Letter #70: Paul Buchheit and Jessica Livingston & Carolynn Levy (2023)

Creator of Gmail and YC Founder & YC Head of People Ops | The Social Radars Podcast

Apr 04, 2023

Today’s letter is the transcript of a conversation between Paul Buchheit (the creator of Gmail, the founder of FriendFeed, and a Partner at YC) and Jessica Livingston (Cofounder of YC) & Carolynn Levy (MD, Chief Legal Officer, and Head of People Ops at YC).

Paul was employee #23 at Google, where he created Gmail. He then founded FriendFeed, which is known for creating the “like” button, which was acquired by Facebook. After that, he joined YC as a Partner. Jessica Livingston is the cofounder of Y Combinator and author of Founders at Work. At YC, she was nicknamed “The Social Radar” as the Partners turned to her to get a final check on whether they should fund the team if they qualified in technical abilities. Carolynn Levy is an MD, the Chief Legal Officer, and the Head of the People Operations team at YC, where she created the SAFE.

I hope you enjoy this conversation as much as I did!

(Transcription and any errors are mine.)

Relevant Links:
Jessica Livingston Compilation (173 pages)

Transcript

Jessica Livingston/Carolynn Levy (JL/CL): I'm Jessica Livingston, and Carolynn Levy and I are the social radars. In this podcast, we talk to some of the most successful founders in Silicon Valley about how they did it. Carolynn and I have been working together to help thousands of startups at Y Combinator for almost 20 years. Come be a fly on the wall as we talk to founders and learn their true stories. Today, we're talking to Paul Buchheit. Anyone who uses Gmail has Paul Buchheit to thank for it. Paul, who by the way, is also known as PB, not to be confused with PG, who is my husband and cofounder, Paul Graham, created Gmail in 2004 while he was employee number 23 at Google. Today, we're going to do a deep dive into the history of Gmail, including the fact that it might never have launched if it weren't for a leak to the New York Times. After Google, PB went on to found a startup called FriendFeed, which was a social media aggregator. It was acquired by Facebook in 2009. After that, he joined Y Combinator as a partner, and also became a prolific angel investor. And now, well, we'll find out what he's up to now as we catch up with him on the very first episode of The Social Radars. I am so excited, because today we have someone on the show, who is one of the smartest people I know in Silicon Valley, and someone from whom I've personally learned a lot over the years. I first met him back in, I think it was 2005, when I interviewed him for my book Founders at Work. He had just created a new email app called Gmail. Of course, it's Paul Buchheit. And then you and I were both lucky enough to work with him at Y Combinator. So we're psyched to be catching up with PB. So, welcome.

Paul Buchheit (PB): Wow, thank you. Thanks, Jessica.

JL/CL: Carolyn and I were sort of talking about your background. And we're thinking it is... it's so sort of long and varied in tech that we wanted to kind of do this chronologically. We want you to take us back to upstate New York. Isn't that where you grew up?

PB: I did. Outside of Rochester. Webster, New York.

JL/CL: So you grew up outside of Rochester, Upstate New York, where did you start learning things from? Because you're one of the most independent minded people I know. What signs were there that you were like this growing up?

PB: You know, I just started learning things, I guess, when I was a little kid. My father was an engineer, and he just loved building things and whatever. And so, you know, he worked at Xerox, which he didn't particularly like, but then he would come home and continue working on his own projects all the time. So, he was always building something or whatever. He would give us all his tools for Christmas, so I have a picture of me when I'm four or five years old, in my footie pajamas, with a bench vice. I was four or five years old, what I got for Christmas was a bench vice.

JL/CL: He was a mechanical engineer?

PB: Electrical engineer.

JL/CL: Electrical engineer. Okay.

PB: But he was interested in all kinds of engineering. So, he built and... you know, designed and built an addition on our house. He just liked building things, and he liked tools. He had sort of a compulsion for collecting tools, and giving them to people for Christmas. So, you know, from an early age, there would just always be lots of stuff around to take apart. He would find... he would find good things in the garbage, he liked bringing home things that they were throwing out at work, or there was an electronic surplus store that we would sometimes go to, or a ham fest, and we'd go get weird old equipment. Or I might just find something that he left in my workbench, or some old computers that he had in the basement that I would take a part and smash. So, there was always interesting things to play with and explore.

JL/CL: To take a part or build.

PB: Yeah, so from you know, pretty much my whole life we're always building things, or taking things apart, putting them back together again, repairing them.

JL/CL: Were you good student? Or was it one of those situations where your parents are always being told, "Paul's very curious, but he doesn't pay attention. He seems bored." Were you one of those kids?

PB: I was an okay student. My limitation was that I was a little bit lazy. And so, I would tend to maybe not always do the work, but I could do pretty good on tests. I think, in my high school, I just went to public high school, I graduated like 40th in the class or something like that. So I wasn't like the top grades, but I did okay.

JL/CL: And then you went on to Case Western, which is, a lot of engineers go there, right?

PB: Yeah, again, I wasn't necessarily like top student or whatever, but Case had a thing back then where they would give out scholarships based on SAT scores. And so the first thing I ever got from them was actually just like a one page letter that said, we will give you this much money if you go to school here. And my family growing up, we were always kind of like, my parents were very frugal. Like, we never went on fancy vacations, or fancy cars, or that kind of thing. They were very much not about spending money. I learned, growing up, also to have an aversion to debt or something like that. So the idea of spending a whole lot of money to go to some kind of college, just sort of... I grew up with the idea that that was not a good idea. And so with Case, I was able to get a scholarship that mostly covered tuition, and it's a pretty good school, and it's relatively close. So that was, that was how I ended up there.

JL/CL: And there were five kids in your family, right?

PB: Yeah. Yeah. So there were five of us.

JL/CL: So a lot of people do put through college.

PB: Yeah, exactly. Yeah. I mean there's no way they were going to send me to some... also, I probably wouldn't have been able to get accepted at a really top school or something.

JL/CL: Where are you in the birth order?

PB: It was funny, though, this one time, you know, when you... when you're applying to colleges, at least back then, they send you these mountains of junk mail. The envelope from Harvard came, and my mother's like, "Oh, you're not going to Harvard." She was very anti... just like, too elitist or something like that. I was like, "Don't worry, Mom, I'm not going to Harvard."

JL/CL: They still send all that mail. That's still a thing.

JL/CL: So you go to Case Western, and then that leads you to Intel, right?

PB: Maybe to back I was slightly, when I was a teenager, at some point, stuff was going on with computers. I was always interested in businesses and making money and stuff like that. Again, it's partially having sort of watched my father who was a very talented engineer, but worked at this sort of dysfunctional, big company who's sort of like... my father was Dilbert, or something, was kind of the sense of... and so, I kind of had this notion growing up that it would be better if you didn't have to work at some sort of terrible big company, and also would do little kind of money making ventures on the side. So I would like... I did door to door sales.

JL/CL: Oh, wow.

PB: There used to be this thing, I'm sure it doesn't exist anymore, but I actually went back and bought some old copies of this like Boys' Life magazine, it was like the Boy Scouts magazine. On the back side of it, there was always an advertisement for either the Olympic sales club or the American sales club, I think there were two of these things that you would sign up. And basically, they would send you like a catalog, and then you would go like door to door or to your relatives, whatever it was, it was this way that kids could go do sales and you would sell like wrapping paper or like greeting cards or like paper goods mainly. But were usually priced kind of like, you know, $4 to $6 price point, and then you would get to keep like $1 for each item you sold. And so, I would actually like ride around the neighborhood on my bicycle and I'm like, I don't know 10 or 11 years old or something like that. And it's like knock on people's doors and be like, "Hi, would you like to buy some stuff?" Or whatever. They go, "Oh, are you raising money for school?" I'm like, "No, I'm raising money for me." So, I did that until, unfortunately, one time, a guy like opens the door and then he is this... he wasn't even living there. He was house sitting there and all of a sudden this guy's got a pitbull and the thing jumps up and bites me in the face.

JL/CL: Oh, what? How?

PB: Harshly mauled by Pitbull. So, I had to like go get stitches and whatnot. After that my parents sort of thought that was not something I should do anymore. So that was the end of my door to door sales career. But you know, it's good to know how to sell, and we would do... we would sell things at hamfest. I would buy, you know, find good deals on some, you know, some sort of pricing and efficiency or something where I could buy something cheap and resell it.

JL/CL: What is hamfest? Is that a ham radio thing?

PB: Amateur radio. For some reason it's called ham like hamateur, I don't know. They're kind of like big electronics flea markets. At least they were back then. I don't know if they still exist or not, kind of got ruined by the internet. But it would be great because you just find all sorts of weird old stuff or you could sell. It's like a little flea market for nerds basically.

JL/CL: Got it. But you... it sounds like you are an entrepreneurial nerd.

