Caroline, I'm really excited. It has taken a couple years, but we have Paul Graham back on the show for part two.
Part two. Yay. Hey, Paul.
You'd think it would have been easier for me since I'm your wife, but here we are.
The shoe maker's children.
The shoemaker's children. Exactly.
Oh, I say that all the time. I say that all the time about YC. How shoeless. We're always shoeless.
So, we had a great conversation the last time and we kind of had left off. We had talked a lot about the history of Y Combinator and we talked a little bit about what you had been up to after Y Combinator but we didn't get into the growth of Y Combinator and what happened once we had, you know, successfully, once it had gotten airborne. What happened? We had eight startups in that very first batch and we always say we grew it organically. But let's talk about how we grew it.
Well, we would just have an application deadline and people would apply and like more and more people applied. We never did anything all that consciously to get more people to apply, right? I mean, we'd announced that there was going to be an application deadline. We would do startup school every year and that's I think that probably caused some people to apply but other than that we didn't have any sort of conscious effort at marketing. Oh, and I would write essays and people would read the essays and I would have this link 'interested in starting a startup apply to Y Combinator' right. So that probably got more people to apply but there was never any deliberate plan to grow it. We would just get more applications each time.
How much overlap with Hacker News was there? Because I have to imagine, oh, I know for a fact actually that Hacker News was a huge source of applicants.
Hacker News, I think I started it in like 2007. So, for the first couple years, it wasn't a factor.
Oh, okay. Yeah. And in fact, I think when we were first getting going, there was very little press. But then we were in the New York Times and that kind of got us out there. And you know, honestly, the only kind of, for something like YC, what this is something people don't realize. For something like YC, the only kind of out there that you care about is whether potential applicants know about you. Do 22-year-old programmers know about you? And so, if you get an article about you in the New York Times and 22-year-old programmers aren't reading the New York Times, it might seem to you've gotten a lot of publicity, but actually it didn't really help much, you know? Yeah, I think founders talk to, I think applicants would talk to one another.
I think that was why we grew.
Did that New York Times article have a like a theme or a thesis other than just like check out this new thing? Like I don't, I probably read the article at some point, but I don't remember it.
Yeah, it was good actually. It described YC as this new phenomenon in the investment world, which at that point it was. So that was actually a pretty useful sort of good historical article.
And kudos to Jenny Lee for getting it. No one else did.
She could claim to understand early on what was going on.
You were talking about how that's not the right demographic, right? Because you want 22-year-old programmers to know about YC. But and we may talk about this a little bit later, but we also had the tiger mom problem. Remember when...
Yeah. Adora actually was the first person to explain this I think to me in an articulate way like you also need the parents to know about it. So maybe the Times article...
Parents would read the New York Times.
Exactly. That was my point. Yeah. Yeah. You know, and in fact, part early on, this is probably not true so much now because now startups are a more accepted thing to do, but back then in 2005, parents would freak out at the idea that their kids would leave school and like not get a job. It wouldn't seem like they were founders. It would merely seem like they were unemployed. And so once YC started to have a big brand, then the founders could say, 'Look, Mom, I'm not just working on my loser startup. I got into YC. There's this well-known thing that I got into.' I think it helped a lot back in the day.
So, what was the secret to our success then in your opinion?
Um, I think we actually helped the startups a lot and so they would tell one another like if we had sucked it would be like you know your friends go to a restaurant, right? And if the food's really good, they tell you, 'Hey, you should go to this restaurant. The food's really good.' I think people who did YC would tell their friends, 'Yeah, we did this thing. It was really good. They helped us a lot. They helped us get investment. Look how they incorporated our startup. They did all this like stuff for us.' I think it was simply that it helped them a lot and they would tell their friends and then their friends would do it too and then you know that's a recipe for exponential growth because then their friends do it and then there's more people out there telling people that it helped them. The alumni were critical, I think, in the growth of YC because young programmer types know one another. They went to school together, they hang out together, you know. Um, so word spreads quickly. If you please that group, word spreads quickly. And back when we weren't space constrained, I mean, we used to have alumni come to the dinner sometimes drop by and they'd introduce the new startups to their investors and it was just this really great community.
Yeah. And sometimes become their customers along the way. I mean, there were so many ups, I mean, I wouldn't say there were dramatic ups and downs, but there were some crazy stories. And Carolyn and I were...
