Garry Tan is a software engineer turned entrepreneur turned investor. On his popular YouTube channel he offers advice to aspiring entrepreneurs. He is intensely aware of the privileged position he is in as a tech-savvy investor in Silicon Valley, and is eager to share his knowledge with the next generation.
Garry Tan is a software engineer turned entrepreneur turned investor. On his popular YouTube channel he offers advice to aspiring entrepreneurs. He is intensely aware of the privileged position he is in as a tech-savvy investor in Silicon Valley, and is eager to share his knowledge with the next generation.
Intro: [00:00:00] From Silicon Valley, the heart of startup land, it's Getting2Alpha, the show about creating innovative, compelling experiences that people love. And now, here's your host, game designer, entrepreneur, and startup coach, Amy Jo Kim.
Amy: Gary Tan is a software engineer turned entrepreneur and investor. On his popular YouTube channel, Gary offers hard won advice to inspiring entrepreneurs.
He's intensely aware of the privileged position he's in as a tech savvy Silicon Valley investor, and he's eager to share his knowledge with the next generation.
Garry: All of society is changing in such a fundamental way because of software and because of the things that you all can build. It's like unimaginable amount of leverage in being able to write a line of code.
How do we get a lot more people to make things for a billion?
Amy: I got a chance to chat with [00:01:00] Gary on the day of the biggest exit of his career, Coinbase. We talked about what's going on in crypto, how he came to be an investor and how he chooses which companies to invest in. Have a listen.
Thank you so much for being here.
Garry: Oh yeah, no, thanks for having me.
Amy: I'll start with just, what is this moment in time like for you? We know you run initialized. We've looked at your background. You just had the biggest exit of your career, but you're also in the middle of working with startups and making decisions, figuring out how much to get into crypto.
Can you just give us a snapshot of what this moment is like for you?
Garry: Yeah, absolutely. I've been working in tech for a really long time, I guess going on north of, I don't know if you can count my teenage years, something like 20 some odd years for sure. And maybe mid 20, 20, 23, 24 [00:02:00] years, something like that, at least.
It feels the same in a lot of ways because it's the same thing that has brought all of you here, which is all of society is changing in such a fundamental way because of software and because of the things that you all can build. And that was something I knew when I was like 14 or 15 years old, when I first cold called the Yellow Pages to get my first web page design job as a 15 year old.
And I think the other thing that's different or unique for me is I grew up the children of immigrants and we didn't have a lot growing up as sometimes food insecure. So I actually was cold calling from, I think it was a two bedroom apartment in Fremont. I always grew up in apartments, latchkey kid. And tech gave me everything that I have today.
And the idea that as a job, I can meet people who are great builders and they can tell me what they think the future should be, and I can write them first a little check [00:03:00] and then a big check later and have an outcome like this happen. I'm just like stunned. I'm stunned that it. happened to me. On the other hand, like, it's what I wanted.
And then I'm stunned and I have great gratitude. And then all I really want is for this grand magic to happen a lot more for all of us. And then that's where I'm at today, April 21st a. m. Pacific, uh, coming at you from. My house in San Francisco, I feel like tech takes a lot of hits right now. Probably rightly so, but at the same time, this story is very present to hand for me because how do we get a lot more people to make things for a billion people?
And then it comes out of directly a lot of the stuff that Amy Jo has been teaching and working on, how do you help people? Find product market fit. How do you find what a user actually wants? And if one user wants it, a hundred will want it, then a thousand will want it, then ten thousand, then a hundred thousand, then a million, then ten million, [00:04:00] then a hundred million, then a billion, then seven billion.
And it really does seem to progress that way because it's like unimaginable amount of leverage in being able to write a line of code. Beyond the product itself, we're talking about like, how do you build organizations and organizations that allow us to be greater and bigger than ourselves? So anyway, I've been rambling a little bit, but it's like, you know, you have broad thoughts and that's where it goes for me.
How do we make this grand magic happen like a thousand times? More often, and it should, and money is one part of it. A lot of people focus on the money part, but one thing that I think I'm realizing more and more is there's actually crazy amount of money in the world. And it's just sitting in bank accounts, doing nothing.
It's being lent at negative interest rates. If there is injustice in the world, that is one of the primary injustices, which is like infinite number of problems, infinite, very smart people who can solve those problems. And then why is the money just sitting there? And then I think the key here is actually communities and teaching and fellowship and learning.
Amy: [00:05:00] Almost like guilds in old fashioned times.
Garry: Y Combinator did that for me. And I think that's one of the models that I want to see multiplied out in the world. That's the highest leverage thing that the world really needs.
Amy: So Coinbase exit is in crypto. Crypto is one of the reasons that you invested is because You saw the need, you felt the need, you were the target market, and then as you just explained, it spread.