PB: Yeah, so I mean we would, you know, go to sell things. My father like selling things at the ham fast or buying stuff. And like I said, he always liked to find a good deal, like if he could find something in the garbage and then repair it. That was like a big... he would find like old broken oscilloscope. So when he when he died, unfortunately died last year, and we're having to, like try to clear out the houses, just houses and houses full of stuff. But he had like 40 oscilloscope or something like that. Because he loved like these old Tektronix equipment and he would get an old oscilloscope and then debug them, figure out what was wrong and repair the electronics. And then he'd have them fixed. And in theory, he was gonna sell them or something, but he never did.

JL/CL: Did you find any treasures amongst all this stuff he collected over the years?

PB: It's kind of subjective. It's kind of weird stuff, right? Like, not everyone is interested in 40 year old Tektronix oscilloscope or something like that. And then, when I was a teenager, I kind of got interested in computers, I bought my first computer. I got a 386 DX 25 MHz, one megabyte of memory, and taught myself how to program. But my father said that the programmers... he was an electrical engineer, he didn't really know how to program, but he said that the programmers that work used C. So I was like, okay, I'll learn C. So actually at a hamfest, I bought a piler, you know, they really didn't have all the free software back then, but I got a copy of Turbo C for $11.

JL/CL: Was that a deal? Carolyn and I don't know if that's a deal.

JL/CL: Sounds like a deal.

PB: $11 for a compiler? It's good. You know, software, you used to have to pay money for it. And then I taught myself how to program in C. And that actually ended up getting a little job programming in high school for this guy that my father knew who had a machine repair business, and so actually, redid the control software out of pick and place machine. So, that was kind of cool. I was actually in a factory...

JL/CL: Oh, wow.

PB:...debugging and I built all the software for this machine that would take metal bars out of a hopper and put them in a grinder and grind them.

JL/CL: Wow, so while in high school, you did this?

PB: When I was like, 16 I was doing that.

JL/CL: Was this like a minimum wage job?

PB: I got $8 an hour. It was more than minimum wage.

JL/CL: That's more than the minimum wage.

PB: I didn't always get paid on time, this guy was not always great at paying bills on time, but I'd eventually get paid.

JL/CL: So then you went to college, and did anything important happen there in terms of your interest in working, building things or working...?

PB: Yeah, I was always interested in building. One of the nice things about Case was that they were pretty early with the internet in terms of having good internet connection and so forth. So one of the things I was looking at when I was looking at colleges, was I was interested in internet. This was like early 90s. And so I kind of knew that this stuff existed, but this is before the web had really taken off. And so, I can remember, I went to one college, I was like, "Oh, do you have internet access here?" And they're like, "Let me check." That was the state of things and she's like... she went and asked someone she said, "Oh, I think we have Ethernet. Is that the same as internet?" I said, "Not exactly." People didn't know the difference between Ethernet and internet at the time. And then when I went to Case, I was like, "Oh, do you have internet?" They're like, "Yes, every room on campus is wired with high speed fiber." I was like, "Cool." So they had actually... it was one of the sort of most forward looking things they had done is that they had actually, in the early 90s, completely did every room on campus, every classroom, every dorm room, in the fraternity houses, every single room had this fiber jack on the wall. Kind of went as a weird technology because it turned out to not really be a technology that took off. We don't really use fiber today. But they wired up the whole campus with fiber.

JL/CL: Were there other like like minded people like you there that were into this stuff?

PB: Yeah, I mean, the people were interested. So when I showed up, it was like fall of '94 when I started. Things were just kind of taking off. So, I also got interested in Linux when I was in high school. So again, at a hamfest, I bought a Linux CD. And this was like the very early days of Linux. It didn't work very well, at all. So, I installed Linux on my computer and was learning... I was interested in learning Unix and that kind of stuff. So when I got to campus, I started looking for people who you know, would know something about Linux. The first day, I actually met someone in my dorm and he's like, "I'm trying to compile the new Linux kernel." I'm like, "Oh, you're my new friend." He was showing me how to figure out how to get Linux onto this weird network because we had special like, you had to have a weird Ethernet adapter to get on the network. It wasn't standard equipment. So yeah, there was that, and then there was an ACM programming team. And so then, I showed up for that. So I was excited. I didn't really know anyone who could program besides myself, so I was excited to kind of beat other people who I could hopefully learn from. So, I showed up to the programming competition. And we actually did pretty good. We had a good programming team.

JL/CL: And then what led you to Silicon Valley?

PB: Well, so I was interested in startups, you know, I was again, kind of like finding some way to not just be a cog in some giant machine, and the idea that you could create something new and exciting... I had actually tried, at one point in college, to make my own web based email service. And this is actually before Hotmail even came out. Summer of '96, I had accidentally ended up not having a job. I maybe only applied for a couple thinking that they would hire me, and then they didn't, so I was just hanging out in Cleveland, in this basement underneath a sub shop, and I was like, I'm just gonna make my own... I had this idea that I want to be able to access email anywhere. And at the time, you know, it always lived on your computer. So people would be like, I need to go back to my dorm room to check my email or something. In Linux, I didn't have that problem, because I could just like telnet into my computer, but for all the Windows people or whatever, email was this thing that lived on your computer. And so, I had gotten very excited about Java and the possibility of mobile code. Code that would be downloaded from the internet. And Java was very exciting. As history turns out, it never delivered on that promise. But JavaScript, of course, did deliver exactly that, which is that the browser downloads the code from the server, and then it executes locally. And that way, you can actually build a really great experience that doesn't require constant roundtrips to the server, but kind of like I ran into my problem of a somewhat limited attention span. And so I got bored with that project and started watching lots of Dukes of Hazzard and going for bike rides and stuff like that.

JL/CL: Oh, no. You could have been the first sort of email.

PB: That's fine.

JL/CL: Well, wait, did you have an AOL address at the time? Or did you have like a...?

PB: No, I had a university address.

JL/CL: Okay. Okay.

PB: Never used AOL. So yeah, I was interested in startups. And then when I graduated college, there was no Y Combinator. There's no startup anything. Like, you look on the internet, I mean, there was no Google. The internet was still very young. And so it wasn't like I could just go find a listing of startups. Like it didn't seem to be anywhere to begin even, but it did seem like they were generally located in California. So just sort of geographically, you know, Silicon Valley, I could like locate on a map. And so my strategy was like, okay, I'll just sort of moved to Silicon Valley and then presumably just bump into a startup or something like that.

JL/CL: Good plan.

PB: Yeah. How hard is it gonna be? In Intel, I had a friend who was a year ahead who had taken a job at Intel, and they had this rotation program that sounded kind of cool. It was basically just sort of a series, almost like a series of three internships in a row. You kind of get to try out different jobs within the company. That sounded kind of fun, so I took that job with Intel. And actually, just coincidentally, my second rotation there was in the mobile marketing division, or group, or whatever. And one of my co-workers in that group was Susan Wojcicki.

JL/CL: No.

PB: While Google was at her garage, she was working at Intel in the same group as I was.

JL/CL: No way. Is that how you sort of...?

PB: No, it was just a coincidence. We didn't realize that until later on. I mean, I didn't know her super well, it wasn't like we chatted a lot or something. I was there for about a year. And Intel was, I just found it, it wasn't bad, but it was just sort of like draining. Like, I would find myself, I'd be at work and I'm just like, oh, I'm so tired. I think I'll go home and take a nap. And then I'd go home and I'm like, you know, I'm not really tired anymore.

JL/CL: Do you watch Dukes of Hazzard?

PB: Yeah, I just found it to be an uninspiring and kind of draining environment, which is very gray, and just like slow, and it just wasn't fun. It just... there was just nothing fun about it. And again, I was still really interested in Linux. And back then, you know, big companies were still very suspicious of Linux. It was like this cancer or something. They didn't like Linux. So the two things I was interested in was startups and Linux. And so, basically I was like, okay, I'm just going to apply for some... try to find a startup and I can't say that I did like a really in depth research. I just made a list of what I could find, which was only like 10 companies.

JL/CL: Where'd you even find them?

JL/CL: Yeah, where were you living, and how did you find?

PB: Well, I lived in Sunnyvale, and working at Intel, I would always read Slashdot. So Slashdot was... that was really the site. And so, Slashdot would cover Linux all the time because it was like a fan's site, practically. But Google would show up on Slashdot periodically, because Google was built on this Linux cluster, which was like, what could be more exciting than like this cluster of Linux machines. And they had this Linux specific search. So, I would see Google show up on Slashdot, as well as, there were a handful of other, basically like, Linux companies would show up on Slashdot. So that was how I would know about them. I forget, it was like 6 or 10 companies, something like that, not like a huge number. So I just emailed my resume to them. And ironically, when I first emailed Google, my resume had bounced. Their mail server was misconfigured. They weren't able to receive email. It was sort of a temporary thing. I had to resend it the next day, but it was kind of funny. The first time I applied to Google, the email bounced.

JL/CL: Oh, wow.

