Nov. 4, 2022

Dan Romero: Farcaster, building a decentralized social network

This is my conversation with Dan Romero about Farcaster - a decentralized social network being developed as an open protocol.

Apple Podcasts podcast player icon
Spotify podcast player icon
PocketCasts podcast player icon
Overcast podcast player icon
Castbox podcast player icon
RSS Feed podcast player icon

This is my conversation with Dan Romero about Farcaster - a decentralized social network being developed as an open protocol.

We talked about how product decisions in social networks have ripple effects on society, Farcaster’s strategy in the highly competitive world of social products, and Dan's personal philosophies around hiring and team building.


0:00 intro
2:08 why this problem?
12:54 both product and protocol
23:41 the algorithmic feed
29:40 Farcaster’s strategy for competing with Twitter
1:00:52 approach to team building
1:14:41 how to use social networks, and meme’ing

Relevant links:

Dan Romero -
Farcaster -
Farcaster docs -
Varun -
Keybase -


[00:00:09] Sina:  Hey everyone. This is Sina with another episode of Into the Bytecode. Today. I sat down with Dan Romero, one of the co-founders of Farcaster. Farcaster is a decentralized social network. It's being developed as an open protocol that is credibly neutral, like email. And it can grow to support many clients and products built on top of it over time.

Personally, I deeply resonate with the level of thought and intentionality that Dan and team are putting into developing this protocol and I really do think it has a chance at becoming foundational infrastructure and making a meaningful difference in the world.

This was an interesting conversation and I had some personal light bulb moments throughout. For instance, talking about how small product decisions in these social networks have ripple effects in society. And how there's a deep path dependency and how things evolve. how farcaster really has to win over users. In one of the most ruthless arenas for competition in the world today, which is competing with other social networks that have had years to mature and thousands of engineers working on them. , and we talked about their strategy for doing this, which I find to be quite brilliant.

We also got into some of Dan's personal philosophies around. how he's building the team and why they're bringing together a small lean group of senior engineers and intentionally not building a large team.

With that said i'll leave you to it and i hope you enjoy the conversation.

[00:02:09] Sina: one of the things I was actually personally very curious about was, when you left Coinbase, right, I know that you spent some time really kind of zooming out.

I I think in our first conversation you told me a bit of, you know, the political science, like history, kind of like deep dive that you did in that interim period. And, I'm curious,how, how that period was for you and then how you were thinking about what to work on next and why you chose this as the thing.

[00:02:40] dwr: So I'd been at Coinbase for five years. I left May 1st, 2019, and with an explicit goal of, of take it a full year off, I, I wanted to get bored. That was advice I had. Uh, you know, I talked to a bunch of different people. How, how do you spend your kind of a sabbatical after working in a company for a while and.

So I think I was generally pretty successful in doing that. I, I think Covid kind of made it a little less fun for the, maybe the second half, but for the first part traveled. Um, but my wife was off as well, so it was the two of us and took a kind of wide aperture in terms of the type of stuff that I wanted to be exposing myself to, especially on the history side.

And one, one thing is like, when I was traveling different countries, I tried to read a history about that country and a little bit more understanding of the world. So we started our trip in the Middle East. I read this ex incredibly dense book on the history of the Arab people and some, you know, 1200 pages or something like that.

And I, I, I think it was just exposing myself to a variety of new kind of ideas and, and just new information. And I think it was a bit of also a kind of decompression from five years in crypto and Silicon Valley. Right. And I think a little bit of time and, and space away from the day to day grind.

Left my mind a little bored and a little bit more creative in terms of like, okay, what, what do I wanna work on next? I have had this idea for building Twitter as a protocol, and, and a lot of people have had this idea, but, uh, there's an influential essay that Paul Graham wrote in 2009, and it's on his website.

So if you go to paul, it's, it's one of his shortest essays. And he makes this really interesting observation that Twitter is a protocol that somehow ended up within a company. And you could maybe say that for other companies, but specifically Twitter if, if you kind of think of.

Email is this decentralized protocol where I can specify the recipients and then they reasonably will receive that email. Yeah, you have spam filtering and email deliverability, but for the most part it's kind of this one to a defined number. Um, and then the web is this kind of, I can put information up there, but then people have to kind of find it and discover it.

And then Twitter is this kind of hybrid where I can put a thought. Uh, and it's not specified to the number of, of people from my end, but other people have specified that they wanna actually receive that information. And so the, the actual kind of original name for what forecaster was, was RSS Plus, because RSS in some ways has a very similar aspect to that where it's, I can tell you which, which blogs I wanna get updates from, and then you do this kind of pull model.

But the, the thing that's beautiful about Twitter is, is a vertically integrated experience. And so the feed discovery, the kind of easy navigation of the UI that you can kind of follow rabbit holes and click through usernames, find profiles, get all their content, all in this kind of really nice, slick app experience.

I think that has always kind of intrigued me is like, Well, could you actually build that level of UX polish but also have the backend be truly decentralized? And, and part of that is also I, I, my first kind of foray is into building. Was building on top of the Twitter api. So I was an early Twitter user in college and I built a polling app.

And, and so I think the idea that Twitter could function as a protocol, those first five, six years were the, the API was much more open. And you had all these third party clients. We actually have a counterfactual. Like it was a pretty exciting time. There was a lot of innovation that happened and then reasonably, the company needed a way to monetize.

They didn't want this fragmented ecosystem. And so they kind of put a cap on what you could do in the third party and, and gotta start working through their, their primary client and which I could show ads and things like that. So that idea had been kind of rattling around my head and I think with a fresh set of eyes on like how, how to potentially go build, it started to kind of formulate this, this general idea of, uh, what if you made Twitter.

Similar to something like Keybase, which for those who might not know what Keybase is, is kind of this nerdy cryptography startup that ended up selling to Zoom. But this really interesting idea where you could associate your other identities online, your, your Twitter, your GitHub, your Reddit with a cryptographic proof.

And then not that every user would need necessarily verify this, but you, you had the option if you were technical, to actually verify that the, the proof that the person's Twitter was, you know, DWR on Twitter, it did actually kind of compute. And so thinking that maybe you could do this and actually use a blockchain, uh, as that kind of public key to identity pairing, and then everything else could actually just magically happen off chain.

And, and then that way you could kind of like bring the whole system together so that you got the kind of best aspects of web two. Right. Speed, usability, no cost to do anything with this kind of, Sovereignty that you could get by actually having stuff on chain. That, that I think is the thing that clicked while I was traveling and a little bit bored to the point where I started spending more time thinking about the idea.

MAs around that, uh, did an exploration of like, how, how could you just improve RSS feels like RSS is actually pretty old school at this point, even though all podcasts, this podcast is gonna be distributed via rss. And I think that the, the summation of things, plus I reconnected with a colleague of mine, uh, from Coinbase we're in Trini Vaan, who also seemed kind of interested in working on this.

And, and we actually just then started working on prototypes because I think that there's benefit, especially if you don't have other kind of like we, we were self-funding it, like just action produces information. So we've said like, let's actually try to get a prototype working that simple system where we put the stuff on chain Testnet and actually get everything else working off chain with a client.

And that initial system ended up turning into what is now far caster.

[00:08:31] Sina: Yeah, it's, it's pretty wild how, simple the, primitives involved in Twitter and now forecaster can be, and how, like the kind of emergent. Things that they can lead to, like at a base level, it's really kind of like rss, like you're saying, right? You're publishing something and other people on the other side can say that they're interested in hearing it.

Um, was it through this process of kind of starting to work with varoon and prototyping things, was it clear to you that this was the idea that you're gonna go after? I'm also asking, knowing that you have a very long term kind of point of view on this, right? I've heard you say elsewhere that you're gonna grow to the first million users, you know, compounding 5% a week or a week or month.

Uh, but in either case, like it's gonna take multiple years just to get to that level. So you are, you're obviously at this point, committed to working on this for a long time to come. Like how did that happen for you?