Some pretty dramatic ups and downs, actually. We have forgotten.
But I can remember feeling, I can remember feeling at the end of the day, Jesus, you couldn't make this [ __ ] up. And...
And I remember also feeling like this was not an unfamiliar thing. Like stuff was, there was always disasters. Stuff was always breaking, you know, like just Hacker News all the time. You forget how many times did I have to run off and like run to my office and deal with some disaster happening to Hacker News, right? That alone...
I hated Hacker News, I have to say.
Me, too. It was such an unbelievable source of stress. In retrospect, I think maybe that was a mistake. Maybe I shouldn't have done that because really, it really was more than half of the total stress of running YC.
Oh. Oh, yeah. Oh, well, wait. So do you mean technical ups like technical dramatic things or community based dramatic or both?
Yeah. The community. Yes. That actually, servers running out of file space, you know, denial of service attacks, people getting into like terrible fights on the site. I mean, it was like in retrospect that was like a startup in itself. It was like I was running a startup at the same time as doing YC.
Do you think Twitter impacted our growth?
When did that come around? 2007.
Something like that. Yeah, I think it probably did a bit. Yeah, people probably because early on especially all the users, you know, the users have gone through phases. In the beginning, it was like computer nerds and then it was SJWs and now it's like, you know, Trump supporters or bots pretending to be Trump supporters. They're hard to tell apart, right? And so that first phase, we forget now, but that first phase, it was pretty technical.
Who were some of your favorite speakers? And I asked that because I remember the Twitter founders, all of them, you know, Ev Williams, Jack Dorsey, and Biz Stone used to come to our dinners to be guest speakers. And I used to love when they came. Who do you remember as some of your favorite speakers?
You know, Zuck actually was a pretty good speaker because he was so young. He was so young when he started Facebook. He was probably still at that point, even though he was now famous. He was probably still younger than some of the founders in the batch. So, they could really relate to him. I could sort of hear them all becoming more ambitious as they thought, okay, here's this guy. He's like, he's super famous and very rich, but he doesn't seem like he's a different species of animal from me. Like, well, I could, why can't I do what he did, right? You can sort of hear them all. And he was pretty candid.
Like back in those early days, I mean, I hope this is still mostly true, but back in the early days, the speakers used to be completely candid. Nobody worried much about whether something they said would get out because even if it did get out, that was, this was before cancel mobs were invented, right? What was going to happen?
What was the first year? Do you guys remember the first time he came? Like what time that came?
Well, it was in California, obviously. It was pretty early on.
It must have been like 2007.
Yeah. Seven or eight, I would think.
Yeah, pretty early. I remember we met him at a restaurant in Palo Alto. Do you remember that, Paul?
Well, this was before he had learned to imitate a normal person, right? Um like early Zuck, you know, I mean, I don't know if Zuck publicly describes himself as being an aspie, but he was doing a good imitation of it at any rate. And so he had this weird thing where if he didn't have anything to say, he wouldn't fill the gap with, you know, passing conversation. He had no concept of small talk. If there wasn't anything he felt like saying at the moment, he would just go like this. And it was surprisingly disconcerting. I didn't realize how important small talk was until I met the lack of it. But he would just stare at you if there wasn't anything. He's like, 'I don't have any words I want to say. Why should I say words?' You know, um...
Ron, well, Ron, in fact, warned me about this specifically. He said, 'All right, you're going to meet Mark. Now, I warn you, there's going to be some big gaps in the conversation, right?' When he said that, I didn't realize what he meant. Um, but then I felt it.
I thought he was drawing imitation also. I just want to compliment you. Sounded just like him.
Yeah. I've got a lot of practice imitating Ron.
Zuck used to come and he would speak at our startup school events and that drew a lot of people.
Yeah. He was a big draw. Startup world back then people didn't care so much like outsiders didn't care so much about the startup world this was before like in, you know, even now we think like if something big happens in the startup world it's a news story but why should it be if something big happens in the hospitality industry or the mining industry right or you know like the advertising industry it's not automatically some kind of big story for everybody it's only interesting to people who are in that industry, right? The trade press, right, might write about it. And it was so weird to me when startups became something that the general public would be interested in. I mean, I understand why they're interested in the private lives of movie stars. Why would they be interested in the private lives of startup founders, but they started to be, and they still are. It's so weird. But back then, this was before they cared. This was before that phenomenon had started. And so, we were just like our own little people working in the, you know, we were the computer club.