Yeah. And so then I never heard back from most of the companies, but I interviewed at Google and they asked, actually really smart questions, which I think like, one advice for... I sometimes give to people talking to... interviewing at startups or whatever, is actually what is the quality of what they're asking you, right? Because that's how they're finding people. So, I did interview at this one other startup that was doing Linux stuff. And like, their questions were just dumb. It was kind of like trivia. What are the layers of the OSI stack, or something like that. At Google, they were actually asking some pretty intelligent questions. Like I remember, Urs asked this, like, sort of hypothetical, okay, you have a server that's running slow. How do you figure out why? Or whatever, right? And kind of like talking through the process of debugging, like, okay, is this like... are you limited on CPU? What's going on? Is it an IO problem? Tracking down where is the bottleneck, and like thinking through system level problems.

JL/CL: Tell people who Urs was.

PB: Oh, Urs was the... I think he is still sort of one of the main vice president of engineering, but he was like the original VP of Engineering, and he ran engineering for a very long time.

JL/CL: Because this is now back in what, 199... What?

PB: This is 1999.

JL/CL: 1999. And was it just Urs in the interview? Or did Larry and Sergey...?

PB: At the time, I basically interviewed with most of the company, and there weren't very many people. So yeah, it was a series of different engineers. But this was before the Series A even, and there was only maybe seven or eight engineers or something. It wasn't a very big team. So yeah, interviewed with all of them. Yeah, Google was the only offer I got, so, you know, that made it easier.

JL/CL: Are any of the companies that rejected you still in business?

PB: I mean, I wasn't rejected. They just never replied.

JL/CL: Never replied. Yeah.

PB: I think one of them was probably like Red Hat or whatever, Red Hat. They're around in some form. None of them are terribly successful, probably Red Hat is the most successful. I don't know the other ones. There was like VA Linux. I don't know what happened to them. I think they died.

JL/CL: You interviewed at one company and it was Google.

PB: Yeah.

JL/CL: I mean, honestly. Oh, I mean, I'm glad no one else contacted you back. I'm glad that Google was it.

PB: Yeah, it makes it's easier that way. You don't have a lot of hard decisions to make.

JL/CL: So you were like number 22... employee number 22. Or am I wrong?

PB: 23 I think it is.

JL/CL: 23. Okay. So very, very early on. What was it like early on? I mean, you obviously joined a couple weeks after the interview?

PB: Yeah, I ended up joining... for some reason, it ended up being maybe closer to a month after the interview. They do that quite... but anyway, yeah. So in the intervening time, actually, it was kind of interesting. So I interviewed, and then they called me back in to make the offer. I think that was when Larry and Sergey, like, presented the offer. And then they're like, Oh, and also like, watch the news or whatever, because it was like, I think the next day was maybe the Series A announcement or something like that. So I got my offer, right, basically, I think it was like the day before the Series A announcement. And then I also had to try to figure out what it even meant, because like, the offer, you know, I'd never, I didn't really know about stock options and stuff, and of course, being kind of like, you know how, it's just like the people on Hacker News are, you're like paranoid, somehow, I'm gonna get screwed on this or whatever, right? Where I'm like, what is this? Why are they priced so low? You know, because like, my original Google shares, I think, like if you account for all of the stock splits that have happened between now and then, I paid like two cents a share for my...

JL/CL: Oh, wow. And I think it would have paid less if the series they had been later. Like if you've been hired earlier because the series they probably impacted that.

PB: Oh, yeah. I mean, if I had gotten hired, you know, a couple months earlier, I'd be much better off, obviously. I mean, it does impact you, you know, because you only get a little, you know, it's just a little fraction of a percent of the company, right, like, but it turns out a small fraction of a percent of Google is actually still pretty good. Yeah, one of the not obvious things about startups is it's much better to have a small slice of a big company than a big slice of a small company. And it's counterintuitive, but the reason is, you can only have as much as 100% of a company. So you know, the orders of magnitude there are limited. But the number of zeros in the valuation can go pretty high. So it's actually easier to make money, I think, like, you know, when you have a really hugely successful company, it makes a lot of people pretty wealthy versus a small success that makes maybe one or two people.

JL/CL: Yeah, which we've seen, obviously, at Y Combinator, we've seen that. So you're at Google, what are you working on when you first join?

PB: Initially, actually, they gave me two projects. One was to build an ad system, because they didn't have any ads. And then the second was shopping search, because like, they also thought that would be sort of a cool thing. Well, one of the things that was kind of funny about joining Google was like, there was just, the level of ambition there was really, it was sort of intoxicating, right? And Larry was just sort of crazy. He was just like, "We're gonna do this, we're gonna do this, we're gonna do this." You know, it was like, here we are just this little tiny company, and it's like, okay, we have a new employee, let's give you two projects. And then actually, I was working on that with Marissa. Marissa had started, I think, the week before me.

JL/CL: Marissa Mayer.

PB: Marissa Mayer, yeah. So, we kind of shared a corner of the office there and started working on these projects. I ended up... the ads thing ended up being kind of a lot of debate about what the ads product should even be. So a lot of people don't really know anymore, but it took four iterations before Google actually had a successful ads project. Like, the very first ad product that came out was this thing that would recommend books on Amazon, and then collect the affiliate revenue, and we made like $20. It was both like... it was both like, problematic in that it would make offensive suggestions, like someone would search for something, and then it would come up with a book recommendation that you would look at be like, ooh, no, that's not really cool. And then on top of that, it never really made any money, because most of the time, people don't really want book recommendations when they search for something on Google. But anyway, I ended up focusing more on the product search thing. And so, I built out like a product search, I mean, all of this was happening over the course of a few months because it's a startup, so everything goes quickly. But I ended up building this product search for a while. It never was quite good enough that we launched it. So, that never launched. Though along the way, while I was working on that, one of the things I ran into is just the fact that I'm not very good at spelling. And then, looking at our query logs, I discovered that that's also true of most of our users. And so, I added the 'Did you mean' feature.

JL/CL: Oh. You did that?

PB: I added the original 'Did you mean' feature because I'm like, this is such an easy way, because they were trying to figure out how to improve quality. And I'm like, the easiest way to improve quality is just to spell your words correctly.

JL/CL: Oh, my God.

PB: Like me, and half the other humans can't spell very well. So, it's a really easy way to boost the quality.

JL/CL: You created 'Did you mean'. I'm in shock, because I was gonna ask you, you created such distinctive things within Google like AdWords, Gmail, and the whole 'Don't be evil' thing. But you also did 'Did you mean.'

PB: Yeah, well, I didn't create AdWords. AdWords was...

JL/CL: I thought you did.

PB: No, I made the first version of AdSense. I sort of... AdSense was the contextual targeting, and that was kind of like a side project of Gmail, because initially, I built it for Gmail. And then people were like, "Hey, this would work on the rest of the web, too." Yeah, the 'Did you mean,' I did... again, I would... I'm pretty good at making a lot of times a really quick and dirty version of something, and so, it was actually while I was building the product search, I was just like, this is kind of dumb that it doesn't have a spell corrector, so I just threw that in. But actually pushing it out to the main web search took a little bit longer because you need to make it... I had just done some sort of quick hack with like Aspell or something like that. So then, we had to kind of license something. The initial 'Did you mean' was not really that great. And so, I was working on making a smarter 'Did you mean,' because it was just kind of a dumb spell corrector that we had licensed from someone and then I did some code to try to make it less dumb. Because for example, it would want to correct words like TurboTax. It would say, "Did you mean Turbot Axe*?" Like a stupid spell corrector would do back then. So at the time, I was working on the spell... this was maybe a year later, when I was working on a better spell corrector, I decided to use it as an interview question. So when I was interviewing people, I'm like, "Well, how would you build a good spell corrector?" And kind of like talk them through it. And most engineers, most people are just terrible. They have no idea. Some people, I would say, two thirds of people are are just bad. One third of people are kind of okay. But then there was this one guy who gave this amazing answer, where he was like... he already went past what I was thinking, like in the interview, he already was... his thinking was already ahead of me. And so I was like, "Wow, that's great." So we hired him, this guy Noam Shazeer, and he started... it was actually a few days before Christmas, or whatever, and I set him up like, "Okay, here's where I am in the code. See what you can do." And I went away to visit family for a couple of weeks. And when I came back, two weeks later, he had built the thing that we know now, the 'Did you mean' feature, which is like the world's greatest spell corrector that had ever existed. It was his first two weeks on the job.

JL/CL: Oh, my gosh, that's amazing. Then when did you start getting involved with Gmail? Wasn't it a side project for you there?

PB: No, not exactly. A lot of times people think it was a side project. But it was actually sort of like an assigned project. I did have a lot of side projects, but that wasn't one of them. Summer of 2001, inside of Google, there was kind of this reorg thing where Larry got frustrated because he thought people weren't moving fast enough. So his solution was to get rid of all of the engineering managers. They had hired a number of managers, and it was organized into these groups, and each group would kind of have its own thing. But they weren't working on whatever he thought was important. So his idea was, we should just not have managers; engineers should just have projects. So instead of having groups with managers, there'd be projects with engineers. And so all of the engineers were given a project, or there'd be a couple of engineers on a project, and so, I met with... you know, when they were kind of meeting with... I sat down with Larry, and then Wayne Rosen, who is the VP of Engineering at that point, and they're like, "We want you to build some sort of email product." So that was kind of the spec: some kind of email thing.

