Paul Buchheit - Startup Investor School Day 2

Meeting founders and making decisions is way more of an art than a science. And as Dalton says, unfortunately in this game, I think you have to lose some money before you can really become an expert, as much as anyone is an expert at answering the questions you need to answer to make an investment. The next person who's about to come up has more wisdom and knowledge about this than perhaps anybody I know.

Before I introduce PB, however, first for the live streamers, I apologize for the problems we're having with the live stream. As I said, this is the first time we've done this here in Mountain View, so we're bound to have kinks. The good news is that everything is going to show up on this site, and once it shows up on the site, it will be more immune from networking issues. They're less likely to be ours and more likely to be yours. At that point, all of the slides will appear on the site within 24 hours after the session, so I believe it's—all of the slides from yesterday are now up with the transcripts from yesterday at investor.startupschool.org. We'll do our best to make the live stream work as well as possible.

So the next investor, the next presenter, excuse me, he's also an investor. Obviously, it's Paul Buchheit, who has been a partner at YC with me as long as I've been here. He's been at YC as long as anyone here. I think it's probably seven years he's been hanging around YC. For even Loveland, he has also notably the inventor of Gmail, which sucks for me because I was part of the team that invented Yahoo Mail. So we have a lot of good conversations. He's still my friend. I think he also created a company called FriendFeed, which he sold to Facebook. He knows a lot about a lot. I guess that's the best way to talk about Paul. But one of the things he does know a lot about is investing. Here's a quote from Paul that is actually an important one. I would think about it a lot as you think about investing: I think the best kind of morality is a product that people like so much that they just want to tell people about it.

So here's one of the smartest and best investors I know in the world, Paul Buchheit, to talk about that.

"All right, so thank you, Jeff. That was probably studying the expectations too high. So, as Jeff mentioned, I've been involved in startups for a while, probably about 20 years now. I've been investing for the past 12 years. I've invested in hundreds of companies. I've seen thousands here at YC. I've interviewed thousands of founders. And so last year, I kind of came to the realization that I needed to stop and actually think about what I had done because I spent most of those years just making somewhat haphazard decisions.

So I spent a bunch of time trying to actually figure out what are the patterns that gave rise to favorable outcomes for me and what are the patterns that gave rise to unfavorable outcomes, with the idea that in the future, I can maybe do better. So really, this is like something I more did for myself in my own investing process. It may or may not work for anyone else, but it's what I intend to do.

And so the structure of this is I'll tell a series of stories, first of favorable outcomes—things that went good for me—and then some unfavorable outcomes—things that did not go good for me—and then finally followed by some conclusions about how I plan to invest going forward.

So first, there's this idea I'm not limiting myself to good ideas, and this actually started out as a joke about 20 years ago, but then I kind of realized it was true. A lot of my ideas are not good, but it's also really relevant to startups because oftentimes the bad ideas, I mean, the good ideas look bad. And the reason for this is that if the idea were obviously good, like the opportunity wouldn’t exist. Big companies are not that dumb. They're capable of executing. They have lots of resources; they have market power. Like, if they thought there is this, you know, ten billion dollar business sitting there, they would just go do it themselves.

So necessarily, for it to be a good startup, it kind of has to be something that inside of a big company would get shot down for being, you know, stupid or unimportant. And actually, the example I like the best is this company—I don’t know if who recognizes that logo—that's the original Google logo. When Google was a research project at Stanford called Backrub, it’s Larry Page's hand in a flatbed scanner. And so, you know, Google started out as this research project at Stanford; Larry and Sergey were PhD students there, and they actually wanted to continue being PhD students and like finish their degree.

So they went off and tried to sell Google to one of the big internet companies at the time, so Yahoo, Excite, Infoseek. They made the rounds to all these companies; they wanted a million dollars for Google, and fortunately, everyone said no. No one thought it was worth a million dollars. They thought search was unimportant. They thought it was just like a commodity, sort of a flavor of the month thing. They're like, 'Oh, you're hot now, but next month, it'll be, you know, some other startup, and we'll just hop from one to the next.'

So again, like if this opportunity were obvious, Google wouldn’t exist. So necessarily, it has to look kind of bad. And so I'll start my story at Google. And the reason here is I think I really want to emphasize the absolute importance in this game of being extremely lucky. So the good news is you're all accredited investors, supposedly. That means you're already among the luckiest people on earth.

But I also like to think about this a lot because, you know, being a startup employee is actually a lot like being an investor, except you can really only do one at a time. And so I've asked myself, like how did I end up at Google? Like what took me there? The first thing is just that they actually had a product that I used. I was a very early user, and so I just knew about it. Like they had made something people want—namely me—they made something I want.

And the second thing, the second reason is Linux. So I was really, really, really into Linux. A friend in college once put it, 'Shut up. Not everything is about Linux.' So when I went looking for a startup to go work at, really my only criteria was that they'd be doing cool stuff with Linux, and so that was how I ended up at Google. They were building these clusters of Linux machines. There’s like nothing could be more exciting than that.