[00:09:36] dwr: Yeah, it, it, it's good to kind of step back and I think one of the other things, having left Coinbase, I kind of had this thing in the back of my mind that Coinbase would go public and there was gonna be a decent economic outcome there. And that the thing that I wanted to work on next was going to have a, a long time horizon.

Elon is a very inspirational in that sense. I, I think the level of his work ethic is something that I, I was never gonna be able to go reach. But I do think that there is something that all people who, who can. Look at that and say, Okay, I have a long time preference here. Uh, can work on something that's really ambitious and know that it potentially takes a long time to get there.

And so, kind of going into it, I I, I didn't wanna build a company that I was gonna work on for three years and then sell to Coinbase or Google or, or something like that. I, I, I really wanted to work on something that I could see myself working on for 10 or 20 years and building a credly neutral public square for the internet as a protocol, which is the best way to describe Twitter.

Seems pretty reasonable to, to potentially work on for 10 or 20 years, especially given, I think a lot of the conservation that now exists with centralized social media and, and kind of, you know, it, it brought several billion people online and connected them. So I think to me, that's actually net positive.

Like, as much as people wanna say social media is net negative, but I do think that there are a lot of challenges when one company owns the entire stack. As opposed to something like the web or email where I think that there's a lot more competition and, and we, you know, could talk through any of those models.

But I think in terms of what I wanted to be working on is saying, Okay, what will I be proud of 10, 20 years later if, if it works? It's big f I think internet scale protocol that you could sit alongside T C P I P, http, smtp, dns and and, and then forecaster potentially as that. This is the social graph protocol that's incredibly neutral like Ethereum and like all this kind of stuff that, that lives in the crypto world, but now trying to kind of blend it because much of forecaster is off chain, right?

It, it, it it's much more of a traditional web protocol. I, I think that that's motivating and, and allows us also to make decisions that. Oriented towards the long term versus trying to hit some arbitrary number in a compressed period of time. So the, the invite model where we wanna be consistently growing but not growing too fast, has downsides in the sense that maybe it's something else pops up that grows a lot faster and ends up getting people to use It could happen.

But I think our, our point of view is quality is underrated. And in a kind of culture that puts a lot of emphasis on doing things extremely fast. Um, and I think that there is benefits to speed and, and kind of urgency, but in terms of building a social network, especially given that there are web to social networks at scale that already exist, we're not, we're not dealing with the land grab scenario.

We are dealing with the, the next best alternative is actually a really mature, polished product where people already have existing audiences. So I think taking a different tack of saying, Okay, we're gonna actually double down on quality. , um, is something that's a lot harder for an at scale network to do because they're already dealing with the issues of, of quality.

When you're at scale is, is extremely difficult.

[00:12:54] Sina: Yeah, it's, it's, it's very interesting. I think there's a deeper kind of meta point here, which I sometimes think about in the context of building anything new or anything new in crypto, which is, you know, if we're really building these long lasting protocols that are meant to be, you know, woven into the fabric of the internet and the world and outlast our lives, like they really, like, they need to be built on solid foundations.

But then on the other side, a lot of kind of discovering what crypto is able to do comes from running these experiments without any kind of like, thinking about like, how's this gonna evolve? Like, you can't really see around these corners in some ways. this is actually one of the things I've really resonated with how you're building forecaster, which is, it seems like you're being very kind of methodical about separating out the pieces that need to be robust and reliable.

And then being like, okay, there's this, you know, there's a way we can experiment with having, uh, you know, there can be different hub clients or there can be different algorithms for sorting the staff and there can be different applications, but they can plug into something that is going to be robust and sustainable.

And I think that's a very interesting thing to think about as we're building these protocols.

[00:14:13] dwr: one frame we have is if you start from the, the goal of what we're trying to build, this is an internet scale protocol that's credibly neutral that anyone can use to build a new social network. Right? That, that's the, that's the end, end state goal. In order to get there, in order to meet that bar, you need many different apps and services clients, open source projects built on top of the protocol that requires developers and developers.

Of effort and energy into building that set of, you know, apps and services. You're not gonna get that just from a technical specification. So, so what is actually going to attract the developers? You need a baseline set of competent technical decisions. But the, the primary thing that developers who are building, at least in social, they're looking for daily active users like that, that that's the, the liquidity, right?

If Defi is tl, like Dow, daily active users is the TL equivalent for a social protocol. And daily active users don't use protocols, they use apps. And so from our point of view, the only way we are going to even have a shot at bootstrapping these ecosystem, and it unclear if it'll, it'll work, right? Like, we're gonna do our best and, and try to take a long term view.

But you, you have to get an engaged, high quality group of people using a single app every day. To generate at least the initial amount of data that actually at a protocol level from the kind of data and APIs that you can access, starts to become interesting enough that other people wanna start to experiment.

Right? And as a result of that, if you're trying to do everything in this kind of decentralized, permissionless, like ship it and then do governance votes, your rate of innovation is just gonna be a lot slower. and I think that you need to just have a pragmatic point of view as saying, Here's what our goal is.

Here are the things that we've done so far to actually make it decentralize. And actually decentralize the core Primitives is a really important thing for people to start to trust you. But part of that is also just a, it's a narrative and a, and a belief in that team that, do I trust that these guys are going to do what they say they're going to do?

And a lot of that is earned through actions. Right. And I think, uh, examples here also, we, we have this initial client. But we're not trying to go gobble up every interesting idea that's happening in the ecosystem because it potentially drives more usage in our client. We, we actually need to be protocol stewards with this.

And, and so for us it's this kind of balance, which is not a web two model, right? Web two is all about bringing that stuff in house so you can better monetize. And it's not a pure protocol model because we actually have an opinionated point of view having built this initial client both on what the protocol should actually be doing as a result of running it for a year and realizing all the shortcomings that happened with the technical side of the protocol, but also we, we have daily active users and when they're not liking something, we need to actually make changes to our client that then we need to think through, Hey, is there a way to actually make this this default or this set of decisions we've made?

Can we just actually codify that in at the protocol level so that another developer, when they're showing up, they don't even have to run into the same issue. They, it just like kind of works in a. That they might not even realize, but like, that's actually been optimized at our client level and we figured out the set of trade offs to, to make it the protocol level, to make it easier for other people to build clients.

And I think that is the, the thing we're, we're gonna run into a bunch of pushback in terms of the architectural decisions we've made for forecaster. And I think that hey, it's healthy, right? Free market. Like if you have a better idea, go, go build. And if people use it, that that's all that matters. But I think what we've realized a year in is that the protocol is, should be in service to new developers, right?

So the, the, the marginal developer who's showing up, you should try to make the protocol as easy as possible for that, that developer to boot up a new app service client and actually be on a level playing field relative. The, the incumbents or the people who got there first. And so you can make decisions on how much storage there should be for, for a hub required, or, uh, how, how many we think is reasonable for decentralization, right?

You could get to a model where if you have five, is that decentralized? Some people would say yes, but from our point of view, we actually want many different clients and experiences. So as a result, we wanna actually make those hubs be excellent from a developer experience standpoint and reasonable to run for anyone who watch, wants to be able to play around with it.

And, and we have to deal with those trade offs versus other, other networks might be able to actually exceed our, our ability to do certain things. But then they're gonna have other trade offs, right? Like they might be more on chain, so there's actually gas costs that comes as a result of using it, or, uh, they will be less centralized or, you know, less decentralized because the requirement to run a hub is, is.

[00:18:52] Sina: Yeah. The, the relationship between the product and the protocol is a very interesting one. How, how, how are you thinking about, the product long term? Are you, do you see it as a core, like you're going to build it into a compelling, application that people use in the future, or is it more for the sake of fleshing out the protocol in some sense?