JL/CL: Some kind of email thing. Okay.

PB: Yeah. So initially, I was still kind of wrapping up another project. So it was, to some extent, part time to begin with, because I was finishing up the original Google Groups. So we had bought... I don't know if you know what Usenet, you know what Usenet is, right? The old... the original Internet Message Board.

JL/CL: Yes, yes.

PB: And it's an archive that goes all the way back to the 70s. It's the oldest Reddit, essentially.

JL/CL: And Google had bought that?

PB: Well, we had bought... it's an open thing. But what we had bought was this site called Deja News that had the largest Usenet archive, but they were... it was a startup that was dying. They'd run out of money, and their systems were kind of shutting down, and so we got like a good deal on it. And then we would be able to own their archive. But we didn't want to try to maintain their code, because it was some completely different code base written in, I think, Tickle, I want to say, like, a weird language, and we had a deadline, because their whole... their data center was shutting down or something like that. So we only had like one month from the time that we acquired the company until their servers were gonna shut down. So we had... it was this really great project, because we had a really hard deadline, which was that we had to launch in a month. There was no wiggle room on that. So, a group of us very quickly built this Google Groups, the original Google Groups project to index and serve this Usenet content. And so, I was finishing up on that when I started Gmail. And so the very first version of Gmail that I built was built based off of just Usenet code that I had been working on previously. So I mentioned earlier, the first time I tried to build an email project, I ran into this problem where I was writing a bunch of code for a while, and then I got bored and distracted. So, one of the things I had learned over the years is that, just my own psychology, I have to always be keeping myself sort of engaged in the product. And so for me, what I really like is to always be launching things. So with Gmail, I was able to build the first version in a day and then launch it. So what I did was I just took this Usenet indexing code that I had from the previous project, and then I took my email, my mailbox, and I just wrote a script to convert the email to look like Usenet data. And then I just shoved my email into this Usenet search. And then I just launched it. I just sent it out to engineering, like, "Hey guys, I built this email search. Let me know what you think." And so yeah, that's like the first version, day zero of Gmail, when I launched this product.

JL/CL: What did people say?

PB: People said, "Well, it's kind of useful, but it'd be better if it had my email instead of yours." So, that's like a feature request. It was all just iterative. So I built the first version in one day and then I got a feature request, which is like, people want to search their own email, not just my email. So then I'm like, okay, that's like a now-I-know-what-to-do. So like, I went to work and I made this thing that would crawl through the home directories and suck up people's email and index it and then partition it so that it would just get one person at a time. And then I had like... that was like version two. So I launched it again, "Okay, now I have email search with your email." And so then people would be like, "Okay, this is cool. And now I want to reply to an email." I'm like, okay. So then I went in, and added a reply feature, and just iteratively would launch a new version every day or two. And then just constantly, people would complain about, now they want to send email, or whatever, they want an address book. And just kept adding feature by feature or, you know, it's too slow, or, you know, whatever the problem is, and try to keep making the product better and better. And we continued through this process. And there are a lot of complicated ups and downs in there as well, but ultimately, we came up with this goal that we'd be ready to launch to the world once we had 100 happy users. And so... and this is something we still kind of use sometimes in advising startups, because it's actually a pretty good way of working on a product, because it's easy to get people to tolerate a product, but you want them to actually, some of them, to actually like it, but you don't need all of them to like it. And so I would do... I actually embedded, just like a little one question survey, inside of Gmail, which is like a little box that would pop up and say, "Are you happy with Gmail? Yes or no." And there's just like, yes or no, and I would get a list of names then, and I could go through the list of no's. I would just go talk to them.

JL/CL: Because they're all part of Google's engineering team.

PB: Yeah, it was just inside. So, I would just go down the hall and be like, "Okay, says you're not happy with Gmail. What's it gonna take to make you happy?" And so just, you know, it's basically door to door sales. It's like, what's it gonna take to get you to be a happy user? And for some people, they'd be like, "Well, basically, it needs to be exactly like Outlook." And I'm like, "Well, that's never gonna happen." So, cross this person off the list. But other people, you know, it's something much more attainable, like they just needed one feature, or there's something they didn't like, or whatever. So basically, there's some people who are easy to make happy, and some people who are difficult, and so you just focus on the ones that are easy. And then eventually, we made it to 100. And then sometimes, you know, eventually, I would get the Outlook people too, because back then, Outlook had this bug where once your mailbox, on disk, would reach, I think, two gigabytes, it would corrupt the data and lose everything.

JL/CL: Oh, no.

PB: So then I had the only copy of their email. Outlook would crash and destroy their copy, I would have the remaining copies and then they would start using Gmail.

JL/CL: When did you decide to start calling it Gmail? Was that immediate or...?

PB: The idea was around very early on. But at the time, we actually... we were into codenames because we were trying to keep things secret. Like they didn't want Yahoo to find out. Which was inconvenient, because sometimes Yahoo would be there. Like one time at dinner, I was sitting there with like, Larry and Sergey, I think were there. And then Dave Filo from Yahoo was sitting there with Larry. And so I'm kind of sitting there at dinner and people kept coming up and asking me questions about... so we had a codename. The codename was Caribou. But people keep asking me basic product questions. And finally, I realized I need to just like get up and leave, or whatever, because I can't answer everyone's questions while I'm sitting next to the founder of Yahoo. So yeah, it was called Caribou internally. Actually, getting the domain was difficult. We just barely managed to get it.

JL/CL: Why?

PB: I don't know. I think we maybe just procrastinated. We ended up actually not quite getting... there's a second domain with a different spelling that we didn't manage to get in time. Maybe they have it by now. I haven't checked. T-m-a-l-e, it was a porn site.

JL/CL: Good gracious. Can I just as an aside, say, listening to you tell the story and saying that you were the kind of person who would keep going, you always wanted to be launching and you would talk to users, you're like the dream founder. I have to say, like, if I were fund-- I wish I could have funded you.

PB: That feedback loop is what makes it fun. So if I work on a project just by myself without that loop, I sort of lose motivation. And so, it's also part of the problem that have, fast forward, like, by the time I left Google in 2006, the release cycles on Gmail were like, you could check something in, but it's gonna be months by the time it makes it all the way out to the world, or whatever. And I just like fast iterations. So when we had FriendFeed later on, I would like, write the code, I would push it live. And then if it's good, I would check it in. I like to see... I just like immediate feedback loops, where you put something out there, and then people love it, or they hate it or whatever. Then you learn and you go forward. This ponderously slow movement where you do something and then you don't see the effects for many months, to me is just the demotivational.

JL/CL: Can you talk at all about... didn't... wasn't it true that Gmail almost didn't see the light of day? It almost did not get launched within Google?

PB: Yeah, I mean, so there were definitely different opinions. There were a lot of people who thought the project should be scrapped. So, there were basically like a couple of objections. One objection was that it was too weird, because it was quite a bit different. Like this one particular VP, he was like, "I really love AOL mail, you should just clone AOL." Like, I don't know how to respond to that. One segment of people said, this thing is really weird and no one's ever gonna get it. Because we did do, I mean, some kind of unusual things, the conversation view was very innovative. It has key bindings that I borrowed from VI, which is a very unusual thing to do to put an obscure Unix text editor, borrow ideas, put them in a webpage. And the way that it behaved as an application was unusual at the time as well, because it was the first of that thing. So even doing the whole thing in JavaScript is very controversial. When I first wanted to do... started doing a lot of things in JavaScript, people were like, "You can't do this. If you write this in JavaScript, Microsoft is just gonna change the browser to break it." Because they were all very afraid of Microsoft at the time. And a lot of the executives like Eric Schmidt had come from Sun. And Eric had spent his career getting crushed by Microsoft, so he was very paranoid about that. But I was like, "No, I really don't think they can just like break JavaScript. And if they do make some small change, we can just change the code on our end to work around it." And the thing that's actually really funny historically is it turns out, not only did they not change the browser to break Gmail, they actually did an emergency push to IE6 at one point, or a high priority patch, to make Gmail work better. Because Gmail was trashing... breaking the garbage collector in IE6. So Microsoft actually made an improvement to IE6 to better accommodate Gmail. The JavaScript thing, yeah, people were opposed to that, because it was really kind of like bleeding edge at the time, the browsers didn't quite work. So one of the things I was afraid of that might keep us from launching is I had trouble crashing the browser a lot. So I had to learn a lot of tricks to not crash the browser, because it was really easy to push the browsers too far. It was IE6 and then, like, Firefox 1.0, or something like that. And they could just barely handle what we were doing. And I'd have to be careful, because if you push too far, it would just trigger some bug and the whole thing would die.

