The Anatomy of Belief
I was talking to a founder the other day. She was running a small biotech startup, one of those companies with a name that sounds like it was generated by concatenating two random Greco-Latin roots. They claimed to be doing something ambitious with AI and protein folding.
Ordinarily, this is where my eyes start to glaze over. I've probably seen a thousand pitches like this. They usually involve a slide with a big, impressive-sounding number, followed by a diagram of a neural network, and a lot of abstract nouns like "synergy," "platform," and "revolutionize." The founders talk about their grand vision, their total addressable market, and the paradigm-shifting nature of their technology. It's all very grand, and it all sounds exactly the same. They're trying to describe the size of a forest by showing you a map of the continent.
This founder, however, did something different. After the obligatory two sentences about changing the world, she put up a slide showing a single, tangled ribbon of a molecule.
"This is p53," she said. "It's a tumor suppressor protein. The handbrake on cancer. When it works, you don't get tumors. When it misfolds, you do. Everyone in this field tests their model on p53. And their models are all pretty good. They get it right 98% of the time."
She paused and zoomed in on one tiny corner of the molecular diagram, a spot where the ribbon made a sharp, awkward-looking turn.
"Our model kept failing on this one specific bit," she said, pointing. "The seventh lysine residue in the DNA-binding domain. We spent six weeks trying to figure out why. It turns out that at normal body temperature, the ionic charge from the surrounding solution causes this residue to twist in a way that every textbook and every existing simulation gets wrong. It's not a big twist. But it's the one that matters. We had to rewrite a part of our physics engine just to account for the electrostatic interference on that one amino acid. That's our secret. We don't just predict the fold; we predict the fold in a crowded, messy, 37-degree Celsius cell."
In that moment, I went from skeptical to sold. I still don't fully understand the biophysics, but I didn't need to. Her description of that one, single, impossibly specific problem was more convincing than any chart of performance metrics or any grand vision statement. It was proof that she and her team had gone deeper than anyone else. They had lived inside the problem at the atomic level.
What she was doing, without perhaps even knowing it, was demonstrating what I've come to call the specificity principle.
The rule is this: to make people believe a large, abstract claim, you don't need to provide exhaustive evidence; you need to share one or two hyper-specific, hard-won details that could only have been discovered by someone who was truly there.
The Maker's Mindset
Why is this so effective? Because specificity is a proxy for expertise. Our brains are prediction machines, constantly looking for shortcuts to determine who knows what they're talking about and who is just repeating things they've heard. Grand, abstract statements are easy to fake. Anyone can say they want to "use AI to cure disease." That's just a string of words. But the number of people who can talk about the electrostatic charge on the seventh lysine residue of p53 is vanishingly small.
That one hyper-specific detail acts as a core sample. If you show me a sample drilled from a thousand feet down, and I can see the distinct layers of sediment and granite, I don't need you to show me the rest of the mountain. I'll infer it. The detail proves the depth of the work.
This is something hackers and makers understand intuitively. When you're debugging a program, you don't say "the application is broken." You say, "there's a race condition in the multithreading code that corrupts the cache when two specific threads try to write to the same memory address simultaneously." Specificity is the currency of people who actually build things. It's the opposite of the managerial abstraction that so often hides a lack of understanding.
The conventional wisdom, especially the kind taught in business schools, will tell you this is all wrong. Focus on the big picture, they say. The market size. The 30,000-foot view. And perhaps that's good advice if you're trying to convince other people who live at 30,000 feet. But companies aren't built up there. For people who actually have to build the thing, or for the kind of investors who like to fund them, the most interesting things happen at ground level.
I see this with AI engineers all the time. You can spot the real builders by how they diagnose a failing agent. The mediocre ones will blame the model's non-determinism. But the great ones get quiet, lean in, and ask for the logs. They'll stare at the screen for a moment and then say something like, "I bet the user's query had a negative constraint, and the agent is ignoring it because the system prompt doesn't explicitly tell it how to handle 'what not to do'." That level of specificity, digging past the surface-level failure to find the subtle flaw in the agent's core instructions, is the hallmark of someone who has actually done the work.
Think of it like this: a truly great craftsman doesn't need to show you his entire workshop to convince you he's skilled. He can just show you a single, perfect dovetail joint. The precision and care evident in that one small connection tells you everything you need to know about the quality of the rest of the cabinet. Specificity is a founder's dovetail joint.
This principle extends far beyond startup pitches. It applies to writing, design, and even thinking. When you're trying to explain a complex idea, don't just describe the whole system. Find the most counter-intuitive and difficult component of that system and explain the hell out of it. If you can make someone understand the weirdest part, they'll trust you on all the easy parts.
I've noticed that people who are faking it often do the opposite. They will confidently describe the general outline of a subject, but become vague and evasive as soon as you drill down. They've learned the map, but they've never walked the territory. The person who has done the real work is often the reverse: they might be slightly hesitant to make grand pronouncements, but they can talk for hours about the gnarliest edge cases and the obstacles they hit along the way.
So if you're working on something hard, stop trying to make your vision sound bigger. That's a losing game; there's always someone willing to sound more bombastic than you. Instead, find your moment of specificity. What's the detail you lost a month on? What's the ugly, messy problem that no one else in your field has the patience to solve? What is the one small, hard-won truth that is yours and yours alone?
So catalog them. Write them down. The next time you're building a pitch deck, make a slide called 'The Hardest Problem We Solved' and tell that story. The grand vision is just the shadow cast by that one, tiny, perfect detail.