And, both sides of this can have kind of good reasons for them, right? There's, Unop, for instance, where the front end app is still one of the best ways to interface with the protocol. But then just in the last couple of days, there's like nouns, you know, disbanding their discord because they don't want to have this like, original thing that was somehow well positioned to, to begin with.

So I'm curious, how do you think about that question?

[00:19:43] dwr: I think Unis Swap is a great example of this. If Unop didn't have, would it be as successful

as it is? I, I, I think most people would say no. And so as a result of that, I think where we stand is having a really high quality client that has the compounding. Assuming we have competent product thinking design, you know, ux, if we compound that client over 3, 4, 5 years, that just becomes that much better for onboarding people into this overall ecosystem.

Now, the thing that we are gonna have to earn trust of is that, going back to this idea that we're not grabbing the best features from the ecosystem and trying to keep people, you can kind of view us, uh, I, I did imperfect analogy. We're like the New York City, Ellis Island of forecaster, right? And so you show up plenty of people who immigrated to the United States when Ellis Island exists, lived in New York, and it's the reason it's the largest city in the country today.

But there are also a lot of people who move to other places and they got to make new dis decisions and, and they belong to the, the protocol that is the, the United States, right? And so I think that that is much more of our role and we need to kind of realize the things that we're going to be good at and actually keeping focused and not expanding the scope of that app over, over time is probably the right call.

And I, and I think one thing that's really inspirational as a company like WhatsApp, whereas if you actually look at what the original WhatsApp was versus where we are today, yeah, there are been incremental product features, but the core of what WhatsApp does, Is more or less the same. And, and, and they were known for being like a relentlessly focused culture of like, we're gonna just be insanely reliable at doing messaging.

And I think that that's a lot of the product ethos that we wanna bring to our own client. Again, separate from the protocol. And I think as in, in doing so, we can actually help onboard more, more users to forecaster than, you know, anyone else, at least for a while. And as a result, I think that that benefits the protocol, right?

Because every daily active user of our client becomes an addressable daily active user for another developer building on top of forecaster. And in, in kind of viewing it like that, we, we have to kind of go again, earn our right, our, our, our reputation as being a steward. And, and we're like, Wow, they, they really mean it when they.

Every user that they onboard is an addressable user for me as an app developer, even if I'm building a competing client, and, and I fully expect if, if forecaster is successful, right? Talk at a hundred plus million people using it, billion plus people using it, we will not have 90% market share. Like if, if we do it like a, I don't think we would get to that scale and be, um,

[00:22:24] Sina: It hasn't succeeded as a protocol.

[00:22:26] dwr: e exactly.

So, so people can trust me or not. It's like I have a long time preference on this. I really, really care about the protocol succeeding and I think I'm just gonna have to repeat that message for years on end. And at a certain point, some people will trust me and other people will, will not. And that's fine.

And so like, I think we have to do that through actions. And if we, if we successfully do that and we can, can, you know, every time we do it, it actually starts to compound our reputation as like, no, no, they really do seem interested in the protocol. This is not all big one, big rug. That like, oh, sorry.

Corporate APIs, like we now run the protocol. Wait. Which I actually think if you, if you look at the protocol and we actually achieve the, the, the trade offs that we're, we're claiming to make with kind of our architecture, you don't even have to trust me cuz you, you actually are running a hub that has access to all the same data that we are.

And, and so I think that, that that's the, the challenge over the next couple of years for us to solve. And I think the early community believes in what we're doing and what we're saying, but uh, to get it to a much larger scale is just gonna require a lot of continual decisions that we make that people start to trust that like, hey, this is actually a protocol that, uh, I'm willing to go spend time building a business on, knowing that it's going to be a incredibly neutral, stable foundation.

[00:23:41] Sina: Totally. And I, I totally agree with you that the application compounding over time is the best gateway for onboarding new people. Like I've even found with myself, like when you added the algorithmic feed, that was just a huge game changer for me cuz it was starting to get to the point where it was just chaos in there and I couldn't follow what was happening.

And it just takes time to build a compelling product experience.

[00:24:08] dwr: Yeah, it, that's a great example where I think we were really hesitant to have an algorithm early on. Partially, cuz the way I like to use Twitter is reverse chronological,

Like I've been using it for 15 plus years. Yeah. And, and you know, well, up until recently I only followed a hundred people on Twitter.

And so for me, I'm a completionist. I, I would actually go through and try to get back to the part of the timeline that I had left off at. Like, it was like that, that, that's how I wanted to use it. I, I grew up reading rss, right. Like that, that's how an RSS reader works. I think more recently I've been trying to follow more people within the crypto world, especially people who are already on forecaster because it's actually good signal when people DM me to get on the network, which is currently the, the way you get into the.

I can actually see mutuals. So if a hundred plus people that I'm following across this pretty wide swath of people in the crypto world, then I have a pretty good sense of like, okay, you're probably someone who's gonna, you know, people are gonna find interesting on Farer. And so I've actually seen a lot more of the algo with Twitter, and there are definitely some parts that I'm like, Wow, this is actually pretty good.

Like, I'm getting a lot more discovery. But then you have the threads and, and some of the things that kind of take advantage of the algo. And so I think one of the things we were really sensitive on is, should we add an algo? Um, because it, that's actually also where a lot of people get frustrated. It's like, well, okay, what is the algorithm rewarding?

And I think we got comfortable about a year in adding a very lightweight one. We call it highlights. And, and you, you, you know, you're actually there. It's like the, the default feed is following. And then if you tap into highlights, you're kind of opting into seeing this. And one thing that we've emphasized in highlights compared to maybe something like Twitter, is that the primary way that we kind of bring conversations up in this algorithm, Is active conversations and, and it's, it's not the kind of like one banger, we call them cast, but like tweet that that is like a dunk or a quote that has a lot harder of a time getting into that algo.

Whereas if you get something that actually has depth, if you think of the conversation as a tree, that you have all these kind of interesting things happening that that's, that's factoring into the algorithm a lot more. And what's important is because this is the default client that most people are using that is actually nudging the, the behavior and norms of the network towards a more conversational platform.

And I think that this is one thing that abstractly I kind of understood, but now that we're seeing it, is the product decisions we make in this initial client have a huge impact on the shape and behavior of the network and the protocol generally, especially in this kind of constrain environment where we are doing a lot of the on we we were doing, we're the only way to onboard right now.

And so I, I think one of the beliefs we have is if we, if we keep it invite only for. Somewhere close to a million, right? It might be a little less, might be a little bit more. And we move to a model where people in the network are actually doing the inviting rather than everything kind of coming through me.

I think that those defaults are really important because whereas right now you onboard with me kind of like in this chat and, and I can kind of reach back out to you if something, hey, like that

[00:27:10] Sina: Totally. And it's like, I, I actually spoke with this person. I can't just go and be a total asshole on the other side of this.

[00:27:16] dwr: it's like being invited to a dinner party. At dinner party with 4,000 people. It's like you do feel a little bit less like, okay, I should be a little bit more civil. But inevitably people are gonna invite people, they're not gonna have any other contexts are gonna show up in the network and they're gonna behave the way they wanna behave.

And so actually having solved for those defaults that reward kind of call it constructive conversation. And I don't think we've, we've figured that out completely. But if, if you can actually do. By iterating while the network is small. Then I think as you scale the, the average experience for a new user as well as kind of power users still oriented towards quality because you, you've kind of made the set of defaults in the interface oriented towards what you want people to do rather than what an algorithm might.

So I think that, that, that's how we thought about it. And, uh, we're not perfect, like we're gonna make mistakes, but I think part of the perk of like I use the app a lot, um, is if things aren't going well, like I'm gonna get a pretty quick sense of that, and then we can make the decision that, like how to change it.

And so I think if we were just working on the protocol and then we had other developers or the ones building clients, it's, it's way harder to convince someone to cha, like you have to do a lot more influencing. and you could criticize that and you could say, Well, okay, so then it's just basically you're making the decisions of who gets into the network.