JL/CL: Oh, wow. I never knew that.

PB: So yeah, there was the technical side of it. At one point, there was a push to actually just go back to being totally HTML; that kind of distracted us for a while. And then, by the time of launch, one objection was that it's too weird. The other objection was that it's not weird enough. It's too conventional. Like, we're Google. People expect something really revolutionary.

JL/CL: Wow. Was Gmail not revolutionary? I mean, I think it was.

PB: It depends on your opinion. Some people thought it should just be like more different. Other people thought it should be more normal. Some people thought the gig was too much. They're like, "Why are we doing a gigabyte? Why don't we just do 100 megabytes? That's still more than anyone else." I was like, "No." Because Yahoo was offering 4 megabytes. And Hotmail was offering 2 megabytes. I was like, it's so much better if we come up with a 1000. That would be real. And then the other problem we were running into is that actually, like the engineering was a little bit bumpy. Still, not everything worked really well. So we were building on top of web search infrastructure that had been... just had different constraints. Like, if it would encounter an error, it would just throw away the data, because it'd be like, well, we'll just get it on the next crawl. So actually, modifying all of our infrastructure to not lose data was really hard. We had weird performance problems. The code was like pretty shaky at the time we launched, and I was very afraid of losing data. And so one of the strategies I took was that we would just keep multiple copies in different locations in different formats and stuff. But consequently, it was like every email ended up getting stored like 13 times or something weird like that in different systems. And the reason for that was that way, when there was a bug in one part of the system, I could replay it from a different part of the system to fix it. So we had this... the way the data was stored was this log structured file system that we built. And the cool thing about that was that you could basically replay--you could fix bugs, and then replay history to undo, so sort of like retroactively fix bugs in the system and some cool things like that to try to make it more responsive. But then, yeah, it almost didn't launch. It actually... the thing that, it might not have made it out the door, except someone inside of Google actually leaked to the New York Times that we were launching this email product on April 1. The thing was that it wasn't quite ready, but there was then this impending story. And so, we actually rushed it out early to try to beat the New York Times stories. If you go back and look at the press release, it launched April 1 at midnight, UTC. So it actually launched the night before. It wasn't even April 1 in the United States yet. We sort of pushed it to UTC so that we could get a jump on The New York Times. The only problem was that we hadn't actually finished writing the code yet. We actually launched Gmail before it was written. So I didn't get a lot of sleep those first couple of days.

JL/CL: And did you like finish in the next... the next day?

PB: Yeah, we worked really quickly. So when we put out the press release, it was such a rushed thing, the DNS didn't even resolve. So if you tried to go to gmail.com, there was nothing there. We launched with less than a landing page. If you're not embarrassed when you launch, you know, you waited too long, so we definitely didn't wait too long.

JL/CL: And what happened when you launched? Then The New York Times article goes out, and do people immediately embrace it or...?

PB: No, no. Well, so part of the problem was that we didn't really have capacity either. So there was this chicken and egg problem, where I couldn't get hardware because the product wasn't really ready to launch and the product wasn't really ready to launch because we didn't really have hardware. So the only machines we had were, I think it was like 300 of these old, I want to say like Pentium IIIs or something like that, that no one else in the company wanted. They had less than a gig of memory, even, I think, like 256 mega, they're like these old, low memory machines. And we had it coupled with the inefficiency of the system. We had enough capacity for 10,000 users when we launched, which was enough for Google employees and their friends, basically, which is also part of the reason we had the invite system. So the invite system was also one of these things where they said, "Well, you can't launch until you're ready for everyone to show up." I'm like, "Well, that will never happen." Because it's so hard to scale something like this. So I was like, "Let's do this invite system. Well, that way we can limit the rate of growth." And Larry was like, "Well, no. That's like planning for failure." I was like, "Okay, well, in that case, it's a growth hack. This will make it viral!" Somehow... somehow saying that it was to limit our growth, it was like... it was a no go. But if I repitched it as making it viral, then it was an acceptable strategy. The real reason for the Gmail invite thing was simply that the entire system would have collapsed otherwise. We just didn't have capacity. But yeah, Gmail didn't even exist when we launched it. To make the situation more difficult, my brother had just died on March 9. So yeah, you know, the last night I was there with him, actually. He was only half alive at that point, kind of struggling to breathe. And I was on my laptop doing some account integration code, like trying to finish up some stuff. So you know, when he died, I took off for a couple of weeks. We had a funeral here, and a funeral back in New York. And so, I was completely away from everything for a couple of weeks. So when I got back, things were not all going exactly according to what would have been an ideal schedule. So there was... it was a little bit of a shitshow back on the team. So we really weren't quite ready to launch. And like I said, actually the code wasn't quite all written yet. So after we launched there on April 1, I had to work to... it was actually... the remaining code was the part that integrates account creation and everything. When we launched, there were zero Gmail accounts in existence. So, I created the first Gmail account. Honestly, I should try to go find the timestamp. It was sometime later that day, so I created the very first Gmail account. They're like, okay, that works, push the code. You're checking the code...

JL/CL: Was it paul@gmail or pb@gmail?

PB: helloworld@gmail.com is actually the very first. And so then, that worked. So then, I sent an invite to myself. And then, my personal email account is like the second ever Gmail account. And then that worked, so that I invited the rest of the team and the rest of the company. But it was only later that day that people inside of Google even had Gmail accounts.

JL/CL: On the day that it launched?

PB: Yeah, it had already launched, and no one, including myself, had an account.

JL/CL: Wow. Oh, my gosh, that's pretty crazy. I don't think I knew the extent of the craziness. And I certainly don't think I realized that it was so soon after your brother passed away.

PB: Yeah, I mean, it was a pretty... it was a pretty out of control launch. They actually, after we launched, implemented a whole bunch of new policies to make sure that a product like Gmail never launches again.

JL/CL: Well, did they ever figure out who leaked it? Like that's pretty weird. Or was it not weird?

PB: I think they did. It was... it was someone in a different part of the company. So I... it wasn't someone I know.

JL/CL: How did like Larry and Sergey respond? Like, let's just say, a week later, were they like, "Great job, Paul." Or were they...?

PB: Well, so initially, we had this problem. We launched this thing, but no one had access to it, right? So eventually, we got around... once, you know, day two or three, we finally invite reporters, people get to try out the product. But initially, some people weren't even sure if it was a real thing, because we launched on April 1, with sort of implausibly good specs. So like, if you go look at Slashdot, people were like, "Come on people, this is an obvious joke. You can't... you can't fall for this, there's no way they're really giving away 1000 megabytes for free."

JL/CL: Oh, my God.

PB: You know, because if you remember back then, people working inside of a corporation or something that had like an Outlook account would usually have maybe a 30 megabyte quota, or something like that. And so actually, interestingly enough, one of the groups where Gmail really took off early on was reporters. Like reporters all loved Gmail. Because their official email address that the New York Times would give them or whatever, only allowed them to save 30 megabytes of email, which, if you're a reporter, you really want more than that. And so they loved the fact that they could just keep everything and that they didn't have to deal with these ridiculous quotas. And for years and years, email administrators had gotten really good at making excuses for why it's not possible to give people more than that. "Oh, you don't understand. Email is really hard. You can't just like give people hundreds of megabytes of storage." And then of course, that excuse falls apart when we do it for free.

JL/CL: Yeah. Oh, my word. I had no idea about some of these details. Remind me, PB, I met you when I asked you to be in Founders at Work, which you did. And I think you were just leaving Google around that time.

PB: Yeah, it was around that time, though. Yeah. Because I was interested... you know, I had reached out to PG because I was interested in Y Combinator. So again, you know, on Slashdot, which is where I get all of my news, I'd seen the original announcement for the summer founders program. And I thought that seemed like super cool. Like, I just... I just thought that was like the best idea ever, because it sort of spoke to my own experience, which is that, I was interested in startups and I didn't go to like Harvard, or Stanford, or whatever, where it's like, "Oh, yeah, my roommate started Cisco", or something, right? I feel like if you go to one of these environments, where everyone is doing it, it's like, no big deal. But my only strategy for finding startups was to physically move to California. I didn't know startup founders who came from my school or whatever. And so, I thought the idea of just having this open program where anyone could apply, and then you would just teach them everything that they need to know and connect them, and like, anyone could start a startup, not just someone who's already super connected. To me, that was, I thought, just such a great idea. And something that is really powerful, because you're opening it up beyond just like the chosen elite or whatever.

JL/CL: You were our target market when we first launched. It was like, smart programmers who might not have any clue about the business side of startups, but they want to build something and they can now build something, because all they really need is a computer. Come study with us, basically, for the summer. So I remember we invited you to speak for that first summer.

PB: Yeah, it was actually... not for the first summer. So I contacted... I think it was in the fall... That summer, there was still a lot of stuff going on. But I... in the fall of 2005, what happened actually was that you announced that you were going to do the second batch in Mountain View. And so, that's close by. And so, I reached out and just introduced myself and like, "Hey, I love what you're doing. Let me know if I can help out", or whatever. And so, that was how I ended up talking to you. Because then, PG introduced me to you.