Now, it took a while. Sometimes it doesn't spread like that. In crypto, it is spreading like that. There's a lot of things about crypto that are hard to understand. There's a lot about crypto that can feel like an MLM or a shell game in terms of the dynamics. There's a lot of going on. But crypto is clearly turning into something more like that.
Then when you first invested, when you were at Y Combinator, one of the questions that came in from our community really taps into a question that so many of us have, many of my clients are currently [00:06:00] struggling with NFTs, NFTs are big, they're clearly some kind of bubble, but there's a real long term value with NFTs.
And we're all struggling together to figure out what it is and how to actually build lasting things versus money grabs, Jorge Briones from our community asks. For startups that are involved in mobile games and arts in the creator economy, this emerging creator economy. What are your thoughts on adding NFTs to various products that we produce as an additional revenue source?
And we get this question a lot. Perhaps you get this question too.
Garry: So, let's see. I'm not the NFTs. You know, I, I would look to people like, um, Chris Dixon at Andreessen. What I'm looking for on the NFT side is I link it back to meaning. And then I think Amy, Joe, you have a great concept that's useful here, which is like your super fans.
And the way to think about NFTs is who are the super fans for that thing. And then what drives them [00:07:00] is really the reason to be for those NFTs, the NBA, NFT, that whole universe, that ecosystem, like that comes out of something that maps very directly to something that people really care about, which is the trading cards for sports.
And that's like a very natural affinity group. And then the power from NFTs is, uh, actually from the community. And so I think of it really is that don't focus on the capability, the idea that you can have trustless ownership of an NFT. Is merely a capability, but that's like being obsessed about SQL databases.
Like that's just a mechanism. That's why I sometimes I'm a little confused by it. You don't see people running around being like, I'm really excited about B trees. Like how to store data in this particular way. That's the equivalent of saying like being really into NF, NFTs. I like to focus on what can the user have.
And just as owning an original Rothko or a Warhol. The value of it is not in having it in your house. The [00:08:00] value of it is being able to display it and have other people know that you have it. The value is in like the brains of people and in the mimetic desire of that owning and wanting and being wanted in that context. Right.
Amy: And so what I love about what you're saying is how much you're bringing in the importance of mapping it to a real world mental model that you can understand that helps you predict what's what might happen. And right now the NFT ecosystem is really missing that shared way of displaying and like there's various ways to display it.
Like my version from the game design point of view, what can you have? What can you do with it? And right now, Most of what's in NFTs, yeah, there's ways to display it, but it's not very well integrated. What people can do now is they can sell it. They can buy it. And then it's, Oh, here's a marketplace. You can sell it.
And if that's the only thing you can do, you're communicating structurally that it's a speculative asset and [00:09:00] you should think of it as a speculative asset, because that's the only thing you can do. Everyone's trying to figure out what can you do with the NFT.
Garry: Yeah, that's right. If it's a multiplayer game, for instance, that starts to mean something within that community.
One of the companies that we funded that's a gaming company where NFTs are at its core is called Skyweaver. So it's Magic the Gathering style online web game that's built on Ethereum. And there are provably rare mega cards and you can like, by playing it, you can unlock those. That's an example of a dynamic where you need a game with a community, and then the NFTs mean something in that context.
Amy: Bingo. So, uh, several of our community members, speaking of communities, Gerard Sands and Adam Dharajji, um, Both wanted to dive into early adopters, what I call superfans. Superfans can be early adopters. Superfans can also be a particular slice of a more mature market. And what people are really wrestling with on the [00:10:00] startup front is how to integrate that into development.
Now, of course, I teach a certain way of doing it. You are on the front lines. You are watching companies over five and eight years mature and either succeed or not. You are funding new startups coming in. How do you think about that? And of course there's a lot of nuance, right? How you do it in a particular situation.
But are there any kind of general principles out of that nuance from your own experience that you would give as advice or share stories with the people that you counsel and want to see succeed?
Garry: Yeah, I guess the examples that come to mind are my old friend and colleague from Y Combinator, Paul Buchheit created Gmail.
And when he created Gmail, he could have easily just turned on the distribution spigot. And you could argue that Gmail would have been popular one way or another, but I would argue that the way they did caused it to be 10 times more successful than your typical Google product. Probably the majority of people use Gmail.
[00:11:00] Like I would venture to guess like the majority of people on this call use Gmail as their primary email. And what powered that was getting to a hundred people who absolutely loved it. Like, they really just held on to it. And this was like, also scratching your own itch. It was like, internal to Google that whole time.
They waited and kept, like, honing the product and fixing bugs, and making it better to the point where Paul Bukite felt like there were a hundred people who just Couldn't live without it. And that sort of starts with yourself. Even it's 10, 000 times easier to build something for yourself. If you're a product person, then to build it for someone who is totally different than you.