It was such happy times. And actually it was still happy when they did start taking notice but it became very unhappy once I don't know when the turning...
There's two things that happened right there was then the when the press started taking notice of Silicon Valley and then almost at the same time wokeness started like in around 2012 and then so they would be busting YC, you know, you'd have not YC but just everybody in Silicon Valley, you'd felt like there was this mob, this incipient mob ready to form at any moment, right? If you said or did the wrong thing, they would come and bust you and hound you. So, that wasn't fun. That combination of being visible plus wokeness. I feel like if I look back on my history of YC, the thing that not Hacker News, but what gave me the most heartburn in life was all the negative press around us and all the Twitter mobs that were based on lies.
Yeah. Yeah. You know, we're all nervous.
Peter Thiel is secretly running the company, you know.
Oh my gosh. Wait. Oh, are we, I think we have to tell that story. Peter Thiel story.
Sure. So, Peter Thiel. Oh my god. What year was this? What year was this? It was after Sam took over, but it was soon after Sam took over.
Yes, it was. He was good friends, 2015 or something.
2015. And he didn't he say, 'Hey, Peter, will you be a part-time partner?' I put it in quotes because he like did one office hours or something.
Okay. No, I'll tell you the whole story. Right.
Okay. So that we had this thing, you probably still have something like it. Um, we had this thing that I named mostly for alliterative purposes, part-time partners, which is people who weren't YC partners, but would come in and like for no money. They had no duties. They got paid nothing. They weren't really supposed to do anything, right? But experts in some aspect of the startup world would come in and they would do office hours with the startups even though they weren't actual partners. And we called these people part-time partners, right? And we had him listed on the website. And so Peter Thiel at one point was a part-time partner. And I got into a fight on Twitter with somebody um who I don't want to name because he's since apologized. I don't want to give him grief, but in the process of this fight, he picked up as if to throw at me like a brick the fact that Peter Thiel was a part-time partner. And when were we going to get rid of Peter Thiel? Because Peter Thiel had committed what was in 2015 an unforgivable like cancellable sin which was to be openly Republican. Um and so there was this mob formed to try and force YC to fire Peter Thiel. And this story grew to the point where Peter Thiel was supposedly like not, I mean part-time partners did nothing, right? Um, and the story was that Peter Thiel was like running YC or you know he was like the head of the chairman of the board or something like that and honestly...
Best friends with you. You'd met him once, right?
Yeah. The funny thing was so there was all this mob telling YC we had to fire Peter Thiel, right? Well, the funny thing was Peter Thiel like in the previous year had done like four office hours and so Sam had been planning to fire him anyway. Fire in the sense of say, 'Okay, you don't seem to be actually doing anything. Maybe we'll stop listing you as a part-time partner.' So Sam had been planning to fire him and then there was this mob saying that we had to fire him for being a Republican. And so it seemed like it would be a really bad idea. I told Sam about this was like you do not want to be the one who sets a precedent in this sense. You do not want to be the company that like fires somebody for supporting the wrong one of two political parties, right? You know, much as I dislike Trump, you don't want to fire somebody for that. And so, Sam, even though he'd been planning to fire him, now had to not fire him because that would have seemed like giving in to the mob.
Well, and to be clear, I think it was like, 'Hey, Peter, come help some startups when you have time.' And he literally didn't have time. He probably just didn't have time to do it. So, I'm not even willing to say like the guy was bad at being a part-time partner. I think he just was busy. So, he wasn't coming around. So, Sam was going to be like, 'Hey, dude, this isn't working out. Like, you know, we're going to take you off the website.' That was literally what it was going to be. And then it turned into this whole thing. And it was like, 'You have to fire this guy who's like, like I said, practically I think the narrative was he's practically running YC.' Well, no, please. Like, the guy was too busy to even be around. It was so frustrating. And by the way, I'm happy to stand corrected if I'm wrong about this, but like the whole partnership, I remember this day. We were all sitting around that big table in 365 and everyone was like, 'No, this is like,' it was emotional because we felt very attacked and it's like no way we're going to take him off the website just because these people are saying we have to. Like that's messed up. So...
Or more like the other way around. We're going to keep him on the website because they're telling us we have to take him off. Now that we're remembering all the like what it felt like to have a cancel mob after us and like that probably happened 20 times, right?
That was pretty stressful.
