Strip an early-stage investment of the deck, the model, and the metrics, and what's left is a question of faith: do I believe these people can build the thing they said, and find the people who will buy it. The industry dresses that question in spreadsheets, but at the early stage the spreadsheet is mostly assumption, and the assumption is a bet on people. There is less holding an early bet up than we like to admit — and naming what actually holds it up is the beginning of doing it well.

Diligence looks for signs. Early-stage has few.

Later-stage diligence half-reads itself. Revenue, retention, cohort curves, a customer list, a track record — the signs are there, and you can lean on the evidence and let the numbers carry a good part of the judgment. Early-stage offers almost none of that: no revenue to speak of, no history, no proof the thing works or that anyone wants it. A few people and a claim. The signs are absent, which means the read falls back onto the one thing still in the room — the people, and whether you have faith in them. Early diligence is not a thinner version of the real thing. It is the real thing with the instruments removed: you are reading the ground directly, because there is no evidence sitting on top of it to read instead.

There is no bedrock — the same as the dollar

An early bet has nothing under it but conviction about people, sustained. That should sound familiar, because it is exactly how the dollar works. Since 1971 the world's reserve currency has been backed by no metal and no hard floor — it holds its value because enough people believe it does, and the belief is self-fulfilling.^[1] You cannot prove a dollar is worth a dollar; you can only observe that the faith holds. An early-stage company is the same object at a smaller scale. There is no proof it will work — there is a read of the people, and the faith you place on the strength of that read. The placing of the faith is the entire act.

Turtles all the way down

It is tempting to treat that missing floor as a flaw peculiar to early investing — the thing you tolerate until the metrics show up. It is not. It is the structure of the whole stack. The dollar rests on faith; the company rests on faith in the dollar and in its people; the customer buys on faith the thing will deliver; the market rises on faith the customers will keep coming. Pull on any layer and the one beneath it is another belief, not a slab of stone. It is turtles all the way down, and there is no ground turtle. What holds the stack up is not a bottom — there is no bottom — it is that the faith at every level keeps holding. Once you see that an early bet has no bedrock, you notice that nothing does. The question was never how to find the floor. It is how to read whether the faith will hold.

So you read the people, not the machine

With few signs, the object of diligence stops being the company and becomes the founders. You are reading the ground they stand on: can they actually build what they described, will they hold when it gets hard — and it always gets hard — is the reason they are doing it real enough to survive the years when nothing works, and does their word match what they can do. None of that is on the deck. The deck is the postcard, the face built to be judged well. The read that matters goes underneath it to the people, because at the early stage the people are the asset and everything else is a projection of them. It is the oldest correction there is: not the machine, the people behind it.

Faith twice

And it is faith stacked on faith. You have to believe the founders can build it; the founders are betting the customers will believe it is worth buying. The market's belief is exactly as unbacked as yours — a customer buys because they are convinced there is value there, which is the same faith you are underwriting one level up. So an early investment is a wager that your faith in a few people will be ratified by the faith of many strangers. There is no floor under either belief. There is only whether both hold.

But the read is only as good as the reader

All of this puts the weight on one thing: your read of the people. Which raises the question nobody in the pitch is asking — what is your read standing on? You are judging whether founders are grounded, whether their word matches their work, whether they will hold when it gets hard. You cannot judge firm ground you have never stood on. A reader who has never had their own conviction tested has no instrument for detecting whether someone else's will survive a bad year; they mistake confidence for depth, polish for foundation, the postcard for the place. The read is only as trustworthy as the ground the reader stands on. So the real precondition for underwriting other people's faith is having done the work of grounding your own.

How one gets there

You do not acquire that read in the moment you need it. You build it over a life, the way anything load-bearing gets built — quietly, before the storm, so it is already there when the storm comes. A few things make it.

Touchstones that do not waver. You set the handful of things you will not trade — what you actually believe, how you keep your word, what you owe the people who depend on you — and you set them when it is calm, on purpose, so that hardship finds them already fixed. The test of a touchstone is not that it feels good; it is that pressure does not move it. A person carrying real touchstones reads differently across the table than one improvising values in real time, and after enough reads you feel the difference in seconds.

Ritual over belief. What keeps the touchstones fixed is not a mood or a fresh burst of conviction each morning — it is practice, repeated whether or not you feel it. The continuity is the ritual. You keep the practice through the years the belief runs thin, and the practice carries you until the belief comes back. An investor who only has conviction on the good days has nothing on the bad ones; the one who has a discipline has something that runs at idle, always on, carrying load before anyone speaks.

Sources you did not make yourself. Nobody grounds themselves from nothing. The read is assembled out of what formed you — the books you actually consumed rather than skimmed, the mentors who corrected you, the lineage you came up through and the people who vouched for you before you had earned it. Their whole utility is to ground you: to give you fixed points outside your own preferences to check yourself against, so that when you are wrong — and being wrong feels exactly like being right^[2] — there is something firmer than your own certainty to catch it. You do not argue from soft ground once you have read the source material and sat under people who knew it cold.

And transmittal. The final proof that the ground is real is that it moves — received from the people above you who deemed you worthy of it, and handed, in turn, to the ones coming up who you judge worthy of it now. Ground that cannot be passed down was never solid; it was personality. The read that can be transmitted — taught, entrusted, given away without losing it — is the one that was built on something. It is the same faculty you are using across the table: you are deciding whether these founders are people you would hand something to, because you have been on both ends of that handoff and you know what it costs to get it wrong.

The counter-case, held open

Calling it faith sounds reckless next to a model with a hundred rows, and the objection deserves a real answer. The answer is that the model is faith in a lab coat. Every early-stage projection rests on assumptions about people — that they will execute, that a market will show up — and dressing those assumptions in decimals does not make them evidence. Worse, it feels like evidence, and being wrong feels, from the inside, exactly like being right, so false precision is one of the more dangerous things you can carry into an early-stage decision. The honest move is not to pretend you are not on faith. It is to have disciplined faith: read the ground rigorously, verify everything that can actually be verified, earn the read instead of bluffing it — and then place the bet knowing it can fail, because a faith that cannot be wrong is not faith, it is denial. Everyone at the early stage is operating on faith. The only question is whether they admit it and discipline it, or bury it under numbers and call it rigor.

Which is why the skill was never calling the company that already has traction — the signs do that work, and by then the faith is cheap. The skill is the read you make when there is little to point at: past the deck, to the people, deciding whether you have faith in what they are standing on — a read only as good as the ground you have built to make it from. That is the whole of early diligence, and it is why a serious platform reads people and their formation, not only their numbers, because the thing being underwritten is a set of human beings and the numbers are downstream of them. Investment, early, is faith. The best investors do not hold more proof than the rest. They hold a better read of the ground — earned by grounding their own — and the nerve to place the faith on it.

1. The US dollar has been a pure fiat currency since 1971, when the United States ended the dollar's convertibility to gold; its standing as the world's reserve currency rests on confidence rather than any hard backing.
2. Kathryn Schulz, Being Wrong: Adventures in the Margin of Error (2010) — on "error-blindness," the fact that being wrong feels, from the inside, no different from being right.