Amy: One of my favorite quotes, and you just told us the story behind the quote.
Garry: Yeah, totally. And then on my end, I think. My story for my startup in 2008, I wish it was as cool as Paul's, but I just wanted to start a startup actually. And then it wasn't really my idea. It was my co founder Sachin Agarwal's idea.
He had been blogging and he built a simple JavaScript that would [00:12:00] listen to an email account via IMAP. And then if you, you know, attached attachments. To the email, it would convert it into a blog post for you and do is literally like a 500 line Java app ran on a server. And, uh, we turned that into, I wrote the rails app.
I did the design. We basically were like, what's the minimal number of things to turn this like useful script that's basically open source that he wrote into something that a lay person could use. And then we realized that we didn't want to make it a connector. We wanted it to, you know, also have its own place.
So we decided to make the world's crudest blog platform that had this one feature that nobody else had. And then that was like the beginning of Posterous. I didn't know people would like it. It was just like, I was interested in it. I loved photography. I loved blogging. This is interesting for us. We'd probably use it.
And then when we released it, I got, Actually, I guess the magic number is around a hundred people. Like I just went and got a hundred of my friends and about 20 of them of the [00:13:00] hundred people I bugged to use it, used it, and then kept using it all the time. And then this very early moment, I realized now as an investor, we use long term retention cohorts as like a gold standard for figuring out whether or not something has legs or has product market fit. Weekly is probably the most useful because you can't have cohorts of two people. It's not predictive enough. Maybe 10 to a hundred is like pretty useful. And if you have 10 to a hundred signups in a week, that starts to be pretty good.
Useful, like enough sample. You're going to have a standard drop off. Like people try it and then they're just like, this doesn't work for me. And then, so you'll see a hundred percent week one, like 50 percent week two, what most products do is they just drop all the way to 4 percent or 3 percent or lower, even at week three or four.
It's just like, they try it, it didn't work, they just never come back again. And then how you can tell you have some legs is if it goes down, but then it flattens out. If it asymptotes at like 20 percent or above, that's sort of the gold [00:14:00] standard I use at this point for can this be good or great. And then the astonishing thing is if you can get 20 and asymptote and then stay there for six to 12 months, then you could actually be a top 500, if not top 100 app in the app store.
I mean, you can have a real business and you can have something that can touch a billion people. That's the closest thing to a divining rod. If you're in a big field, you just bought this, you know, big field of land and you need to find water on it. And you're like, I could just start digging anywhere.
Like, how do I know if I have water? Long term retention is basically the divining rod, as far as I can tell. The funny thing about my startup story is like, we barely made it. Like our, our long term cohorts basically hit 20%, like at six to six months, 12 months, 18 months. So we were like basically on the low end of what would work.
Amy: So that connects so much with what I've learned and what we teach is how do you then figure out who those 20 percent are and tighten up [00:15:00] your marketing, tighten up your whole funnel. Pulling in the people who are going to convert and that's this whole feedback loop. We have a question now from someone who's here, Ian Brand.
Ian wants to know, Gary, if you were back in time when you started Posterous, but it's today, 2021, what startup would you start? And then also following up on that. What are the product design challenges you see most often in the startups you worked with? So, Ian, introduce yourself and elaborate on that a little.
Ian: I'm a big fan, Gary. Just love all your generosity with your time and the content you put out. It's just been really inspiring getting to know you. I work for General Motors and my job is to find those new ideas and to continually expand our growth into new products. So, launching new products, validating new ideas.
So I take a lot of inspiration from Amy Jo and you and the startup world. So those were my questions.
Garry: I think we're super early in the crypto revolution. Obviously I am like deeply pot committed to that narrative at this point, but so far [00:16:00] every super cycle, once you survive a couple of winters, you're like that winter will come again.
And then that's the time to build. I don't think we've hit our current super cycle peak. As long as you're fine with a lot of BS and a lot of hype and a lot of things that are annoying about crypto, as long as you can zoom out and have longer duration than everyone else and you build, that's actually probably the most important thing.
If you're going to do a crypto related startup today is you have to have more belief than anyone else. And it's definitely something that we saw in, you know, 2016, 2017, that was Supercycle and then like 18, 19, a lot of 20, even like at the beginning of 20. Was just like winter, like darkest, darkest winter.
There's sort of like the set of people who give up and like once the last person gives up, that's when like everything rockets back, which is fascinating. I think in terms of where it's going, I'm still always interested in software eating marketplaces. There's sort of a fever dream for especially [00:17:00] engineers who are a little bit antisocial.
