constantly. That’s exactly why it works. You’re learning very quickly and breaking things very quickly, and every broken thing tells you what to fix next.

The gym has a version of this too: beginners progress fastest. A new lifter can add weight almost every session, while a veteran fights for months to add five pounds. Your first year as a founder works the same way. You will never adapt as fast as you do at the start, so starting from zero and pacing yourself politely wastes the steepest part of the curve. Load the bar heavier than feels reasonable.

Nobody assigns your reps

Here’s what makes all of this harder for founders than for lifters. In the gym, you choose the weight. In a company, the weight mostly chooses you. Customers show up with problems you haven’t solved before. The team grows past the point where you know everyone’s week. The invoices gain zeros, and so do the mistakes. The company is the calf, and it doesn’t check whether you’re ready before it grows.

So the question isn’t whether the load will increase. It will. The question is whether you’ll be under it every day while the increments are still small, or whether you’ll look up one quarter and meet the bull all at once.

That’s why habits matter so much here. Nobody assigns your reps. No boss, no syllabus, no review that shows up on time. The only accountability a founder gets is the accountability they build, and habits are how you build it. James Clear’s line is that you don’t rise to the level of your goals, “you fall to the level of your systems.” The founders we’ve seen last the longest aren’t the most motivated ones. They’re the ones who stopped relying on motivation at all.

The warning

Now the part that made us want to write this.

Muscle adapts to a fixed weight in a few weeks, then stops changing. Founders adapt the same way. If your weeks have looked identical for months, same tasks, same conversations, same size of problem, be wary. You’re not compounding anymore. You’re maintaining.

The dangerous part is that maintaining feels like discipline. You’re showing up. You’re working hard. Nothing about it feels like quitting. Bezos was pointing at the same thing with Day 1 and Day 2 at Amazon: “Day 2 is stasis. Followed by irrelevance. Followed by excruciating, painful decline. Followed by death.” He was talking about companies, but it starts with the founder, and it never announces itself. A plateau just feels like a routine that finally works.

The readout we trust is the look back. Every stretch of real growth at Moichor ended with that same disoriented question: wow, what was I even doing before this? So run it in reverse. Look at what filled your days six months ago. If it doesn’t seem small from where you’re standing, worry. It’s one of the first things Shevy and I check as investors now too. When we meet a founder a second time, a few months after the first, the real question under the update is whether the problems got bigger.

What to do when you catch it

Say you run that check and the answer is bad. Your weeks are on repeat. In the gym the fix is simple: add weight. For a founder, the move has two parts, and this is the most practical paragraph in the essay.

First, the task you’ve mastered doesn’t get heavier. It leaves your plate. Molly Graham’s version is “give away your Legos”: the job you finally got good at is exactly the job you should hand to someone on your team, even though it stings. And in 2026 the bar is lower than a hire. If you’re doing the same task the same way every week, that’s no longer your job. That’s a workflow, and you should be automating it with AI.

Second, take the capacity you just freed and point it at something you’re currently bad at. That’s the full cycle: master a rep, hand it off or automate it, pick up a heavier one. Delegation isn’t a reward you earn at scale. It’s how a founder adds weight to the bar.

The other failure mode

Exhaustion is the opposite signal, and founders read it exactly backwards.

In the gym, nobody confuses pain with progress for long. If your joints ache for weeks, you added too much weight too fast, and the fix is built into every real program: deload. Pull back, recover, then keep going. Rest isn’t a break from training. It’s the half of training where the muscle actually gets built.

Founders treat burnout as proof of commitment. It isn’t. It’s information, and what it’s telling you is that the increments got too big, or that you kept the starting pace long past the start. The unscalable sprint works in year one because everything is small, including the cost of your mistakes. Keep sprinting like that under a much heavier load and something tears.

I don’t have a formula for the right increment. I’m not sure anyone does. But it’s smaller than your ambition wants: one conversation above your weight class, one ask earlier than feels comfortable, one system built this week instead of a task repeated. Naval likes quoting his weightlifting coach, Jerzy Gregorek: “Hard choices, easy life. Easy choices, hard life.” Pick the slightly heavier week on purpose, then actually recover from it.

The gut check

Paul Graham’s definition of a startup is a company built to grow fast, steered by one number: the growth rate. That’s the external readout, and it’s a good one. But you’ll feel a plateau long before any chart shows it, so here are the internal checks we’d actually run:

• Does what filled your days six months ago look small from here? If your old weeks look the same size as your current ones, the weight stopped going up.

• Are the problems getting bigger, or just more numerous? New categories of problem mean the overload is working. The same problems at higher volume is a plateau.

• What did you hand off or automate this quarter? If the answer is nothing, you’re hoarding reps you’ve already adapted to.

• Is anything on your plate slightly out of reach? There should always be one thing you’re currently bad at. That’s the new muscle.

• Which kind of tired are you? Tired from heavier work recovers. Tired from the same work doesn’t. One is training. The other is a warning.

Keep carrying it

So that’s the program. Show up daily, because the calf grows whether you do or not. Start heavy and unscalable while breaking things is still cheap. Hand off the reps you’ve mastered, automate the ones a machine can do, and keep one thing on your plate you’re still bad at. Read exhaustion as data, not proof.

Our calf was a set of renderings in a kitchen. The bull was a full production lab. In between, the effort never felt like it changed, and that’s the part nobody tells you: growth doesn’t feel like growth while it’s happening. It mostly feels like Tuesday. That’s how you build the muscle.

Carry the calf. It’s growing either way.

We have lived both sides of this. We went through it building Moichor, and now we sit on the other side of the table backing founders at NPU. The longer we do this, the more convinced we are that the gap between the feed and the job is where most would-be founders get lost.

So this is our attempt to separate the two. Not what being a founder looks like, but what it actually takes.

1. What Is Your Drive?

Not passion in the abstract. A specific itch that will not leave you alone.

For us at Moichor, it was this: veterinary diagnostics was controlled by a duopoly, IDEXX and Antech. The incumbents had no real incentive to innovate. Why would they? The playbook was simple: lock clinics into contracts, then raise prices on customers who cannot leave. Vets paid more year after year, and ultimately pets and their owners paid more, for a service that had not meaningfully improved in decades.

That bothered us. Not in a “this is an interesting market” way. In a “we cannot stop thinking about this” way. If the incumbents were never going to modernize, someone from the outside had to. We wanted vets to finally have a modern alternative.

We now watch this from the other side of the table, and the pattern is consistent: startups rarely die from competition on day one. They die from founders quitting in year three, usually because they picked a problem they never actually cared about. The problem has to be something you are willing to live with for five-plus years.

Your drive is not a launch-day vibe. It is your fuel tank for the stretch when nothing is working. If you cannot name your itch in one sentence, keep looking.

2. Do You Want to Make Money?

Founders get shy about this one. We think that is a mistake.

We understand where the shyness comes from. Nobody wants to look like they are in it for the money. But there is a difference between flashing wealth and wanting it.

The best thing we have read here is Naval Ravikant’s How to Get Rich (without getting lucky). His core line is that you get rich by “giving society what it wants but does not yet know how to get,” at scale. That is all a startup really is. Find the problem society has no solution for, solve it, and society rewards you handsomely. The money is the byproduct of building something generational, not the reason it exists. Wanting that outcome is not corrupt. It is honest.

Naval also warns in the same thread that quietly resenting wealth is a good way to guarantee you never build any. We see this play out with founders all the time. The ones who are squeamish about money undercharge. They avoid sales. They pitch impact instead of a business. Your customers, your employees, and your investors all need you to want to win, and in business, winning is denominated in dollars.

Own it quietly. Then go build.

3. You Have “Nothing to Lose.” Burn the Boats Anyway.

Let us be honest about this one, because the cliché version is a lie.

We had things to lose. After graduating, Matt passed on grad school and the stable career path. Shevy walked away from med school. Our parents were not exactly comfortable with any of it in the beginning, and we understood why. Those were real costs.

But when we actually ran the numbers on the downside, it was survivable. Worst case, the company fails, we end up a few years behind our peers on paper, and we get jobs, probably as far more capable people than when we started. The downside of a startup is capped. The upside is not. Once you see that clearly, the leap of faith gets a lot smaller than your parents fear.

Then comes the second move, and this is the part people miss. Once you are in, you act as if there is no way out. Our mindset was “we have to make this work.” But if we are being honest, that energy did not come from fear of the downside. It came from belief. We were fully bought in. We truly thought this would happen. Looking back, was that level of certainty justified? Probably not. That is exactly the point. No rational forecast generates that kind of energy. Belief does, and it has to be real.

And underneath the belief was something simpler. We wanted to control our own destiny. Neither of us wanted a 9 to 5. Once you know that about yourself, the boats burn themselves. The way back exists. You just have no interest in taking it.

Michael Seibel wrote an essay called Why Should I Start a Startup? and his answer stuck with us. There is a certain type of person who “only works at their peak capacity when there is no predictable path to follow,” where the odds are low and failure is personally theirs to own. That last part is the whole game. When you are at the lowest of lows, who is going to summon the courage to go get that next customer? Nobody but you. No manager, no playbook, no one coming to help. And that same muscle powers everything else: your selling, your hiring, your pitching, your ability to execute the vision.

Refusing to quit is not just a personality trait. It is a strategy.

4. The Unglamorous 95 Percent

The feed shows the raise. The job is everything underneath it.

Paul Graham’s Do Things That Don’t Scale is the canonical essay here. Startups do not take off on their own. Founders drag them off the ground manually, one user at a time. Airbnb’s founders famously went door to door photographing listings themselves.

Ours looked like this. When we were closing our first set of customers at Moichor, Shevy would book the cheapest Spirit or Frontier redeye he could find, fly in overnight, visit the clinic, and fly back the same day. No hotel. No lost day of work. Just two bad flights and one conversation that mattered, over and over.

That type of drive is what you need. And it raises the obvious question: how are you supposed to do that if you are living in the hype, building in a space you are not actually into? You cannot fake your way onto that flight. The redeye is downstream of the itch.

Most of building a company looks like this. Unsexy, repetitive, manual work that nobody will ever applaud. If the montage version of being a founder is what attracts you, the actual job will grind you down. If the schlep does not scare you, you might be the right kind of person for this.

5. The Emotional Tax

Here is the part the feed will never show you.

Jensen Huang built Nvidia into one of the most valuable companies in the world. Asked on the Acquired podcast whether he would start it again knowing what he knows now, his answer was no. Building the company turned out to be a million times harder than anyone expected, and if founders understood the pain, vulnerability, and embarrassment in advance, “nobody in their right mind would do it.” He has called a founder’s ignorance a kind of superpower.

The 2am payroll math. The customer who churns the week you thought you had turned the corner. The pitch that goes nowhere. Every one of those problems is ultimately yours, and there is no one above you to escalate to.

That is the tax. Know the number before you sign.

What Makes a Founder Unstoppable

Put all five together and something interesting happens.

People look at great founders and see a set of superpowers. Early employees who leave safe jobs to join them. Pitches that land. A vision they can see so clearly it feels inevitable. From the outside, these look like separate talents. They are not. They are all downstream of the mindset.

Your early employees do not join for the salary or the deck. They join because conviction is impossible to fake. They can tell when the itch is real, when you have already accepted the downside, and when you would get on the redeye yourself. Nobody bets their career on a founder who is playing startup.

You can pitch well for the same reason. A great pitch is not a performance. It is a description of something you have already seen up close. You know the problem better than anyone in the room because the itch made you study it obsessively. When you have flown the redeye and sat across from the customer, you do not need to rehearse conviction.

And you can see the vision clearly because you are not guessing. You have lived inside the problem long enough that the future state is obvious to you before it is obvious to anyone else. Execution is just closing the gap between what you see and what already exists.

None of this comes from the feed. All of it leads back to the founder mindset.

A Quick Recap

The question is not whether you want to be a founder. The title, the feed, the announcement graphic. Plenty of people want that.

The question is whether you want to do founder things:

• Is there a specific itch you cannot stop thinking about?

• Are you honest with yourself about wanting to win, in dollars?

• Have you looked at the real downside, accepted it, and burned the boats anyway?

• Would you get on the redeye?

• Do you know the emotional tax, and are you willing to pay it?