You know, the funny thing is wokeness has now decreased so much. I mean on Twitter especially, but all just in general that people can't imagine the [ __ ] we went through like growing YC like the years where we were really growing YC fast was also the exact same years that wokeness was really growing fast, right? And so and you know why YC got so much [ __ ]? Because we were the only like consumer brand in the venture business. You know, there's all these prestigious venture firms, but like ordinary people don't know their names. YC takes open applications from like essentially job applicants, people who are just graduating from college. So YC had to have like a consumer brand. We had to be known by everybody. We're not just like doing series A's, right? And so these people wanted to blame Silicon Valley and since they were outsiders, they didn't know who the participants were. The only people they knew in Silicon Valley were YC. So everything like all the problems in the world were Silicon Valley's fault. The only person in Silicon Valley that they knew was me. So all the problems in the world were my fault. You know...
So you remember how we started talking about this? Was it actually stressful? I'm thinking, you know, I bet we're forgetting something. It probably was pretty stressful because I remember feeling really beaten up all the time. And I'm trying to, I think it was more stressful because of all the attacks than the actual startup implosions.
Actually doing the job was much less stress than being attacked for doing the job in the press.
Right. Exactly. There was still a lot of stress doing the job too. I mean...
There was a good deal of stress. It's like you know there was a lot of fighting that we had to do behind the scenes on behalf of the startups because people would maltreat the startups and then you have to go basically it's like you know two kids have a fight and the parents go and talk to each other. If someone tried to maltreat the startups we would have to go and talk to their investors and say come on this isn't right. They have to back off.
Paul did that all the time. Do you remember the Adrocks? Do you remember them?
Well Antonio wrote a book about this. This story will be preserved. Um because yeah, they weren't a super super successful startup that this is one of the weird things about YC. People imagine because the famous successful startups like Airbnb and DoorDash and Coinbase are what become known later. They imagine that during the batch we were spending all our time talking to Airbnb and DoorDash and Coinbase. No, actually, you know, Airbnb was probably in a batch of like 25 startups and they got 1/25th of the time at any given time. I was talking to some startup nobody's ever heard of now, right? And so the amount of trouble caused by startups is not remotely in proportion to how big they end up being. In fact, there's probably an inverse proportion. So you spend basically YC means you spend all of your time dealing with these really naughty problems. Naughty like K-N-O-T-Y naughty not naughty naughty. Um you spend all your time dealing with these contorted miserable problems for companies that are never going to amount to much anyway.
But it's the nature of the business.
I want to flip this conversation because I'm feeling very negative. I want to talk about so many of the positive things that happened. Do you remember giving any advice you gave to some of our most successful startups that really helped them? Do you remember?
Oh, Airbnb. Airbnb. Absolutely. They would have died without YC. I mean, they were dead when they showed up. Um, though we didn't know it. Um, but like telling them to focus on New York, like just focus on that one initial market because you need to get this ball rolling, you know, that ended up making all the difference for the company. And so that's the best case scenario when you can tell a startup something that makes them take off and then they become huge. That's the dream situation. That's not how it usually worked. Usually you would tell people something and they would ignore you and then they would fail. That was the usual situation. That was the median case.
What was that famous quote that you gave to some startup? You're like, 'Thank God we fund, you know, 50 startups because you're driving into off a cliff right now.'
If you want to drive off a cliff, that's fine with me. Um, I think they're actually doing pretty well.
What was the company called? The founder's name was Brandon. And then I think he...
Yeah, I think it was Sift.
I think they're quite successful.
I didn't know you said that to them. Oh gosh, that story was printed in a book. That book, The Launchpad.
Yes, that will be preserved too.
Carolyn, were there any moments, this is a very naughty N-A-U-G-H-T-Y question. Were there any emails that happened that Paul sent that you were like, 'Oh jeez, oh gosh.'
Uh, well, probably, but no, actually, I don't think I thought that like, oh, this is going to get us into legal trouble kind of emails.
No, I never would say things that we...
Legal troubles, but just like...
Yeah. Just like, oh, that's really going to...
Oh, I'm sure that I was like, this is exactly like the harsh truth that this founder needs to hear. Like, I don't think I ever had the reaction like, oh, that's going to go over poorly.
Okay, that's good. I think people sometimes just need sort of the unvarnished truth from someone who's like, 'Look, this is what you need to hear.'