You have the client and you have the protocol. But I think going back to this idea of earning the trusted developers, which we have an early group of developers who've been building on top of forecaster despite, I think, an okay dev experience, But, but the idea is that they're realizing that the community is really high quality.

The data and APIs are actually open. And they're seeing me say on a consistent basis, No, no, no. Like we're, we're actively trying to figure out ways to actually support other developers building and like, these are the tradeoffs that we're making. I, I'm optimistic that we can actually get to a place where we have a competing client in the next year that has a reasonable level of ux.

Cuz that's, that's the other thing. There's the theoretical decentralization. Oh, well you could go build your own client if you want, but like if we kick you off our client, you effectively are kicked off the network. So I think that there's a practical level of decentralization and practical bar that is much higher threshold than theoretical.

Like, yeah, sure, go write your own client. Like no one, no one's gonna do that. They're just gonna go back to Twitter. Whereas if there were two legit clients above a certain baseline of ux, then I actually think you have something

[00:29:40] Sina: For sure, and, and the cra The crazy thing is this is making me realize what I had missed, which I do think about sometimes, which is I'd been focused on the protocol and the protocol is what matters, but the client is actually. You know, yes, there is, uh, the, the design and the architecture of the protocol will deeply influence the things that emerge out of this over the next 10, 20, 50 years.

But really, it's the client, the early client, and the decisions around small product tweaks, like the shape of the algorithmic feed, the what you prioritize, the people who seat the community, these things are designing the kind of social layer of this entire kind of wave. Right?

And in some ways, I mean the, the cra like planning for the success case, right? If this thing actually does grow, to be used by billions of people, like these decisions that you're making today could have ripple effects into the shape of society in the future, the same way that, you know, Facebook and Instagram and TikTok like this product decisions that they make ripple through, like, you know, just how, how people feel and think on a personal level, how societies relate to each other.

And that's the super fucking crazy thing about, about what you're doing.

[00:31:08] dwr: Yeah, I think it's well said. Actually, biology, uh, casted yesterday after being off for a little while. He came back. And he made an observation where forecaster I is effectively community first, not technology first. And I would quibble a little bit to say that, that we've actually had to make some interesting technology choices informed by usage.

But I do think that there is some truth to that in that you can make all the right protocol choices, like perfectly designed protocol thought through everything like beautifully written software. But if, if you can't actually bootstrap the initial community, which is a little harder than people just choose to use the technically best thing, they don't, they, they choose to use the thing that is most interesting.

And the thing that makes it most interesting most of the time is other people. Right? And so I think the story of like, how did we get from, you know, 10 people using it every day to over a thousand. I don't have like a perfect, you know, we thought through this and we knew it was gonna happen. It was a bit of.

Onboard a bunch of people have a bunch of people turn, continue to make week over week progress. People refer other people because it's kind of interesting. And at a certain point you hit a conversational liquidity point that the the network, I can open up the app and there's always something new and interesting assuming you're following a sufficient number of people.

And so we're now at a level of, I actually think we have something where people are referring people to forecaster. You should check out forecaster, not because of the, the product decisions and the technology

they influence, it's the people. So that, that becomes another question is like, well, okay, how, how are you actually going to compete against something like Twitter?

Right? It's like the 200, 300 million people who are using it every day and we have a thousand plus. But what I would actually say is you gotta break that problem down differently. So something like Twitter is actually, call it ninety five, ninety 8% entertainment and news for people who use it. And then a very small subset who are these kind of like very online public figure.

very much a tech or crypto who actually use this as this like weird public chatroom social network, right? Like my, my parents use Twitter, but they don't tweet. Whereas, you know, I, I, I'm having conversations with strangers all the time and they, I, I think that's completely normal. So they're actually a very small percentage of people who use Twitter as a social network.

So right away you can actually go from, Oh, this is a, a network of two or 300 million people using it every day. How could you compete with them too? There are several million people who use it as a social network every day so that that becomes a, a more achievable number if you think about it. And then if you actually say, we're only gonna focus on one community to start.

which is what we, we have, and it's, let's just focus on the people who might on the margin, and I don't even think this is a primary driver, might be finding that this is a decentralized protocol ideologically kind of interesting, some NFT features. None of those are game changers. But I think it, if you, if you focus in on that, what, what is the actual size of crypto?

Twitter, a hundred thousand daily actives, Maybe less so now. Now if you, if you actually say it's, well, if I have a thousand people using the app every single day and there are a hundred thousand people in my addressable market that are at least trying to go after, I'm 1% of the way there. So like that, that like changes the game in terms of like, okay.

And I actually think we've been even more specific. I'm not even trying to win all crypto people. Like I love Bitcoin. I own Bitcoin. I, I, I don't think Bitcoiners are gonna wanna use anything that hast anything that a theory, maybe at some point, if it's interesting enough, fine. So what, what I'm actually focused on as Ethereum people and, and people who maybe are like, you know, in a different ecosystem but don't necessarily hate Ethereum.

That's, I think, even smaller number. And so like that, that's the goal. It's just like, just focus on people who are building and intellectually curious and Ethereum make that the best social network for those people and, and permissionless innovation and, and nice integrations. And then from there, see if you can actually convince another concentric circle of people and, and, and kind of work your way out.

And so I think that that has, uh, been informative in the sense. We don't have to, work on like virality and, and all of these things that I think people just take for granted with social networks. And if anything, it's actually much more kin to how Facebook, which if you look at any social network, it is by far the largest, even if it is kind of declining in dau.

And what do they do? They started at a one college and then they expanded to another set of colleges, all, all kind of permission based, and then eventually they, they expanded to everyone with email address. You can think of that as like, almost like dot e, but it was invite only in the sense that you had to have a, a, a credential to access Facebook for multiple years.

So the most successful social network in all of web two was invite only for two years. And, and then people are saying the way to grow social network is to grow as fast as possible with virality. It's like

[00:35:53] Sina: Right.

[00:35:53] dwr: maybe, maybe there was a lesson there.

[00:35:55] Sina: university students turn out to be a very interesting cohort that's getting older and other people wanna be around them, which could also be potentially set about the current cohort of like Ethereum, crypto, Twitter. Like this group is going to become way more influential and nerd sniped the rest of the world potentially.


[00:36:14] dwr: Yeah. That, that's, that's the belief. Like, so for me it's, it's dot e are, are the equivalent. And it's like start your seed, your social network with the optimistic, the people who are curious building the future in e And I actually think that, that, that is a, that is an interesting group of people to go on a journey to see if you can build a, a incredibly

neutral, decentralized social network. Yeah. And, and, and maybe, maybe it never achieves that scale, but we're gonna, we're gonna try to do it.

[00:36:40] Sina: This is, it's actually, it's, it's, it's easy to not, like there's, there's like real kind of strategic brilliance here I feel of like not, not to, not to,

[00:36:51] dwr: the hypothesis. How about that?

[00:36:53] Sina: you too much, but, yeah, carving down this problem of how do you compete with these incredibly mature social networks that are, it's basically the most ruthless like arena of competition in the world, right?

Like the way TikTok and Instagram are competing, Like, it's insane. And to carve that down into, We don't need to get everyone like size it down to the dot East people and then design, the protocol I imagine would, I mean some of the protocol features are probably being prioritized cuz they unlock certain abilities of the product layer, but the protocol's gonna be what it's gonna be.

But the product showing NFTs and like these sorts of things and the community you're seating it with. So it's not only like a, you know, target user who we're going after, but it's also like the product features that you're working on is all designed around getting this initial cohort. Um, I feel that's super fascinating.

[00:37:54] dwr: look, I, we'll see if we can do it. I think that there are credible other projects out there that maybe their strategy ends up appealing more to these, these types of folks. So that, that's like one way. One other thing is the stated preference is we want this, but the reveal preference is, I like having a million followers on Twitter and I'm not really at risk of getting D platform, so it doesn't matter.

