Paul Graham, the father of Silicon Valley startups, on how to achieve superlinear returns?
TechFlow Selected TechFlow Selected
Paul Graham, the father of Silicon Valley startups, on how to achieve superlinear returns?
10x growth is non-linear; massive growth doesn't require more effort, but often less effort—just better execution.
Navigating Web3 tides with focused insightsBy: Paul Graham
Compiled by: Youxin Newin
Paul Graham published a 10,000-word essay today—still packed with insights as always. He pointed out that linear growth cannot yield extraordinary returns; superlinear returns are rarely proportional to “your effort.” This echoes a book I read earlier, 10x is easier than 2x, which I highly recommend: A 2x improvement is linear and requires increasing effort—it means working harder, not smarter. A 10x improvement is nonlinear; massive gains don’t require more effort but often less, provided you do things better.
Paul also noted that times have changed: we no longer need to join prestigious organizations or institutions to access resources or achieve exponential growth. More people now can enjoy the freedom once reserved for 20th-century artists and writers. Many ambitious projects require little initial capital. Especially for Indie Makers, leveraging social media to build in public enables early cold-start growth and customer acquisition—an approach Magineer has believed in from day one.
Additionally, Paul discussed how to identify domains offering superlinear returns and who is best positioned to capture them: look for fields where only a few players dominate others. In areas where everyone performs similarly, superlinear returns are unlikely. Your early results may seem insignificant, but the reward curve for superlinear returns rises extremely steeply. He reiterated the well-known Silicon Valley mantra: Do things that don't scale—focus intensely on a small group of customers initially, and you’ll achieve exponential growth through word-of-mouth. The same applies to learning, research, and investing.
Among all this, the point that resonated most with me was this: Don’t equate work with a job—be driven by curiosity, not professionalism; give your curiosity room to breathe instead of doing only what you’re supposed to do.
The full text is below:
When I was a child, I didn’t understand that one of the most important facts about the world is the degree to which performance returns are superlinear.
Teachers and coaches implicitly taught us that returns are linear. “You get,” I heard countless times, “what you put in.” Their intentions were good, but this is rarely true. If your product is only half as good as your competitor’s, you won’t get half the customers—you’ll get none, and go out of business.
In business, it’s obviously true that performance returns are superlinear. Some think this is a flaw of capitalism—that if we changed the rules, it wouldn’t hold. But superlinear returns are a feature of the world, not an artifact of our invented rules. We see the same pattern in fame, power, military victories, knowledge, and even contributions to human welfare. In all these areas, the rich get richer.
If you don’t grasp the concept of superlinear returns, you cannot understand the world. If you’re ambitious, you absolutely must understand it, because this will be the wave you ride.
There appear to be many situations with superlinear returns, but as far as I can tell, they reduce to two fundamental causes: exponential growth and thresholds.
One of the clearest cases of superlinear returns occurs when dealing with something that grows exponentially—like bacterial cultures. Once they start growing, they grow exponentially. But culturing bacteria isn’t easy. This means the difference in outcomes between skilled and unskilled practitioners is enormous.
Startups can also grow exponentially, and we see the same pattern here. Some companies achieve high growth rates; most do not. The result is qualitatively different outcomes: high-growth companies tend to become very valuable, while low-growth ones may not survive at all.
Y Combinator encourages founders to focus on growth rate rather than absolute numbers. This prevents discouragement during early stages when absolute figures are still low. It also helps guide priorities: You can use growth rate as a compass for developing your company. But the main advantage is that focusing on growth rate often leads you to build something that grows exponentially.
YC doesn’t explicitly tell founders that growth rate matches “how hard you work,” but it’s not far off. If return is proportional to performance p over time t, then the return becomes proportional to p^t.
Even after decades of reflection, I find this formula astonishing.
Whenever your outcome depends on how well you perform, you’ll experience exponential growth. But neither our genes nor our customs prepare us for this. No one naturally grasps exponential growth. Every child is surprised the first time they hear the story of someone asking a king for one grain of rice on day one, doubling each day thereafter.
For things we can’t intuitively grasp, we develop habits to cope—but we lack such habits for exponential growth because it was rare throughout human history. In principle, herding animals should be an example: the more animals you have, the more offspring they produce. But in practice, land area limits expansion, making sustained exponential growth impossible without special plans.
Or more precisely, without universally applicable plans. One way to make territory grow exponentially is conquest: the more territory you control, the stronger your army becomes, and the easier it is to conquer new lands. That’s why history is full of empires. Yet few people created or managed empires, so their experiences didn’t shape cultural norms. Emperors remained distant, fearsome figures—not role models for everyday life.
Before industrialization, the most common case of exponential growth might have been scholarly research. The more you know, the easier it becomes to learn new things. As a result, just as today, some individuals accumulated significantly deeper knowledge than others in certain subjects. Still, this had little impact on mental habits. Although the idea of empire-building could be repeated—and thus more empires imagined—before industrialization, this form of empire had little practical effect.
This has changed over the past few centuries. Now, intellectuals design bombs capable of defeating territorial emperors. But the phenomenon remains so new that we haven’t fully absorbed it. Even participants rarely realize they benefit from exponential growth, let alone reflect on its broader implications.