At NPU, this is a big part of what we look for, because we cannot diligence a feed. We look for founders who wanted the job before they wanted the title.

One last thing. There is no shame in reading all of this and realizing the job is not for you. Plenty of brilliant people build great careers without ever founding anything. Knowing that about yourself early is its own win.

But if you read this and felt the itch, we would love to talk.

You close a round, land your first customers, and go heads-down. Building mode. Every hour is spoken for. Coffee with a stranger feels like a luxury you cannot afford. Networking feels like something people do when they are not serious about the work.

So you stop. And for a while, nothing bad happens. That is what makes it dangerous.

Your network does not break loudly. It freezes quietly, at exactly the size it was the day you went heads-down.

The Irony We Lived

Our company only existed because of cold outreach.

As seniors at Penn State, we flew to San Francisco, cold messaged YC founders, and asked to learn from them. Those conversations shaped everything that came after. Some of those people are still our mentors seven years later.

Then we started building Moichor, and we did the exact thing this post is warning you about. Board, employees, customers, vendors. Repeat. From the inside it felt like discipline. Looking back, it was a slow leak. The people we met in year one were still the people we knew in year four. The company had grown. Our world had not.

The Other Side Looks Exactly the Same

The lesson did not fully land until after Moichor, when we crossed the table and started building NPU Ventures.

Raising a fund means building relationships with LPs. And within weeks we noticed something that reframed everything we thought we knew about this industry.

Startups, VC, LP. It is the same business at every altitude. All of it is relationship driven, and all of it runs on more feeling and emotion than anyone admits.

Founders think investors decide on metrics. Investors think LPs decide on track record. Both are only half true. People invest in people they have watched over time. Conviction is built across touchpoints, not in a single meeting. Mark Suster wrote the canonical version of this fifteen years ago in Invest in Lines, Not Dots: a single meeting is a dot, and investors cannot commit to a dot. They commit to the line your dots draw over time. What we did not appreciate until we were on this side is that the same is true one level up, between GPs and LPs.

A VC meets hundreds of founders a year. One meeting, even a great one, fades in weeks. The founders who stay in view are the ones who keep in touch. Not the loudest ones. The consistent ones. Keeping in touch beats the first impression, every time.

When It Clicks, Pay Attention

Inside all those meetings, GP to founder or GP to LP, something happens occasionally that no deck can manufacture. It just clicks. The conversation runs long. You get each other without translating.

Do not treat that as small talk. Treat it as signal. Chemistry is data.

The people who click with you understand what they are actually doing: believing in you and in how big your market can be. Not renting your metrics for a quarter. You see the difference the first time the plan changes. To a spreadsheet investor, a pivot is a red flag. To someone who invested in you, a pivot is not negative. It is a path.

And you cannot detect click over a wire transfer. You detect it across conversations, over time, before there is a deal on the table. One more reason the relationships have to exist before you need them.

You Cannot Sprint a Relationship

Founders treat connections like a task you can batch. Go dark for two years, then blitz a conference when you need to raise.

It does not work, because relationships run on a lag. Trust compounds like revenue: slowly, then all at once, and only if you keep depositing.

When you suddenly need your network, whether for a raise, a key hire, or a customer intro, it is already too late to start building it. You are harvesting whatever you planted two years ago. If you planted nothing, you harvest nothing.

The good news is the bar is lower than you think. Chris Fralic of First Round spent a career studying this, and his most repeated finding in How to Become Insanely Well-Connected is almost embarrassing: simply following up, and doing what you said you would do, already puts you ahead of most people.

We have said before that the founder journey is a marathon. Relationships are the purest example. Fundraising, hiring, and sales are all lagging indicators of how you spent your time when you did not need anything.

Come With Intention, Not Just a Calendar Link

Now the caveat, because this advice gets misused.

Staying connected does not mean spraying meeting requests. Nobody owes you a coffee. A random “can I pick your brain” out of nowhere is not relationship building. It is noise, and people can feel the difference immediately.

You need a reason to chat. Something you shipped. Something you learned. A milestone worth sharing. A question only that person can answer. An intro you can make for them. The reason is your forcing function: if you cannot name why this conversation should happen now, you are not ready to send the message.

And the best reasons come from the thing that should already be number one: your customers. When you are making real progress with customers, the update writes itself. Traction is the most natural excuse to reconnect that exists.

Sam Altman traces most of his best career opportunities, and three of his four best investments, to a decade-long habit of helping people without keeping score.

Come without an ask. Never without a reason.

It Is Up to You to Build It

We used to think connections were something that happened to lucky people with warm networks. Ours started with cold messages from two seniors with no leverage and no track record.

If you had asked me on day one of Moichor whether we would ever have the network we have today, I would have been shocked. It did not exist, and nothing about where we started suggested it ever would.

That is the point. Nobody hands you a network. Nobody builds it for you. Ultimately it is up to you, and you have to start somewhere. The start is almost always small, awkward, and cold. Every network you admire started the same way.

Relationships are not the thing you do after the real work. At every level of this business, they are the real work.

The network you will need in two years is being built, or not built, this week. It is a marathon pace, not a sprint. Small, boring, consistent.

Do not be lazy about it. Be diligent. Come with a reason. Pay attention when it clicks.

Make the connection before you need it.

What founders call a pre-seed has changed, and the data makes it obvious. Among SAFEs that raised at least a million dollars, the average deal reached $1.4M in 2025, up from $1.1M in 2024. Carta puts the median seed fundraising amount in 2024 at $2.5M.

But the more interesting story is the split underneath it. More than a third of pre-seed rounds in the last quarter of 2025 came in under $250K. The market is pulling apart into the lean and the loaded, and the loaded ones are increasingly raising big before they have proven anything.

Founders treat the bigger number as a win. But it carries a cost they do not anticipate.

Before you have early adopters, a bigger round does not lower your risk. It raises the price of being wrong.

A big pre-seed is not dangerous because the money is bad. It is dangerous because it can turn a cheap learning cycle into a multi-year commitment.

Everyone models the upside. Nobody models the downside.

When a founder decides how much to raise, the math is almost always about the good case. What does $3M buy that $300K does not? More engineers, more runway, more go-to-market. Every line of that math quietly assumes the thing works.

The case that actually decides your life is the other one. What does the money cost you if it does not work?

Run that math instead.

Raise $300K. It lasts you twelve months. You stay close to the customer because there is nobody else to do it, you ship, you test the thesis, and at the end of the year, the signal is still not there. That is a painful outcome. It is also a cheap one. You spent twelve months and $300K to learn that the thesis was wrong. You have a working prototype, a real story, and most of your time and savings still intact. You can walk into your next thing, or your next raise, almost immediately. Being wrong costs you a year.

Now raise $3M on the same unproven thesis. It lasts you two, maybe three years. You hire the team that the deck promised. You build the broad version of the product. And the entire time, the signal is lukewarm. Customers are interested but not urgent. Pilots are active but not expanding. Nothing is failing outright, so nothing forces you to stop. At the end of it, you reach the same conclusion you would have reached in the first scenario: this is not working, except now it costs you three years, a team you have to let go, and a cap table priced for an outcome that never came.

Same answer. Wildly different price.

That is the part founders miss. The size of your raise is the size of the bill when you are wrong. And before there is proof, you do not yet know whether you are wrong.

You can see the later-stage version of the same mistake in the companies that raised enormous rounds before demand was proven. Quibi raised $1.75B before launch, projected seven million subscribers in its first year, landed around five hundred thousand, and was gone in six months. Fast reportedly burned through more than $120M of investor money against roughly $600K in revenue. Color Labs raised $41M before its app was ever in users’ hands, then opened to two-star reviews. These are not pre-seed stories. They are what the same mistake looks like with two more zeros. Capital can let a team be wrong for a very long time, very loudly, before the market ever says yes.

Lukewarm is the trap. Not failure.

We made this point in When Promising Becomes Dangerous. The startups that quietly kill founders are not the ones where nothing works. Those are clear. You run out of money, you face the truth, you move on. The dangerous ones are where something is kind of working, and keeps kind of working, for years.

Failure gives you an answer. Lukewarm permits you to keep going.

A big round is what funds the “for years” part.

A small round has a forcing function built into it. The money runs out. That deadline is brutal, but it is honest. It makes you confront whether the signal is real while you still have the time and the energy to do something about it.

A big round removes that forcing function at the exact moment you need it most. You can sustain a lukewarm signal for a long time because you can afford to. You can keep hiring against it, keep building against it, keep telling yourself the next feature or the next segment is the one. The money becomes the reason you stop looking for the truth.

Paul Graham calls the end state the fatal pinch: default, dead, slow growth, and not enough time to fix it. The cruel part is that founders rarely see it coming, because the money still in the bank feels like proof they are fine.

Running out of money forces clarity. A big round lets you buy your way out of clarity, and clarity is the one thing you cannot afford to avoid this early.

The quieter version is more common, and it does not need a famous logo to go wrong. Noogata raised $28M and counted PepsiCo and Colgate as clients, the kind of names that look like proof, then failed to hit its milestones, could not raise again, and wound down. Brand-name pilots are not a fundable business, and capital is very good at hiding the difference. A logo is not the same thing as pull. A pilot is not the same thing as urgency. One founder put the trap precisely: too expensive for new investors, without the traction to justify the next round. That is not a problem money solves. It is one money that creates.

The thing you are actually spending is time.

Money is a recoverable resource. You can raise again. You can take a job and rebuild your savings. Founders treat capital as a scarce thing because it is the thing they are negotiating over. It is not the scarce thing.

The scarce things are your time, your optionality, and the specific window you are building into.

A bigger raise on an untested thesis commits more of all three to a bet you have not validated yet. This is the part that should sit with you. You are the only person at that table underwriting the downside with your life. Your investor wrote a check as one position in a portfolio. That is the model, and it is the right one. They are supposed to diversify. You are not diversified. If the company does not work, they lose one bet. You lose years.

That is not a reason to distrust investors. It is a reason to remember that the person with the most to lose from a bigger round is you, and to size the round like it.

Why being wrong was cheap for us

At Moichor we started with bloodwork for humans. Within about a week and a half of talking to vets, pet owners, and agriculture experts, we had pivoted twice, first to livestock, then to pets. We have told that story before as a story about listening to customers. It is also a story about cost. The reason we could listen was not just that we were open-minded. It was that we had not made the wrong answer expensive yet.

Those pivots were cheap because we had not over-committed capital to the first thesis. We had not hired a team to build the human product. We had not built the broad version of anything. Changing direction cost us a week and a few honest conversations among the three of us. It cost nothing on a cap table, and it cost nobody their job.

If we had raised a few million on the human-bloodwork thesis and staffed up against it first, that same pivot is not a Tuesday. It is a layoff, a write-down, and a very uncomfortable conversation with investors who priced the round on the thing you are now walking away from. Same insight. Same correctness. Completely different price for acting on it.

We raised reasonably at every stage, and what that bought us, every time, was the ability to be wrong quickly and cheaply. Things always took longer than we modeled. Being able to absorb that without a near-death event was worth more than any amount of extra runway.

Identified early adopters change the math

This is the line that matters, because the point is not raise small forever.

Before demand is proven, a big round is a bet on a thesis. You are spending money to find out whether you are right, and the more you spend, the more it costs to find out you were not.

After you have early adopters, real people with an urgent and expensive problem who are pulling the product out of your hands, a big round becomes something else entirely. Now you are not buying discovery. You are buying acceleration of a thing that already works. The same $3M that is reckless before the pull exists is exactly right after it.

So the rule is not a number. It is a ratio. Match the size of your raise to the strength of your proof. When the signal is thin, raise the amount that keeps being wrong cheap. When the pull is undeniable, raise the amount that makes being right fast.

Raise to match signal, not ambition.

We argued the small end of that in Raise Small, Stay Sharp. This is the other half of the same argument. Raising less is not discipline for its own sake. It is refusing to make being wrong any more expensive than it has to be, until you have earned the right to spend more.

Final thought

You cannot control whether your first thesis is right. Most are not.

What you can control is how expensive it becomes to find out.

Before early adopters, raise the number that keeps being wrong survivable. Once the signal is undeniable, raise the number that makes being right fast.

We think most founders are preparing for the wrong thing.