And I think that that's also pretty reasonable from the individual's perspective. I'm optimistic that if you focus on quality and try to just make the average conversation on ER is one of the highest quality places that anyone is talking about anything, Crypto or Web three, and I don't even think we're there yet.

I think we have quality, but I think it can get even better if, if that can happen, then I actually think you can attract a very interesting group of people to be using at least a client, hopefully ours, at least a. Every day, which then developers all of a sudden say, This is actually a really interesting audience to go build an app for.

Because not only, not only are they interesting people within a community

[00:38:57] Sina: Right. You bootstrap the other side of the marketplace too, cuz these are people who are early adopters who have money to spend, who have like, you know, creativity to lean into new use cases.

[00:39:09] dwr: with, with also the primitive of, you know, every user on forecaster has to have a public and private key, which as a primitive by itself is really powerful, right? Like, I think it's one of the reasons people are so excited about building an Ethereum is like when you can identify every user by a public and private key, all of these zero knowledge use cases start to pop up.

It, it moves from the realm of theoretical because you've solved the, the like cold start problem of, well how do you get public and private key infrastructure for every individual? It's like, turns out, it's like some of the financial speculation that I think has driven cryptocurrency adoption has built out a huge ecosystem.

That you can now play around with some of these ideas that if you go back to the web of trust, G ppg, pgp, like all, all of that stuff, that was always kind of this dream. And it goes back to this idea that keybase, I think very noble in what they were pursuing. I one of the most like mind bending products to think about like if, if they had actually been successful at hitting some amount of the escape velocity.

I think you can kind of walk through the idea as to say maybe it will never be a consumer use case, but the idea of actually getting everyone to be identified by a public and private key, let alone a cryptocurrency address, just, just simply a public and private key, I think unlocks a completely different set of experiences that you as a developer can build and it, it, it fundamentally would change how the way humans communicate.

Right? So one easy example and to and encrypted experiences get way easier to manage. If, if you can make some assumptions. Whereas basically the only, the only at scale and, and encrypted you have signal and then you had WhatsApp. And the only reason they did it is kind of one of the co-founders was really passionate about implementing it, which is an amazing achievement.

Um, but basically everything else is kind of like done in this surveillance state, uh, infrastructure, which is kind of wild if you think about it, especially in a place like the US where we have very strong pro protections around freedom, speech and privacy. And because we just decided that as a society, reasonable individual choices, that convenience trump's kind of some of these core tenants.

Now basically all of our communication belongs to a bunch of big companies that you, you, you have no, you have no reasonable expectation or privacy with. And, and I think if you spend more and more time online, like you, you should be able to use experiences that aren't clunky, but also give you your kind of like constitutional rights, uh, in a variety of different ways.

[00:41:35] Sina: Yeah. it's hard because there's a path dependency to how it develops, Right? And the trade offs that each individual makes at their own level of, I don't wanna use this app. That's like, hard to use, gonna take like five seconds to load all that stuff. I, I mean, even me with forecaster, like I haven't actually, like, I'm just kind of, I'm in the cohort of users who's, who's like converting finally I feel, and for sure with this conversation.

But, so the, so the user makes the choice of I wanna use this thing that's easy to use. And then the net effect of that compounded on itself for five years is that the entire world is using Facebook and all of our data can be read by, you know, by Facebook and,

[00:42:20] dwr: Yeah. And, here's the thing. You can either get mad at that or you can just, It's, it's a, it's basically a law of human psychology. People are gonna use what's convenient and easy. Like they, they love free. That's why ad models work. And so you can, you can get mad at that or you can say, How can I build something competitive with that so that the user looks at it and goes, Oh, this, this feels basically the same, but could potentially have a very different set of underlying principles that, again, might not even matter to that user, but they're using it because it's, it's solving a problem.

Or, or, or it's providing some utility to them. But all that stuff underneath the hood is abstracted away, but has enabled a whole bunch of other use cases because it's fundamentally built on, on a protocol. And you, you have these public and private keys that you can actually use as a developer to create new experiences.

I think. I think that is, , Um, I'm excited about if, if we do hit some amount of scale in, in many ways it's actually very similar to what Apple does, at least from their marketing standpoint is, and, and, and to their credit, I think a lot of their decisions are actually privacy enhancing in, in a good way.

But then it's like, Oh, well we need to increase revenue, so let's build an advertising business and kill everyone else's advertising business and ask me not to track, but the Apple can track me unless you go shut it off in this esoteric menu. So it's like, okay, so fine. I don't really have many alternatives, but you can imagine a world where that becomes the kind of base level is, is norms are around kind of privacy encryption.

And then, you know, people can compete with that being a, like a core primitive of what you can do rather than, you know, maybe a, a corporate marketing strategy, which they changed

[00:44:02] Sina: like, I feel like the gravitational pull of the other, or the existing model is so strong both from a, you are the, you are an entrepreneur, developer, you're building something, or from the user's point of view, and like the ripple effects that these things have in the shape of the world and society over time.

And what's crazy to me is, one, how it comes down to. nuanced product decisions and like concrete execution right now, you know, and like onboarding these users and getting something that works like Keybase didn't work because of, you know, slight things about their model. You know, it, it's not enough.

If you believe this is a better world, you have to actually like, out execute the other people who are running downhill. And, because the other way is easier. it also requires, you know, a founding team and early people who are working on it, who actually hold these beliefs, um, truly right, and are principled about it because the other, like, it, there's just so many off ramps where you can sacrifice them in some way.

And this is the same thing that I've always admired about Vitalic, for example, right? Like Ethereum has. Like the, the staunch, like, no, we're building this decentralized infrastructure that could be. Like a protocol that much of the world taps into in the future. Um, you can't be kind of optimizing for short term wins or personal financial gain, even though, like those things will also happen naturally as a byproduct.

But it's, it's. Almost wild. Wild how I feel like if you ran multiple, you know, simulations of the world, like if there wasn't Monte Carlo simulation of the world, like the world where something like this actually exists and takes hold is maybe a rare one. Um, but then once it does like get past a certain point, it's like so strong and robust cuz it's gonna suck in all the liquidity.

It's gonna be something everyone trusts, like can't be rugged. Uh, it's, it's very interesting in that way. And I feel like, I mean, just even thinking about that, I feel like in the journey of forecaster there will probably be this inflection point, right where it switches to, okay, now this is the credibly neutral, you know, communications protocol of the web.

And if, if it succeeds and then it starts just like out competing everyone else much, much more easily.

[00:46:36] dwr: So two things there. One, I think Ethereum and Vitalic are, are deeply influential on how Verin and I think about this, and there's a reason we're using Ethereum ML one despite people saying, Oh, the gas costs or this or that. I think for, for the identity primitive that makes forecaster work, outside of what's off chain, we couldn't think of a better network to go put it on.

And I actually think optimizing for cheaper transactions per second or any of that stuff is, especially if your architecture actually doesn't need most of the stuff on chain, you only need to do it. What better chain to go put it on, because it's actually capturing a lot of how we approach things is this pragmatic approach to building, but also a deep philosophical belief in decentralization.

And I know there are plenty of people who wanna criticize Ethereum for not being decentralized. I'm not gonna convince those people. But I think for, for us, building on that infrastructure like that, that's the same thing that we wanna achieve for our own protocol, is that we want people to deeply believe that, that they're doing this because they actually want this to hit a level of scale.

Incredible neutrality, similar to Ethereum. the thing is, is like, will, will it work or not? I'm not sure, But we control our own destiny in the sense that we don't have to rely on anyone else. We don't have to convince anyone else other than the users to use the protocol.

because we, we've built our own client and if we can continue to grow, the number of people who wanna show up every day have conversations, even if it's just one client, we do that for two or three years and then people start to get interested in building on top of it cuz we've hit some sufficient number of people building.