When I interviewed there, the people were really smart; they asked good questions. I interviewed at another startup where people asked dumb questions, and you could really tell the difference. What I didn't know was that it was a good business. I actually expected that Google would get squashed by like Alta Vista or something. I didn't go in thinking I was gonna make money. I went in thinking I would learn something about startups, and in fact, I did learn some stuff, which was good.

So I had a good time there. Jeff mentioned Lost Gmail, but eventually it was kind of turning into a big company, and I'm interested in startups, so I kind of got the itch again to get back involved in startups. So I'm looking around, and then in early 2005, I see this thing—the Summer Founders Program. And this is sort of funny because I don’t know a lot of you probably don’t know this. The very first batch of Y Combinator was pitched as like a replacement for a summer job. So the idea was college students, instead of going to work at Google as an intern, we just come to YC and start a company.

Missile Eli K. really bad idea to most people in 2005. In fact, their original lawyers they tried to work with refused to do the work. He said, 'This is terrible; we're not going to do it.' Investors all thought it was a joke. They were like, 'No good company is ever gonna come out of this.' So how did I end up here? How did I end up connected to YC? And the answer is that Paul Graham had written a lot of really good essays, and so I had read a lot of his stuff. He was really smart; I didn't know him; I thought he was really smart. He had really good insights, and I thought this was a really intriguing idea—not necessarily a good idea, but like intriguing, interesting.

And for me, I think interesting kind of includes this element of unpredictable. Like if I know what the outcome is, then there's no point in getting involved. It's the things where I don’t know the outcome where I can learn. So basically, I just cold email PG and it was like, 'Hey, you know, can I help out?' And so I started hanging out here, got to know the first, or the first batch that was in this building, which was the winter '06 batch. And there was about 12 years ago. I actually since then haven’t met with him; I invested in every batch. I'm a repeat offender.

But my very first angel investment was in a YC company called Wufoo. And here's what the internet thought about it. For those of you who can't read it—which is probably most of them—one person says, 'Is anyone keeping a list of all these new Y Combinator startups? Seems like it might be entertaining a year from now after all the hype has died down to go back and look at it.' I mean, sure, it seems like it's written in Python and it uses XML HTTP requests, as being confused with stuff like, 'I have a business plan, and my product satisfies a demand that consumers actually have.'

And I like this because it's both kind of like making fun of YC and Wufoo, and it was very characteristic of the thinking at that time. And in fact, you know, Wufoo never did IPO, but my investment on that returned like 44X when SurveyMonkey bought them, and the founder, Kevin Hale, who's extremely talented, ended up being a partner here at YC some years later and no doubt there for help earn me more money by helping our startups.

So, you know, there’s both an immediate return and then there’s like the long-term return. The following year, I invested in this YC company; it’s called Justin.tv. The idea was that Justin here was gonna stick a camera on his head and then live stream it to the internet—that was like the idea. And at first, I was like, 'Are you joking? Like now we're gonna do it?' I'm like, 'Awesome!' Because, again, this is like really unpredictable. What's gonna happen when you start live streaming your life 24/7?

And I guarantee this is not the idea that would come out of Google, right? Like you're not going to see Larry Page with a camera strapped to his head 24/7. And nor should it come out of Google. So again, this was just like an unpredictable thing. And in fact, the company struggled for quite a while, which I mention because someone asked earlier, Dalton, you know, what if the company is kind of going sideways? Should I— I mean, I think the implication was should I like kick the founders until they give me my money back? The answer is no.

So this kind of money—they actually even tried to do an aqua hire several years in, but no one would aqua hire them because they thought they didn’t have any talent. They couldn’t do a talent acquisition because the acquirers thought there was no talent. Fortunately, they eventually pivoted to Twitch, which was focusing on just the one part of their product. People were streaming all kinds of things on Justin.tv, but Emmitt became interested in specifically video game streaming. People were streaming their games, which I never would have thought was a business that was a total surprise to me.

But that turned out to be a great business which they sold to Amazon. Unfortunately for a billion dollars, it’s probably worth way more than that right now. But the other interesting thing I want to point out here is this company had four founders. So Emmet, who is the CEO of Twitch, is still the CEO of Twitch at Amazon, where it's estimated to be like a twenty billion dollar business at this point. Justin Kan has started a bunch more startups; he was here as a partner at YC for a couple of years. Now he has a new company called Atrium, which if you invest in any of our companies going forward, you’ll probably interact with because they’re doing kind of like the default legal provider now for YC companies.

Michael Seibel started a company called Socialcam, which actually was a Justin.tv spin-out, sold that to Autodesk for sixty million, and then became a YC partner, and actually he’s now the CEO of Y Combinator. And finally, Kyle Voight started a company called Cruise, which you may have heard of because GM bought them for a billion dollars. So like that four group of people who had so little talent that no one would talent acquire them have done some pretty remarkable things.