The other source of superlinear returns is captured in the phrase “winner-takes-all.” In sports, the relationship between performance and reward is a step function: the winning team gets the victory regardless of whether they performed slightly or significantly better.
The origin of step functions isn’t competition per se, but the presence of thresholds in outcomes. You don’t need competition to encounter thresholds. Even as a sole actor, you may face thresholds—such as proving a theorem or hitting a target.
Fascinatingly, situations with superlinear returns often involve both sources: crossing a threshold triggers exponential growth. The winning side in battle usually suffers fewer losses, increasing their chances of future victories. Conversely, exponential growth helps you cross thresholds: companies growing fast enough in network-effect markets can exclude potential competitors.
Fame is an interesting case combining both sources of superlinear returns. Fame grows exponentially because existing fans attract new ones. But its concentration stems fundamentally from thresholds: there are only so many spots on the A-list in the average person’s mind.
Perhaps the most important domain combining both sources is learning. Knowledge grows exponentially, yet it contains thresholds. Learning to ride a bike is one example. Some thresholds act like machine tools: once you learn to read, you can learn anything else faster. But the most significant thresholds represent new discoveries. Knowledge appears fractal; if you push hard at the edge of a field, sometimes you discover an entirely new domain. If you do, you gain first access to everything within it. Newton did this; so did Durer and Darwin.
Are there general rules for finding situations with superlinear returns? The most obvious is to seek work that compounds.
Work compounds in two ways. It can compound directly: performing well in one cycle improves your performance in the next. This happens when building infrastructure, acquiring customer bases, or marketing brands. Or it can compound via education, since learning itself compounds. The latter is particularly interesting because during the process, you might feel you’re underperforming or failing immediate goals. But if you’re learning a lot, you’re still achieving exponential growth.
This partly explains Silicon Valley’s tolerance for failure. People here don’t blindly tolerate failure—they maintain confidence only if you’ve learned from it. But if you’re learning, it might actually be a good sign: Your company might not be growing as planned, but you yourself are growing, and that should eventually pay off.
Indeed, exponential growth unrelated to learning is rare. We should probably treat this interplay as the rule, not the exception. This suggests another heuristic: Always keep learning. If you’re not learning, you’re likely not on the path to superlinear returns.
But don’t overly optimize what you learn—don’t limit yourself to things known to be valuable. While learning, you can’t predict what will matter later. If you’re too rigid, you’ll miss outliers.
What about step functions? Are heuristics like “seek thresholds” or “seek competition” useful? Here it gets tricky. The existence of a threshold doesn’t guarantee the game is worth playing. In Russian roulette, you certainly face a threshold, but at best, you gain nothing. “Seek competition” is equally flawed—if the prize isn’t worth competing for? Sufficiently fast exponential growth guarantees both the shape and magnitude of the return curve: if something grows fast enough, even trivial beginnings become huge. Thresholds only guarantee shape.
Any principle leveraging thresholds must include a test ensuring the game is worthwhile. Here’s one: if you encounter something mediocre yet popular, replacing it might be smart. For instance, if a company sells products people dislike but still buy, creating a better alternative could win customers.
It would be great to find a method for identifying promising knowledge thresholds. Can we predict which problems lead to entirely new fields?
I suspect we’ll never predict this exactly, but the reward is so immense that even slightly better-than-random prediction would help—and we may find such methods. We can sometimes predict when a research question is unlikely to yield discovery: when it seems reasonable but dull. Problems that truly lead to breakthroughs often seem mysterious but perhaps unimportant. (If they were both mysterious and obviously important, they’d be famous unsolved puzzles pursued by many.) Thus, a key heuristic is to follow curiosity, not professionalism—let your curiosity roam freely instead of doing only what you’re supposed to do.
For ambitious individuals, the prospect of superlinear returns is exciting. There’s good news: this space is expanding in two directions. More types of work now offer superlinear returns, and the returns themselves are increasing.
Though two reasons exist, they’re closely linked—almost one-and-a-half: technological progress and declining organizational importance.
Fifty years ago, joining an organization was far more necessary for ambitious projects—the only way to access resources, colleagues, and distribution channels. In 1970, your reputation largely depended on your organization’s prestige. Reputation was an accurate predictor because outside organizations, achievement was unlikely. Exceptions existed—artists and writers working alone with cheap tools, owning their brands. But even they relied on organizations to reach audiences.
An organization-dominated world suppressed variation in performance returns. But this world has weakened dramatically within my lifetime. Now more people can enjoy the freedom once unique to 20th-century artists and writers. Many ambitious projects need little upfront funding, and new ways to learn, earn, collaborate, and reach audiences abound.
Despite the persistence of the old world, the pace of change is historically unprecedented—especially given the stakes. It’s hard to imagine a more fundamental shift than in performance returns.
Without organizational constraints, outcomes will be more varied. This doesn’t mean everyone benefits: top performers will do better, but poor performers worse. An important caveat. Exposing yourself to superlinear returns isn’t for everyone. Most people fare better as part of a collective. Who should pursue superlinear returns? Two types: the ambitious who know they’re exceptional and believe they’ll net more in a more variable world, and especially young people who can afford to take risks to find out.