I think that's totally fine. Like, like that. I think that's a reasonable path. And people might criticize, Oh, well you, you were, you didn't really have a, a thriving protocol until then. But I think what matters is, does the protocol have actual people using it? And like, do I actually wanna build a product or service that go after those people?

I think that that's it. And so I think just take the time preference of saying like, this is what I wanna go do and, and be willing to realize that it's gonna take a long time. you know, easier said than done, but I, I think for what we were optimizing for, Going back to working on a problem for a really long time and having a clear goal of like, where we want to head this, right?

Like there is no version of selling this to Facebook or Coinbase or whatever. It, it's, the protocol should outlast any company if, if it is successful. And so I, I think one analogy here is when the internet happened in the nineties, AOL in the US was this huge platform in terms of getting people online, but AOL, in the early days, didn't have support for the internet.

The internet was purely in, you know, academic thing. And then Mosaic came out and, and then people had other providers that would provide access to the internet. And then eventually AOL added the support from, instead of just having this closed network, you could actually go use the open network. And I think that the point you bring up of there will be an inflection point if it's successful, where, because the amount of kind of innovation that's starting to happen on this open ecosystem and the kind of ability for anyone to go build a business and not have to worry about gatekeepers over here, I do think has a shot of happening.

But I, but I would suspect it's going to be a while. Uh, and I think having the mentality that you just know it, it's like, all right, there's probably another year or another, two years or three years before that happens, helps you manage through the cycles of the ups and downs because you, you don't expect it to happen tomorrow.

And it's not gonna be like this sudden, You know, magic inflection point as much as it, it, it kind of flips. And then that in itself starts to be a compounding curve. and so if you actually just think through, you know, whether it's an S-curve or you know, like exponential growth, you, you realize these things go for a very long time where they're small, if it's actually on a curve, and it's just hard to know where you are on that curve, right.

[00:50:33] Sina: Right, right. One of the things I was gonna ask about is in relation to competition. So we, we talked about, um, Yeah. How strategically you are competing with these existing social networks, but I'm curious more as a, more like as an operator and the person who's doing this right, from a personal, like emotional point of view, how do you think about the notion of competition?

How do you, how do you relate to it? And I imagine, like, one of the things I was kind of trying to put myself in your shoes was seeing all of this stuff happen with Elon Musk and Twitter, right? Who is like one of the most formidable, like the most formidable entrepreneur of our time, uh, has bought this company.

I think it's going through is like letting go of 75% of people supposedly. So it's a new kind of like force that's interest entering this, this dynamic that has a lot of moving parts already and. part of what I'm, what I've taken away from what you've said, which resonates with me too, is if you are really just focused on your own, if you really believe in what you're doing and you have a long term point of view, and you are in control of your own destiny in some ways, right?

You can, you can just kind of shut all of that out. Um, but I'm curious, just also you having been someone who's, you know, was at Coinbase through all these years, and just as, as a, as a person who's doing this right now, how do you think about these things and at what can you share about that?

[00:52:11] dwr: So I think I learned a few things at Coinbase. Um, one, and I think people don't appreciate this from the web two element, cuz Coinbase is very much a Silicon Valley company and, and you know, big in the sense that it has a lot of users. Coinbase doesn't have a network effect. Coinbase is a brand, a brand that people trust.

There's convenience, uh, there's, you know, a lot of, lot of features, uh, you know, payment methods, all, all the countries that it serves. But you can buy the same ether Bitcoin at a competitor and probably for a cheaper fee. And so for the, 10 plus years people have been saying, Oh, Coinbase fee compression.

Uh, the only thing Coinbase has done is double prices like in 2016 went from 1% to 2% and everyone said, uh, you're gonna have all this fee compression. And, and you might even look at their public financials and, and say, Oh look, the average fee is going down. It's actually just cuz the institutional volume is increasing and so that's cheaper.

And so it's a mixed thing. Prices probably come down over time, but, but fundamentally, Coinbase actually is much more kin to what, at least initially drove a lot of what people picked Apple for in the sense that they, there was a brand, it was kind of a set of D decisions if you did every feature. You know, the classic example is the iPod compared to some other MP3 player.

If you're just strictly doing it on storage functionality, something better. I think, um, Ben Thompson's talked quite a bit about this of like the, the thing that a lot of the, the kind of analysts when they, they do everything data driven, miss is, is taste and consumer preference. And the idea that like, people like things that, they, they can resonate with.

And I wouldn't actually say Coinbase has that level of love that compared to something like Apple, but I, but I think actually what people underrate is, most people who get into crypto don't wanna have a hardware wallet. They don't want a user control wallet. They actually just want something that feels safe.

And the reality is that for a lot of people, Coinbase just is super easy and it, and it makes them feel safe and the company is taken security really seriously. And, and I think that that's paid off for them. So, shifting over to kind of thinking about that for forecaster, the reality is I. The network effect.

A, if the protocol's successful, it's actually you've commoditized the network effect. You've said the network kind of is a neutral ground and anyone can kind of tap into it. And so as we think about building our client, our client is very much assuming we actually follow through and, and don't do things that are adding features that create kind of lock in.

But I think the, the reality for our client is it's always going to be based on do consumers or, or you know, if we do a subscription model, do our subscribers, do they find value in this client? Like are we adding the features that people actually want and providing some amount of value to them to, to actually use it every day, otherwise they're gonna switch to something that's better, right?

And so I think that that's like a mentality that, a lot of web two marketplace companies are with the network effects think that they have, but the reality is, is they can get very far off of just taking the same kind of network effect with proprietary liquidity and, and run for a really long time.

The one I love to point out is Airbnb. Amazing design product is, I think generally really great stand at plenty of Airbnbs, but the fact that people complain about this cleaning fee and they have made this decision, AB testing or conversion rate, that no, we're gonna hide the cleaning fee until the end.

If there was a real, I think, marketplace out there of like you assume Airbnb was a protocol that as a client, I don't think would cut it. And so they're able to kind of exist on the fact that the way that that model is structured, and they did a lot of hard work to build that proprietary liquidity, but it isn't a protocol mindset.

And so I think that that's like one thing that is from Coinbase that we very much resonate with is that we're only as good as like, what's the last feature we shipped or what value are we providing to, to our customers? I think the second thing on competitors, and I, and this is a little bit more of a philosophical point, but I very much believe in Paul Graham's philosophy, and it's a little crude, but you're more likely to dive as a startup from suicide, not homicide.

And, and I think it, it's a little bit of a Hollywood version of startups where you're in the arena competing with others and it's like the competitor launches a product and you gotta follow. And, and, and that's how people think. And, and I think that the reality is all people think that they're competitors are out executing them or, or they, they're perfectly polished when the reality is, if you know anything about a startup, they're pretty chaotic inside.

And I think that if you spend too much time worrying about competitors versus just if you have users, like what do your users want? And a really simple example of this is every so often on forecast, I just say product annoyances. Like go at it. Like tell me all the things that you can't stand about the client.

And if I was worried about like what competitors are launching, I could like go build a product strategy based on like trying to compete there versus just asking that simple question, getting a hundred responses. And then you kind of realize like, Oh shit. Like there's a lot of stuff that we can just improve with our own thing, let alone any new features.

I think that is a lot more of the type of thinking that people should be orienting towards.

[00:57:04] Sina: It's interesting talking about that on forecaster is like the extreme version of tell people about your idea. They're not gonna steal it, they're gonna give you good feedback, but you're literally posting it publicly for all potential competitors to come and see the list of your deficiencies and where you know what they can do and, and all of that.

[00:57:24] dwr: Yeah, if you, if you're not intellectually honest with that and you can't actually even let your competitors look at it, I, it's, I think it's gonna be very hard for you to win, especially with software, right? Like, I, I like other industries might have slightly different models, but the, the industry I know is like internet software.