All right, so eventually, I wanted to start my own company. I was like, 'REI, I think I see how this is done. I can do better.' So that leads me to FriendFeed. Now, when I decided to start a startup, I actually had a lot of ideas. Probably most of them were genuinely not good, but the one idea that was totally correct is that I thought the most important thing was having a really great team. And so that was sort of the original focus—just like figuring out who I can get to join on to this.

And fortunately, I managed to convince this guy, Brett Taylor, to co-found FriendFeed with us. Brett is someone that I knew from Google. I hadn't worked with him, but I kind of knew him; you know, we had chatted. I knew him by reputation. And the story I like to tell about Brett is he was the PM on maps—not even—or maybe like an APM, not even an engineer—and he was frustrated that the product was so slow and clunky. This was before it was launched. And so finally, one weekend, he was just like, 'You know what? I'm just gonna fix it.' And so he rewrote the entire JavaScript front-end over a weekend, made it a third the size and ten times as fast as what an entire team of developers had been working on for a long time.

This is the kind of person you want to invest in, right? Like that is an absolutely great bet. Let’s see, so FriendFeed generated a lot of buzz. Unfortunately, it didn’t generate a lot of growth in part because we ended up competing with this company, Facebook. So they turned out to be a very formidable competitor. I got a chance to meet Zach; he’s like really impressive. And eventually, he came to believe that Facebook was unbeatable—not by us, not by Google, not by anyone—and that it would therefore be like a Google-sized success, and therefore we should sell. So we sold FriendFeed to Facebook for about half a percent of Facebook's equity, which turned out to be a pretty good deal.

And so the lesson here I really want to emphasize is find your betters. Like you want to find people who are smarter in everything—more insightful than you are. And I think Brett and Zach and the team at Facebook are great examples of this. They were just clearly better at it than I am. So that’s like one of my criteria when I'm meeting a founder. If I’m thinking like, 'Oh, I could do this job better than you,' like I shouldn't fund, right? I only want to fund people who are better than I am, right?

And so this is why it’s not that I think idea in market don’t matter—like those things are of critical importance. You're not gonna build a big company in a tiny market. But my belief is that if the founders are actually better than me, then they're gonna have better insights, they're gonna know more than me. So I can kind of outsource that part of the thinking to the founder, and then I just have to figure out like does the founder believe it? And so what you want to do here, I believe, is use your knowledge to probe—to ask intelligent questions to see if they really do know what they're talking about, right? But not using it to decide—not being like, 'Oh, you know, I'm an expert on air mattresses; I know there's no business here.'

The challenge, of course, is finding these great founders. So I want to go through a couple more examples that I think highlight some traits. The first is Meraki. This is another early investment in 2006. And when I went in to meet them in their little office in Mountain View, they showed me around; they introduced me to the team; they showed me some of the hardware they were building—these little mesh Wi-Fi boxes at the time. And I was like, 'Wow, this is really impressive! Like how much money have you guys raised?' And they're like, 'We don’t raise any; that's why we're talking to you, right?'

And I was shocked because they had like hardware there; they had been selling these devices, and they did it all by kind of like scraping together money here and there. They got somehow a deal on this and a deal on that, and someone lent them money, and they were able to string the whole thing together for zero dollars, which is exactly what I like to see. And so this is another one I like to think about as an example of like what you want.

And then just kind of for laughs, I'll give you what would be the opposite of Meraki. I would say is Juice Arrow. Like they spent a hundred million dollars on their like juice bag squeezer without ever talking to a customer, without ever trying to sell someone a bag of juice, which, by the way, is what it is—it's just a bag of juice, and it squeezes it out. It’s horrible, right? So you want the opposite of Juice Arrow; you want Meraki. And this was Cisco bought this for $1.2 billion, which was like a seventy-three X return. Once again, they sold too early. That company's worth ten, twenty billion dollars now. Founders always sell too early because a billion dollars—I make the same mistake; I sell too early.

Next up is Cruise. Cruise, I mentioned earlier, one of the founders of Justin.tv, Kyle Voight, started this company. And my favorite Kyle story is actually from the early days of Justin.tv, where they came to me with this problem. I had invested and then they came to me with this problem where they were burning up all of their money on streaming because it turned out streaming video was very expensive. They could either use the streaming CDNs that were really expensive or they could like license some kind of software to run their own streaming CDN, but that was also really expensive. So they're like, 'What should we do? You know, we're gonna burn all of your money,' which I don’t like.

