These are my notes from “A Conversation with Paul Graham”, moderated by Geoff Ralston. An entry of my Startup School series where I follow the curriculum in Y-Combinator’s Startup School
Geoff Ralston is a current partner at Y-Combinator. These are notes on what I thought were the more interesting parts of his interview with Paul Graham, one of the original founders of Y-Combinator. Geoff’s questions will be under “G.R.”, Paul’s replies will be under “P.G.”. This isn’t an exact transcript, it’s just my best attempt at summarizing what was said.
G.R. - How do you pick the startups to fund
P.G. - It’s not just that outcomes of startups are hard to predict, they’re actually indeterminate. There’s a huge amount of luck involved.
Because of the role luck plays in a successful, even if you were able to pick perfectly, you’d still need to pick 20 to get 5 successes. It’s just impossible to know. It’s like poker, even if you played your hand perfectly every time, the 8 other players at the table will still win some hands because of the luck involved.
G.R. - Robert Morris, Trevor Blackwell were your co-founders. Were they good co-founders? how did you pick them?
- Robert was not into it, wanted to sell it at every offer so they could work on something else
- I picked Robert because he was my co-conspirator. I would do everything with him. We always had schemes. And he was a really good programmer. If he didn’t like what the OS was doing, he would edit the source, re-compile and it would do what he wanted. He was exceptional.
- I picked Trevor because I asked Robert who was the smartest person he knew. So we recruited him. He was exceptional. I didn’t know, but I trusted Robert.
Intelligent people can judge other intelligent people. But trustworthy people can’t always judge other trustworthy people. It works for some qualities but not others.
do things that don’t scale - do things in a manual way. In the back of your mind, you know you can’t do it this way when you get a little bigger. But do it early anyway. If you don’t you won’t become big anyway, so you have nothing to lose.
For example, we would go to people we thought might want to be customers of our online store building app. We asked them if they wanted to use our store builder, and they would say no! So we would ask them “well can we just build your store for you, and you pay us to do it”. They would say okay, and then we would build it for them using our own app. Obviously this wouldn’t scale, but we did it, and we learned a ton about customers doing it this way.
If you think you know the solution and you build lots of stuff for it. You’re wasting time.
G.R. - Thoughts on competitors?
P.G. - Y-combinator has helped about 1900 companies, 1 died because of competitors. Don’t worry about competitors. You already have built in protection. It’s the same protection a small airplane has against crashing into other small airplanes. SPACE!
G.R. - What makes startups succeed?
P.G. - Start with your team. It’s too difficult by yourself. You need a good founding team
- people where you know each other well
- people you can work together with
- much more important to be determined than smart.
- But only one of you has to be the determined one that will drag everyone else with him.
G.R - What will they get wrong?
P.G. - You will not pay enough attention to users. You’ll make up some idea in your own head, thinking about your vision by yourself. You would be way better off, finding someone, anyone that would pay you to solve there problem. Then fix it for them, and then find more people like that. You’ll shrink away from contact from your users, because you’re embarrassed about what you’ve built so far, or you think you know what they’ll say and your product is just not ready. that’s a mistake.
We launched YC within 1 week of having the idea. We just had a site with a form, asked people to apply by email. We would print all the emails out, and grade them.
G.R. - How do you deal with different commitment levels from founders?
P.G. - Ask yourself would I rather have 25% of this person, or 100% of another person.
G.R. - Is launching to private beta good enough to be “launched”?
P.G. - I’ll take it, it’s closer to what you need to do than not launching at all
G.R. - How do you user interviews?
P.G. - Start by asking them questions about what they want, what there problem is. Then try to figure out what the real problem by asking what if you could do this, what about this, etc.
G.R. - How do you justify the risk of launching too early, the bugs that come with it.
P.G. - It’s justifiable because there’s a bigger risk to launching late. It’s not one sided. If you can do one thing that makes one person super happy, then launch it. Build from that one user.
G.R. - How to price during launch?
P.G. - Don’t sweat it too much. if you want to raise it later, just grandfather old users and raise it. if you have exponential growth, then it’ll be a small portion of your total users. if you want to drop prices no one will complain. Don’t price too high, get users, they’ll teach you how to get better.
full video - https://www.youtube.com/watch?v=4WO5kJChg3w