JL/CL: Okay, and we must have invited you to speak at the... during the...

PB: Yeah, yeah, so then I spoke at the second batch, with Wufoo and all of them.

JL/CL: Oh my gosh, that seems like so long ago, with the paint drying on the walls of the building. So, you leave... why did you leave Google? Like, what was your plan?

PB: So I mean, I didn't exactly have a plan. But I... you know what? What really happened was that our first child was born in 2005. Severely premature, she was 100 days early, and almost didn't survive. Had lots of surgeries, and so forth, which is why the whole summer of 2005, I was actually in the hospital again. We spent the summer of 2005 in UCSF's neonatal intensive care unit. So I had taken a seven month leave from Google. And when I went back to Google in the fall, I was actually really excited, because I always loved working and launching things. And just like, I like the excitement of movement, and getting things done. And I got there and I was so excited, I was having trouble sleeping the night before, and I got there, and it was that thing where the life just drained out of me. It was like I was back at Intel.

JL/CL: Aww. Too many people?

PB: My computer had stopped working, and I tried to get a new one. And then I ran into sort of... some sort of IT bureaucracy where they gave me a new computer, but actually it was broken, that I spent time debugging it. And then when I finally figured out that they had misconfigured, the DMA or something, they were like, "Oh, yeah. We knew about that. You just have to fix it." I'm like, "Why are you giving engineers computers that you know are broken?" They're like, "Well, we don't have time to fix it", or whatever. I just immediately got this feeling like I'm stuck back at Intel or in some big company.

JL/CL: What happened in those seven months? Like that's a seven month time period?

PB: Personally, it was just like the frog boiling water kind of thing. You know, you become more aware of it. But also, the company was growing, so I had end up in some meeting, people I don't know. I don't know. I kind of had this realization, which may or may not be accurate, but I kind of felt like, okay, I can either... if I stay here, I need to become an effective big company person, because a lot of what needed to be done was a little bit more, kind of politicking, like, the last thing I did was manage to get more hardware for Gmail. We were always running out of hardware because we were growing. We were resource-contrained, but we were running the system at close to 100% capacity, which is a terrible thing to do. It was like grinding the engineers, because they were constantly scrambling to keep the system from dying, because we were so close to the limit. And so, I was always trying to get like... the last thing I was doing was just constantly trying to get more hardware. And so finally, like my tactic was, I announced, we're gonna shut down new signups or whatever, and just sort of an internal bluffing thing, or whatever, to try to embarrass them. And then finally, we got approved for $100 million in new hardware, something which admittedly is a lot. The product could have been actually a lot more successful, but we were hardware starved for years. But anyway, I kind of had this realization, I could either become a really effective big company person, or I could just leave. So, I chose the latter.

JL/CL: And then when did FriendFeed come about? Like a couple years later?

PB: Yeah, so I ended up just kind of taking off some time. And you know, we still had this baby to take care of and everything. But yeah, I was talking to my friend Sanjeev, who is like the second person on Gmail. And we'd get together, and he was kind of interested in starting something. And so... so, we decided to start something. But we also... my biggest thing is always like, you just got to find really great people. And so, we were keeping an eye on other people who we thought were really great, who we thought might be interested in leaving Google to join the team. So we didn't have a specific product in mind, we actually had more of a culture in mind. We wanted to make a really great engineering company where it was just really fun to work. Basically, I was just trying to make a job for myself. That would be like... I wanted someplace that would just be like a really fun place for engineers who like shipping things, you know? Like there's different kinds of engineers. There's some engineers who are just like arguing on email all day, you know? The people who spend all day arguing about politics or whatever. They are not the ones we want. But like people who love to like ship products. And so, some other friends from Google, Bret Taylor and Jim Norris, they left, and they were actually looking to start a startup as well. So we started kind of having lunch with them and talking to them.

JL/CL: They were Google Maps, right?

PB: Yeah, so Bret was the original PM on Google Maps, and Google Maps, people, you'll see, again, on Hacker News, like, Oh, Google just acquired it and launched it. But the thing that they acquired was actually a Windows app. The original... the startup that they acquired had a Windows app that was the mapping thing. It wasn't even web based. So, they converted it to be a web based thing. And then, the story I always remember with Bret that I think is super impressive, is that, internally, the Google Maps wasn't very good. It was just like this horribly slow and clunky thing. And the person who had been building it was one of these... one of the engineers, I think, had fallen in love with XML. It had way too much XML bloat. Anyway, Bret got frustrated with this, and over a weekend, rewrote their entire JavaScript for Google Maps. And he made it... if I remember correctly, like a third the size and 10 times as fast. They show up on Monday and he's just completely replaced all of the code. That's the kind of engineer Bret... and the thing is, it isn't bad code. Like, it isn't like sloppy stuff. He'll crank out really high quality code very quickly. So he's... he's like... he's not even a 10x engineer. He's like, 100x, or 1,000x engineer. He's just really good.

JL/CL: That's the person you want to be at your company that you're starting.

PB: You definitely want that on your team. So yeah, we managed to convince Bret and Jim to join forces. And so we launched FriendFeed. Then, I think it was August 2007, if I'm remembering it right.

JL/CL: And tell me a little bit about FriendFeed. I mean, I think it was you who created the idea of like, following people.

PB: No, I mean, I think following was already around. So FriendFeed, it kind of started off as, it was, I guess, sort of the Cambrian era of social. There were like a million different social things at the time, and they all had these feeds, so you can have like an activity feed off of videos that you post on YouTube, and like things on Twitter, things on... so Twitter was already out with a follower graph. And so there were all these different RSS feeds and things like that. And so, the original idea was that you could kind of like pull all of them together, all of this social activity from different places and aggregate it together, and then create this environment where you'd see what your friends are doing across these various social products. And then very quickly, we added comments, and then actually, we also... I think it was like October 2007, we launched the "Like" button. So FriendFeed is actually the original like button. Though, this claim is somewhat disputed by Facebook's version of the story, which is that they had created it but just forgot to launch it. The actual story is that they had been at a very similar feature internally called "the awesome button," but then they never launched it because it was lame.

JL/CL: Because it was not... not awesome.

JL/CL: Because it was not awesome.

PB: Yeah, calling it the like button. So after we launched the like button, they sort of like dusted off the awesome button.

JL/CL: And remind me, didn't Facebook sort of copy everything you did once you guys launched it? Or am I wrong?

PB: Yeah, I mean, so they... I don't know about everything, but they definitely kept a close eye on us. And Facebook, you know, I didn't really know too much about them at the time. I sort of hoped that they would be too successful to be nimble. But that's definitely not the case. Like, they were very willing to move fast and break things and whatever. They were a very aggressive competitor. And so one of the things that happened, you know, being in the social space for that period of time, is you start to sort of realize, like, which people are really good and which people are clowns. So like, at one point, we were talking to MySpace. Like, we flew down to LA to meet with the people down there. And it was just like, it was just a disaster. Like, they went silent on us for a while, and then all of a sudden, he gets back to us. He's like, "Okay, I think we can do a deal, but you're gonna have to throw a lot of people under the bus." And like, just weird... just like weird clown show stuff. And you know, Twitter was always in constant disarray. And, you know, worst is like going in to meet the team at Facebook. They just clearly understood what was going on so much better than everyone else, that it was just like, no one else stands a chance, you know? People at Google still had no idea what they were doing with social. So it just became readily apparent that basically Facebook was gonna win it, because no one else was even close.

JL/CL: Well, to take you back for just a second, why did you start talking to these companies? Like did you know that you didn't want to run it independently or...?

PB: No, I'd never wanted to sell. Initially, I mean, I was trying to create a place that was a really fun place to work. Facebook, again, kind of to their credit, had been basically reaching out to us from the start. Like, "Hey, we should work together. Hey..." So they were always wanting to talk to us. And so they would ping us pretty frequently. We, with FriendFeed, kind of reached the point where our growth had kind of plateaued. And so the problem... ultimately, you know, our product, we never really had the right product, is basically the right answer. We were too... we were too similar to Facebook and Twitter, who were both much larger and more successful. And so like, we had really great growth when Twitter would have the fail whales. So like when Twitter was down, we would have great growth. But that's not a good long term strategy. Like eventually, they figured out how to keep the service up. And so, we needed to either pivot or sell. Looking back, you know, one of our pivot ideas was basically to build Slack or something, because one version of the product that worked was we had these things we called Internal Rooms that were basically like Slack channels. So like, we could have done that or whatever. But given that we were sort of at this inflection point, you know, we were like, well, okay, maybe we should, you know... Facebook reached out to us, we're like, okay, well, if you're so interested in acquiring us, make an offer. And they're like, "Well, how much do you want?" And so we had this whole ridiculous back and forth, and we're like, well, you're the ones who reached out, so, you know, you have to do it. And, so they're like, "Well, if we make an offer, you know, we'll just lowball you." I was like, well then, so don't lowball us, alright? Okay, okay. So there was a whole back and forth, it was kind of a crazy negotiation. So then when they finally did come make the offer, it was actually this really funny meeting. They come over to the office, and then we make tea or coffee, and sit down, and they slide the offer over, and we look at it. And it, we were like, they sort of low balled us. So we're like, "Okay, well, thank you for your time." That was it. Then we sit there quietly, and then, Chris is like, "Can we finish our tea?" I'm like, sure. And so we just get an awkward couple of minutes, silently waiting while everyone finishes drinking their tea.