This week made the point better than we could. Apple, the most valuable company in the world, effectively conceded it could not win the model race on its own. It shipped its new AI built on foundation models it developed with Google, and bet its future on the experience layer instead. If leaning on a frontier lab does not doom Apple, it should tell you something about your own odds.

We have heard a version of this question in every season of our journey. Paul Graham made the same point this week. “What if the model companies do this” is just the new “what if Google does this.” He is right. There is a question like this in every era.

Every Era Has Its Boogeyman Question

The form changes. The function does not. In 2000 it was “isn’t this just Pets.com?” A decade later it was “what if Google or Facebook builds this?” After 2019 it was “is this another WeWork?” After FTX it was “is this going to blow up like crypto?” Today it is the labs.

It is the lazy pattern match an investor reaches for when something makes them skeptical and they do not yet know why. It feels like diligence. It is not.

Mine Was Theranos

When Matt and I were raising for our diagnostics company, mine was Theranos. We were building animal diagnostics in 2021, and every few meetings someone would lean back and ask, “so, are you like Theranos?”

I would roll my eyes. Theranos claimed to run more than thirty diagnostics out of a single mysterious box. We were doing one diagnostic. We were applying computer vision to the narrow problem of identifying cells. The only thing the two companies shared was a loose association with blood.

Anyone who had actually read what we were building would never have asked. That was the tell.

The Question Is Usually a Tell

This is the part founders get backwards. The boogeyman question is what someone asks instead of doing the work. The honest answer to defensibility was never a clever line. It lives in the details. How you go to market. The proprietary data you accumulate as you serve customers. The reputation you compound over time.

Someone asking the lazy version has not read your deck. They have not researched your company. Sometimes they have not really been listening on the call. If they had, the question would have answered itself.

So do not contort your pitch into a soundbite built to defeat “why can’t OpenAI just do this?” You will spend your best energy on the people who were never going to dig in. Know your real moat cold, and bring the conversation back to the specifics. The investors worth having will follow you there.

Know When It Is a Fair Question

Now be honest with yourself, because there is a version of this question that is completely fair.

If you are building a frontier lab, competing directly with OpenAI and Anthropic to train the most capable general model, then “what if the labs do this” is not lazy at all. It is the question, and you had better be able to answer it. But that is a different point, and a different company.

Most startups are not on that path. If your edge is a remarkable customer experience and the data you acquire as you actually service customers, the labs are not on your path, and they are not going to be. They are building the broad magic. You are building the specific, compounding thing in front of a customer who trusts you. Theranos was a box that promised everything. We were a narrow tool that did one real thing. That is the same gap that separates a frontier lab from almost every AI startup being asked about one today.

Final Thought

The boogeyman changes every cycle. Pets.com, Google, Theranos, WeWork, FTX, the labs. There will be a new one the moment this one gets boring. The discipline does not change. Build the narrow, real thing only you are positioned to build. Compound the data and the trust you earn serving customers. Answer the fear underneath the question, not the costume it is wearing this year.

And when someone leans back and reaches for this era’s version of that question, let it tell you how closely they were actually listening.

Those are painful, but at least they are clear.

The harder ones are the startups where something is kind of working.

A customer is interested, but not urgent. A pilot is active, but not expanding. Revenue exists, but every dollar requires custom work. Users like the product, but only when the founder is personally involved. Investors are excited, but keep saying, “Let’s stay close.”

None of these are hard no’s.

That is what makes them dangerous.

Promising is not the same as working.

Promising means there may be something there. Working means the signal is getting stronger, clearer, and easier to repeat.

The danger is that promising can keep a founder busy long after the company should have forced clarity.

The trap is partial success

If nothing is working, the answer becomes obvious: you have to change.

But if something is somewhat working, the decision gets harder.

A founder signs a $10K paid pilot with a mid-market customer. It is not life-changing revenue, but it is real. The team celebrates because for the first time, the market feels like it is saying yes.

Three months later, the pilot is still “going well.” The champion likes the product. The weekly calls are positive. But no one owns the budget, the VP has not engaged, and every conversation ends with one more feature request.

The founder keeps building because the customer feels strategic.

Six months in, the company has revenue, but not repeatability. The product is better for one customer and harder to sell to the next ten.

That is when promising becomes dangerous.

Not because the founder is lazy. Usually the opposite. The founder is working constantly.

The problem is that all of that work is going into extending the current version instead of asking whether the current version deserves to continue.

The better question is not: “Is anything working?”

The better question is: “Is this working enough to become the company we are trying to build?”

Big promises raise the bar

Raising big early rounds at high valuations makes this trap worse.

A premium valuation can feel like validation. It tells the founder, the team, and the market that the company is special.

But premium valuations require premium results.

We learned a version of this at Moichor after our seed round closed.

At the time, I thought fundraising was about getting people to believe in the biggest version of the vision. Think big, sell the upside, show where the company could go if everything worked.

That part matters. Investors need to believe the upside is worth the risk.

But once the round closes, everyone starts watching the gap between what you said and what the company is actually doing.

If you raise on a huge promise, normal startup messiness starts to look like underperformance. A slower sales cycle. A delayed launch. A customer that needs more hand-holding. A product wedge that needs to shift. None of these are unusual at seed. But if the story was framed as breakout momentum, every lukewarm signal gets judged against that promise.

The better way is to present the business in two layers:

• Base case: what you believe the company can realistically prove with the capital you are raising

• Optimistic case: what could happen if the strongest signals compound

Then talk openly about the barriers to the base case. What has to go right? What could slow you down? What assumptions are still unproven?

That does not make the pitch weaker. It makes it more credible.

Because after the round, investors are not only watching whether you hit the goal. They are watching how you set the goal in the first place.

This matters a lot at pre-seed. If you raise at a $20M pre-seed, the next round is not judged like a normal pre-seed company. Investors are not just asking whether there is something interesting here. They are asking whether the company has grown into the price.

Investors almost always leave the door open. They will say they like the market, want to track progress, and would love to take another look when you raise the next round.

None of that is bad. It is often genuine.

But it is not conviction.

If you raised at a $20M pre-seed and you do not get to something like $1M ARR, or at least show a very clear path to that level of traction, your next round just became much harder. The investor who was “excited to stay close” may still like you. They may still think the team is strong. They may still believe the market is big.

But the question has changed.

It is no longer: “Could this become something?”

It becomes: “Why has this not broken out yet relative to the price?”

That pressure can make founders defend the original story right when they should be most open to changing it. The market may be saying, “There is something here, but not exactly this.” The valuation may be saying, “Prove the thing you already pitched.”

That tension is dangerous.

Force clarity earlier

The clearest way to separate promising from working is to ask whether learning is compounding.

Is each customer making the next customer easier? Not easy. Easier.

Are objections becoming more predictable? Is onboarding getting lighter? Is the product becoming easier to explain? Are customers pulling you toward the same use case? Is usage becoming less dependent on founder hand-holding?

These are signs that the company is learning its way into a repeatable motion.

The opposite is also revealing. If every customer requires a different explanation, a different product, a different onboarding flow, a different pricing structure, and a different definition of success, the company may still be searching.

That is okay. But founders need to be honest about the difference between search and scale.

Search is when you are still trying to understand the pattern.

Scale is when the pattern is clear enough to push harder.

A lot of startups get in trouble by scaling while they are still searching. They hire too early. They raise too much. They build too broadly. They turn every customer request into roadmap.

The move is not to become pessimistic. The move is to force clarity earlier.

Ask customers:

• What would need to be true for this to become a yes?

• Is this a top-three priority right now?

• What happens if you do nothing?

• Who owns the budget for this?

• Would you pay for this today?

The answers may be uncomfortable, but they are useful.

If the customer has no budget, no urgency, no owner, and no consequence for doing nothing, that is not traction yet. It may be research. It may be a future customer. It may be a useful relationship. But it should not be treated like proof.

The same is true with investors. A thoughtful maybe is still a maybe. A long diligence process with no clear next step is still a maybe. A partner who “really likes what you are building” but will not commit is still a maybe.

A clean no is underrated because it gives the founder time back.

The job is to find out what is real

Every great company starts with weak signals: a few users, a weird customer, a small pilot, a product that barely works but solves one painful problem.

The best founders do not ignore those signals. They study them, but they also test them.

They ask whether interest is turning into urgency. Whether usage is turning into habit. Whether pilots are turning into expansion. Whether conversations are turning into decisions. Whether founder effort is turning into systems. Whether one customer is teaching them how to win the next ten.

That is the difference between promising and dangerous.

Promising is useful when it moves you closer to truth. Promising is dangerous when it becomes the reason you stop looking for one.

At pre-seed, we do not expect everything to be working. The company is still early, the product is still changing, and the market may still be forming.

But we do look for founders who can tell the difference between something that is getting closer to working and something that is simply getting better at consuming their time.

Because in the beginning, the job is not to look like you are making progress.

The job is to find out what is real.

For most of the founders we meet, it is not.

The principle is simple. Raise less than you can. Use the money to learn what works, not to staff what you think might. The specific dollar amount is less important than the discipline of taking only what you need.

The hiring risk nobody priced in

Pre-seed means you have not found product-market fit. You probably have not even found product-market direction. You do not know who your customer is, what they will pay, or what the product needs to do to make them pay it.

Now imagine you just raised a couple million. The deck said “team of six by month nine.” Your investors believe that. Your co-founder is excited. So you start hiring.

Here is what we have actually seen happen. You hire two engineers before you know what to build, so they build the wrong thing. You hire a head of growth before there is a product to grow, so they run experiments on noise. Every wrong hire costs you three to six months of runway and a culture rewrite.

Early hires before product-market fit are the most expensive mistake a pre-seed company can make. They do not just burn cash. They burn focus. And focus is the only real asset a company has before product-market fit.

A smaller round forces discipline. You cannot hire a team. You can barely hire one person. You have to stay close to the customer because there is nobody else to do it.

More capital, higher bar

Every dollar you raise raises the bar you have to clear at the next round. That is the part founders almost never weigh against the upside of the bigger check.

When your post-money lands at the high end of what the market will price you at, you have priced the company to perfection. The investors who came in at that price are underwriting a specific story. The product works. The team executes. The next round happens at a meaningful markup. If any of those three assumptions slips, the next round gets much harder. A flat round looks like a failure. A down round is a near-death event for both the cap table and the team.

The reality at pre-seed is that the first plan rarely survives contact with the customer. You will pivot. The first product will not work. The first GTM motion will stall. Those are normal events at this stage. The question is whether your cap table can absorb them.

More capital also means faster growth expected. A bigger team to justify. A steeper proof point your seed lead will ask you to hit. You are not just raising money. You are raising the bar you have to clear, and you are doing it at the moment you have the least signal about what that bar should be. Ask yourself if the extra capital is worth the higher expectations it locks in.

A smaller round at a reasonable cap has the opposite shape. If the first plan does not work, the next conversation is still easy. You can raise another small SAFE at a similar cap. You can extend with a bridge. You can take the time you need to get to consistent monthly revenue, which is the single best thing you can do to make the company fundable on its own merits.

At Moichor, our post-money was never ridiculous or inflated. We took the price the market gave us at each stage and moved on. What it bought us, every time, was the ability to raise an extension when we needed one. And we needed them. Things always took longer than we modeled. Equipment broke, sales cycles stretched, customers showed up in different segments than we expected. Because our cap was reasonable each round, the next conversation was always available to us. If our pre-seed post had been stretched, those extensions do not happen.

The hidden cost is the fundraise itself

There is one more cost to the bigger round that founders never model into the math. The fundraise itself.

A multi-million round is not a few conversations. It is a process. You have two ways to get to the number. You find a pre-seed lead willing to write a notable size, which is a slow conversation with a handful of funds and a high failure rate at this stage. Or you build the round out of angel and small-check capital, which means landing roughly twenty $100K checks. Either path is months of work, and most founders end up running both in parallel because neither one is moving fast enough on its own.

A $300K to $500K round looks completely different. You can close it in fewer than ten conversations. A lead angel writing $100K, three or four follow-on angels at $25K to $50K, and one or two small funds is enough. Most of those people decide in a single meeting. The whole round can close in three to four weeks if you move with intent.