So I was like, 'You know what? This doesn't seem like a hard problem to me. Like bits go in, bits go out. Like what's the big deal? You should just write your own streaming media server.' And Kyle, you know, took me seriously. I was like, 'Well, okay, I guess if PB says it's possible.' And actually I told him he should be able to do it in an evening. I said, 'You should just build this tonight, launch it tomorrow.' Like I don’t get what's so hard, just make it work. It turns out it took him maybe like a whole weekend, but he did in fact write their own streaming media server, and this is actually the reason that Justin.tv survived at all, right?

Because there was no way that the business was viable if they were paying so much for streaming. And again, it's a founder who's willing to do something that a more reasonable person would say is impossible. Like 99% of Engineers, if I told them, 'Oh, yeah, you should replicate this other company's product, and you should do it tonight,' okay? They'd just be like, 'You're an idiot. Like you don't know what you're talking about.' It would give me a big long list of reasons why it’s—and but instead he just did it.

And so the Cruise idea actually seemed four years ago kind of in a similar boat. Now self-driving cars are really hot; car startups are really hot. But four years ago it actually seemed like an absurd idea. Like the technology is so hard; Google has like a huge team working on it; there's no way, like this is too much for a little startup. There’s no way it can work. But of course, it had Kyle invest it anyway, and as mentioned, GM bought them for like a billion, which was about a 50X return—not too bad. Again, probably sold too early.

This company you may or may not have heard of—they get a lot of press because it's awesome. They came out of YC two years ago. They're building supersonic jets! In ten years, when you want to go to Tokyo, you're gonna get there in like six hours. Like that’s pretty exciting. And so this is another great example. The founder, Blake, he's actually a software guy; he doesn’t know anything about aeronautics or anything. He just became determined. He's like, 'You know, it's the future. We should have supersonic travel.' He’s like, 'I'm just gonna do it.'

And so he took the time to just go teach himself everything and figure out why like the Concorde it was such a huge failure and what was different now. And what was so remarkable, I remember when he was in YC, is that he was able to very clearly explain to me why he was gonna do this thing that seems really impossible, but he could very clearly outline to me essentially like the three things that were wrong with the Concorde that he was fixing—the reasons that this was actually a viable business, and like work through the economics of it all, and he laid out a very clean picture.

And again, this is one of the traits of really great founders: they're able to communicate what's a fairly complex idea. I don’t know anything about supersonic aircraft. I'm not an expert in supersonic aircraft, right? But he's able to explain this to me in a way that I understand very easily. He’s got good insights and just absurd ambition—just really ambitious. Something that again, people would be like, 'There's no way a startup is gonna build a supersonic airliner!' I believe they are, you know, it's not yet proven, but they're making amazing progress.

I always love to read their investor updates; he also sends the best investor updates of any company I’ve ever invested in. And so the other point I want to make here actually is don’t be afraid to invest in things that aren't software. If you're working according to this philosophy that like you want a really great founder and the founder brings the expertise, then you're able to invest in so much more. And we have so many great companies that are like in bio. I think we're gonna have some just giant exits, you know, ten years from now. And I see every time at demo day, 'cause I work with a lot of the bio companies because I really like them, I see, you know, at least half the investors just like cross it off the list on principle because like, 'I don’t know anything about supersonic aircraft!' So don’t do that.

All right, now as promised I will talk about some of the less flattering stories. So in honor of the Dropbox IPO, let's talk about Dropbox. So this email, again, for those who can't read it, there's me replying to Drew. This is September 14th, 2007: 'Sorry, I wasn't able to make it today; are you gonna be around this weekend?' So I had actually gone to Boston to meet the Summer 7 batch, which was only like 19 companies, so it wasn’t that hard. Drew seemed really good; he was definitely, you know, seemed like one of the strongest and best founders in the batch.

But you know, I had some questions about competition, like 'What are you gonna do about Google Drive?' And all that, so I set up a meeting to meet with him when we were back in San Francisco. And then on the day of the meeting, my previous meeting ran long, and I canceled on Drew. And I want to make a couple of points here. One is that that's really shitty investor behavior, because you're wasting the founder's time—don't do that.

And two, karma kind of got me back. I missed out on what would have been like a thousand X return and an exciting IPO because I didn’t show up. All right, so I missed Airbnb. I mean, Dropbox? But what about Airbnb? So here's Michael Seibel introducing me, this was June 26, 2008, to this company, AirBed and Breakfast. And then just sort of like humorously, four years later, something, three years later, I respond. I should have responded. The thing that actually triggered this was that the Airbnb founders were like standing right here telling their story, and I was like, 'Ah, this hurts so much! I can't believe I like missed this company.'

In my defense, I had my own startup at the time, and I think if you are a startup founder, you need to give your own company absolute priority over angel investing. The thing that actually makes it more painful is that this was just the first time—this was before they were in YC. Once they're in YC, PG is like, 'Oh, you gotta talk to the AirBnB guys; they're so good. They're like the best company.' He was just constantly hyping them. And generally, if PG gives that strong of an endorsement, just like, you just invest. I'm like, 'Cool, you know, I'll invest. I'll watch for them on Angel Day—which is what you used to call Demo Day.'