This idea that you could just sit in a cave and write code and then have it go out into the world. Of course, like that's not really going to work, but I think that's the origin of my obsession with the idea of like marketplaces and crypto software eating money. And then of course, Airbnb, Uber, all of these other things are merely like the largest slices of the pie of software eating these other markets and large two sided marketplaces, people who need this thing, they need housing, they need transportation, or the funniest thing is on our end, we're seeing every industry, right?
Like literally we have a company that we funded that's growing super fast. That is cornering the market on a giant rolls of aluminum, which it turns out like hundreds of billions of dollars market. It's not crazy high margin because it's a B2B market, but it is one of the largest markets in the planet because it's like what builds our buildings and makes all of our stuff.
And that marketplace was totally pick up the phone, like fax machine, take people out for steak dinners, and then [00:18:00] it's totally being converted over to all online, all data marketplace that is better, faster, cheaper. And we're seeing that across every market in the world. And that's interesting. And we're doing a lot of work on that side.
And then on the flip side, crypto is happening before our eyes. And so if you merge those two things, on the one hand, most of the vertical marketplaces like Uber or Airbnb, because like they require really great customer service and user experience, they're still being made by centralized companies. But there is an argument longer term that if you take it to the logical conclusion, you could imagine a society that is mediated by an open protocol.
Like for instance, Origin Token, which I'm a really small personal investor in, I think is very interesting because they're trying to carry that torch of what is the infrastructure you need to be able to stand up this other type of marketplace. It might be B2B, it might be B2C, right? You're seeing things like BitCloud.
That are pure, you know, semi decentralized. I can't tell if BitClout is [00:19:00] Facebook. It doesn't feel like Facebook to me. It feels like six degrees, but there are concepts embedded in their implementation and what they've done that will inspire what really works. So that's sort of where my head goes. I'm like, okay, we're actually in the first inning of the first pitch of crypto.
And then we're also maybe in the second or third inning of. Software eating, like, all human activity.
Ian: I was really fascinated by your comment that you're going to fund the next 100 Dekacorns through crypto.
Garry: It might be crypto, it might be this other file cabinet, eat all of GDP sort of way. Basically, both of the things are happening before our eyes in our portfolio.
And so if I had to project out on what's worked and what's not worked, it might be one, the other or both. And if I'm hyper optimistic, I think that it's going to be both.
Amy: So we also have a question from Dustin Block who works in media about where you see local journalism and all journalism headed and how [00:20:00] do we actually protect democracy and deliver a diet of fact based information in our climate, which is a really hard question.
Dustin, please introduce yourself and maybe add a little color to that question because it really, some of it Gary comes right down to community stuff, which I know you're really interested in.
Dustin: That's it. Yeah, that's exactly right. Gary, for your time. I love your YouTube channel. So coming from local TV, we're fascinated in trying to build a future, right?
And we're watching our audience literally die. I work for Graham Media Group. We call it a boutique local broadcaster. Part of the Graham family, the Washington Post family, but they spun it years ago. I spend a lot of time in local TV newsrooms, kind of on this digital strategy side. And I came to Amy's class because we're really, um, trying to build community, focusing on engagement.
And a lot of people want to do gamification, but in a bad way. And so Amy's helping us do it in the right way. So we're really concerned about the future of local media. I'm really fascinated. There's a lot of interest in [00:21:00] NFTs. Right after this call, though, I go to a climate collaborative discussion where the newsrooms all over the country are going to be working together on covering climate change.
And so that makes me really nervous about crypto.
Garry: Crypto and climate change. Yeah, yeah. That's a secondary question. Yeah. Let's see. The first one's obviously really big. Take my, what I say with a super grain of salt, because I don't know, this is like the most internet positive view of it. And this is just my opinion.
I don't know if I'm totally right, but my view is we live in this really insane time where we don't trust the institutions anymore. And then the tricky thing is now that the internet is here, it's very hard to know who to trust. And so now there's like polarization, bifurcation based on who you trust.
Like people trust the people they know directly more than they do a national or state level institutions I realize are valuable and they're being abandoned. So that's happening in society. I'm not someone who is happy about that. I think that's dangerous, obviously. [00:22:00] And then the question there though, is like, how do you create, how do you create new institutions?
But I think the pro internet criticism is that like. The old systems and the old institutions are decrepit, and it's tricky because you can spot how some of the institutions that people used to have rock solid faith in, they're just getting rocked, right? Because there wasn't an alternative view before, and now the alternative view is It's present to hand and often being spouted at you by someone you knew in high school.
And then that's where it's difficult because who do you trust? It wasn't Facebook. It would be someone else. It would be someone else's platform that was dealing with this. And the platforms are all dealing with it in their own way. I never came up through any of the media platforms, so I don't know how the media platform should react.