JL/CL: Like, there was no even discussion. It was... it was such a lame offer...

PB: It was just like, "No, okay, okay, thank you for your time." And they did a terrible job of pitching it too. They didn't pitch it. So anyway, so yeah, just goodbye. And then we just sort of stopped answering. We just stopped responding to them. And then, but we were also talking to Twitter, and actually had gotten pretty close to selling to them. I had sort of like, concerns about Twitter. And I'll tell you, this is sort of an interesting story. When we went in to talk to them, it was a different experience from Facebook. There were like 10 people in the room, or something like that. And they're telling us about the structure of the company. And it's kind of like... it's almost like engineering is like one of 10 things. And the thing that's hilarious for me now is like, with sort of the Elon situation, it's all coming right back to the surface. It's never been an engineering company, right? And so for the first time ever, he wants to make it into this engineering-centric company. So, it should be fun to watch. You know, I... it was a really weird thing. I got this, like, just sense of hostility from their VP of Engineering, like I don't think he liked us. But the part that ended up being, really, I thought, funny in hindsight, was a year later, every one of those 10 people we met with, were gone. The entire management team had turned over within a year. It was just like a very unstable place. So I don't know all the drama of what was going on behind the scenes, but it was just like... it was like weird stuff. So yeah, I sort of like, then had a little bit of this epiphany where I'm like, well, if Facebook is obviously going to be the winner of this whole thing... well, first of all, it would just be interesting to see, you know, because I'm a sort of a curious person. I'd seen Google and I wanted to kind of know what another one of those things looked like. And then also, maybe some level of arrogance. I thought we could be like a positive influence or something like that, like, Oh, we can then make Facebook good. So yeah, I then responded to Zuck on a Friday. I'm like, "Okay, can you chat?" Because we'd been ignoring their emails for a few weeks. And so I went over on a Friday afternoon, and it was just like, okay, look, there's three reasons we can't sell to Facebook. First of all, everyone hates Facebook. Secondly, your offer is terrible. And third, I don't want a job. So I'm like, alright, here's where we stand. So we spent about four hours or something walking around Palo Alto and kind of talking, and then eventually, reconstructed the deal in a way that I thought was like a really good deal for everyone. And so, he basically came up with better terms, higher value, and some way of revaluing Facebook to make the deal more attractive. And so then, basically did a handshake deal that night. And then on Saturday, we discovered that we didn't have a lawyer because our lawyer had gone fishing. Literally, because we'd just had some weird little lone proprietor or something like that. And he was gone, he was offline. And here we are trying to sell our company. And so, we scramble all Saturday to try to find a lawyer, and no one wants the deal because it's like, you're gonna have to work really hard and then you're gonna have like a permanent conflict with Facebook or whatever. But we finally found some lawyers by like Saturday night. And then all Sunday, they're working like crazy to get preliminary docs in place, and then we actually sign the documents Sunday night, announce the deal on Monday. So, we went from... from Friday morning, we woke up with nothing. I email Zuck. We have a handshake deal Friday evening. Saturday, we look for lawyers. Sunday night we signed the deal. And Monday morning, we announced it to the world.

JL/CL: Carolynn, are you stunned and impressed right now?

JL/CL: Yes, and I feel terrible for that legal team. But honestly, like, what was the rush?

PB: You know, just like speed. We believe in speed.

JL/CL: It was just your thing. Okay.

PB: There's this like, sort of rule of thumb that one of our advisors gave us. I like to tell a startup sometimes is just like, every day the deal doesn't close, the odds that it closes goes down by 10%, or something like that. Like, if you're gonna do the deal, do it now, right? Like, every day that goes along is an opportunity for something to go against you.

JL/CL: That's true, that's true. I just have to say, PB, I miss whole of your keen insight. You just like distill these little tidbits of knowledge that you drop that are so profound and important. Like startups should listen to you... just make a book of like, all of your advice, from start to finish. But back to Facebook, was Zuck super persuasive?

PB: Not super, but, you know, like I said, I was impressed by what they had achieved. They were just clearly more competent than everyone else. And so I'm like, okay, if we're going to sell our company, it'd be interesting to go someplace where the people actually know more than we do. Like, we could have gone back to Google and lead one of Google's, you know, additional failed social efforts or whatever. But like, why would that be fun to just go like, fail at this thing again? Like we could have gone and built... what did they call it? Taco Time, or something. Buzz was what they were working on at the time. Like Google had a series of these failed social efforts, like Google Plus was just like the last of several.

JL/CL: I don't even remember what they were. So the FriendFeed team goes to Facebook. So you're now at like, the second greatest tech company of that generation. And what's that like for you guys?

PB: I started to regret my decision, honestly, when I found myself in a windowless room in the basement. I'm very sensitive to lighting conditions. And I was like, what happened? Last week, I had my own startup, and everything was wonderful.

JL/CL: And I had sunlight.

PB: Yeah, I had sunlight. Now, I'm in a basement. You know, it was interesting. So I, again, I was just kind of curious about it. So I was trying to figure out what sort of role I would even want inside of the company. So I decided to just go around and just kind of interview lots of different people and try to understand what was working and what wasn't working, because you start getting all these different groups inside of the company who are unhappy with this or that. So, I spent a lot of time just talking to different engineers and getting the history of the company, and what they thought was good or bad. So I spent a lot of time just like learning about it. So it was actually pretty interesting; it has a very different history from Google. Google hired really smart, talented people from the start. Facebook had a really hard time hiring technical people early on, and the founders were kind of amateur. They weren't like great engineers, or anything like that. I remember talking to the one guy who had, blanking on his name, but he was kind of the first sort of like, adult engineer hire that they had, who had to try to convince them to use revision control. Like they didn't use source control, they were just sort of like... they would just sort of edit on the Harvard server, and then, if it looked good, they would SCP it to the other servers. And they didn't bother to do any revision... they didn't use, you know, obviously, they didn't use Git, but they didn't use anything of that sort. So it was a very different... the company had a very different history. And it was very, very hacker in the sort of, like, messy hack kind of sense of things. But they really valued launching things quickly, and hacking, and that kind of stuff. So it was it was kind of interesting. Very different culture from Google, though.

JL/CL: Really?

PB: I thought so.

JL/CL: Is one better than the other?

PB: You know, it depends on what you like, but I... maybe it was just because I spent longer there, but I generally liked the Google culture more. Facebook... yeah, I don't know how to put it in terms that I'm comfortable stating. I just think like, there was a higher degree of like, technical ability, and just, I think just sort of the systems thinking at Google was more interesting. You had a lot of really good leaders, people like Jeff Dean, who all Google engineering really respected, because he's just like, really smart, really thoughtful, like you had a lot of that kind of leadership that Facebook didn't have as much of that history behind it. And Facebook had more of a kind of like, their version of hacker is a little bit closer to Trickster. They were clever at getting people to do what they wanted them to do. Facebook is a very good mousetrap.

JL/CL: Well, interesting.

JL/CL: Yeah. How long did you stay there?

PB: So I was there for about a year and just like, never really found anything that I was super interested in working on, versus just like, not working. To me, especially, I guess, you know, partially in light of my brother's death, I always kind of had this thought in mind, like, is this really what I want to do with the rest of my life? And so the idea of sitting in a windowless room when I could be, for example, outdoors, I have to ask myself, why am I doing this? Like, why is this what I'm doing with my life? Because I have no economic need, obviously, like, why is this the thing that I'm choosing to do? And so if I, like, really deeply believe in what I'm doing, then I can sort of rationalize it that way. But if it's just kind of like, well, it's so that I can get people hooked on Facebook slightly better, whatever, I'm like, ehh. So, the only feature I ever launched actually at Facebook was the account export and download feature, which became popular a couple years ago. Long after I left. So if you ever do the thing where you download your Facebook as a zip file, that was my one and only addition to Facebook.

JL/CL: I've done all my photos, but not all the stuff. Just...

PB: There's like a thing where you can click to download, you kind of have to dig through, but you can click to download a zip file that contains all of your Facebook.

JL/CL: Just wrap up your whole thing.

PB: Yep, turns it all into a zip file with a static HTML. You can click around.

JL/CL: So PB, then, you came to Y Combinator, right?

