"The greatest danger for most of us is not that our aim is too high and we miss it, but that it is too low and we reach it." — Michelangelo
Scaling something to make it useful to millions or billions of people, usually with new technology, is the best way to grow an outsized business and have an outsized impact on the world.
Paul Graham’s Do Things That Don't Scale is an essay often cited by tech founders. In it, Graham argues that founders need to start scrappy and grow their companies one user at a time by doing things that they won't have time for as their companies grow. For example, he suggests that a CEO might send a personal thank you notes to early customers. He contrasts this with Tim Cook, who would never be able to do that for every Apple customer.
To most founders in Silicon Valley, the “natural” approach to building a business is roughly:
Design a product that can be used by a million (or billion) people
Build the product
Market it in a way that can reach that massive audience (go big!)
Graham pushes back on this approach. He argues that it’s best to start really small: get five, ten, or twenty users, delight them, and learn from their feedback. Small size is an advantage — it allows for agility and direct connection with early adopters. This idea runs counter to the “natural” Silicon Valley instinct.
By contrast, most people outside of Silicon Valley do only things that don't scale. A typical high school teacher doesn't ask "how can my teaching help a million students rather than 100?" A plumber doesn't wonder "how can I service ten times as many houses next year?" An accountant doesn't aspire to grow her client base from 100 to 100,000.
But Graham's essay doesn’t apply to most people — it’s aimed purely at founders who want to scale their companies. His message is clear: to scale, you must first do unscalable things. Founders need to market, sell, and deliver in ways that are intensely personal and labor-intensive — and then they can transition to approaches that scale.
If Graham had written this essay for a general, non-startup audience, he might have titled it: Do Things That Don’t Scale, Then Figure Out How to Make Them Scale. He would also have needed to spell out one of Silicon Valley’s core approaches: growing to serve millions or billions of people.
Over the past few years, I've been going back and forth — physically and metaphorically, personally and professionally — between Silicon Valley and the rest of the world. And I’ve realized that you can explain what separates Silicon Valley by what is — and isn’t — in Paul Graham's famous essay.
I believe that scaling — whether inside a for-profit business, a nonprofit, or outside of an organizational context — is the best way for individuals to maximize their impact on the world. This style of thinking has become core to who I am and how I approach my work.
Yet in conversations and projects with people in other geographies and other sectors, I often encounter a form of culture shock. “Why,” some people wonder, “must we scale everything? Can’t we just create something reliable that works for a small number of people, with less risk?”
Silicon Valley is mostly about maximizing expected value: the odds of success multiplied by the size of the outcome (whether measured in dollars made, lives saved, or something else). The rest of the world tends to focus on clear, predictable outcomes and on minimizing risk.
Silicon Valley wants to scale. Most everyone else is just fine staying small.
Of course, the world needs both approaches. Some problems call for carefully crafted solutions or businesses that serve a smaller audience with lower risk. Other challenges demand bold, scalable ideas that can transform entire industries or societies.
For me, scaling is the most exciting and impactful approach — but both have their place.
Not wanting to focus on scale may not be just about being risk-averse. Scale is homogenizing and impersonal and may require large amounts of venture capital at a loss of founder control. In many cases it is motivated not for maximizing impact but instead for maximizing rapid wealth extraction while ignoring societal impact. This latter motivation is how I would characterize Silicon Valley.
Ugh - this frustrated me so much. I apologize for my Ill made and jumbled response.
Your point about risk is poorly considered. For many creators, the goal isn’t to reach many but to earn respect. Respect thrives on individuality, not group appeal. Only a single customer is needed to provide that respect.
Scaling forces trade-offs: efficiency over nuance. Creators value direct connection. The compromises brought on by scaling feel wrong. They didn’t start a business to ruthlessly optimize output. Fundamentally, most aren’t avoiding risk, they are avoiding compromising, frankly adopting risk.
SW products are inherently easy to scale, because as autonomous agents, they can’t drop in quality when they are replicated. This isn’t true for real products, thus furthering lowering the risk of scaling.