rhyssullivan How to earn a billion dollars
Apply this framework when generating or evaluating startup ideas, advising on what to build, or estimating how big something could get. It distills the mechanics of why startups make founders wealthy and where good ideas actually come from.
Attribution
Adapted from Paul Graham's essay "How to Earn a Billion Dollars" (June 2026), based on a talk he gave at the Oxford Union. The framing, examples, and prescriptions below are his; quoted phrases are kept short.
Two numbers determine the outcome
A startup's eventual size — and its founders' wealth — comes down to two numbers, and nothing else matters nearly as much:
- Growth rate — how fast revenue (or usage) compounds, usually measured per month.
- Duration — how long that growth can continue, which is set by market size.
Everything below is in service of pushing one of these two numbers up. When evaluating an idea or a company, anchor on them first.
Exponential growth is the whole story, and it's deeply unintuitive
People assume vast wealth requires cheating because they can't feel exponential math. Don't reason about it by intuition — compute it:
- A startup growing 15% per month grows
4,384× over five years ($526M/year** — its founders are billionaires, honestly.1.15^60). A company making $10K/month becomes one making ** - The gap between "a couple million in revenue" and "a billion-dollar company" at 93%/month is about nine and a half months (
log(500, 1.93)). Small-but-fast and enormous are closer than they look.
The practical takeaway: a modest, sustained monthly growth rate is the difference, not a one-time stroke of luck or fraud. When someone says an outcome "seems impossible," check whether they're just failing to multiply.
Why this needs no cheating
Look at the two numbers:
- Growth rate is hit honestly "all the time" — by making something users love enough to recommend. Referral-driven growth is exponential, and it's the opposite of exploitation.
- Market size sets duration — and there's no way to cheat to make a market bigger. So the lever that determines how long you grow is fundamentally honest.
If you're advising someone whose plan to grow relies on extracting from or tricking users, that plan is fighting the math. Love-driven referral growth compounds; coerced growth doesn't.
Ask the growth rate first
When evaluating any startup or idea-in-progress, the single most diagnostic question is "what's your growth rate?" It reveals whether they've built the right thing better than any pitch does. A real, sustained weekly/monthly growth number is signal; a story about why it'll grow later is not.
Where good ideas come from
Make something people tell their friends about
In a market economy, needs get satisfied quickly, so it's hard to make something people want that they don't already have. The win is to discover a need no one else knows about yet — then serve it so well that users bring you the next users for free.
Feel the need yourself — especially if you're young
- Your intuitions about other people's needs are a weak signal. Don't build for a hypothetical user you've imagined.
- Your intuitions about your own needs are strong signal, because your needs predict future demand. What a small group (often young people) adopts now, the mainstream tends to adopt in roughly ten years.
- So: build something you and your friends genuinely want. It's fine — good, even — if it's niche and most people don't get it yet (a tool for molecular biologists, a thing for drone hobbyists). Niche-but-loved beats broad-but-tepid.
Don't look for startup ideas — build cool projects
Consciously searching for "a startup idea" filters out exactly the ideas worth having. The best ones sound lame at first (Apple, Facebook, Airbnb all did), and a deliberate search lops off those outliers because they fail the "does this sound like a real business" test. YC funded Airbnb despite disliking the idea — they liked the founders.
Instead:
- Work on projects with friends that you think would be cool to build. Apple, Google, and Facebook were not meant to be companies at first.
- Trust your unconscious. Anything that genuinely seems like it'd be a cool thing to build has a high probability of leading to a good startup idea, "no matter how preposterous." (A guy live-streaming his life from a head-mounted camera became Twitch.)
Don't worry about market size up front
Because you're predicting future demand, the market will grow into you. You only need a beachhead in some territory of unsatisfied need that you can expand from, plus adjacent markets to grow into later. A small market today that's growing fast beats a large static one.
The key is empathy, not exploitation
The core skill of a successful startup is understanding a group of users so deeply that you can build exactly what they want. The question to keep asking is: what would make these users' lives dramatically better? For young or inexperienced founders, the reliable hack is to make something for yourself — then you have a user (you) whose needs you understand completely.
Using this to evaluate an idea
Run an idea or company through these checks:
- Growth: Is there a real, recent growth rate? Does the product create referral-driven (exponential) growth, or does growth depend on spend/coercion?
- Love: Do current users love it enough to tell friends, or merely tolerate it?
- Duration / market: Is there a beachhead of genuine unmet need, with room to expand? Is the market growing?
- Empathy: Do the founders deeply understand the users — ideally by being users themselves?
- Origin: Did this come from a real itch / cool project, or from a top-down hunt for "a business"? Be more interested, not less, in ideas that sound lame but that the founders clearly want to exist.
If the answers are strong, don't be put off because the current numbers are small or the idea sounds unserious. Get a high growth rate by making something users love enough to share, get duration by being in a big (or fast-growing) market, and the wealth "will happen automatically if you just keep making customers happy."