You know, I wanted them all to succeed. I wasn't going to be shooting them down.
Paul, no one worked harder than you, I swear, at helping startups who were failing. You would not give up on them. Other people would be like, 'Whoa, just let that one go.' And Paul would be like, 'No, I'm going to help them with a new idea.' Or you would just work so hard. This time will be different.
When did you start taking office hours sort of remember when you built the software that allowed startups to book office hours? The...
I wrote every bit of our software. Don't you realize like up until the point where I retired, I wrote every bit of our software except the JavaScript that did the up and down up and down buttons on Hacker News, which PB wrote because I didn't know JavaScript. I wrote everything. If we had software, I wrote it.
And I wrote every word on our website.
Yeah. Yeah. Our website was awesome. I love...
One of my memories um of that era is how you would always um well, you would frequently help startups with their name. And I think we might have talked about this in...
No, no, no. I know. So my question for you is do you have a name you're particularly proud of where you're like why don't you call yourselves this and they're like amazing...
Carolyn has no idea who Imgur is.
No I do. I remember the name. What batch is it though?
Oh god I don't remember batches. Find P-H-I-N-D.
Oh I remember that one. I feel like if I had a dollar for every time you were at instant domain search with founders during an office hours, right? Like right on your kitchen counter.
Because you need them to stand around and look over your shoulder. So I do it right on the island in the kitchen. So many names, so many startups names have been found at that one corner of the island.
So you're kind of famous for blowing up people's ideas like taking their idea...
Blowing up in the good sense.
In the good sense. Making them, here's how you can be the next Google.
Can you and you used to do this and some startups would take your advice, some others wouldn't. Can you remember an instance of that? Because you'd come home and you'd be like, 'Oh, wow. If you had been there, you wouldn't believe the idea we cooked up.'
Yeah. It's going to be the next Google.
You know, I can't remember specific instances because I would do it so much. I mean, I would do it like maybe one-third of the office hours I did. We would figure out how to make an idea much bigger.
Yeah, it happened all the time and still does happen.
Yeah, when I do office hours now, I still do that. And you know, I see these startups right in the beginning of the batch sometimes and so there's a lot of room to expand their idea. I want to also talk about your, you know, you're really good at working, you were good at working with the startups on their demo day pitches. Tell the listeners about sort of what used to happen before demo day with the founders.
I would write their demo day pitches. One of those little secrets we would do this thing called vertebrae office hours where we would figure out what were the vertebrae of their demo day presentation because the thing about demo day is you're up there with 50 or now like 300 other startups. And so the investors are not going to remember anything about your presentation by the time they get to the end. Their brain will be wiped clean. And so you really can only hope that they'll remember about five sentences worth of stuff in the very best case. And so you should decide in advance what those sentences will be. And those are the vertebrae of your demo day presentation. The rest is just filler essentially. And so we would try and figure out what are the five sentences that everybody ought to remember about this startup that would cause someone to invest in them. There are five sentences with a specific purpose. So like this is going to be a giant market. We are already solving the problem better than anybody else. It's ours to lose and when we get there there's network effects and so that means it's winner takes all. Right. That's a pretty, you're already wanting to invest, don't you?
Yeah. Sounds great. And...
That's the thing. Demo day presentation shouldn't be like fast-talking marketing speak. They should be almost like logical proofs of why this company's good to invest in. You follow if you agree with this step and this step and this step, they all follow from one and the other. And the end result is the expected value is really high. That's a good demo day presentation.
But I feel like you did more than just vertebrae. You would spend hours and hours and hours at YC and people would come in.
I would do the vertebrae with them and sit down and think, 'Okay, this is what the core of your presentation should be.' And then they would go off and make something that said what I said to say. And then I would sit there and they would practice before demo day. They would come to YC and take turns standing there projecting their slides and talking. And I would sit there in the audience essentially almost like a playwright saying, 'No, no, cut that, right? Say that.'
You were like a playwright. You'd be like, 'No, say this. Don't use that word. Say this.' It was so such specific...
That part down. We were always trying to get people to talk slower.
Yeah. See, you remember? I can remember. Wasn't there somebody who wrote 'talk slower' on their shoes so that when they looked down...
They would, it was like 'talk' on one shoe and 'slower' on the other or something like that.
I think there's a, I think online I think there's a recording of me sitting talking to Brian Armstrong about his demo day presentation. I think that might be online on YouTube.