The math nobody runs is what you traded for the extra capital. Four months of CEO time spent on a fundraise instead of the product is four months your competitor spent on the product instead of a fundraise. The smaller round buys you back the thing you actually need most at this stage. Time.

What 10% month-over-month actually buys you

10% month-over-month is not a vanity metric. It is the threshold at which a business stops being a science experiment and starts being a company. Compound it. $10K MRR becomes $31K in twelve months and $108K in twenty-four months. That is a fundable seed at month twelve and a fundable Series A at month twenty-four, both on your terms.

More importantly, hitting 10% month-over-month tells you, and every future investor, that you solved the only problem that matters at this stage. Customers want this thing. The thing keeps working as you scale it. Every other metric is downstream of that one.

The path to 10% month-over-month almost never goes through a bigger team. It goes through the founders sitting closer to the customer, shipping faster, and saying no to everything that is not growth. Extra capital makes all three of those harder, not easier.

The one key hire

If a smaller round is right, the question becomes what you spend it on.

Mostly runway. But one slot in the budget should be reserved for a single key hire. The person who removes the founders’ biggest bottleneck. For a technical founder selling into B2B, that is usually a founding GTM person. For a non-technical founder, it is usually a founding engineer who can ship without supervision. For a product-led founder, it is sometimes an ops generalist who absorbs everything that is not product.

The criteria are narrow. They have to compound. They have to make the founders twice as productive on whatever the founders are uniquely good at. If a candidate does not do that, do not hire them. Extend the runway instead.

This connects directly to the point we made in Your First Hire Is Not Someone Who Will Sell Your Product. The first hire is not the person who does the thing you do not want to do. It is the person who multiplies the thing you do uniquely well.

When the bigger round actually makes sense

There are real exceptions and they are worth naming so founders can be honest about which bucket they are in.

Capital-intensive businesses. Hardware. Biotech. Deep tech. Anything that physically cannot be built in a garage. Those genuinely need bigger early rounds. So do businesses where speed to share is decisive and the winner is whoever can spend fastest on launch. Regulated industries with long approval cycles sometimes qualify. Second-time operators with the track record to justify a stretched valuation also qualify, because for them the math we walked through above looks completely different.

There is one more case worth calling out. If you can draw a clear line from the bigger round to 10% month-over-month growth, and the path is real and repeatable, raise it. That is a different conversation. What we are pushing back on is the version where the extra capital is there to fund discovery, where the team plan in the deck is a guess at what an organized company looks like rather than a derivative of a motion that already works.

This post is for founders who are just starting out. The product is not yet doing what it needs to do. The sales motion is not yet repeatable. Nothing about the business is on rails. That is the moment when raising less is almost always the better call. The further along you are, the more the math shifts the other way.

The thesis, in one line

Capital is the easiest problem to solve at pre-seed. Focus is the hardest. Every dollar above what you need solves the easy problem and makes the hard one worse.

Raise less than you can. Hire one person. Get to 10% month-over-month. Then go raise the round you actually wanted.

You raised four post-money SAFEs at caps that looked founder-friendly. You thought you sold about 25% of the company. You actually committed 38% of your fully diluted cap table. The bill shows up at your priced round, when the Series Seed lead asks for an option pool refresh and suddenly each founder owns less than the new investor.

We’ve been through this ourselves, and we’ve watched plenty of pre-seed founders walk into the same trap since. It’s the most common cap-table catastrophe at this stage, and the most preventable. Until you price a round, your issued cap table still shows 100% founders. But the day you sit down with a Series Seed or Series A lead, every SAFE you’ve signed already counts at its cap on the model in front of them. The math stays invisible until it isn’t, and the YC post-money SAFE template makes each individual round feel small.

We built a calculator. Link at the bottom of this post. Use it before your next conversation.

A note from us

When I did my own first priced round, after two rounds of SAFEs, before NPU, I was genuinely shocked at the dilution number that came back from the model. Raising on SAFEs is one of the easiest fundraising motions there is. That’s why everyone does it, and that’s exactly why it bites you. The cap on each individual round looks reasonable. The cumulative bill does not.

Two things compound the problem when you finally price your round. Neither one matters the day you sign a SAFE. Both will bite you when you sit down with your seed or Series A lead, and almost nobody mentions them until it’s too late.

Pro rata. Many of your SAFE investors will have pro rata rights via side letter, and the professional ones will assume informal pro rata even without one. At your priced round, every one of them gets to maintain their %, which means the dilution from the new check lands on you, not on them. Track exactly who has pro rata, and what it costs you at conversion, before you sign your next note.

Employee option pools. The option pool is not free equity. Pre-seed founders often treat it like it is. Going from 0% to a 10% post-money pool at your priced round means 10% of the company is coming out of you and your SAFE holders, not your new lead. Build the pool into the dilution model from day one, not as an afterthought.

Neither of these is a problem you fix at pre-seed. They are problems you bake in at pre-seed, and pay for at your first priced round. Run the numbers before you take the next check, not after.

How it happens

Nobody sets out to over-dilute. The path is always the same.

The dangerous part is that each round feels small on its own.

You take $250K from friends and family on a $5M post-money cap (YC standard since 2018; every cap below is post-money too). Tiny round, easy yes. 5%.

An angel comes in for $750K on a $7M cap to anchor the pre-seed. 10.7%.

Three months later you raise $1M from a small fund on a $9M cap. 11.1%.

One last “extension.” $1.25M from a multi-stage fund’s scout program on an $11M cap. 11.4%.

Each conversation looks like a win. The cap goes up each round. You never lower a price. You never give a “preferred” term. You feel disciplined.

Total raised: $3.25M. Total dilution from SAFEs alone: 38.2%.

You haven’t priced a round yet.

Why post-money is the trap

Pre-money SAFEs (the original 2013 version) diluted each other. Each new SAFE pushed prior SAFE holders down with you, the founder. Investors hated it because it created adversarial dynamics with each new check, so YC switched to post-money SAFEs in 2018.

Post-money SAFEs solve the investor-vs-investor problem by making each SAFE holder’s percentage locked and additive. SAFE holder #1 doesn’t get diluted by SAFE holder #4.

The downside, which nobody puts on the cover page: every percent of dilution from new SAFEs comes out of the founders. None of it is shared with prior SAFE holders.

This is the asymmetry founders never see until the cap-table model lands on the table at their Series Seed.

The bill at your priced round

Now you raise: $4M Series Seed at a $16M post-money valuation, with a 10% option pool refresh.

The math, in order: the Series Seed investor takes 25% of the post-money cap table. The option pool refresh takes another 10%, sourced from the pre-money holders. That leaves 65% for everyone who was on the cap table before the round.

Your SAFE holders convert at their locked percentages, collectively 38.2%, and now hold 24.8% of the post-priced company.

The 38.2% is locked at conversion. Once the SAFEs convert, the new investor’s check and the option pool refresh dilute the SAFE block and the founders proportionally, which is how 38.2% becomes 24.8%.

You and your co-founder, who entered the round owning 61.8% of the pre-priced cap table, walk out of the priced round with 40.2% between you. Split it down the middle: about 20% each, which is less than the new investor’s 25%.

This is what the calculator shows when you plug in those numbers.

Default scenario in the SAFE Stack Calculator.

You haven’t shipped a Series A. You haven’t hired a CTO from outside. You haven’t done a single follow-on without pro rata. And you’re already at the equity floor where a Series A board starts asking who’s running the company in five years.

We all want to create unicorns. And maintaining ownership matters because it’s the thing that pulls you through year five or six of building, when the early excitement is long gone and the work is hardest. Ownership is also the one variable you can’t re-earn once you’ve given it up.

Three rules to follow

1. Model the priced round before you sign your second SAFE. The number that matters is not the cap on your next note. It’s your founder ownership the day you sign your Series Seed term sheet. Build the model, or use the calculator, before you take a second check.

2. Cap your cumulative SAFE dilution at 20%, not your individual round. Founders track each round’s dilution in isolation. Smart investors track the cumulative. At 25% cumulative SAFE dilution, the same priced round leaves you below 50% combined; at 20% you walk out with comfortable founder control and cushion for a slightly bigger pool or seed check.

3. The cap is the price. Stop pretending it isn’t. “This is just a SAFE, we’ll figure out price later” is the most expensive sentence at pre-seed. Every SAFE is a priced round in disguise; you just don’t see it priced until conversion. Treat each cap with the seriousness you’d treat a priced valuation, because that is exactly what it is.

Try the calculator

Plug in your real SAFE stack and your projected priced round. If the founder ownership it spits out makes you wince, you have time to fix it. The number you see today is the number your Series A lead will underwrite later.

→ Open the SAFE Stack Calculator

Or reach out if you want us to run the model with you!

The first is a no. That is what people think of when they hear rejection. Someone looks at what you built, what you are asking for, or who you are, and they pass. It stings. But a no is clean. You process it and you move on.

The second is the maybe. The weeks of follow up calls that go nowhere. The investor who keeps asking for one more data point. The customer who says they love it but never signs. That is the one nobody warns you about. And that is the one that quietly does the most damage.

It is like dating someone who responds to your texts every other day. Deep down you know what that means. You just do not want to admit it yet. That is what both forms of rejection have in common as a founder. Whether it is a sharp no or a slow maybe, you have to learn to see them clearly and move.

I have been getting rejected since high school. Cut from the soccer team. Turned down for dates. Told no by customers, investors, people I respected. None of those nos broke me. What nearly broke me was the space between yes and no where you cannot plan, cannot move, and cannot stop hoping.

The No That Rewired Me

At Moichor we were deep in a raise that could have reached fifteen million. One firm in particular felt like the one. Weeks of diligence. Multiple calls, detailed financials, reference checks. They asked every question you could imagine. We answered all of them.

Then they passed.

The reason, which I found out later, was that I did not come across as having enough energy.

Not the product. Not the traction. Not the market. Me.

I could not get over it for a week. If I had simply smiled more, been a little louder, we could have closed a fifteen million dollar round. It felt like one expression on my face was the difference.

But here is what I would tell any founder who hears something like that: those people do not get you. Do not over adjust. Do not become someone you are not.

VCs are investing in you. If who you are is not enough for them, they are not your investor.

The months we spent in their pipeline were painful, but the no itself was what stayed with me. It was personal in a way that product feedback never is.

Fast forward to today. I am raising a fund, and I deal with at least three rejections every single day. Every single one still irks me for a few seconds. And that is okay. That sting is part of the job. The day it stops stinging is the day you have stopped caring, and that is far more dangerous than any no.

I have not forgotten that feedback. I just no longer build around it.

The Maybe That Quietly Bled Us

The VC story was about the no. This one is about the maybe.

Customer rejection is sneakier than investor rejection. Not because it hits harder, but because it never actually arrives. You just slowly realize it is not going to happen.

At Moichor I had this happen more than once. You meet a customer with multiple locations. You run the numbers and realize that one account alone could carry your quarter. Suddenly that customer is not just a deal. They are your entire quarter.

Your forecast assumes they close. Your roadmap shifts to accommodate their requests. Your team talks about them like they are already signed. But they never actually sign. They do not say no either. They just stop responding as quickly. The calls get shorter. The timeline keeps sliding.

The floor drops out. Not because you lost a customer, but because you spent months building around someone who was never in.

Compress the Time to Clarity

The most expensive thing in a startup is not a no. It is the time you spend in limbo waiting for one.

The founders who struggle most are the ones who let a single maybe sit in their pipeline for months, afraid that following up too aggressively will kill the deal. The deal was probably already dead. They just did not know yet.

You are either in or you are out. I will follow up every few days until I get one of those two answers. That is not being pushy. That is respecting both of our time.

If you are talking to five investors and one says maybe, that maybe owns you. If you are talking to fifty, a maybe is just a row in a spreadsheet. Same with customers. Same with hiring. The math only works when you have enough at the top to absorb the losses as you move down.

Every no I have received sharpened something. Getting passed on by that VC pushed me to stop performing and start finding investors who valued how I actually operate. Losing customers at Moichor pushed us to never build a forecast around a single name again.

The nos made me better. The maybes just burned time.

If you are building something right now and you feel stuck, ask yourself: did you get a no, or are you stuck in a maybe? If it is a no, take the lesson and move. If it is a maybe, force the answer.