Now I show up on Angel Day, and there's no Airbnb. So I email Brian—I'm like, 'Hey, missed you guys today; was hoping to chat and invest.' He's like, 'Yeah, sorry, we just signed a term sheet with Sequoia.' So again, I had basically in my head decided to invest, but then just kind of was lax today, cycle and took some time. Fortunately, I was able to get into a later round, but at a much higher price. But again, the point I want to make here is that if you move slow, you're going to get adverse selection. Like the companies that you can string along for months—give a hundred different meetings and like you need to talk to this person—probably aren't going to be the best companies.

So like be very deliberate in your decision-making—like do or do not, write. Don't just be like, 'Yeah, you know, I’ll see what happens.' Like when you identify a company as being really promising, as something you want to invest in, you could spend like another week thinking about it or you could pull the trigger and get Airbnb.

All right, what else doesn’t work? Ah, really good price—this one I've been duped into a bunch where I'm like, 'Oh, it's such a—' you know, 'they're giving me a bargain. Like, I'm Tremp today, a bargain. All the valuations are so low; it's a steal. Like, I'm gonna—I'm getting a great deal!' I've never gotten any return from something that I invested in because it was a good deal. There really is no such thing as value investing in startups, as much as it seems like there should be. Every time I've done this, it's been a mistake.

Another mistake: not investing because of price. Clearly, there must be some limit at which this isn't true, but I've never encountered it. Anytime I've been like excited about something, but then there's like, 'Ah, the price is so high—yeah, you know, I'll invest in something else.' I've always regretted it, like so now actually in my head, I kind of turn around the rule. I'm like if I'm disappointed by their price, I should probably invest more; I need to write a larger check.

This one is probably kind of obvious. I make this mistake a lot because I work with the founders; I get emotionally attached, and I just want to help them out. They seem helpless if you're investing out of pity. But it also has never made any money. I've always turned up at because again, like the really great founders just are so formidable. You know, someone like Blake, get Boom, is just sort of like a force of nature. Like he’s somehow just going to will this thing into existence. The people who come to me with like a sob story—'Oh, no one wants to invest, maybe if you write a check, they will.'

Like no, I haven't made anything. So I'm gonna try to keep myself from making this mistake. I'll probably do it. Another one is cynical ideas. This one is something where I'm actually not really that excited about what they're doing and what it means for the future, but it seems like it could make a lot of money. And so, you know, if they can make a lot of money, I’d like some of it. So far, so far, and I'm kind of actually happy about this—I've never made money on these companies.

Somehow the cynical ideas, I don't know why, but they haven't worked out for me. Maybe I have the wrong kind of cynicism, but the principle here I would suggest is invest in optimism—invest in the future you want to see, not the future you fear. And these are basically ideas where I invested in the future I fear. I had this one idea at one point that I should just invest in anything that would not be out of place in the movie Idiocracy. Fortunately, those investments didn't work.

All right, another mistake: really impressive numbers. I think Sam touched on this yesterday. It’s really easy to get excited, and good numbers on their own are not a bad thing. Obviously, sometimes really good companies have really good numbers, but the thing that happens is you kind of get blinded by the numbers. You're like, 'Wow, it just keeps going up!' Ignore the fact that like somehow the founders are kind of a or like the product doesn't really make sense to me.

Again, I never made money on these. If I'm just investing because of the numbers, it turns out to be a bad investment. And I think the reason for this, the question to ask is, is this company actually creating value? Like are they actually making the world a wealthier place or are they maybe just exploiting some temporary inefficiency in like Facebook? Facebook’s gonna change their album algorithm, and that company's there, be dead, right?

All right, so I've tried to boil that all down again just more for my own use to a checklist. Because again, my thinking on startups is that it's really like the founders are absolutely critical, and everything else kind of derives from that. So if I'm just asking myself: Is this person a great founder? Here's some of the traits: clear concise communication. And I want to be clear; this doesn't mean slick salesy kind of communication. I mean people who can demonstrate a real depth of understanding and clarity of thought.

In the example, actually, probably one of the best examples of this is just Paul Graham with Y Combinator. Like the whole reason Y Combinator really exists is because he'd been writing these essays for years that were sort of like he's kind of like the Pied Piper of nerds. Everyone would just read these things and get excited. And that was, to this day, I believe it's like the number one or number two source of leads to people applying to Y Combinator. We ask a question on the application, 'How did you hear about Y Combinator?' And they say, 'From Paul Graham's essays.'

But also, you know, essentially all of the other founders fall into this category, right? Like Kyle was able to tell me a story about how he's gonna make a self-driving car boom. Again, he's able to tell me the story about how they’re gonna do this impossible thing. And so if people—if I find myself struggling to understand or like doing the work for them, it's always a bad time.