And I think that the, the kind of like, look, all our metrics are public, All our competitors, like in the web two sense or web three sense, their, their metrics are public. Web two, they're public. So you get a sense. So, so, so in some ways it's like everyone kind of knows ev all the dirty laundry that exists.

And the thing I think you're like, instead of spending time on your competitors, I think what you should actually be focused on is again, another, Paul Graham built something people. And so for us it was, and another thing he said is, you know, find a hundred people who love your product than a thousand people or 10,000 people who like

[00:58:11] Sina: So much wisdom in that.

[00:58:12] dwr: Yeah. I, I, I think like any entrepreneur who is building something new, Paul Graham is like, that's the zero to one bible in my view, because it helps you get the right mindset. And a lot of them sound trite, but, but the reality is I think that they are the, the correct, uh, points of view hard to follow. Cuz sometimes when you're going through a day to day of the emotional rollercoaster that can, uh, you know, it's like, okay, I don't need any more trite advice from Paul Graham, but, but if you actually step back and, and remove some of the emotion from it, it's, it's probably right.

And I think the, the miss is people tend to think that, like doing a startup is like this kind of like strategy chess thing that you're thinking about competitors and where to kind of, that is much more at the one to 10, or frankly 10 to a hundred scale. That now you have like big, big groups of people and it's like, okay, where do you direct them?

Whereas as an early stage startup, it's just you, you, you wanna be as close to the metal as possible, which is what are your users saying? How are they feeling? And so it's, you know, talk to users, find people who absolutely love the product, double down on what seems to be working there, hopefully on end, but a local maximum.

But like overall, your, your competitors are not gonna be the way to, to kind of like find the unlock. It's, it's actually in your community of users. And so I I I think people


rotate on

[00:59:28] Sina: also don't know what they're doing. And you've got, you've got as good of an intuition or better one with the users that you have and what they love.

[00:59:36] dwr: Yeah. I would actually say it's, it's, you start worrying about your competitors. You're starting to let them drive your

[00:59:43] Sina: Roadmap.

[00:59:44] dwr: roadmap, which if you have good intuition, now you're actually imported someone else's intuition, which you don't know if that's actually good or not. And, and yeah, there can be some growth numbers that let look exciting.

But if you're not confident even enough, like the the point like why did you even start working on the thing in the first place? If, if like basically what you're, you're doing is looking at another thing and saying, Oh, well they're doing this, like we should do this too. Now, asterisk, like sometimes someone actually unlocks like fundamentally something new and you know, it's probably useful to know what other people are doing, but outside of a rare instance, I don't actually think.

And, and in that case, even if you try to go copy it, it's unlikely if someone gets this breakthrough thing, product, strategy, whatever, that somehow you're gonna copy it and then out execute them on the thing that they already have the insight. And so I, I think, you know, call it whatever percentage of time people spend thinking about competitors, it probably should be a 10th of the time, if not, you know, even lower.

And then that extra time can go back to thinking about your community, your users, your product, your protocol.

[01:00:51] Sina: Totally. And, and then the, the kind of, um, mirror to this question is with this, uh, mindset of everything's open, we're gonna talk about things openly. It, relies on an abil, like in a, in a knowledge that you can execute and you can, you can move really fast and you can, you can actually like ship the things you're gonna do.

And on this point, your. Actively hiring kind of like staff, you know, engineers, and you're keeping the team very lean.

And this is, you know, this is not what everyone does. There's people who just go hire like 50 people immediately as soon as they raise money. Um, so how do you think about the team architecture and how that reflects into how you're operating and building what you're building?

[01:01:49] dwr: So Varun and I both worked at Coinbase. I was employee 20 and you know, I think Varun was around play 50. We saw a thousand plus people get hired, and I can tell you from my experience is when we were a thousand people, I don't think we were moving any faster than we were when we were 50 and. I think there's a variety of reasons for that, but, but fundamentally adding more people adds more complexity to doing anything with a group.

That's why I think generally great things are achieved with much smaller groups of people, like in terms of like great creative output, because managing a large group of people, you know, the, the top of the pyramid talking down to the bottom, it, it just kind of doesn't work. And so I think where we are approaching this from is we still have a lot of stuff to figure out, both at the protocol level, but also the product level.

And having a large organization means I'm outsourcing a lot of those decisions to people who have been, you know, they might be smart, but they just have been working and thinking less about this problem than I have. And I think that, like you, you can tell like forecaster like it, it's very deliberate in the sense is like every decision that we've made in the app was made by me, but also, By Veron and the team's input, right?

Like they're there. It's, it's, it's actually multiple people who have kind of like really put an imprint on it versus I think in a larger organization you end up getting a lot of individuals make decisions and then it kind of like filters up. And I think that the other general point of view is I think most companies just hire people because that's what other companies do.

It's not a first principal's approach. So example, Instagram got to 27 million users, I think more or less with 13 people in 2012. Arguably one of the most successful products of ever, if not, you know, any type of software network, period. I mean, it's basically where all Facebook's value comes from today, like in terms of the growth and like monetization.

And so that happened in 2012 and, and we're in 2022. And you're not saying 10 years of developer productivity, like with all of these cloud services and tools and, and you know, improvements that you can't go build. Uh, a social app with that many users, with that few people. And, and, and the other example is I think WhatsApp.

And there was a viral tweet today, JD Ross, who's the co-founder of Royal, had it, Elon responded to this by the way, so he, he brought out the fact that WhatsApp had 35 engineers with 450 million users. That's more than Twitter. And so Twitter has 7,000 plus people. Today they're talking about doing a 75% reduction.

Look, anyone losing their job like that sucks. Like you, you know, you've got bills pay, et cetera. But just if, if you take the abstract of going from 7,000 people to 2000 people, and WhatsApp again, almost a decade ago, was at 450 million people with potentially way higher daily active usage in terms of what you're kind of sending around that, That's, that's an insane ratio.

And, and I think the last point, and to specifically go to crypto, Coinbase got up to 6,000 people last year. Fdx is 200, maybe, maybe a little higher. Roughly the same trading volume. And yeah, there, there's other things, like Coinbase has a hundred million consumers and like there's a lot more there in the us there's regular, but I, I can promise you that if, if one has 200 and you have 6,000 and you're doing roughly the same volume, the one that has 6,000 has too many people.

And so I, I think just if you, if you approach it like that, you, you should be really, really hesitant to add headcount, uh, if, if you don't need it. Because what you're doing is you're, you're decreasing the ability for your organization to move fast because you're adding overhead. And some businesses are gonna need that, right?

Like if you have an operations heavy business, like a business like Bordash or like, yeah, of course you, you're gonna need to do that. But in a case of building a protocol, Like if, if you actually build it correctly and then you have this client that's on top of it, you, you should be able to get a lot of leverage and scale with software and, and, and these tools.

And the, the, the other examples I'll give you is I don't think Unis swap was that many people for all the volume that's gone through Unis Swap and then, and then open C even, right? Like yeah, Open C last year had to end up hiring a lot of people. Very similar to Coinbase in 2017, but they hired a bunch of people and now volumes have come back down.

So in, in some ways , like in, in hindsight 2020, you might not have hired as many people cause they, they, I think they had to do a layoff. So it it, it's this, you, if you, if you think about the systems that you wanna design up front, you have senior people who actually know how to build those systems rather than build a system that works today and that has to be replaced two or three times along the way.

I think you can get a lot of leverage with a really small.

And frankly, it's just more fun, right? Like Jeff Bezos has this concept two pizza teams. Like there's, there's kind of like a fun aspect of like, we, we had the team on site, everyone's remote a couple weeks ago and, and we had a night where we just played out at the board game Terraforming Mars.