A no costs you a few seconds of ego. A maybe costs you weeks of momentum. One stings. The other quietly kills your company.

Get your answer.

TLDR: Start in the East Bay if you want lower rent and focus. Move to Potrero Hill or Dogpatch once proximity to founders, investors, and YC matters more.

When Matt and I were seniors at Penn State, we used to fly out to San Francisco and meet as many YC founders as we could. We had no idea what we were doing. We just knew the startup world was out there, and we wanted to see it up close.

We fell in love immediately. Scootering along the Embarcadero, overhearing conversations about companies that were going to change the world, meeting investors for the first time, and feeling like the entire city was building something bigger than itself. San Francisco had an energy I had never felt anywhere else.

So right after graduation, I moved to the Tenderloin.

I had never been there before and had no idea what I was getting myself into. I stayed at the Oasis hostel and quickly learned my first lesson about moving to San Francisco: do your research on neighborhoods.

I would strongly recommend not living in the Tenderloin. Walk through it if you want to understand the city more fully, but it is not a pleasant place to live. I still remember an Uber driver looking at me and saying, “I can’t believe you walk through here.”

After we got our first $100K check, my cofounders and I flew from Pennsylvania into Oakland, found an apartment in a day, and spent six days in a Berkeley Airbnb before moving in. That kicked off a multiyear journey through the Bay Area that taught me more about where founders should live than any guide on the internet ever could.

Here is what I learned.

Start with the East Bay

This might sound counterintuitive. Most people will tell you to move straight to San Francisco. I don’t think that is always the right move.

Our first real home was on Piedmont Avenue in Oakland, and I loved it. It is quiet, walkable, and full of charm. It is close to MacArthur BART, which gets you into San Francisco quickly. You get the calm of the East Bay with easy access to the city whenever you need it.

Rent is cheaper. The food is better than people expect. And it is easier to focus.

Emeryville is another underrated option. Safe, clean, and convenient. The Amtrak station is a huge plus if you are traveling up and down California. Emeryville feels like a small town sitting right next to one of the most important startup ecosystems in the world.

We eventually spent time near Aquatic Park in Berkeley, which was one of the most beautiful areas we lived in. It is right on the water, peaceful, and ideal if you want space to think and build without constant noise.

If your budget is tighter or you want more breathing room, the East Bay is a smart place to start.

When You’re Ready for San Francisco

Eventually, our company grew to the point where being in the city made more sense.

Our first stop was Dogpatch, and I loved it. It used to be an industrial neighborhood, but it has transformed into one of the most interesting parts of San Francisco. There is great food, waterfront access, and a real sense that the area is still being built in real time.

More importantly, Dogpatch is now at the center of the founder ecosystem. Y Combinator moved its headquarters from Mountain View to Pier 70, and founders are increasingly clustering around Dogpatch and Potrero Hill. If you are fundraising, building, or trying to stay close to the pulse of early-stage tech, this is one of the best places you can be.

My favorite neighborhood in San Francisco, though, was Potrero Hill. It is safe, sunny, and feels like a real neighborhood. When you are grinding through long days, the weather and environment have a real effect on your energy. Potrero Hill gives you skyline views, a calmer residential feel, and easy access to the rest of the city without forcing you into the chaos of it.

It is also right next to Dogpatch, so you get the best of both worlds: quiet streets and proximity to YC and the broader founder scene.

If I were moving back to San Francisco as a founder today, Potrero Hill would be at the top of my list.

Other Neighborhoods Worth Considering

Mission District is one of the most vibrant neighborhoods in the city. Great food, strong cultural identity, better weather than most of SF, and good BART access. A lot of younger founders and startup people end up there because the energy is high.

SoMa is the classic tech neighborhood. It puts you close to offices, coworking spaces, and events, but it can feel a little more industrial and transactional than other parts of the city.

Hayes Valley has become a center of gravity for AI. If you are building in AI and want to maximize density with founders, investors, and operators, Hayes Valley is worth considering.

Mission Bay is another strong option if you want something newer, cleaner, and closer to Caltrain. It feels more polished than some older neighborhoods and is appealing if you value convenience.

The NYC Question

After Covid, San Francisco got rough in ways that made a lot of founders question whether they should still be there. In May 2025, I spent time in New York City to see if it could work as a founder base.

New York has incredible talent and momentum, but it also has more distractions. Every block offers a new restaurant, a new event, a new reason to stop working. That can be amazing for lifestyle, but it is not always ideal when you are trying to build something from zero.

What pushes you in San Francisco is that so many people around you are all in. You sit in a café and hear people talking about revenue growth, hiring, fundraising, or product velocity. That kind of environment raises your own standards. The city’s social pressure is not about nightlife. It is about building.

That is the part people underestimate. San Francisco is not just a place where startups happen. It is a place that makes you think bigger.

Practical Tips for Making the Move

The Bottom Line

I have lived in the Tenderloin, Piedmont, Emeryville, Berkeley, Dogpatch, and Potrero Hill. I have spent enough time moving around the Bay to know that where you live shapes how you build.

San Francisco is still the best place in the world to build an early-stage startup. The talent is here. The capital is here. The ambition is here. And the density of serious builders is still hard to replicate anywhere else.

If you are an aspiring founder thinking about making the move, my advice is simple: optimize for focus first, then optimize for proximity. Start where you can afford to think clearly. Then get as close as you can to the people pushing you to build faster.

And whatever you do, do not move to the Tenderloin.

Shevy is a Co-Founder and General Partner at NPU Ventures, a $5M pre-seed fund backing AI and national security founders.

TLDR: Most startup decisions are reversible. The ones that aren’t almost always come down to people. Here are three that stand above the rest: your co-founders, your investors, and your early hires. Think about the worst case before making these calls. If you can accept the consequences, move forward with confidence. If you can’t, slow down. And for everything else, stop treating it like it’s permanent.

Most startup decisions are reversible. You can pivot the product, change your pricing, scrap a campaign, redo your deck. The cost is time, but you recover.

The decisions that actually shape a startup’s trajectory are different. They involve people. In our experience, three people decisions stand above everything else in terms of how difficult they are to undo: the co-founders you choose to build with, the investors you bring on, and the early hires you make.

Hire slow, fire fast. It is a cliché because it is true. That does not mean you should be paralyzed. But I have always liked to think about worst-case scenarios before making these calls. Not to be pessimistic, but because if I can accept the consequences of the worst outcome, I can move forward with confidence. If I cannot accept them, that tells me I should not move forward yet.

The Co-Founders You Choose to Build With

The decision of who to build with is the single most irreversible choice in a startup. You can fire an employee. You can replace a vendor. But separating from a co-founder is surgery. It touches equity, IP, institutional knowledge, team morale, and investor confidence. Even when it goes well, it never goes cleanly.

Pick people you truly enjoy building with. At Moichor, Matt, Thanh, and I spent literally twenty-four seven with each other for over three years. We worked out of our Oakland apartment, played Overwatch at night, woke up, and did the same thing again the next day. That was our life. If we did not genuinely enjoy being around each other, it would have been unbearable.

Our relationship only strengthened with time. At first we were shy about talking about the hard stuff. Eventually we stopped holding back. We argued, got loud, patched things up, and moved on. That cycle became how we operated. Over time we became a package, and in the beginning that was viewed as a strength. Three founders who were completely aligned, inseparable, and deeply trusted each other. Later, as the company grew and more experienced leadership was wanted at Moichor, being a package was sometimes seen as a weakness. But looking back, that bond is what carried us through the hardest years of the company. It outlasted every other advantage we had.

If you are starting a company with someone you met recently, be honest about what you do not know about them yet. How do they handle bad news? Do they avoid conflict or lean into it? When things go wrong, do they blame or take ownership? You learn these things through shared difficulty, not conversation. If you have not been through something hard together, stress-test the relationship before you sign papers.

The Investors You Bring On

Of the three, investors are the hardest to reverse. The moment someone writes a check, they get shares or some right to your company that is essentially impossible to erase. A co-founder can be bought out. An employee can be let go. But an investor is on your cap table for the life of the company.

Do not take easy money. The excitement of someone wanting to invest can cloud your judgment about whether they are the right partner for a decade-long relationship. Capital is replaceable. The wrong investor is not. One of the most revealing things you can do is talk to founders in their portfolio whose companies did not work out. That is where you see an investor’s true colors. Did they disappear when things got tough, or did they show up with real help?

Once you bring investors on, do not over-promise. This is one of the most common traps, and one we fell into ourselves. You want to impress them. So you set expectations high. But if you put expectations at the ceiling on day one and then under-deliver, you create a trust problem that is very hard to fix. Set expectations you can actually meet. If you beat them, wonderful. But do not put yourself in a position where you are constantly explaining why reality does not match what you told them.

Over-communicate during the rough times. When things are hard, the instinct is to go quiet and try to fix the problem before anyone notices. That is a mistake. Your investors are managing many companies at once. Whether for worse or better, they are not in the weeds with you. They miss key details. This happened to us multiple times, situations where our investors did not have the full picture because we had not shared it proactively. By the time we did communicate, the problem had compounded and the conversation was harder than it needed to be. Silence does not signal strength. It creates surprise, and surprise destroys trust faster than bad news ever does.

The Early Hires You Bring On

Your first few hires are not just filling roles. They are setting the culture, the pace, and the standard for everyone who comes after them. Every early hire carries disproportionate weight because they become the template. How they work becomes how the company works. What they tolerate becomes what the company tolerates.

This is why hire slow, fire fast matters so much at this stage. A bad early hire does not just underperform. They set the wrong precedent. They create dynamics that the next five hires inherit. Before you extend an offer, ask yourself: if this person does not work out, can I live with the consequences?

The best early hires at Moichor, and the ones who stuck with us the longest, did not come from recruiters. They came from our own networks, our own instincts, and our own judgment about who would thrive in the chaos of an early-stage company. VCs often feel comfortable recommending recruiters because they have established relationships with them. And recruiters can be valuable later. But at the earliest stage, trust yourself more than you trust VCs and recruiters when it comes to hiring. They are not the ones who have to live with the consequences every day. You are.

Do not rush to fill a seat just because it feels like you should be growing. Before hiring, ask yourself whether you are solving a real bottleneck or just trying to feel like you are making progress. Leverage beats headcount, especially early on.

The Common Thread

Co-founders, investors, and early hires. All three are people decisions. All three are hard to undo. And all three shape the company in ways that persist long after the original decision stops feeling like a decision at all.

Think about the worst case. If you can accept it, move forward with confidence. If you cannot, slow down and figure out why. The founders who last are not the ones who avoid mistakes. They are the ones who give the irreversible decisions the weight they deserve, and save their speed for everything else.

The Flip Side

Most of the things founders lose sleep over are not actually irreversible. Pricing can be changed. Products can be rebuilt. A missed deadline is not a death sentence. If your product does not look great today, you ship a better version tomorrow. Most doors can be walked back through and fixed.

Do not treat every decision like it is permanent. Save that weight for the ones that actually are.

And those, as we have learned, almost always come down to people.

TLDR: The founders who get the most out of being scrappy are not the ones who stay scrappy the longest. They are the ones who know when each scrappy decision has done its job.

We have been thinking a lot lately about how we ran Moichor in the early days. What worked, what did not, and what we would tell ourselves if we could go back. A lot of it comes down to one thing: being scrappy with intention.

A question we hear often from founders is some version of “don’t we need this to do that?” They assume they need a certain tool, a certain hire, a certain budget line before they can move forward. Almost every time, there is a way to get it done differently. Cheaper. Faster. With what you already have.

Scrappy is a personality. Some founders have it and some do not. The ones who do find a way through problems that look like walls to everyone else. But scrappy without intention is just chaos with good stories. The real skill is knowing which corners to cut, when to stop cutting them, and how to get better at that judgment every time you get it wrong.

Here is what that looked like for us.

Getting in Front of People Without Spending a Fortune

Conferences in our industry were the obvious place to find the right people, but attending was expensive. Flights, multi-day passes, hotels. For a bootstrapped team, the math did not work.

So we showed up anyway. Not to the show floor. Just to the area around it.

We would reach out ahead of time, let people know we were going to be nearby, and schedule meetings on the side. Coffee in the hotel lobby. A quick chat between sessions. Most people were happy to talk. They were already there, already in the mindset of meeting new companies, and they did not care whether we had a badge or not. The total cost was a plane ticket and a few cups of coffee.

