Paul Graham
Paul Graham sold Viaweb to Yahoo for $49 million in 1998, used the money to fund the first Y Combinator batch in 2005, and then wrote essays that reached more founders more directly than anything else in the history of startup advice. He is the most cited thinker on My First Million whose name most casual listeners might not immediately recognize.
The Collison Installation
Paul Graham noticed one thing that the Stripe founders did differently from everyone else in their YC batch. He called it the Collison Installation.
When most founders got a potential customer interested, they’d say: “Great, I’ll send you a link to the beta when we’re ready.” What the Collisons did: “Awesome. Do you have your laptop on you? I can just set you up right now.” And they would brute-force get customers onboarded on the spot — open the laptop, install Stripe, walk through it, onboard them right then and there.
They did this for the first hundred to two hundred customers. Manually.
Graham’s observation: why doesn’t everyone do this? It wasn’t a complex strategy. It wasn’t even that hard. His answer: shyness, or fear of rejection. And a deeper misconception — people think big things come from big things. But actually, big things come from an accumulation of smaller things.
“Startups happen because the founder makes them happen. They take off because the founder makes them take off.” You hand-crank the engine to get it going. Once it’s going, it starts to roll on its own. At the beginning it feels like pushing a boulder to the top of a hill.
The Animal Heuristic
Graham has a framework for evaluating founders that Shaan has returned to repeatedly: can you describe this person as an animal?
Not like “he’s a giraffe, long neck.” Just: “Dude, that guy’s an animal.” The phrase captures tenacity, intensity, and the speed with which someone moves through obstacles. Among the thousands of founders who came through YC, this was one of the most reliable signals.
The point isn’t aggression or competitiveness. It’s a quality of engagement — the sense that this person will not stop until the thing is done, that obstacles are problems to be solved rather than reasons to pause.
Replit, PG, and the Email Relationship
Amjad Masad had an extended email relationship with Graham after Sam Altman connected them. Masad described spending hours crafting each message. Graham had been thinking about similar problems to Replit — coding environments, accessible programming — since selling Viaweb.
When Amjad got into YC on his fifth attempt, after Rick Rolling the application video, Graham had been a champion behind the scenes. Masad described him as “a great writer” who engaged seriously with the technical problems. The email exchanges helped Masad develop his thinking about what Replit was actually for.
At Boom Supersonic’s sound barrier test, Sam mentioned seeing Paul Graham — “who doesn’t seem like a very excitable person” — looking visibly excited. “It almost felt patriotic.”
The Life Mana Essay
In a conversation about startup risk factors, Sam cited a Paul Graham essay about how life choices compound your margin of error.
Graham’s model: entrepreneurs start with roughly seven points of freedom — capacity to take risks and achieve what they want. Each obligation subtracts: minus one for a mortgage, minus one for other obligations, minus two for certain life changes. Your margin of error for pulling off something ambitious gets tighter with each addition.
Sam found the essay in his early twenties and it stayed with him as a framework for thinking about optionality. The point isn’t that obligations are bad — it’s that they’re real costs that entrepreneurs often don’t price accurately when they’re accumulating them.
Founder Mode
Graham wrote an essay called “Founder Mode” that became a reference point in Shaan’s thinking about how great CEOs actually operate.
The concept, as Shaan describes it through an example from Twitch CEO Dan Clancy: great founders operate like a square wave, not a sine wave. They spend time at the 10,000-foot level — priorities, big picture, what are the three things we need to do to win. Then they drop straight down to 10 centimeters — into the pixels, the copy, the specific bug, the thing a single engineer is working on. Then back up. The two modes are not in conflict. They’re the method.
Most managers pick a level and stay there. Founder mode is the ability to hold both — the North Star and the most specific details of the problems that threaten it — and move between them deliberately.
YC as a Cultural Institution
Y Combinator’s impact on the MFM universe is pervasive. Blake Scholl credits YC with saving Boom Supersonic — the demo day format forced him to get his first real customer letters of intent, which led to the Virgin partnership that financed the company’s survival. Amjad Masad describes the YC period as transformational for Replit’s product.
What Graham built at YC, Ben Horowitz noted, is a culture where people tell you what they really think. Michael Seibel telling Blake Scholl his pitch sounded “completely full of shit” is the canonical example. The feedback that founders actually need — direct, honest, without performance — is rare. Graham built an institution that delivers it at scale.
Graham himself was a consistent presence in these stories: the person who noticed Replit was important enough to email Altman about, the person who was visibly moved watching the Boom test flight, the person whose essays founders are still citing fifteen years after he wrote them. He built the institution and then stepped back. The essays remain.