I am super bullish on the new media. Like I think Reddit is a model. Hacker News is a model. Uh, Substack is the model. Looking at what Substack is, it's putting dollars in the hands of the people who create the content. And the difficult thing is it's [00:23:00] been atomized. So you used to have the Washington Post, the Washington Post is still there, obviously, but people don't trust it as much, and I think that's problematic, like I think we need something like that.
And what it's being replaced by is individuals, and the reason why is it feels more pure, and then social platforms really amplify parasocial relationships, and I see this myself on my own YouTube channel, right, like, today you're catching me on a crazy day because I think I'm gonna hit 100, 000 subscriptions later today.
Yeah, which is I'm super like it's liked about it two years ago I had zero subs and what I realized now is if I can get to a hundred and then keep Building and growing I can get to a million and if I'm at a million I can get more views on my videos than like TechCrunch will get on like a million Any given post.
Amy: And then you'll, then you'll level up into the elder game, which is doing stunts.
Garry: Maybe not.
Amy: But if you actually look at the trajectory of YouTubers, when they get to a certain level, there's this [00:24:00] interesting trajectory. And it's also interesting that like you could get to a million and do you want to, or would you rather have a smaller, more focused audience?
It's up to you. You've talked about this really openly. And I think it's meaningful to lots of people, including me, you can play the. Stick a number in front of people and they want it to go up game. You can play the driven by the algorithm game and say, of course I want to go to a million. The question is why?
Garry: Yeah. So I think there's a useful way to think about the world right now from someone who I really respect named Venkatesh. He talks about the world being bifurcated in above the API line and below the API line, and it's a scary concept because like, it's also relevant to us because we're also the people who sometimes build the APIs.
But in this case, for instance, as a content creator, I'm below the API line. So I am subject to the gamification. I am subject to looking at studio. youtube. com and seeing the graph. And they're like, congratulations, like your views are up 570 percent because of this [00:25:00] video. Make more videos like this. Don't make these other psychological videos that nobody wants to click on.
Make this one about like how you made a billion dollars and keep doing that. Like for the rest of your life. Because ad revenue and YouTube has to do that, right? Like the PMs, right? It's the narrative. It's a Delaware C Corp and corporations are built to maximize shareholder value. And so I, all you can really do is try to find as many places in the world where you can be above the API line and you can control the destiny of yourself and others.
And then know when you are below the API line and be hyper conscious of their motivations, desires, and needs. And then...
Amy: What is like above the API line?
Garry: Then you're like running YouTube or Facebook or Uber or Airbnb. And then. The transcendence would be, instead of having a centralized entity that controls it, could it be software that the users choose, right?
Could you have decentralized networks? Could you atomize it to, like, the subreddit? And give people choice, like at [00:26:00] the code level. And so...
Amy: All the reddits that work have leadership and that like the whole thing with decentralized pretends that leadership doesn't matter and it does big time.
So how do you reconcile?
Garry: I guess it's like decentralization. But what's funny is this rhymes with the other problem, right? Like one of the things we're worried about is tribalization. And that's what a subreddit is. So you have to accept the good with the bad. Like the good is that if you can self determine in small groups, then you can have things that work for you or for that group and then The downside to it is like, we are tribal beings that are more or less predisposed to Dunbar's number.
And we're also predisposed to war against each other. Marshall McLuhan talked about this. He's the global village is actually going to be extremely tribal and brutal. And if you look at subreddit, if you look at Twitter, you can see like the patterns. Marshall McLuhan knew like he predicted all this stuff and now we have to live with it.
Amy: I want to probe into this [00:27:00] with this. I think a lot of people here. Are struggling with understanding what decentralized means. And recently I had another guest, Kevin Werbach, who's a professor of business at Wharton and worked a lot with the SEC in the early internet days with policy at that level, and he said something that blew my mind, which is if we want.
Not just. You know, criminals and underhanded and experiments and speculative stuff to happen, but we want really good people doing really good things to be happening on crypto everywhere. It needs to be regulated. in some way. And what do you think about that?
Garry: Yeah, I agree. And I think that was the critical thing that Brian Armstrong recognized very early, which is he had two big things he had to come to.
One was the being converted to Bitcoin and reading the Satoshi Nakamoto white paper. And at that point you go from [00:28:00] normie to a fringe. And what's funny is we talk about going from normie to fringe as maybe a bad thing. But when you're trying to create new things, it's actually an amazing thing. So that was the first step, but then even within that fringe, there was an idea that was a self limiting belief, right?
The self limiting belief in crypto in 2000, you know, 11 and 12 was that crypto and decentralization is anti government and anti institution. And you've talked to most people, like the most crypto converted today, they still believe that. Uh, I don't believe that. And I think Brian was very smart to realize, I mean, it was funny.
