Paul Graham's Startup Playbook: 10 Essential Essays
The 10 Paul Graham essays every founder should read
Paul Graham is a co-founder of Y Combinator. Airbnb, Stripe, Dropbox, and Reddit were shaped by his thinking. The essays he's been publishing since 2001 are the closest thing to a startup manual that exists — and they're free.
1. Do Things That Don't Scale
The thesis: startups don't take off on their own. You need to "jump-start" them manually, like an old hand-cranked car.
Airbnb went city by city photographing the apartments of its first hosts. Stripe invented the "Collison installation": when someone said "I'll give it a try," the Collison brothers would grab the person's laptop and install it right there. They didn't send a link, they didn't send a doc — they did it in front of you.
The classic mistake: founders believe all they have to do is ship, run paid traffic, and wait. It's not. You recruit the first 10 users by hand, one by one, treating each as a VIP.
If you're not yet doing things that don't scale, you haven't passed the initial stage. Period.
2. Startup = Growth
The thesis: a startup is growth. It's not having an office, it's not having funding, it's not having a brand-new company registration. It's a company designed to grow fast.
The numbers YC uses as a benchmark:
- 1% per week = you haven't figured out what you're doing yet
- 5-7% per week = healthy, you're in the game
- 10% per week = exceptional
Compounded over 4 years, the difference is brutal: a company generating R$1,000/month growing at 1% per month reaches R$7,900/month. At 5% per week, it reaches R$25 million/month.
The practical implication: if you pick a weekly growth target and chase only that metric, practically every decision becomes obvious. "Should I go to this event?" Only if it helps hit the target. "Should I build this feature?" Only if it helps hit the target.
Growth is the compass.
3. Startups in 13 Sentences
The thesis: 13 principles that condense everything YC has learned. The most direct cheat sheet out there.
Let me summarize the ones that matter most:
- Pick good co-founders (the number one factor)
- Launch fast
- Let the idea evolve
- Understand your users
- Better to make a few people love you than many think you're okay
- Offer surprisingly good customer service
- You become what you measure
- Spend little
- Stay "ramen profitable" (i.e., enough revenue to cover the founders' basic bills)
- Avoid distractions
- Don't get discouraged
- Don't give up
- Deals will fall through — don't count on them until they're signed
Print it and tape it to your monitor.
👉 paulgraham.com/13sentences.html
4. Founder Mode
The thesis (from 2024, which became a meme in the Valley): the conventional advice of "hire good managers and give them autonomy" is wrong for founder-led companies.
The essay was born from a talk by Brian Chesky (Airbnb). He said he followed the classic playbook to the letter — he delegated, created layers, gave autonomy — and almost broke the company. He only returned to growth once he started operating Steve Jobs-style: skip-level meetings, deep involvement in specific areas, breaking through hierarchy when needed.
Graham calls this founder mode. The rest (managing by proxy, treating areas as black boxes, blindly trusting manager reports) is manager mode — and it works for hired CEOs, not for founders.
The implication: you don't have to turn into a "professional CEO" as you grow. In fact, that can kill the company. What got you here (obsession with detail, direct contact with the product, fast decisions) is exactly what needs to continue.
👉 paulgraham.com/foundermode.html
5. Be Good
The thesis: no growth hack beats a product people love. And products people love usually come from founders genuinely trying to help, not extract value.
Graham observes that many startups early on are indistinguishable from a nonprofit. Craigslist and Google are examples — they spent years focused on helping users before focusing on monetization, and precisely because of that they became giants.
The counter-intuitive implication: instead of obsessing over monetization on day 1, you should obsess over "is this genuinely useful?". If it is, the money comes later. If it isn't, no sales funnel will save you.
"Be good" here isn't hippie moralism. It's strategy. Companies that try to cheat the user (dark patterns, hidden fees, abusive lock-in) rarely survive long-term because they lose trust — which is the most expensive asset to rebuild.
6. How to Make Wealth
The thesis: wealth isn't taken from other people, it's created. And a startup is the fastest way to do this because it allows leverage: reaching lots of people with a small team.
The mental math Graham does:
- The average employee generates, say, X per year for the company
- In a startup you work 3-4 years as 3-4 people (measured in intensity and focus)
- And with leverage (technology + big market) your impact can be 30-100x greater
- Therefore, in 4 years of a startup you can compress what would take 30-40 years of a traditional career
The cost is high: intensity, risk, no guarantees. But the mechanism is mathematically simple.