And it’s the trap I sometimes fall into is I'll talk to them for so long I'll come up with a story that makes sense, and then I'll just sell myself on my own story. You know, I think maybe I'm just not smart enough, but no, if they're a good founder, they're gonna be able to communicate to me. Moves fast. Again, think of Brett rewriting maps in a weekend or Kyle making a media server or, you know, the Meraki team building that product with no money. Moving fast is absolutely critical.

And again, it's something that I think most people don't do. It's really easy to come up with the reasons why this should take longer or, you know, I really want to get this thing right or that thing right. Founders will come up with infinite reasons to move slow, and that's fine; I'm just not going to invest. Accomplishes a lot with a little. And so again, like Meraki is the great example of this. No money in there shipping hardware, somehow Juice Arrow is the anti-story of that. Most startups unfortunately accomplish very little with a lot, implausibly ambitious or frivolous ideas.

And the reason that we're kind of at these two extremes is that if the idea is in the middle, as I mentioned at the start, probably there's like a big company somewhere working on it or a hundred other startups. And so you want to see something that’s either it sounds impossible like, you know, a supersonic jet or just straight-up stupid like putting a camera on Justin's head, right? Things that seem like toys—actually Aaron Harris had a good blog post about this, I think last week. You should look up on the YC blog. Ideas that look like toys are oftentimes really great.

And the founders have the conviction to pursue these, obviously bad ideas. It's a talent magnet, so the best people want to work with the best people, and somehow the really great founders seem to be able to attract other great people. And actually, YC, this is I think really the secret of YC: it's just a giant top talent magnet. Like we could bring them in and do almost anything. I think it was still sort of work; the key thing is that we're attracting the best people, and the best people want to come be a part of a Y Combinator because this is where all the other best people are.

So this, and we see this with a lot of our top startups like Stripe, you know, really brilliant founders, and then their team is also really impressive. Make something people want—this is key. This is our slogan: make something people want. Google is a great example. Like I actually was using the product. If they’re building a thing that you actually want, that's a good sign. If you can't identify anyone in the world who would possibly want the thing that they're making—anyone who if you offer it to them, they'll be like, 'No, thank you.' That's a bad sign.

I was gonna say something else about that, but I remember—oh right, like actually another funny story here is DoorDash, which you may have seen in the news; they just raised another half a billion. When they interviewed here, like actually in the room right over there, I like one question for them was like, 'Will you deliver to my house?' Because at the time when they went through YC, there was no food delivery where I live, and I get hungry. And so they're like, 'Yes,' and I was like, 'You're funded.' So I mainly funded that company just because I wanted food for myself. But, you know, they're doing good.

And finally, determined and committed. And in fact, PG says determined is the number one most important trait in a startup founder. And the reason for this is because starting a startup is really, really, really hard and unpleasant. And so if you're not super determined, you'll probably give up at some point. You'll be like, 'Why am I putting myself through hell? I should just go get a job at Google and be like a normal person.'

Unfortunately, I think like this can sometimes be a little bit hard to suss out because the founders will of course be like, 'Oh yeah, I'm totally committed.' But some of the things you can look for is, are they hedging their bets? Like do they have kind of easy Plan B? And actually, the group of founders where I've seen this happen most is founders who are on leave from med school—doctors who realize they didn't really want to be a doctor.

So they’re like, 'I'll try out the startup thing.' And so they come in here; they'll work for six months or whatever, and then think, 'Man, this is hard! I'm scared; I'm gonna go back to medical school,' right? Like they have that easy out. They've already made a huge investment in being a doctor, and they don’t want to throw that all away for some pipe dream startup.

And I think the other thing that I like to look at is, have the people made mistakes in their life? I'm very nervous about people who've always gotten perfect grades—they got into all the right things; they did all the right clubs—the people who have the perfect resume. The people who didn't fail because these are people who've never failed. And startup founder, being a star founder, you're basically failing day after day. And so what these people do, you know, they're these are people who are scared of failure, and so they almost inevitably will take just like an early exit, and then they can put on their resume, 'Oh, you know, started a YC startup and sold it to some other company,' right?

It doesn't say that they sold it for like zero dollars or whatever, but they can add it to their resume, and they can go get a safe job at McKinsey or something. All right, so that is it. Thank you!"

[Applause]

"So the question is, with respect to making things people want, how does—how do you evaluate that for like bio companies and things like that? You know, actually, I find most of the time with those companies, it's pretty obvious. Like you're making a cure for cancer; I want a cure for cancer, right?

The risk on those companies is most often not market risk. It's not the risk that no one wants a cure for cancer; the risk is that you don’t actually have a cure for cancer, right? So with those companies, more often it's one of these things where like the demand is kind of obvious; it's just like a question of whether you can actually create the product."