It's like he was at Airbnb watching like one of these marketplaces happen. And Airbnb itself was secret knowledge. This idea that you could sell space on the internet and have great community and have that be a good, solid experience. That was still a heretical, weird, fringe view. And then that was enabled by regulation and payment technology and anti fraud.
And so [00:29:00] it actually makes a lot of sense. It took someone to join a fringe, but then also have this other secret knowledge that Airbnb and things like Airbnb are going to take over the world. And then he's, he just synthesized it. And then he was like fringe within fringe at that point, but then he was right.
He wrote, he made BitBank. He started working on the software. He started banging the drum to try, try to get, it is no mistake that he spent a lot of time trying to find co founders and trying to hire people to come over to his vision because he knew he couldn't build it alone in a nutshell. Like that's a really useful mental model.
It's like he had to be fringe within fringe and then he had to be right. And then now we're sitting at a 60 billion company, right? Right.
Amy: Or you had to be able to navigate toward like Airbnb did, like they were right, but they also made a lot of early pivoting.
Garry: Yeah.
Amy: So James Young has a followup question exactly on that, which is.
What framework did you use in your head when you decided to make that investment? [00:30:00] James, please introduce yourself and follow up a little. Cause this is very related.
James: Just wanted to understand, like having this idea when it came to betting on Brian early on, just, just kind of what, what are you thinking?
Cause I want to try to figure out the framework as we're doing social crypto now and what creators and what projects. As we're building out our tool to bet on.
Garry: Yeah. First off there's team. And then I was like, this person is a fantastic engineer. Like one of the things that I got referred to Brian from Jason Tan at CIFScience, so CIFScience is a really great anti fraud platform that was growing fast back then.
Brian Armstrong at Airbnb was their best customer. And so one of the stories that Jason told me was that here, Brian Armstrong would stay up until 2 a. m. to debug the anti fraud platform of a vendor, like SIFSigns was a vendor of Airbnb. I'm like, okay, that is like a super good data point for me because I love funding builders who like are legit.
And the funniest [00:31:00] thing is, I don't think investors use that as a criteria because they can't tell because they never built anything. So I look for teams that can really build it. The second is the micro, I think. I'll start with the micro first. Just using the service, it was clean and well lit. A cabinet maker can tell another cabinet maker is good, not by looking at the front, but by looking at the back.
Nobody looks at the back except other cabinet makers. You could tell that there was like thought and effort and good design embedded in what Brian had created even that early when he was alone. And the founders are like the moat of dust that is like, that crystallizes the diamond and embedded in the founder contains multitudes.
And it has to start with someone who is that good. And then Beyond that, there was the macro. And so I was like semi converted in that working for Peter Thiel previously, I had to read the dollar crisis. So at the macroeconomic level had a sense that this was a problem. And then back down to the micro level, I had tried to buy Bitcoin because of that macro, like sort of idea.[00:32:00]
I was like, Oh, this sounds plausible. I don't know if it will really work. It's free to try it out. And I like to play with weird things. And I tried to buy it and it was terrible. And then that popped me back to the micro, which is, Oh, like how I would build it if I were him. And then I don't think it was more complicated, complicated than that.
Amy: Let me just put brackets on what you just said and make sure everybody gets it. Popping between the macro and micro level is a superpower. It's a ninja skill. That's a really important thing. Being able to do that. And it's hard. You can get the bends. Sometimes you can get a team that helps you do that.
Cause teams are about asymmetric skills. That's really where the magic is. We also have Armando.
Armando: Hi Gary, huge fan, Armando. I'm director of product at Papa. So I work very closely with Andrew Parker, which I know you're very familiar with our business. Yeah. That's going to be another IPO.
Yeah, very excited to join Andrew in his mission and things like [00:33:00] that.
And as you're fully aware, we're growing fast and scaling. So I just have one question for you. I'm a builder myself, and we're going to go through some growing pains as we scale and accelerate what we're doing at Papa. With new hires, complexity of things, and we level up the organization. Any advice from you as we go through growing and scaling the business, just some things that kind of come to mind that you've experienced, you've seen with other companies that we should be aware of as we scale and accelerate what we're doing.
Amy: Papa is a marketplace.
Garry: It's a marketplace, it's a marketplace for anti loneliness actually, which is a really big, it's not only a great business, it's also a really big vision. That's a great question. I wish I could be a lot more specific. I think the reality is like every company is so different and if anything, I realize a lot of what companies go through is, it's up to the psychological makeups of the founders and the exact team.
And so, there's not one size fits [00:34:00] all is the tricky thing. I can tell you from my experience, I wish I had a better, I had taken better stock of my own like psychological peccadilloes because they definitely cause problems for me. I talk about it like all the time on my YouTube channel, probably more than I should, but at this point genuinely believe that if I had, um, addressed more of it, like my startup would have done a lot better, maybe 10 or a hundred X better.