Migration away from organizations isn’t just about current members leaving. Many new winners would never have been admitted. Thus, democratization of opportunity will be larger and more authentic than any internal version organizations might devise.
Not everyone welcomes this release of ambition. It threatens entrenched interests and contradicts certain ideologies. But if you’re an ambitious individual, this is excellent news. How should you leverage it?
The most obvious way to harness superlinear returns is to produce exceptionally good work. At the tail end of the curve, incremental effort yields outsized rewards. Especially because competition thins out—not just because excellence is hard, but because the prospect feels so daunting that few dare try. This means producing outstanding work is a bargain—and merely attempting it is too.
Many factors affect work quality. To become an outlier, you must get nearly all right. For example, to excel, you must care deeply—it’s not enough to be merely diligent. Thus, in a world of superlinear returns, knowing your interests and finding ways to work on them becomes increasingly vital. Choosing work suited to your circumstances matters too. For instance, work inherently requiring massive time and energy investment becomes more valuable when you’re young and childless.
Producing excellent work involves considerable craft—it’s not just about trying hard. Let me attempt a paragraph summarizing key principles.
Choose work you’re naturally good at and deeply interested in. Cultivate the habit of pursuing personal projects—whatever excites ambitious curiosity. Work hard, but avoid burnout; this will eventually bring you to the frontier of knowledge. From afar, frontiers look smooth, but up close, they’re full of gaps. Notice and explore these gaps; if lucky, one may open into an entirely new field. Take as much risk as you can—occasional failure means you’re not being too conservative. Seek the best collaborators. Develop strong taste by studying the best examples. Be honest, especially with yourself. Exercise, eat well, sleep sufficiently, and avoid dangerous drugs. When in doubt, follow your curiosity. It never deceives you—it knows better than you what deserves attention.
Of course, you also need one more thing: luck. Luck is always a factor, and it becomes more critical when working independently rather than as part of an organization. While some sayings about luck contain truth, there remains a real element of chance beyond control. The solution is to try multiple times—another reason to start taking risks early.
Science may be the quintessential example of superlinear returns. It features exponential growth through learning and thresholds at the extreme edges of performance—precisely at the limits of knowledge.
This leads to inequality in scientific discovery so extreme that wealth disparities—even in highly stratified societies—pale in comparison. Newton’s discoveries might have mattered more than those of all his contemporaries combined.
This may seem obvious, but spelling it out is worthwhile. Superlinear returns imply inequality—the steeper the return curve, the greater the variation in outcomes.
Indeed, the correlation between superlinear returns and inequality is so strong that it offers another method for identifying such domains: Look for fields where a few big winners dominate everyone else. Because in areas where everyone performs similarly, superlinear returns are unlikely.
Which domains feature a few dominant winners? Clear examples include sports, politics, art, music, acting, directing, writing, mathematics, science, startups, and investing. In sports, external thresholds create dominance—a fraction of a percent speed difference wins races. In politics, power accumulates similarly to imperial times. In several fields—including politics—success is heavily driven by fame, which itself exhibits superlinear dynamics. But remove sports, politics, and fame effects, and a striking pattern emerges: The remaining fields perfectly match those requiring independent thinking—where ideas must not only be correct but novel.
This clearly applies to science. You can’t publish ideas already proposed. Same in investing. Believing a company will succeed is only useful if most other investors disagree; if everyone believes it, the stock price already reflects that, and you can’t profit.
What else can we learn from these fields? In all, initial effort is required. Superlinear returns seem trivial at first—progress feels impossibly slow. But because rewards rise so steeply at the extremes, extraordinary measures to reach them are justified.
In startups, this principle is called “do things that don’t scale.” By giving extraordinary attention to an initial small set of customers, you ideally trigger exponential growth through word-of-mouth. But this applies to any domain with exponential growth—like learning. When starting out, confusion reigns. But investing initial effort to gain footing pays off, because the more you learn, the easier it gets.
Another subtler lesson from superlinear domains: Don’t equate work with a job. For most of the 20th century, the two were synonymous for nearly everyone, so we inherited a habit of equating productivity with employment. Even today, for most people, “your work” still means their job. But for writers, artists, or scientists, it means whatever they’re currently researching or creating. For them, work is something carried across jobs—even when employed, it may serve employers but also belongs to their own intellectual value system.
Entering a winner-take-all domain feels daunting. Some enter deliberately, but you don’t have to. With sufficient talent and by following your curiosity, you’ll naturally arrive in one. Curiosity won’t fixate on boring questions, and interesting questions tend to generate superlinear-return domains.
Domains with superlinear returns are far from static. In fact, the greatest returns come from expanding the domain itself. So while ambition and curiosity can both get you there, curiosity may be more powerful. Ambition leads you to climb existing peaks, but if you dig deep into a sufficiently interesting problem, a mountain may grow beneath your feet.
Join TechFlow official community to stay tuned
Telegram:https://t.me/TechFlowDaily
X (Twitter):https://x.com/TechFlowPost
X (Twitter) EN:https://x.com/BlockFlow_News