We did the same thing with vet clinics. We would show up with a box of donuts. It was cheap, it gave us a reason to walk in the door, and it made us memorable in a way that a cold email never could. The vet techs and veterinarians loved it. We have since seen other founders take the same approach, slapping their logo on a donut box and spending a day going door to door talking to potential users.

The insight is not “bring donuts.” It is that we made a deliberate call to test whether in-person attention was worth more than a polished campaign. Once we knew it was, we knew exactly what to invest in next.

From meeting attendees in a hotel lobby with no badge or booth, to platinum sponsorship at the same conference. Same event. Very different starting point.

Doing the Work Yourself

At Moichor, we needed thousands of images cropped for our cell labeling work. We could have paid for a tool. Instead, the three of us sat down, set a timer for four hours in intervals, and hand-cropped images one by one. One person would finish their shift and go to sleep. The next person would wake up, sit down, and keep going. Then the next. We rotated like that until it was done.

It was tedious. But it was free, and at that volume it was the right call.

The same thinking applied to how we lived. All three of us stayed in the same apartment. The office was the kitchen and the living room. It sounds like every startup story you have heard before, but it was one of the best moves we made. It consolidated our biggest personal expense into one place, and more importantly, it made us faster. You wake up, walk a few steps, and you are already working. The iteration speed that comes from that proximity is hard to replicate any other way. It also made us closer as a team in ways that carried through the entire life of the company.

Neither of those decisions happened by accident. We knew exactly what we were trading. That is what separates scrappy from just being broke.

Our apartment kitchen was the first lab. We were not trying to build something permanent. We were trying to prove the workflow worked. Once it did, the real lab followed.

When Scrappy Starts Working Against You

All of the above worked because we were intentional about it. Where things got harder was when we held on too long.

When we were scaling our lab operations, we needed equipment. Real lab equipment is expensive. So we did what any scrappy team would do. We borrowed what we could, leaned on connections, bought used gear from startups that were liquidating, and spent weekends at lab auctions picking up things for a fraction of retail.

It worked. We got our operation off the ground for way less than it should have cost. Every dollar we did not spend on a shiny new machine was a dollar we could put toward learning and iterating.

But used equipment is a coin flip. Things broke. Repairs took longer than expected. Some of the gear we bought at a discount ended up costing more in maintenance and downtime than buying new would have. We found ourselves negotiating from a weak position because we were not buying direct, and the time we spent troubleshooting was time we were not spending on the product.

The original decision was right. The timing of when we stopped was the mistake. Scrappy decisions that make sense on day one can quietly become the thing that slows you down on day ninety.

Scrappy Proves the Bet. Then You Fund It.

There is a pattern here worth naming. Almost every scrappy decision we made was really a cheap experiment. We were testing whether something mattered before spending real money on it.

T-shirts are a perfect example. We knew branded shirts could be a great marketing tool, but we had no idea if they would actually move the needle for us. Ordering a few hundred custom shirts from a vendor would have cost thousands. So we made our own. We printed them ourselves, kept the quality good enough, and started handing them out. People wore them. They talked about us. It worked. Once we had proof that shirts were driving awareness and conversations, spending real money on a proper run made sense. But we never would have known that without the scrappy version first.

We did the same thing with lab equipment. Before we had a lab, before we even knew if our workflow would hold up at any real volume, we bought used equipment on the cheap and set it up ourselves. The goal was not to build a lab. The goal was to prove that the workflow worked, that the science translated from theory to practice, and that we could actually process samples in a repeatable way. Once we had that proof, investing in a proper lab setup was an obvious decision.

Scrappy is not the destination. It is the experiment. You use it to build conviction, and once you have conviction, you spend the money. Holding on past that point is not discipline. It is just friction.

The Feedback Loop

This is the part we wish we had understood earlier.

If you find yourself going back to fix something scrappy, that is a signal. It means you held onto it too long. Maybe the equipment should have been replaced two months ago. Maybe the manual workflow made sense when it was just the three of you, but now there are ten people relying on it and it is breaking every week. Maybe your customer base outgrew the duct-tape version and you did not notice until things started falling apart.

Every time that happens, it should change how you approach the next scrappy decision. Not to avoid being scrappy, but to get better at identifying the moment when a particular piece of scrappiness has done its job and needs to be replaced with something real. The first time, you might hold on too long. The second time, you catch it faster. Eventually, you start building with a replacement timeline already in mind.

Scrappy is a genuine advantage. But the founders who get the most out of it use it to move fast, build conviction, and then make the call to replace it before it becomes the thing holding them back. Every scrappy decision becomes a data point for the next one. That is how the judgment gets sharper. And the judgment is the whole game.​​​​​​​​​​​​​​​​

It makes sense on the surface. You built the product. You love the product. Talking to customers, pitching, closing deals — that is the part you like least. So you want to hand it off to someone whose job it is to do exactly that.

This is almost always a mistake.

The Founder Has to Sell First

Technical founders tend to want to outsource the component they are least comfortable with, and that is typically talking to customers or selling to them. The instinct is understandable. But the problem is that a sales rep is not a founder. They will not figure out how to sell your product from nothing. That is your job.

A sales rep needs a predictable sales motion. They need to know who the buyer is, what the objections sound like, how long the cycle takes, and what the conversion path looks like. They need training materials, case studies, competitive positioning, and a CRM pipeline that actually reflects reality. Without those things, even a talented rep will struggle.

A founder operates differently. You will get on a call with no script, fumble through the pitch, learn something from the rejection, adjust, and try again the next day. You will figure out the channel through sheer willpower and proximity to the problem. A sales rep will not do that. They should not have to.

It is on the founder to figure out the early sales channel. Once you understand how deals actually close, then you hire a rep to accelerate what already works. Not to discover it for you.

Your First Ten Customers Will Break Everything

There is another reason the founder needs to own early sales. When you are signing your first ten customers, you are not running a stable business. You are pivoting. You are learning. The product is changing week to week. The pitch is changing call to call. What you told customer three might not even be true by the time you onboard customer seven.

That level of chaos is normal at this stage. It is how you find product-market fit. But it is deeply stressful for a sales rep who came from a big company where the product was stable, the territory was defined, and the role did not change every quarter.

A rep from that environment expects a playbook. They expect consistency. When they join and the ground is shifting under them constantly, they either freeze up or fall back on habits that do not apply. Neither outcome helps you.

The early stage demands someone who can absorb ambiguity and keep moving. That person is you. Once the product and the sales motion stabilize enough that a new person can follow a repeatable path, that is when a rep can thrive. Not before.

Sales Reps Are Not What You Think

Here is something most first-time founders do not realize about sales reps: they pay extremely close attention to their commission plan. A great sales rep at their current company is making a lot of money. Their comp plan is dialed in. Their territory is producing. Their pipeline is warm. Leaving all of that to join an early-stage startup with an unproven product is a massive risk for them.

That means if you want to recruit a strong rep, you need to go on full outreach mode. The best reps are not applying to your job posting. You need to find them, pitch them, and convince them that the opportunity is worth the trade-off. It is a sales process in itself.

And here is the part that trips up a lot of founders: do not trust what sales reps tell you at face value. I say this with respect for the profession, but reps are trained to make you feel good. They are trained to build confidence and belief. That is literally their skill set. When you are interviewing them, they are selling you.

Always ask to see their numbers. Where did they rank in their territory? What was their quota attainment? How did they perform relative to the rest of the team? The data tells you what the conversation cannot.

Hire the Generalist Before the Closer

Before we ever hired a sales rep, we hired a generalist marketer. That decision turned out to be one of the smartest early hires we made.

The marketer built out all the materials the sales motion would eventually need. The one-pagers. The case studies. The email sequences. The CRM pipelines. The tracking systems. All of the infrastructure that a sales rep would later rely on to do their job effectively.

By the time we brought on a rep, the foundation was already in place. They did not have to build their own tools. They could focus on selling.

This is the order that works: the founder figures out the channel. A generalist builds the systems and materials around it. Then a sales rep steps into a role that is designed for them to succeed.

If You Cannot Track It, You Cannot Manage It

When you do hire a sales rep, you need tracking and systems in place from day one. How many calls did they make this week? How many demos did they book? How many follow-ups went out? What does the pipeline look like at each stage?

Without this visibility, you have no way to know whether the rep is performing or not. You end up relying on their self-reporting, which brings you back to the same problem — reps are optimistic by nature. They will tell you things are going well even when the numbers do not support it.

Set up the CRM before the hire starts. Define the metrics you care about. Make reporting a non-negotiable part of the role from the beginning. This is not about micromanaging. It is about creating the conditions where both you and the rep can see reality clearly and make decisions based on it.

The Uncomfortable Truth

The uncomfortable truth for technical founders is that selling is your job for longer than you want it to be. You cannot skip it. You cannot delegate it before you understand it. The first version of your sales motion has to come from you, because nobody else has the context, the conviction, or the willingness to sell something that barely exists yet.

Once you have done that work — once you know the channel, the buyer, the objections, and the close — then hiring a rep becomes one of the highest-leverage moves you can make. But not before.

Build the system first. Then bring in the person.

It does not have to be as lonely as we make it.

One of my biggest regrets from building my first company was not leveraging mentors earlier.

As a first time founder, I felt an unspoken pressure that everything had to be perfect. That I was supposed to have the answers. That struggling quietly was part of the job. I waited too long to communicate problems to the people I trusted most.

Silence does not signal strength. It compounds mistakes.

In hindsight, many of the decisions that cost us dearly could have been avoided, or at least softened, by sharing issues earlier. Problems do not get smaller with time. They just get heavier.

This is one of the clearest differences I see between first time and second time founders. Second time founders share earlier. They understand that no journey is linear and that every company breaks in unexpected ways. Asking for help is not a weakness. It is pattern recognition.

Years before I understood this, when we were seniors at Penn State, we did something that felt slightly crazy at the time. We flew to San Francisco, cold messaged founders from YC companies, and asked if we could learn from their experiences. We had no leverage and no track record. Just curiosity and respect for what they had built.

To our surprise, many said yes.

Those early conversations were invaluable. Not because they gave us perfect answers, but because they gave us perspective. Along the way, we met mentors who, seven years later, still shape how we think, how we operate, and how we show up as founders.

Many of the mentors we had while building Moichor are the same mentors we rely on today. When I transitioned away from Moichor, every person I consider a true mentor stayed by our side. That continuity matters.

These relationships did not come from warm introductions or carefully constructed networks. They started with a cold email and a conversation over coffee. Over time, those conversations turned into trusted relationships that genuinely changed the trajectory of our lives.

Better, earlier, I would have optimized for conversations, not appearances.

The right mentor is not just someone you go to for advice. At their best, they become someone you can call a friend. Someone who checks in when you least expect it and shows up without keeping score.

Great mentors can hear the worst version of the problem. The messy, unfinished, uncomfortable truth. They still help lift your perspective when confidence dips.

It matters who you surround yourself with. Choose people who genuinely have your back, regardless of which path you take. You do not need another person enforcing their opinions or dictating every decision. That is what your board is for.

True mentorship shows up at the forks in the road, even when the mentor would not personally make the same choice. The role is not control. It is conviction in you.

As much as you invest in your product, invest in your mentors with the same intention and care.

Just as important as finding great mentors is knowing which ones to avoid.

Mentors are not there to run parts of your business for you. They are not responsible for getting your first sale or fundraising on your behalf. You are the founder. You are doing the work. Time and time again, founders get excited about advisors who promise to do X, Y, and Z, and it rarely plays out that way.

The best mentors will not ask for anything until you offer. They do not lead with equity, titles, or expectations. They lead with time, honesty, and genuine care. The relationship compounds because you want to give back, not because you feel pressured to.

Good mentors also remind you to stay grounded in reality. They will push you to reference your customers constantly. Not the pitch deck version of the story, but the lived one. What customers say. What they pay for. What they complain about. When advice drifts too far from customer truth, it stops being useful.

Predatory mentorship is real. Some people with credibility ask for outsized equity in exchange for advice, and founders feel obligated because of the name attached. Most investors and customers do not care who your advisor is. They care about what you are building and how well you execute.

