It was an ordinary Tuesday evening dinner at Y Combinator. YC S11 — Paul Graham's artisanal era, when there were around 60 companies in a batch, when Paul would sit at the end of a long table and you could actually talk to him, when the dinners felt like a graduate seminar rather than a conference.
Then Yuri Milner and Ron Conway walked in.
They announced that every company in the batch would receive $150,000 as an uncapped SAFE — no negotiation, no diligence, no terms to discuss. Just a check, for everyone, tonight.
The reaction in the room was silence first. Then a kind of disbelieving energy, where people were checking with the people next to them to make sure they had heard correctly.
Nobody had been tipped off. Nobody expected it.
That moment — the palpable shock in the room — is the first time I understood how quickly a funding environment could change. What I didn't fully understand yet was that the change was permanent.
The Start Fund, as it came to be called, did three things that are easy to underestimate from a distance of fifteen years.
It made seed funding institutional. Before the Start Fund, seed rounds were assembled check by check from angel investors who each made individual decisions. The process was slow, relationship-dependent, and highly variable. The Start Fund demonstrated that seed capital could be deployed programmatically — same terms, same amount, every company in the batch. That shift from artisanal to institutional changed the supply side of early-stage funding in a way that cascaded for years.
It put uncapped SAFEs into common usage. The SAFE instrument had existed before S11, but it wasn't the default. The Start Fund used uncapped SAFEs — no valuation cap, just the right to convert at the next priced round. Founders initially worried about what this meant for dilution; I remember conversations that week about whether accepting the money was actually good for us. (It was.) The normalcy of uncapped SAFEs today traces directly back to that evening. It shapes cap tables across the startup ecosystem fifteen years later.
It changed founder psychology overnight. Before the Start Fund, raising a seed round required founder attention from day one. The question "have you closed your seed?" was a proxy question for "are you a real company?" The Start Fund decoupled startup legitimacy from the seed raise, at least for YC companies. The psychological effect — being able to focus on the product before worrying about the next fundraise — is hard to quantify but was genuinely significant.
Walking out of the dinner that night, I found myself in a conversation with another founder from the batch — I won't name him, but he had been in the middle of working his angel network to close a small seed round. He stopped mid-sentence at some point and said something like: "I think we just don't need to do that anymore." Not triumphantly. More like someone realizing that the map they'd been navigating by had just been redrawn. That was the feeling in the room — not celebration exactly, but a kind of recalibration. The question "how do we close our seed?" had been the background anxiety for most of the batch. That evening it simply stopped being the question.
I've thought about that evening many times in the context of fundraising conversations I've had since — both as a founder and from the investor advisory side.
The most important lesson isn't about SAFEs or uncapped instruments. It's about market timing.
Yuri Milner and Ron Conway didn't make the Start Fund because they had solved a theoretical problem with seed funding mechanics. They made it because they identified a moment when the cost of deploying capital into early-stage startups had dropped dramatically relative to the potential upside. The decision looked audacious. The analysis, if you saw what they saw about where the venture capital market was heading, was straightforward.
This is the pattern I've come to call "inflection point recognition" — the specific ability to identify moments when something structural has changed, such that now is different from two years ago in a way that matters.
The Start Fund was right because mobile was beginning to explode (we were in the batch because of Realm's mobile database thesis), cloud was making the cost of starting a software company negligible, and the historical pattern where seed capital dried up before the best YC companies could raise proper Series As was genuinely destroying value. The structural change had happened. The Start Fund was the right response to it.
Most founders think about fundraising as a skill — how to pitch, how to run a process, how to navigate the investor relationship. These things matter, but they're table stakes.
The founders who raise well consistently aren't the best pitchers. They're the ones who understand the market they're raising in — what structural changes have made their thesis timely, why now is different from two years ago, where the next pool of capital is forming and why.
The Start Fund was a large-scale demonstration of a small principle: the investors who are right about early-stage companies are the ones who recognize when the market has changed before the market knows it has changed. The founders who raise efficiently are the ones who can articulate that same insight about their own companies.
What has changed — technically, regulatory, in distribution, in customer behavior — that makes this company's thesis viable now when it wasn't viable before? That question is harder than "how good is the pitch?" But it's the question that matters.
I know because I was in the room when someone answered it correctly and wrote a check to every company in the batch.
Fifteen years on, the Start Fund's specific financial terms look modest — $150,000 doesn't go as far as it did in 2011, and seed rounds have grown substantially since. What hasn't changed is the underlying logic: identify the structural change, recognize the moment, and act with conviction before the market catches up. That's still the game.