The implication: stop trying to "come up with a billion-dollar idea" and start understanding what leverage is. What makes the same hour of your work worth 1x or 100x? Scalable software, a big market, efficient distribution.
7. The Hardest Lessons for Startups to Learn
The thesis: the distractions that silently kill startups. Graham identified them after watching hundreds of companies at YC.
The eight most important:
- Release early. Ship a minimal v1, not a broken v1. More startups die from shipping slowly than from shipping fast.
- Keep shipping features. Users love a product that gets better every week. This is also marketing.
- Make users happy. You can't force anyone to use your product. You sing for your supper.
- Fear the right things. Don't fear the big players (Google, etc.). Fear yourself stopping executing.
- Commitment is a self-fulfilling prophecy. If you go in thinking "I'll try 3 months, if it doesn't work I'm out," you've already lost.
- There's always room. "But someone already does this" is almost never a reason not to do it.
- Don't set expectations. Deals don't exist until signed. Investors don't exist until the money hits.
- Speed, not money. Raising more capital doesn't fix execution problems. It just delays them.
👉 paulgraham.com/startuplessons.html
8. The 18 Mistakes That Kill Startups
The thesis (literally from Graham): there is only one mistake that kills startups — not building something users want. The other 17 are variations of that.
The short list:
- Single founder
- Bad location (being far from other founders)
- Marginal niche (picking a small market to avoid competition)
- Derivative idea (copying something existing)
- Stubbornness (not adapting)
- Hiring bad programmers
- Choosing the wrong platform
- Taking too long to launch
- Launching too early
- Not having a specific user in mind
- Raising too little money
- Spending too much
- Raising too much money
- Poor investor management
- Sacrificing the user for (supposed) profit
- Not getting your hands dirty in the business
- Fights between founders
- Half-hearted effort (staying in the CLT day job "for safety")
Avoid these 18 and you're ahead of 90% of teams. Simple as that.
👉 paulgraham.com/startupmistakes.html
9. How to Convince Investors
The thesis: investors decide in the first few minutes whether you look like a winner or a loser. After that, it's hard to reverse.
What investors look for (in this order):
- Formidable founders — someone who looks like they'll get what they want, regardless of obstacles
- Promising market — big and growing
- Some evidence of traction
And what does it mean to be "formidable"? It's justified confidence. It's not swagger, it's not imitating a Valley CEO speaking American English. It's you understanding the problem deeply and genuinely believing you're going to solve it.
The recipe for an inexperienced founder:
- Build something worth investing in
- Understand why it's worth it
- Explain that clearly
That's it. If you can't convince yourself that your business is worth investing in, no pretty deck will fix it. Investors detect faked confidence from a mile away.
👉 paulgraham.com/convince.html
10. How to Get Startup Ideas
The thesis: good startup ideas aren't thought up — they're noticed. Trying to sit down in a chair and "come up with a billion-dollar idea" is the fastest path to a bad idea.
The essay's most famous formula, a combination of Paul Buchheit and Robert Pirsig:
"Live in the future and build what's missing."
People at the frontier of a technology, industry, or behavior live in the future relative to the rest of the market. They're the ones who notice what doesn't exist yet but already should.
Practical tests for an idea:
- Do you have this problem yourself?
- Is there a small number of people who urgently need it? (Better than millions who'd "think it's cool")
- Does the idea sound boring or like a "schlep" (a pain to do)? Great — less competition.
- Does the idea have a way to grow that you can imagine?
Don't: try to copy a trend ("Uber for X", "Airbnb for Y"). If you're doing that, you're already late.
👉 paulgraham.com/startupideas.html
Why this matters
The founder who reads Paul Graham shows up different. They know that:
- Execution is worth more than the idea (and a good idea is noticed, not invented)
- Growth is the only metric that matters early on (5-7% weekly is the benchmark)
- Doing things that don't scale is exactly what you should be doing in the first 10 users
- Shipping fast and iterating kills fewer startups than perfectionism
- Founder mode > manager mode while you're still the founder
All the essays are at paulgraham.com/articles.html. Free. No paywall, no ebook, no course. One Saturday afternoon and you cover all 10.
If you're building, raising, or scaling anything, this is the best 2-hour ROI out there.
Tell me afterward which one hit hardest.