Big names without time are another trap. It may look impressive on paper, but what matters far more is whether someone actually shows up. Consistent care will always beat occasional prestige.

The strongest mentors are often people who have recently done what you are trying to do, or who are deep in the process right now. Their advice is practical, current, and grounded. Mentors change over time, and that is expected. Different stages require different perspectives.

Do not forget the early mentors who took your call, answered naive questions, and helped when you had very little to offer in return, especially when you reach later stages.

Better, earlier, send the text.
Schedule the coffee.
Make the call before the problem feels urgent.

Most intro calls are thoughtful. The ideas are interesting. The markets are often large. Some founders also bring an infectious level of passion and focus. On paper, many of these companies could work.

What we are usually trying to understand in that first conversation is whether a team is building in a way we believe leads to something durable. These are not patterns we learned from evaluating hundreds of companies. They are things we lived through and now have conviction around.

At the pre-seed stage, we are not expecting everything to be figured out. What gets us excited is seeing early signs that execution is already taking shape. Over time, a few things have stood out. These are what we believe separate teams that ship from teams that stall.

1. Signs of Pacing and Execution

We believe progress matters more than polish.

Many founders can articulate a strong long-term vision. They can explain where the product might go and how different pieces could eventually come together. That matters. What we are usually listening for, though, is what is happening right now.

Especially early on, progress often shows up in small ways. Shipping something rough. Running an experiment. Talking to a handful of customers. Changing direction based on new information.

In our previous startup, we started with bloodwork for humans. Within a week and a half of constantly talking to vets, pet owners, and agriculture experts, we had pivoted twice, first to livestock, then to pets. Each conversation clarified who actually had the problem we could solve and who would pay for it. That willingness to kill assumptions fast is what eventually got us to product-market fit.

Raising money always involves telling a story. But real leverage comes from grounding that story in execution. Tangible progress makes the vision more believable.

2. Real Customer Conversations

We spend a lot of time listening for evidence of direct customer engagement.

We speak with many early teams and often hear founders describe who they believe their customer is. Sometimes that belief is well grounded. Other times, it is still mostly intuitive.

Founders who have spent time talking to users or buyers usually describe the problem differently. Their answers are more specific. They remember objections. They talk about real workflows and real constraints. They often use the same language their customers use.

At this stage, those conversations do not need to be perfect or exhaustive. They just need to be real. Even a small number of direct conversations tends to clarify things in ways that desk research never does.

3. A Thoughtful Starting Point With Real Pain

Market size matters to us, but so does where a founder chooses to begin and whether that first group actually feels the problem acutely enough to pay for it.

We are often drawn to teams that have a clear idea of who they are starting with and why. That first group might be small, but it is intentional. There is usually a reason those users feel the problem more urgently or are easier to reach early. What matters is that the initial wedge is grounded in reality and the pain is expensive to ignore.

Early feedback can be encouraging, but interest alone does not always translate into a business. We try to understand whether the problem is one people are actively trying to solve today, or something that feels nice to have addressed eventually. Founders do not need pricing nailed down at this stage. But having some sense of who would pay, why they would pay, and what happens if the problem goes unsolved often signals that the pain is real.

In our previous startup, it became clear early on who needed the product most. That clarity made everything else easier, from product decisions to early sales. The difference between curiosity and urgency shows up pretty clearly once you start digging. Large outcomes often begin with a focused starting point where the pain is undeniable.

4. How Decisions Get Made With Limited Information

Finally, we pay attention to how founders talk about decisions.

Early-stage companies operate with incomplete data almost by definition. What tends to stand out is whether founders can explain why they made certain choices, what they deprioritized, and what they are still unsure about.

We are less focused on whether every decision was right and more interested in whether decisions were deliberate. In our previous startup, we chose to pause our hardware production even though it was technically impressive and represented months of work. The hardware was expensive to develop and would not get us to our next milestone within our budget. More importantly, when we watched customers actually use the product, we realized they did not care about the cutting-edge tech at all. It was not solving their pain point. Knowing what not to build, even when you have already invested in building it, mattered as much as knowing what to build.

Over time, that kind of intentionality tends to correlate with stronger execution.

At this stage, there are a lot of things we have learned not to focus on too heavily. A perfectly polished deck. A long-term roadmap that assumes everything goes right. Early metrics that feel precise but rest on thin data. Those things become important over time. Early on, we care more about how founders navigate ambiguity and whether real progress is showing up between conversations.

We are building our portfolio around these beliefs. Not because we have proven they predict success, but because we have learned firsthand what it takes to build something that lasts. If you are building this way, we would love to talk.

What do you think early-stage founders should focus on most? We’d love to hear your perspective. Reply to this email or reach out if you’re building something and want to share what you’re learning.

Most founders hear that early, nod, and move on. It sounds obvious. But it doesn’t really land until you’re deep into the journey. Momentum slows. Uncertainty compounds. The cost of just pushing harder starts to show up in ways you didn’t expect.

What no one says clearly enough is this.

For most founders, the biggest enemy is not competition, capital, or market timing. It is yourself.

The Illusion That It Gets Easier

Early on, many founders believe the journey will eventually smooth out.

Once we hire this person, things will calm down.

Once we roll out this software, operations will finally be clean.

Once we close this round, the pressure will ease.

But it rarely works that way.

Every stage of a company solves a set of problems and immediately replaces them with new ones. Often more complex, more novel, and with higher stakes.

Headcount creates coordination issues.

Systems introduce new points of failure.

Growth brings ambiguity instead of clarity.

The work does not disappear. It changes.

Founders who wait for a future moment when things settle down often end up postponing their own well-being indefinitely.

What I Learned the Hard Way at Moichor

During my time building Moichor, especially toward the end, I did not have the same energy I started with.

From the outside, things looked fine. We were executing. The company was growing. But internally, my body was sending signals I chose to ignore.

I let others control nearly every aspect of my life. The board. Employees. Customers. There was very little left that I personally enjoyed.

That showed up in obvious ways. Terrible eating habits. Fast food six days a week. Constant traveling. Very little recovery.

Over six years, I gained more than 80 pounds and dealt with ongoing stomach problems. None of this happened overnight. It happened quietly. Stress stacking on top of stress, all justified by the belief that relief was always one milestone away.

I was not being outworked by competitors.

I was not failing strategically.

I was waiting for things to get easier while slowly burning myself down.

As founders, we usually love what we are building. But at some point, we need a mental break. And that is okay.

Scaling Changes the Pressure

As companies scale, the founder role shifts.

You do less and decide more.

Those decisions carry more weight. The margin for error narrows. The emotional load increases even when the company looks more stable from the outside.

If you do not adapt how you take care of yourself, the pressure shows up somewhere. Health. Relationships. Judgment. Usually all three.

You cannot out-hustle a marathon.

The Importance of Celebrating

One thing I wish we did more of along the journey was celebrate.

There’s a Japanese phrase I love from the movie Perfect Days:

Ima wa ima, kondo wa kondo.

Now is now. Next time is next time.

It’s a reminder to be present. To acknowledge progress instead of constantly moving the goalposts.

Founders are particularly bad at this. Wins get discounted because they don’t feel big enough. We tell ourselves we’ll celebrate later.

But progress is built from small, compounding victories.

Making it through a difficult week.

Surviving a tough quarter.

Growing ten percent month over month.

Shipping something that finally works.

Those moments matter. Not just for morale, but for longevity. If you never let yourself feel progress, the journey starts to feel endless even when you’re moving forward.

One thing I also regret is not documenting more of the journey. Taking photos. Recording short videos. Capturing moments before rushing on to the next milestone.

Looking back, you don’t want just the outcome. You want to see how far you’ve come.

Do Not Fall Into the Comparison Trap

Being a venture-backed founder comes with constant comparison.

Whenever we went to founder events, the same two questions always came up: How much money have you raised, and how many people work at your company? For a long time, we thought those were the questions that defined the best startups. More capital meant more success. Bigger teams meant more progress.

You see peers raising large rounds, chasing top-tier valuations, and posting rapid growth metrics. Some of that is healthy ambition. Some of it is competitive instinct. A lot of it creates unnecessary pressure.

Not all companies are meant to grow at the same pace. An animal diagnostics company will not scale the same way as a B2B SaaS platform selling to startups. Different customers. Different sales cycles. Different constraints.

Yet founders often hold themselves to benchmarks that were never designed for their business. Comparison does not make you sharper. It makes you impatient. And impatience leads to decisions that optimize for optics instead of fundamentals.

Over time, we realized that some of the best companies are not the ones that raise the most money. They are the ones that raise the least and still find a way to build something meaningful. They do more with fewer resources. They stay disciplined longer. They are forced to focus on what actually matters.

Every founder is on a different journey, even if the label says venture-backed.

A Better Founder Narrative

The startup world still celebrates endurance through suffering. That story has produced incredible companies. It has also produced a lot of unnecessary damage.

There is a better way.

One where founders accept that the journey does not get easier. It just changes.

One where sustainability is practiced throughout, not earned at the end.

One where ambition does not require self-destruction.

One where progress is measured against your own path, not someone else’s highlight reel.

Startups are a marathon.

And the companies that matter most are built by founders who are healthy enough and grounded enough to run their own race.

A few patterns kept coming up. Nothing groundbreaking, just small things that tripped up teams, including us.

Here’s what we noticed.

Culture Shows Up Fast

The shift from founders-only to your first hire changes things more than you’d expect.

Suddenly everything becomes visible. How you make decisions, handle conflict, structure the week: it all turns into “how we do things here” whether you planned it or not.

Early employees are joining your way of working as much as your product. They’ll immediately pick up on founder dynamics. If roles aren’t clear between cofounders, you get crossed wires and wasted effort. We watched teams spend months redoing work because nobody clearly owned decisions.

One thing that helped us: making sure founders had comparable equity and identical vesting. Not because it’s “fair,” but because it signals everyone’s committed for the same timeframe.

Founders Should Own Sales Early

At least one of you needs to be in sales conversations, at least initially.

Not forever, but long enough to watch your assumptions get corrected by reality.

We learned this the hard way. We built a microfluidic chip because it was technically elegant. Customers didn’t care: it was expensive and slow to iterate. Then a device. Still too complex. Then a test. Better, but not sustainable without volume. Eventually we ended up running a full reference lab.

Each pivot took months and real money. The common thread? Customers never cared about what impressed us technically.

Those sales conversations shaped our product more than any internal roadmap for two years before we hired anyone.

Worth running multiple experiments at once rather than validating one path completely before testing another. Speed of learning beats perfect execution.

You can hire salespeople later. Can’t really outsource the learning part.

Financial Basics Prevent Weird Problems

Most quiet failures start with messy fundamentals.

Small example: we delayed setting up proper payroll when it was just us. Created more cleanup work than we saved.

Same logic applies as you grow. HR tools and benefits platforms all look fine in demos. The difference shows up when something breaks and you need actual support, not a chatbot.

Edge cases matter. Being able to reach a human quickly often beats saving a few bucks monthly.

If you’re hiring across states, a PEO can remove real headaches. Benefits choices compound too. Premiums rise, employees anchor on what they get early, switching later disrupts people.

Also: every founder should know the burn exactly, not roughly. Keeps everyone accountable and prevents silos.

Spend Deliberately

Free tools and startup credits are helpful, but they accumulate.

We found it useful to keep a simple list of what we’re using and why. Makes it easier to spot what actually matters versus what just stuck around.

Building relationships with key vendors early helped us when we needed flexibility later.

Legal spend: most early docs are solved problems. NDAs, advisor agreements, basic option grants don’t need expensive firms. Deep tech or heavily regulated spaces are different, but don’t confuse ambition with complexity.

On outsourcing: works best after you understand the work yourself. We spent months and real money having experts build that microfluidic chip, only to learn customers didn’t value what it offered. Hard to know what matters if you haven’t done it yourself first.

Use outsourcing to scale what you already understand, not to skip the learning.

That’s It

These are the things we keep coming back to: culture materializes fast, customer conversations teach you what matters, financial basics prevent surprises, and intentional spending beats reacting.

Easy to miss when you’re moving fast. They mattered more than we expected as things grew.

If any of this is useful or you want to talk through your specific situation, happy to chat.

This reflects general observations and isn’t legal, tax, or investment advice.