PB: I did, yes. Yeah, actually, the funny thing was, when I was thinking about leaving Facebook, one of the ideas I had was, at the time, Reddit was sort of, like, withering on the vine or whatever, because it still belonged to Condé Nast, but it wasn't getting the resources. So they were like begging for money to keep their servers running. I'm like, this is really stupid. I wonder if they would just sell Reddit to me. So I initially went to PG and I was like, "Do you think they would sell me Reddit? You know, like, if I asked?" And PG is like, "No, you don't want to buy Reddit." He's like, "You don't want that. You know what you ought to do, you ought to come to Y Combinator."

JL/CL: Oh, well, I can't believe that he dissuaded you from making Reddit an offer. But I'm delighted that he said come to YC.

PB: He's so scarred from running Hacker News that...

JL/CL: True.

PB: Correctly identifies running a community as being sort of like one of the most painful things you can do.

JL/CL: I can't imagine you would have enjoyed that, Paul.

PB: Yeah, I think he was right, actually.

JL/CL: So when you joined YC, we had already had a relationship with you, because you used to invest, you'd come to our demo days and used to invest in a lot of YC startups. So, you were like family when you joined. But did you like working at YC?

PB: Yeah, it was fun. It was very different. It was kind of more... the other realization, I guess I had towards the end of FriendFeed and being there at Facebook was that actually, you know, I have kids, and maybe I want to spend time with my kids, and maybe... maybe like actually, working 100% of time on a startup isn't really the thing I want to do with my life. And so you know, YC was great, because I could be involved with startups, but at the end of the day, none of it was my responsibility. Like the really great thing is, you can stop... you talk to a startup and I'm like, This is what I would do. Here's what I think you should do. You can follow my advice or not, I don't care. You can... right, if you... you know, this is why we fund a lot of startups, right? Like, some of you are going to fail. You know, it's not my responsibility to make you a success. And so that's... that works better for me.

JL/CL: Now, I think the numbers have proven out that you are amongst like the best startup picker at Y Combinator, right?

PB: I don't know. Someone did some statistical thing at one point. That was a long, long time ago. I don't know if it's...

JL/CL: But it leads me to ask, I mean, obviously, at YC we're exposed to so many different startups. We see everything. And you did so many angel investments on your own. Have you gotten better at picking startups? And what does getting better consist of?

PB: I hope so. But it's hard to tell because as always, it's driven by the outliers, so it's easy to... if you just miss one outlier, then you fail. Yeah, I mean, the biggest thing... it always comes down to the founders. And we say this over and over, but unfortunately it's true. The trap I fall into most often is wanting something to work. Even though I know the founders not really that great, but I like... I really want it to be a success. And so that's my number one source of failure, I think, in terms of investing, is that I invest in something that I know the founder's not really that great, but I just sort of like, convince myself, maybe if I just like believe in them harder, it'll work.

JL/CL: Yeah. Maybe it'll work, maybe they can pull it off.

PB: It like, literally never worked for me. Like, the founders just have to be good. No amount of me wanting them to be successful.

JL/CL: Which has been your most successful, financially, angel investment?

PB: I think probably DoorDash. Stripe will be, presumably, someday, but you know, they're not... they're not public, so we won't count that yet.

JL/CL: Oh, interesting.

PB: I'd have to go back and look if there's anything closer, but yeah, I mean, certainly, you can't beat large numbers. Again, that's the, I think that one of the things that's counterintuitive to people is that there's a tendency to fixate on percentage ownership. But it's never the percentage of ownership that makes you a great, really great return. It's the really great return that makes a really great return.

JL/CL: Right. What other key advice would you give to people doing angel investments? Like that's a... that's a big one, it's better to be invest... you know, be an investor in the most successful company than have a bigger chunk of one that fails, obviously. But what other good advice is there?

PB: You know, I mean, I think like, it helps if they're making something that you personally want. I don't know if that's just coincidence, or whatever. But, you know, this kind of goes back to actually just having a great product is like, is there really evidence that people want this thing, because the most common reason a startup fails is that they didn't do what it says on the t-shirt, which is make something people want. And there's a million and one ways that you can kind of, like, fool yourself into thinking that you're making something that people want. As a YC partner working with the companies, right, when they come into YC, I always felt like, my number one thing was to force founders to have the realization that no one wants what they're making. Because most of the time, that's the state of things. And so, the thing I would always push them to do is like, okay, go get LOIs, go... somehow, you got to go get your customers to make some sort of sacrifice on your behalf. Because, like, if they won't even sign a non-binding piece of paper that costs them nothing to sign, how are they ever going to write, you know, a 100k check, or whatever it is you're going to ask them to do down the line? If they really, truly, desperately want the thing you're making, they should be willing to jump through some hoops to help you out or to demonstrate that. This, kind of again, goes back to wanting something to be successful. Sometimes, sort of the other category is like there's a pet idea where you think it's like a cool idea. But it sort of turns out that it's not a thing that anyone really actually wants in reality. The counter example of this would, of course, then be back to DoorDash. So DoorDash was the thing I was looking for. I had been wanting to fund DoorDash. In fact, I had invested in Caviar as well, prior year, but they only were operating in cities. And I lived in the suburbs. So when DoorDash interviewed, I was like, if we fund you, will you deliver food to my house?

JL/CL: I remember this. I remember you were really gung ho.

PB: Their initial delivery area was Mountain View and the part of Los Altos that includes my house. But you know, it's a thing I use constantly. I, you know, it's great. We go out, we order food for the kids, or, you know, I order lunch, something like that. So it's like, I have a regular need to eat.

JL/CL: You were a happy user.

PB: Yeah, I'm a frequent user.

JL/CL: As a fellow suburbanite, we appreciate DoorDash all the time too. Teenagers, all the time. In fact, we got an email from a high school that said, "Please remind your students they are not allowed to have DoorDash delivered to school."

PB: That's funny. I wonder why.

JL/CL: Yeah, my 13 year old, when we come back to Palo Alto, he's on my phone ordering DoorDash for like every meal possible. What else do we... it was... what other... was there any last minute questions we had for PB?

JL/CL: Well, this is not a fully formed question. But I have noticed that you've said the word fun like a whole bunch of times as we've been chatting all this time. And so, and I think I... because you've explained really well kind of what you mean by that, and I think it's satisfying curiosity and being fulfilled in what you're doing and having your work be something that you ship fast, whatever, what do you do now, like, what's your life like now where you get that sense of fun?

PB: Oh, that's a good question. Not having enough fun.

JL/CL: Come back to YC.

JL/CL: I was gonna say. We can make arrangements.

PB: You know, you're right. I spend a lot of time with my kids. You know, like every morning I actually walk to school with... so I'm waking up earliest I've ever woken up, I get up at 6:20 to walk to school with Cam at... so it's like two and a half miles to school, and then...

JL/CL: You walk to school? That far?

PB: Yeah, we walk to school, so every morning, we walk and have a nice discussion. It's great. You know, it's like, that's precious time, right? Like, kids grow up. So I try to spend time with the kids. So yeah, every morning, we walk to school, have a good talk about, you know, she's very curious about the world and likes me explaining whatever. We talk about the ego, and strange things. And Michael Jackson, we talk about Michael Jackson a lot. And then I run home and then, just depending on the day, I usually go to yoga at some point... So I exercise quite a bit. I run a lot, I go to yoga a lot.

JL/CL: You look like you have not aged at all, unlike me who COVID was very unkind to. You look like you haven't aged at all.

JL/CL: I said that to you last year at your birthday. I was like, damn, PB is not getting older.

PB: I try not to. Ageing is a downer, so... we try not to age. I'm working on a handstand. Eventually, I'll have a really great handstand. It's been about 10 years. I think a couple more years, and I'll really have it.

JL/CL: Well, gosh, it... I have to say, this has been the most fun I've had all day. Like this has just been so much fun catching up with you and chatting with you. And like I said, you have... you've had so much experience and have had so many interesting things happen. So, I appreciate your time. We'll hopefully see you in California sometime!

PB: All right.

Wrap-up

If you’ve got any thoughts, questions, or feedback, please drop me a line - I would love to chat! You can find me on twitter at @kevg1412 or my email at kevin@12mv2.com.

If you're a fan of business or technology in general, please check out some of my other projects!

Speedwell Research — Comprehensive research on great public companies including Constellation Software, Floor & Decor, Meta (Facebook) and interesting new frameworks like the Consumer’s Hierarchy of Preferences.
Cloud Valley — Easy to read, in-depth biographies that explore the defining moments, investments, and life decisions of investing, business, and tech legends like Dan Loeb, Bob Iger, Steve Jurvetson, and Cyan Banister.
DJY Research — Comprehensive research on publicly-traded Asian companies like Alibaba, Tencent, Nintendo, Sea Limited (FREE SAMPLE), Coupang (FREE SAMPLE), and more.
Compilations — “A national treasure — for every country.”
Memos — A selection of some of my favorite investor memos.
Bookshelves — Your favorite investors’/operators’ favorite books.

A Letter a Day

Discussion about this post

Ready for more?