Infinite startups

Most career paths have a ceiling. You show up, you perform, you advance along a well-defined ladder. Startups are different. They operate on a completely different set of rules, ones that most people never stop to think about. The math is unusually forgiving. You can try as many times as you want. You only need to get it right once. And when you do, the payoff isn't capped.

Unlimited attempts

The default assumption about startups is that failure is catastrophic. That if your company dies, you're done. But that's not how it works. Founders shut down companies and start new ones all the time. Some of the most successful startups in history were second, third, or fourth attempts by the same people. Justin Kan started Justin.tv after his first YC startup, Kiko, was crushed by Google Calendar. Justin.tv eventually became Twitch, which Amazon acquired for nearly $1 billion. The first failure wasn't the end of the story. It was a prerequisite. According to Harvard Business School research, first-time founders succeed about 21% of the time. Founders who previously failed? 22%. Founders who previously succeeded? 30%. The numbers tick upward because experience compounds, even when the previous venture didn't work out. There's no rule that says you only get one shot. The startup ecosystem is designed for iteration. Y Combinator alone has funded founders on their second and third tries. Investors know that a founder who has been through the cycle before, who has felt the pain of a failed product and lived to tell the tale, is often a better bet than someone fresh.

Each attempt makes you better

This is the part that makes the whole thing rational rather than reckless. Startups aren't lottery tickets. They're skill-based. Every failed attempt teaches you something concrete. You learn how to hire. How to talk to customers. How to manage cash. How to say no. How to build something people actually want. These lessons stick with you, and they transfer directly to your next company. Paul Graham has written extensively about how most startups die because they don't make something people want. That's not a random outcome. It's a solvable problem, and founders who have made that mistake before are far less likely to make it again. They develop better instincts for what real demand looks like versus what just sounds like a good idea. Nassim Taleb describes this as convexity: a setup where you gain more from positive variations than you lose from negative ones. Each startup attempt has a bounded downside (you lose some time and money) but the knowledge you accumulate bends the odds in your favor over time. The losses are small and informative. The potential gains are large and compounding. This isn't unique to startups. It's the same dynamic behind getting better at any craft. A chef who has burned a hundred dishes knows things about heat and timing that a culinary student doesn't. A salesperson who has been rejected a thousand times can read a room better than someone on their first cold call. Repetition, especially repetition that includes failure, builds pattern recognition that no amount of theory can replace.

The upside is uncapped

Most jobs offer a linear return on effort. You work more hours, you earn more money, up to a point. The relationship between input and output is roughly proportional. Startups break this relationship. A company that finds product-market fit can scale in ways that are completely disconnected from the number of hours the founder puts in. Software, in particular, has near-zero marginal cost. One more customer doesn't require one more unit of effort. This is what makes the asymmetry so compelling. Your downside on any single attempt is finite: the time you invested, the savings you spent, the opportunity cost of not working a salaried job. But the upside, in the cases where things work, can be orders of magnitude larger than the input. Venture capital is built on this exact insight. A VC fund expects most of its investments to fail. The entire model works because the winners don't just make up for the losers, they dwarf them. One company returning 100x can carry an entire portfolio. The same logic applies to the founder. You don't need every attempt to succeed. You need one.

The real risk is not trying

People overestimate the risk of starting a company and underestimate the risk of not starting one. A failed startup, in most cases, doesn't ruin your life. You end up with more skills, a better network, and a clearer sense of what you want to build next. Paul Graham has pointed out that even YC founders whose companies died didn't seem traumatized by the experience. Most just moved on to something new. The bigger risk is spending years in a career that has a hard ceiling, accumulating expertise that only compounds within a narrow context, and never testing whether you could build something with real upside. This doesn't mean everyone should drop everything and start a company tomorrow. But it does mean the decision framework most people use is miscalibrated. They treat each potential startup as a standalone bet and ask, "What if this one fails?" The better question is, "What happens if I keep trying?"

Infinite game, infinite upside

The combination is what makes startups unusual as a career path. You get unlimited attempts. Each attempt makes you meaningfully better. And the upside, when it comes, isn't constrained by the same forces that limit most professional outcomes. It's not risk-free. It's not easy. But the structure of the game is more forgiving than it appears from the outside. The founders who succeed aren't usually the ones who got lucky on their first try. They're the ones who kept playing long enough for their accumulated knowledge to meet the right opportunity. You just have to be right once.

References