For this edition, we are featuring Luke Boyer, a software engineer at Google. He is a core developer of LiteRT, the industry leading framework for on-device inference. Luke works at the intersection of machine learning, resource-constrained compute, and over-the-air software distribution.

If you would like to connect with Luke or continue the conversation on on-device ML, you can reach him on LinkedIn.

What is on-device ML?

When most people think about AI, they picture a request going to a big model in a data center and a response coming back over the network.

On-device ML flips that picture, wherein inference is executed directly on the edge, as close as possible to end-user devices or physical sensors.

Luke describes this as the foundation of personalized, “real-life AI”, and a future of systems capable of adapting to their users and interfacing directly with the physical world. Think “a phone without a screen”, what Siri wants to be, or the first steps toward a real-world Jarvis.

Why does it need to be on-device?

Three forces keep pushing intelligence onto devices instead of leaving it in the cloud:

• Latency: Real-time use cases cannot tolerate unpredictable network lag. Braking a car, tracking fast motion, or translating live speech does not work if you are waiting on a round-trip to the server.

• Privacy and personalization: The most powerful agents see calendars, documents, video feeds, and sensor data. In many consumer, industrial, and military settings, streaming all of that to a remote server is a non-starter.

• Cost and reliability: Server compute is expensive at scale. Running optimized models on-device shifts cost into hardware you pay for once and keeps systems working when the network is slow, jammed, or offline.

How NPU thinks about ODML

At NPU, we care about on-device ML because it sits right at the intersection of our thesis: AI that actually interacts with the physical world, and national security or critical infrastructure, where latency, reliability, and autonomy matter most.

We expect a meaningful share of the most important AI companies in the next decade to be built from the ground up around ODML, proliferating user devices, sensors, vehicles, and any systems at the edge.

The rest of this piece is a Q&A with Luke on how founders should approach ODML, what is working in practice, and where the biggest opportunities and traps are.

When you talk about “on-device ML”, what exactly do you mean, and what is the biggest misconception you run into?

On-device ML (ODML) to me refers to the systems that execute and leverage the ability to do inference directly on an edge device. This includes not just the ML, but all the surrounding technologies leveraged to put an experience in the users hand.

The biggest misconception I run into is edge inference being just an alternative technical approach to ML. Edge inference can unlock fresh and novel user experiences that wouldn’t be possible with cloud only inference.

What are some real-world examples of ODML already running in production that most people would be surprised by?

Take your phone out and zoom in as much as possible to something far away. What you’re seeing is not what the camera is capturing, it’s a model predicting what the image might be at a higher resolution. Similar things occur around the phone’s audio inputs, like background noise mitigation in phone calls.

What changed in the last few years that made ODML feel inevitable rather than hypothetical?

There have been progressive advancements in the computing capability of mobile and edge based devices. The CPUs, GPUs and ML specialized NPUs (neural processors) shipped in smartphones today are able to tackle surprisingly sophisticated workloads efficiently. Supporting technology allows most app devs to simply drop ML into their product, simultaneously, creative devs are learning how to get the most of the computing devices for advanced use cases.

This combined with an explosion of interest in AI-powered use cases (especially generative and agentic ones) leads me to believe the market is ready for a swath of new ODML based products.

For a founder who has only worked with cloud models, how should they think differently once they move intelligence onto devices?

Today, ODML can be a more difficult technical problem than the cloud approach, in part due to the high fragmentation of the mobile market, or the model architecture specialization required for optimized use cases.

Where ODML shows its value is in the new product experiences it can create. Creative and astute founders will understand how to turn edge inference into killer features for users that cloud inference will not be able to support.

Where do you expect ODML to create the most value over the next few years?

I think there are three main categories we will see ODML accruing value in the next 5 or so years.

The first is integrated applications for smartphones or wearables. Particularly interesting are applications in a specific vertical, like personal fitness.

The second is industrial applications focused on a single input media. I’m looking at things like computer vision for monitoring or QA control within agricultural or factory production contexts.

The third category relates to AI powered tooling and environments for creators and artists, a re-envisioning of things like photoshop or animation software. The winners will be employing ODML to offload inference cost to user machines and protect proprietary information.

For industrial, defense, and field systems that operate in bandwidth-constrained or adversarial environments, what makes ODML such a good fit? Any concrete examples you'd like to point to?

ODML enables real-time analysis and reactiveness on physical signals in the field. Autonomy and reliability of end points is critical here, especially within an adversarial context. The most robust field systems will be blending ODML on end points with centralized compute for high level monitoring, controlling and sophisticated decision making.

Anduril’s LatticeAI is the canonical example of this approach in practice. Their stack can deploy smaller ML models (like image recognition) directly on the end devices which alert and inform a centralized monitoring service.

What do strong ODML founders look like? How are they different from teams building typical cloud or SaaS products?

On the engineering side, strong ODML founders will be more comfortable with lower level, platform specific considerations that affect how their stack performs in the hands of the user. They will have a holistic view of development, thinking deeply on efficiency and deployment.

On the product side, strong ODML founders will have an intimate understanding of their users’ needs and experience, and how the underlying technology can benefit them. They also will have a connection or interest in a specific domain or vertical where ODML can be applied naturally, like health or agriculture.

What are the most common technical traps you see early ODML teams fall into when they try to go from demo to deployed system?

The most common technical trap I see, at least in the mobile space, is underestimating the fragmentation of devices and systems and the complexity of scaling across it. It is easy to build a demo considering only a single operating system and device, later to realize the approach might not scale to as much of the market as anticipated.

How should teams think about the tradeoff between model quality and real-world constraints like latency, power, and memory on their target devices?

Always think product-first, and holistically, model quality being just one piece of the puzzle. Depending on the experience being tailored, a sacrifice in quality can create an overall better product, if it leads to snappier response times, or even the oft-ignored app download/install time.

If you had to give founders a simple rule of thumb, how should they decide whether a use case really needs to be on-device versus just living in the cloud?

The answer just comes down to the user experience. Running inference on the edge can unlock a suite of experiences regarding personalization, real-time interaction and privacy. If these features give their product a fundamental competitive advantage, they should go all in.

The image above is the first pitch slide we ever created. It was May of 2019 and we were getting ready to present Moichor at a small pitch competition. At the time we were proud of this slide. The bright colors felt modern. The parrot made it memorable. The tagline sounded clean. To us it looked like what a real startup would put on a screen.

We later learned it said almost nothing.

The title read Smarter Complete Blood Counts for Animal Clinics. That sounds neat, but it does not explain the problem. It does not show who feels the pain. It does not create urgency. It does not capture the daily reality of technicians who did manual differentials for twenty minutes per sample. It was a vision without a struggle. A solution without a story.

The rest of the deck followed the same pattern. Nice visuals, vague language. We talked about exotic vets as a category instead of the one technician in the back room counting cells by hand. We talked about automation instead of describing the moment a vet waited too long and a bird declined before results came back. We talked about markets instead of lived pain.

We were pitching from too far away.

High level language cannot replace a specific moment of truth. Investors remember the story of one frustrated technician far more than a slide about a market size. They remember when a test took too long more than a sentence about workflow inefficiency. They remember what is real.

Everything changed when we grounded the story in one person and one problem. The pitch became clearer. The product felt necessary. The company felt like something that had to exist rather than something that might exist if everything went well.

Specific stories create inevitability.

Around this time we also began tightening our pitch process. One mentor gave us advice that stuck. Every title in your pitch deck should tell a story by itself. If someone only reads the slide titles, they should understand the narrative arc from problem to inevitability. Once we made this change the deck felt less like a collection of slides and more like a guided path.

We also stopped practicing in isolation. We started scheduling meetings in batches to stress test the pitch across different conversations. Running ten variations of the same story in a single week makes weaknesses painfully obvious, but it makes the pitch sharper faster. And before speaking with investors, we always pitched founder friendlies. Founders give the best feedback because they listen from a place of lived experience. They also iterate with you in real time if the story resonates with them.

There was another shift that mattered even more. We found the right mentor at exactly the right time. He was our first angel, but he also became our clearest voice of reason. He taught us how to slow down and understand what we were actually building. He helped us see the difference between a feature and a purpose. He believed in us before we fully believed in ourselves.

Every founder deserves that type of person. One honest mentor can reroute the entire trajectory of a company. They do not remove the struggle, but they give it shape. They help you hold the direction when everything around you feels shaky. Their impact lasts longer than any early pitch you put on a slide.

Your story is the product before the product. Tell it with clarity. Tell it through someone real. Build a pitch process that sharpens your thinking through rapid feedback and trusted founders. And when you find the person who helps you see the story clearly, stay close.

Two weeks ago, we released our first conversation with David Stone-Resneck, NPU’s marketing advisor. That discussion focused on how founders can build early traction by staying close to customers and refining their message through conversation.

In this follow-up, we explore how to measure what matters, spot early signals, and stay grounded amid shifting marketing trends.

Measuring Marketing Impact

What are the right things to measure when you’re early and data is limited?

Start qualitatively and move quantitatively. Since you don’t have outcome-based data yet, begin with something simple: record all your interviews. Use whichever tool you like best, but transcripts are essential.

As you review them, you’ll naturally start surfacing new insights. Keep high-level notes that validate or invalidate your hypotheses.

Once you’ve collected multiple transcripts, begin cross-referencing them to find the quant in the qual.

In the early discovery phase, identify the problem sets and understand the depth and pain caused by each. As you continue, look for subgroups within your broader audience that experience the problem more severely. Over time, this helps you refine your signal detection and differentiate which leads truly fit your target audience or ideal customer profile (ICP) and which do not.

You can do this manually, but I prefer building a Claude or ChatGPT project for each stage to create a growing knowledge base of core themes emerging from transcripts. You’re looking for patterns—so use LLMs to help.

Reach out on LinkedIn and I’ll share a set of system prompts I use for tracking patterns across call transcripts.

To validate distribution, set up Google Analytics and Google Search Console as soon as you can. Spend a day or two configuring event tracking for deeper-funnel pages such as whitepapers or foundational blog posts. You can also enable the free plan from RB2B on these high-value pages to get a sense of who’s exploring your solutions.

Trends & Broader Insights

What emerging trends in marketing should early founders pay attention to?

Andrew Chen wrote a great article about what’s working in marketing right now (spoiler: not much). The piece is about six months old and opinions vary, but it highlights some important truths.

@andrewchen

Every marketing channel sucks right now

These days I’m spending a lot of my time with very early stage startups (yes, as part of the program at a16z to invest up to $1M into each, called a16z speedrun) and as part of this, I spend a lot of time talking about launch and marketing strategy for new products…

Read more

a year ago · 529 likes · 81 comments · Andrew Chen

What does work is talking to people—leveraging your own knowledge to convey thought leadership to an audience genuinely interested in the topic, on the platforms they already use. It takes time and consistent effort to build, but that’s the game.

How has founder-led marketing evolved in the last few years?

From my perspective, this has become one of the most effective marketing strategies today. It only costs time to start, and it scales.

As AI has given everyone an efficient secondary-source research assistant—one that synthesizes existing published work—it has also automated much of traditional marketing. That shift makes your primary-source research (talking to people and identifying behavioral patterns) uniquely valuable. Readers are more receptive to this authenticity now than ever before.

What’s one underrated principle or mindset shift you wish more founders understood about marketing?

Founders often misinterpret the goal of 0→1 marketing. They think it’s about brand, advertising, SEO, or virality.

In reality, early-stage marketing is about finding a problem worth solving for a specific group and identifying the best forum and messaging to communicate that value to early users.

Takeaway

Founder-led marketing starts with speed, clarity, and proximity to your audience — but it grows through discipline.

Once the first signals emerge, the next step isn’t to scale—it’s to measure what matters, listen closely, and refine what truly resonates.

The best founders treat marketing like product discovery: they learn from every conversation, use insights as feedback loops, and let authenticity drive traction.

Curiosity compounds faster than any campaign, and staying hands-on keeps your message real long after the first 100 customers.

Connect with David

This conversation was part of our NPU Advisor Series, where we collaborate with operators who help our founders navigate their most critical early decisions.

If you’d like to dive deeper into David’s frameworks or get access to his system prompts for analyzing interview transcripts, you can connect with him directly on LinkedIn.