"I have a question, right? Just wondering, I know Justin; I know Michael; I know Emmet; I know Kyle. What is it about those guys that made you invest? Because, I mean, you didn’t really say it, but you have heard this incredibly stupid idea, and yet you invested in this talented group of four people. Why?"

"I mean, I think they did have talent. I think they told a very simple story, actually. There's a good question—they told a story that really connected with me, which is we're gonna stick a camera on our head, on Justin's head, and then stream it and see what happens. And that's the kind of thing I love because I'm like, I don't know what's gonna happen, right? It’s really unpredictable. But again, the thing that I liked was that they weren't gonna go off and spend a lot of time thinking about it and like making complex plans; they were literally just gonna do it.

They bought a camera, stuck it on Justin's head, and like hacked together some of it so that they could stream it. So they were gonna be able to test the idea very quickly and learn and iterate. And again, this is part of why the—you know, on the checklist here, moves fast, is super important. Like you can almost define a predict the I think success of a startup just based on how quickly they're able to iterate, and the really great companies just like iterate really quickly. The ones who, you know, do a release every two years fail."

"Juice Arrow, sure."

"So the question is, unlike private equity where you can do months and months of diligence, like we kind of have to move fast as investors or, as I mentioned, face adverse selection of not funding Airbnb. So how do we know if the founders move fast? I think is what you're saying."

"As well as all of these other things? And what you can do is you can actually ask them: what have you done? What have you accomplished in the past month, right? What...? There's actually a really good blog post you should all go find where it says a little bit of slope is worth a lot of y-intercept. I think is the title. And basically, the idea is you can have something up here, you know, millions of dollars in revenue, but it's kind of flatlined, or you can have this little toy down here with like almost nothing but it has an upward slope.

And you'll want to invest in the thing with slope. And so this is what we look for in founders and companies is just like how quickly are they moving forward? How soon are they going to do something, right? And so like the example of Justin.tv, it wasn’t like we're gonna launch this thing next year; it was like we're gonna launch this thing next week, right?

And so just talking to them about their schedule, what they've done in the past, and what their anticipated future schedule is. And I want—and actually asking them like, 'Okay, why aren’t you doing this thing faster?' And seeing if they give me, you know, just kind of like excuses or if there's maybe is a legitimate reason for not moving fast. And you know, you can also try pushing the founders; you know, probably more so after you fund them; pushing them to move even faster and see how they respond.

And, you know, like my story with Kyle, I gave him what was maybe an unreasonable thing—like you should rewrite some other company's product in an evening—but he took it and ran with it, right? He didn’t just like...I've worked with a lot of engineers; 99% of engineers would just like tell me I'm smoking crack, and like don’t know what I'm talking about, right? But he just like took it and ran with it."

"So here at YC, we actually fund the companies based on a ten-minute interview, right? So it is possible to get, it's amazing how much you can get out of people in ten minutes. And actually, this is again why I put for me number one on the list is clear concise communication. If I spend ten minutes interviewing a company, and I still don’t know what they do, that's an easy no. Like it’s so easy to say no to that, right? Like if they can't, if they can't tell me what it is—that's probably like, you know, at least a third of the companies that come in here to interview at the end of the interview, we still don’t know what they're doing. And so we don’t fund them."

"So the question is, did I make any mistakes with respect to the fundamentals of the startup? I mean, okay, the thing you have to remember is that most of your investments aren't gonna work out. And it's easy to spend a lot of time kind of making up complex theories, but what I find is the danger when I see investors who do that is they come up with clever ways of talking themselves out of good deals because like, you know, Justin.tv if I give him that like five minutes of thought, I could have come up with you know a hundred reasons why it wasn’t going to work, right? So yeah, did the companies fail for fundamental reasons? Of course, and you sometimes you never even know."

"Right? Like suppose Justin.tv had managed to actually do a talent acquisition. Someone thought they had talent; they wouldn’t be up on the slide, right? It’s kind of random, you know, some random person at Google flipped the bit one way or the other, and I got a completely different outcome. But if I were to then look back on that, let’s say it wasn't a favorable outcome, and then try to draw like deep conclusions about, you know, why I should never again invest in like a streaming video company or something like that, I might miss out on something really good.

So one of my other principles is don't over-learn from past mistakes, and it's really easy because it hurts. It hurts so much to lose money, like it hurts me at least, that you want to avoid it. But if you're gonna be able to do this, you kind of just have to bite the bullet and recognize that fundamentally investing at seed stage involves—and I started with this—a great deal of luck."

"So the question is when I look at ICOs, do I see that as the open source of funding or the G-Sera funding? It's a good question. I'm trying to figure out how do I like politely answer. I think there’s something really interesting going on in crypto. So crypto is another one of these things like I gave Linux as an example where the thing that brought me to Google was my enthusiasm for Linux. And I think if you were paying attention to what like nerdy 21-year-olds were excited about five, six years ago, you would have invested in Coinbase and Bitcoin and all that stuff.