Like maybe our company could have been Instagram when instead of Instagram, if I look at like the tail of the tape, I'm like. All of those things were in the cards, but we didn't play it mainly on the management and conflict resolution side. And it's a little bit scary to say, but I'm like, I wish that I had worked through the things that I was running away from around my own conflict avoidance.
One thing I realized now, for instance, is that. I am incredibly hard on myself, but I am actually very lenient on other people. And that's great as an IC, that's horrible as a manager. I assumed that people wanted, like, high love and low structure. And so I tried to run all of [00:35:00] my organizations. Like, uh, open source projects, and that's not a way to do it really.
Amy: Again, though, that's the whole dream of decentralization, which is that it doesn't actually need leadership. It doesn't actually need somebody reiterating the vision again and again, and actually drawing a line in the sand so everybody understands there is one, right? All the basics. Yeah.
Garry: Yeah, yeah. The, the grand revolution now for crypto is actually tribal leadership.
It's like the greatest experiment in decentralized tribal leadership, right?
Amy: OMG. That's mind blowing.
Garry: Yeah. It's small groups of people sitting around a GitHub and they're trying to define their own protocol, figuring out what other protocols they can ally and then they're out there.
Amy: And then they add voting for conflict resolution and then they grow and then they add representation and then they have democracy.
Garry: Yeah, exactly. That's the only other thing that is definitely a super trend. Every. Because software is eating the world, the world is being governed by software. And so that's why we have [00:36:00] corporations that are software companies that rival nation states. And so we are in an uncomfortable time with that realization.
When a Delaware C Corp can, that's publicly listed, can de platform a head of state, right? Like, how weird is that? Personally, I was like, great!
Amy: Rules are meaningless unless they're publicly defended.
Garry: Yeah. And the tricky thing about that is in that moment, I was like, this is awesome. But like, how would I feel if my side, like my tribe was de platformed in the same way?
Amy: We should do a follow up on this and we should include the people that have run MUDs and online games for years that had tribal clustering as a feature plus voting before,
Garry: That's super interesting.
Amy: Want to know where that leads? There are a lot of people that have run those experiments. Oh yeah. And let me just say.
It brings up even more problems than it's interesting. It's there's, but that's for another day.
Garry: Oh man.
Amy: So we have,
Garry: I'll come for that. I'll come to listen to Amy Jo's talk on that. That's super interesting.
Amy: I want you in that, but it's [00:37:00] like, maybe we get Kevin in that too. But the other thing we have to talk about more just to pin it for the future.
And maybe you've heard of this Gary, you mentioned Dunbar's number. So let's expand on that for a moment. Around social because I think that's really a great way to close this. So Dunbar's number is famously 150 people and it's like a little village and it's a people you can know and recognize. That's the Dunbar's number idea that is there more to it?
How do you think about Dunbar's number? And then I want to tell you about Dunbar's layers.
Garry: Oh, interesting. I hadn't heard about that.
Amy: Ding, ding. How do you use and think about Dunbar's number in practice?
Garry: All of the institutions that run the world to date were created by people who couldn't conceive of smashing Dunbar's number.
And even today, even the way we like build software, Bezos talks about two pizza teams and things like that. The thing about great collaboration and like omnipresent, super high bandwidth internet is that we're smashing Dunbar's number. Like [00:38:00] across the board. The capacity for like collective instant action has like never been greater.
The toy version of this is like the New York City flash mob, but we have flash mobs absolutely all of the time at this point. And all of our institutions, media, education, government, all of our institutions are not built for that. And the software has not become a part of those things yet. And then healthcare, right?
The idea of PAPA, which is pre care, en masse, helping not 150 people in a small group, but you know, thousands, if not tens of thousands, eventually hundreds of thousands and millions of seniors. not be lonely. If you were a health plan, you weren't able to do that. And if someone came and pitched you that, can you imagine how many people would have to work the phone banks?
We'd have to hire a thousand people in the Philippines to just coordinate it. We don't have to do that anymore because everyone has like an infinite computer in their pocket and things like Instacart or Uber or DoorDash or any of [00:39:00] these things are all possible now. And so you can just go into society.
And look at like anything that human beings need from other human beings, and every single one will be replaced by hopefully something that you all build.
Amy: That's a wonderful way to encapsulate it. Dunbar's number is actually one thing from a larger model called Dunbar's layers. That's about how humans meaningfully cluster.
into social groups. Interesting. And 150.
Garry: Very relevant.
Amy: Yeah. And it turns out there's two, five, 15, 30, 35, 70-ish, 150,000 and up. And within the large scale networks that are DoorDash, et cetera, which is very depersonalized. Like you don't necessarily get to know your guy, right? In the Papa model, you absolutely get to know your guy.
