Lecture 1 - How to Start a Startup (Sam Altman, Dustin Moskovitz)
Y Combinator's foundational startup lecture dismantles Silicon Valley myths: ideas matter more than pivots, 100 users who love you beats 10,000 who like you, and you should only start a company if you literally can't stop yourself.
Read OriginalMy Notes (4)
- Successful founders are fanatical about small details, about getting the copy just right, and about customer support.
- YC companies that succeed often hook PagerDuty to their ticketing system so users get a response within an hour, even at 3am. Founders feel physical pain when the product sucks and want to wake up and fix it.
- They don't ship crap. If they do, they fix it immediately.
Get users manually — do things that don't scale
Recruit early users by hand. Don't buy Google ads. You don't need many — just ones who'll give you feedback daily and eventually love the product.
Ben Silbermann recruited Pinterest's first users by chatting up strangers in Palo Alto coffee shops and setting Apple Store display browsers to the Pinterest homepage before getting kicked out.
Get extremely close to early users. Listen to them. They'll tell you how to build a product they'll pay for, and they'll become the advocates who bring your next users.
The feedback loop is everything
Build an engine that turns user feedback into product decisions, ships it, and repeats. Ask what they like, what they'd pay for, if they'd be bummed if you disappeared, and if they've recommended you to anyone.
Make the loop as tight as possible. 10% better every week compounds fast. Software startups can measure this loop in hours. The best companies have the tightest loops.
Great founders do sales and customer support themselves early on. Don't put anyone between you and your users. Stanford startups specifically make the mistake of hiring support and sales people too early.
The company builds whatever the CEO decides to measure. Ignore vanity metrics like total registrations. Track active users, activity levels, cohort retention, revenue, and net promoter scores. Be brutally honest when they aren't trending right.
Why a small number of users loving you beats a large number liking you
For V1, you almost always have to choose: something a lot of users like a little, or something a small number of users love a lot. The total "happiness under the curve" feels equal. It's not.
It's far easier to expand from a small number of people loving you to a lot of people loving you. Going from widespread lukewarm enthusiasm to love almost never works.
If you get this right, you can get a lot of other things wrong. If you don't get this right, you can get everything else right and still fail.
Big companies like Google or Facebook will grab any opportunity where V1 can reach a lot of people with a lot of love. Startups don't get that option. Pick the small-and-intense path.
The only thing that matters at the start is making this work. Everything else can wait.
It's easier to start a hard startup than a easy startup
What "evaluating an idea" actually means
- The idea includes market size, growth, growth strategy, and defensibility — not just the product.
- Business needs to be difficult to replicate.
- Google was the 13th search engine. Facebook was the 10th social network, limited to broke college kids. Airbnb was sleeping on strangers' couches. All sounded awful. All became massive.
- If an idea sounds obviously great, too many people are already working on it.
- You want a monopoly, but you can't get one in a big market right away. Find a small market, dominate it, expand. That's why great ideas look bad early — the starting market looks tiny.
- Unpopular but right is the goal. Most people will think your great idea is bad — that means they won't compete with you, and sharing it is safe because it won't sound worth stealing.
- V1 doesn't need to sound big. It needs to take over a small specific market and expand from there.
How to evaluate the market
- Growth rate of the market matters more than current size. Small and rapidly growing beats big and slow.
- Most investors obsess over today's market size and ignore how it evolves over 10 years — this is a systemic mistake.
- Fast-growing markets have desperate customers who'll tolerate an imperfect but improving product.
- You cannot create a market that doesn't want to exist. You can change everything in a startup except the market.
- Sequoia's "why now?" question is essential: why is this the perfect time, why not two years ago, why would two years from now be too late?
Summary used for search
TLDR
• Contrarian take on ideas: Great companies start with great ideas that sound terrible ("unpopular but right"), not pivots - look for rapidly growing markets, not just big ones
• The love vs. like framework: Build something a small number of users love rather than something many users like - it's far easier to expand from intense love than mild enthusiasm
• The "can't not do it" test: Only start a startup if you're so compelled by a problem that you work on it anyway AND you're uniquely suited to solve it - all other reasons (glamour, flexibility, money) are illusions
• Reality check on returns: Employee #100 at Dropbox made $10M, at Facebook $200M - your startup needs $100M+ valuation to match, and you need more confidence in your non-existent company than proven winners
• Execution fundamentals: Sit at your desk, talk to users obsessively, measure active users not registrations, and make the feedback loop as tight as possible - nothing else matters until you have a great product
In Detail
Sam Altman and Dustin Moskovitz deliver Y Combinator's opening lecture by systematically dismantling popular startup mythology. Altman argues that the "ideas don't matter, just pivot" culture is fundamentally wrong - great companies almost always start with great ideas, not pivots. The best ideas have a specific signature: they sound like bad ideas but are actually good ("unpopular but right"). Google was the 13th search engine, Facebook the 10th social network limited to broke college students - all sounded terrible but captured rapidly growing markets. The key is finding small markets you can monopolize that will become huge, not chasing large markets with entrenched competition.
On product development, Altman introduces the counterintuitive "love vs. like" framework: it's better to build something 100 users love than something 10,000 users like. The area under the curve might seem equivalent, but it's exponentially easier to expand from intense love to mass love than from mild enthusiasm to passion. This manifests as organic word-of-mouth growth - if you're not seeing it, your product isn't good enough yet. The prescription is fanatical focus: sit at your desk, manually recruit initial users, obsess over feedback loops measured in hours not weeks, and ignore everything else (PR, partnerships, conferences) until you have something people love. Measure active users and retention, not vanity metrics like registrations.
Moskovitz delivers the reality check on why NOT to start a startup, systematically destroying common motivations. Glamour? You'll spend years at a desk doing customer support, not partying like The Social Network. Flexibility? You're always on call, you're a role model so you can't take your foot off the gas, and you're committed for 5-10 years minimum. Money and impact? He shows the math: employee #100 at Dropbox made $10M, at Facebook $200M. Your startup needs to hit $100M+ valuation (with your 10% post-dilution stake) to match - and you need to be more confident in your non-existent company than in proven winners. The only valid reason is the "can't not do it" test: you're so passionate about the problem that you work on it anyway, AND the world needs you specifically to solve it. At Asana, he and Justin Rosenstein kept building their task manager at night after full Facebook workdays because the idea was "beating itself out of our chest." That's the feeling you need.