That said, as far as I can tell, the vast majority of the ICOs are scams, or just like these are companies that if they were trying to raise from the people in the room here, maybe they'd be able to raise half a million dollars, but instead they go do an ICO and they raise, let's say, a hundred million dollars. No way, no way those are all going to zero. I don't believe any of those are gonna work. And part of the reason for that - is this counterintuitive thing that I've learned, working with a lot of startups is that actually too much money is a huge liability, because when you have a lot of money, you start doing things the right way.

And that's actually kind of what happened with Juice Arrow, as an example, is they got, you know, a hundred and whatever million dollars. If that company only had a hundred K, they would have had to go try to sell juice bags. They wouldn’t have been able to spend years building this stupid machine, right? They would have had to actually talk to customers. And so having a giant mountain of money enables the founders to insulate themselves from reality to an unhealthy extent."

"Just a couple of questions from cyberspace or in this room, and they're just too shy to speak it out. One was directly to me, but I think it's an interesting question, which is people tend to be nervous about some sort of deep, deep technology companies. Investors especially, how does—how do you get over that? And how do people in general get over not being afraid of some sort of hard tech?"

"I mean, I guess I also just like tend to invest in things that I'm excited about, and it's so much easier for me to get excited about a supersonic jet than some sort of like obscure enterprise software or something like that. And not that an enterprise software is a bad investment, but I don’t know, for me it's not really that hard because I sort of accept the fact that I actually don’t know very much about almost anything.

So it’s not just the hard tech companies; like I don’t know anything about almost any of what the company is doing. That's why I have to invest in people who are better than me, right? And so it really just comes down to like do I believe in this founder and their expertise? And then therefore, I can just rely—I can like outsource my thinking to the founder."

"I don’t think you answered this, but if you’re in an hour meeting and I had a startup and you were trying to determine whether I was determined, how would you do that?"

"Actually to talk about this a little bit, but I—you know, a lot of it is kind of probing for, you know, do they have any plan B, right? Are they gonna go back to medical school? Have they quit their job yet? Do they still have a job at some other company? Like, and we get this a lot of times where we’ll ask founders, 'Are you going to leave your job?' And they’re say, 'Well, if I raise enough money that I can pay myself XYZ,' like they have these conditions on it.

And ideally, you want the people who are just willing to charge straight into a brick wall. That’s necessary. I think this is really deep; Paul Graham says this is the most important quality of a founder you should look for. But it's extraordinarily difficult to know when someone has it. One of the reasons, by the way, we ask questions on the YC application about your past is because that is one way I try to determine if someone's determined. You know, has that happened to them in the past? Paul mentioned people who have never ever failed. That doesn’t mean that they’re not gonna fail this time, but you just don’t have very much information about what happens when they get punched in the face."

"Right? Actually, if they've never failed, they're gonna be afraid of failure, and therefore, they're gonna take a safe exit."

"Yeah, and actually like to follow up on the what you can do to examine determination is actually also just ask about the history of the company. So like Airbnb is another one where they actually worked on that for a long time. And you know, Michael actually helped them out a lot with the pitch. They pitched a number of investors; everyone said no. When they came to interview at YC, Paul Graham actually tried to talk them out of the idea. They're like, 'You know, you guys seem really good, but this idea is terrible. Can you work on something else?' And they were just like no, right?

Like they he tried to talk them out of it. If he had talked to them out of it, that probably would mean it was a bad idea."

"Okay, one more question. So the question is, what about like, I invest in things I don't know about? What about things I do know about?"

"I have a trick here, which is I don’t know about anything. I mean, honestly, like I think it's really easy to overestimate your own knowledge and experience. So like I said, I try to avoid learning too much from the past. Specifically asked about email, yeah, I am a little bit burnt out on email startups. So there's like maybe kind of a high bar just because I know this is a case where I can take my knowledge and I put this earlier when you're talking to the founder, use your knowledge to probe and see if they have a good response.

And so, you know, a lot of the email stuff they're like tweaking Gmail, and I’m like, 'Okay, what if Google just adds this feature? What are you gonna do? You know, right? Like, okay, that's not a good answer.' But I have had good email startups. Front, it was actually recently in the news; they just raised a big round from Sequoia, I think. Really great founder, again, Mathilde, she's great. I, you know, that was an email thing, but she could tell me a really clear story about why it wasn't just like a feature in Gmail."

"Laughs."

"I have to unlearn things that you think you know. You know, I ran one of the very first ever personal services, and I might have thought it was done and never invested in any personal service again. And I think that would have been a pretty bad mistake. So I think we give, I think a little bit contradictory advice here all the time. It's part of the game. You should completely learn everything you can from your past mistakes and ignore the past at the same time.

So it's not very easy to be a good investor, but that's why we're all here, right? So thanks a lot, Paul!"

"Thank you."

[Applause]