You want. To form longer term [00:40:00] relationships, there are failed marketplaces because that part didn't work. So there are social groupings that we all understand intuitively that have different qualities. There's a reason that 15 is like an Among Us squad. Then 5 is different. 5 is a family size. It can be 4.
4 is a common multiplayer number. These groupings have a lot of social meaning. That's fascinating. All one right. grouping. It really depends on what you're trying to do. Sometimes if it's an old person and a young person, the duo model, what is that? That's two player co op, right? That's really fun. There's a lot of examples of that.
So those come right out of the same Dunbar model. It was just stretched and had more to say about those different groupings.
Garry: It's like the future was foretold by Marshall McLuhan and Dunbar.
Amy: Boom. All right, so Ian, you have one quick follow up question. It looks like, um,
Ian: I was on a clubhouse. You were talking about compound startups and it [00:41:00] was following up on our manos question about when do you start to think about going beyond your core mission and, and exploring new opportunities as a startup?
How do you do that? When do you do it?
Garry: Oh, yeah. I don't know. Maybe this is like an American thing. I'm not really sure. American startups can't really walk, chew gum at the same time. The flippant answer is don't do it. The, the real answer is do it only after you have one thing that definitely has product market fit.
And then you have an adjacency that you have an unfair advantage around. And then even then, you should only do it if you find a barrel, which is like a leader who can hire and manage and have the OKR on that new initiative. And then you can give it to them fully. And if you don't have that person who you trust, you need to find them or you need to not do it.
Amy: That was gold. That was so good. Thanks.
Garry: I don't know. I'm just standing on the shoulders of giants.
Amy: It's really specific though. And let's tie it back to the divining rod and long term retention. And you said, [00:42:00] look for a high retention cohort. Part of what we do in game thinking is get you to look for it before you launch through these low fidelity techniques and high iteration, blah, blah, blah.
When you've got that diviny rod and you're still doing it after you've launched, right? It's like, where's the longterm? I thought it was going to be this. I ran an experiment. If nobody sticks around and you're dependent on long term retention, then no, you don't have product market fit. If you have one person and you're building a marketplace, no, you don't have product market fit.
But it seems like as a rough rule of thumb, can you get like 20 people out of a hundred and you did a decent job finding the hundred to stick around. For a few weeks or a few months and reuse the service multiple times to some metric. And if you can grow from there.
Garry: Yeah. And then the danger is if you don't have that retention, then it's a leaky bucket.
And then that's where people get lost because you can always spend more money to buy more users. And then if it's a leaky bucket, they're just falling out the bottom. And you'll never get the money [00:43:00] back. And then that's not a business. That's a money losing live action role play.
Amy: No. And if you throw metrics at it to juice it, you'll never figure out why it's a leaky bucket.
It's like a local maxima, but then you don't see the whole, you don't step back and be able to get to that higher peak.
Tons of very well funded. Very high profile startups end up on my doorstep because they built an amazing leaky bucket. Like it's actually pretty common.
Garry: This is the tail wagging the dog.
Like there's infinite money in the world. And then there's also infinite principal agent problem from like GPs at these funds that are investing other people's money. So they literally get paid to like take risk, which is not a bad thing. That's why some of these money losing startups are so like astonishing because at IPO you're reading the S1 and you're like.
The principal question about whether or not they can have a successful IPO is actually, well, it's been a live [00:44:00] action role play for a while. We'll make it up on volume, right? At some point we will start generating earnings and it's a land grab and the hard part is anyone's guess that the truth. will come out and then at the public market level, like that's a pretty late stage to do it.
But that's where we live in 2021 today is you can't really tell sometimes.
Amy: Is that what happened with WeWork?
Garry: I think you could argue that I, you know, never worked on like that transaction. So I don't know enough about it as an outsider, uh, watching the documentaries. If you look at that and that doesn't make sense, think about retention and leaky bucket and like how much money there is in the world.
And it really is. Some of these startups are just trying to weave a narrative that someday it will be valuable. In the WeWork case, I actually think it's quite plausible because it is actually driven by the same mental model we just talked about. Which is all transaction and all goods and services that people have in the world will be mediated by some form of software startup.
And then WeWork is way out on the edge because you can [00:45:00] either believe or disbelieve that's a software driven business.
Amy: Bingo.
Garry: Thank you guys. I really appreciate all the questions. Like, your community is the best.
Amy: You're the best. We are so grateful. Thank you so much. Have a great day, Gary.
Garry: Thank you. Goodbye everyone.
Outro: Thanks for listening to Getting2Alpha with Amy Jo Kim, the shows that help you innovate faster and smarter.
Be sure to check out our website, getting2alpha. com. That's getting2alpha.com for more great resources and podcast episodes.