There are only six of us. Or you know, and so it, it like, that's pretty magical when the group of people that's working on something that hopefully ends up being scale, can sit around a table and play board game and, and you know, have two pizzas. And so I, I think like for us, like getting much bigger in this early stage where we don't know if we've actually figured it out. It's just your, your carbon colting Google, which is what I think most startup things in Silicon Valley. Look at Google. Which is the best business model of all time. And they go, Oh, well we should offer these perks and we should do this, and that, and that. And it's like, yeah, if, well, if you have a search engine that kicks off a hundred billion a year, I would offer these things too, and have all these, you know, moonshot and venture firms and all this other kind of stuff.

Whereas the rallies, most businesses aren't like that. So maybe be a little bit more ju judicious in hiring.

and does that, does that flip at some point? Like when you get it to a particular scale, does the mindset shift I mean, I could, I could see it also not if FTX is an example and I don't see from first principles reasoning why you need to get to like thousands of people.

[01:07:48] dwr: and look like I, I grew up on the operations in BD side at Coinbase, I don't have any interest in hiring a BD person. I, I have very good friends who are, you know, business development. I don't, I don't see it adding a ton of value. I know people talk a lot about BD and like crypto and like telling you it really, like users don't really care.

so that's like a whole function you don't need, uh, I think you can be really smart about finance, like in terms of just like building out, know, I think one other thing that we've done is like, most of our stuff is off chain, so we don't need to deal with all these like money transmitter licenses and all the other stuff.

That's something like Coinbase. So, so being deliberate in terms of what you're doing on the technical side of things can actually flow through with the shape of your organization. But, but yeah, look, if, if you had. 10 million daily active users like you need, you need an operations team to help with the customer support side of things.

Believe me, I felt that at Coinbase public failures of, of not scaling customer support fast enough. Uh, I think you, you're going to need the ability to handle like a DMCA notice. And it's like, yeah, it's a decentralized network, but your client actually is a company, so you need to comply. Like all that stuff happens, but doesn't happen for a while.

And I think part of where we need to focus is not building like the infrastructure for a thing that might happen. It's actually make that thing happen. And if you're a little bit more experienced and get a little lucky, you can kind of start to sense that that's the trajectory is actually happening and then you can actually start to hire those people in.

And I think I'm fortunate in the sense that I, I had to do that on the fly at Coinbase and there were a lot of failures associated with that. But I think if that, we were fortunate enough to have that happen again. I think I would be able to kind of execute on that a lot better. Cause I had already done it.

[01:09:31] Sina: And it, it's a, it's a, it's in some ways a better problem to have than a bloated team that's ahead of the problems it's trying to solve, to like get hit in the face with like a ton of work and everyone just has to like, do work 24 hours a day until you can catch up.

[01:09:46] dwr: Yeah. I, I think, uh, another interesting thing that I, I associate with these headcount numbers is it's, it's pre covid Silicon Valley culture where people really don't talk about revenue and metrics. Like, it, it's kind of this, this thing where you, that, that, it's weird. People will talk about how much they raise and at the valuation cuz that now became public thing used s not be, and now people don't talk about revenue or profitability or, or metrics while they're a private company.

So the proxy to measure your progress is to tell these startup parties, What your most recent round was. So it's like, Oh, look at these fancy investors telling me it's worth this on paper. And then how many people? How many, Oh yeah, we're a hundred now. Like, Oh, that sounds impressive compared to a two, two or three person startup.

What I think is more interesting is like, no, no, no, we're we're 12. Yeah. We have got 10 million daily active users like that, that you don't hear that. Right? Like you hear plenty of companies that are a hundred people.

So I I think, I think it's just a norm thing and, and I think Silicon Valley, late stage zero interest rate policy money is cheap.

Like that was the culture.

[01:10:53] Sina: yeah, it's interesting that you know, these are people who. the role should have self-selected for people with conviction, with an ability to think, you know, outside of the herd mentality. And these sorts of things still happen, One other question that came up cuz I'm very curious to hear how you think is remote.

Um, why are you building the team remote? and, and how do you think about that?

[01:11:18] dwr: So there's a practical reason and then a, the philosophical reason. So the practical reason is I'm in la my, we moved down here to LA from SF after 10 years there, during Covid, my wife got a job at SpaceX,

liked it. So we're, we're gonna stay. And my co-founder Veron moved down. So the two of us are actually here.

We work in the office every day, sitting next to each other. It's like a garage. And we could do the model where we actually say, Hey, you need to move and co-locate. But I think the reality is la despite being a large city, and there are definitely talented engineers here, the pool of of senior engineers in LA is gonna be way smaller than somewhere like the Bay Area.

So we could either move back to the Bay Area or we can say, Hey, let's, let's build remote. And I think that the, the thing that works for remote in my view is if you hire senior people, they by definition need less management. Like they, they're more experienced. Like if you say, Hey, we're trying to accomplish X week over week, they, they're actually good enough to to, to push back up front.

Like, Hey, that's too much work for this week. And you actually can trust that judgment. And then assuming you kind of scope something correctly, you can give 'em that work. And they need minimal input. Yeah. There's instances where you're brainstorming, but they get the work done week over week. I think a more junior person and, and junior people are my favorite people to work with in the sense that they just tend to have a lot more energy.

They're earlier in their career, they're hungry to learn. And that's not to say more experienced people don't have it, but on, on, on the average. Those people actually I think need to be in office. And, and my, my view is, I think it, it, you learn a lot by kind of osmosis being around people and, and like I went through that experience for the first 10 years of my career where, you know, in an office, like being around other people.

And so I think that if we do start hiring more junior people, we're gonna have a very kind of focused in office culture here in LA because I think that that will be the best way for those people to get maximum amount of learning. And also their effort, which they tend to be very high energy, high effort, can kind of be channeled in the right way to get the most efficiency out of.

Um, and so that's, that's our general approach is we're gonna hire remote for senior and then for, for CO and, and engineers and, and designers. Like, I don't think operations people generally should be remote. Like I think they should be in person like it. And I grew up on that side of the business. So. and then I think that the last point is, and we may change this over time, But, uh, we're pretty strongly in favor of keeping the time zones to North America.

We're here on the West coast. Our view is like plus three minus three, which there's really not much minus three Hawaii think would qualify. But that's it. That that's all we we do. Because even if people tell you that they can work those hours, it's really inconvenient when it's three o'clock in, you know, LA and you ping your colleague in New York and they say they're working those hours, but the reality is, you know, their kids are home from school.

Like now they wanna be having dinner and, and, and the two o'clock or three o'clock people on in the West Coast, they're like, Hey, like, we wanna actually get this resolved now. And so I think getting farther outside that just gets, that much more challenging. And so I think we're willing to compromise, uh, with like east coast to West Coast, but not much else.

[01:14:39] Sina: Yeah. Awesome man. well, I know we're coming to the end, so I thought we would close with more of a fun question, which I, I think will still be instructive for me, which is, uh, this, this actually came to mind reading some of Maroon's casts where, uh, he recently did this AMA and someone. Veron is the dark horse comedian of forecaster, and he said that you basically force him to do five Joe casts a day as an , as an engagements metric.

And the, the question there is, maybe related to that, by prompted by that, which is if you're talking with someone who's thoughtful has things they're thinking about in their head, but is not effective at using forecaster or Twitter or like social media in general, what are the things you tell that person, whether it's Varoon or me or anyone else?

[01:15:37] dwr: Yeah, so the, to clarify that is a joke from Veron. I do not require him to make jokes.

Um, he's actually a pretty funny guy. I think


[01:15:48] Sina: a good like kpi,

[01:15:49] dwr: Yeah, yeah. Look, I, I think naturally and, you know, it's, people have asked me this, oh, you should just hire a community manager and have them do the community stuff, and that way you can free your time up.

And I view that as like antithetical. I'm, I'm trying to build a social network, like if I'm not the most active user or close to one of the most active, Clearly I'm, we're doing something wrong.

[01:16:09] Sina: So you

[01:16:10] dwr: is ironic because


[01:16:11] Sina: actually teach this to the people who work at

[01:16:13] dwr: Yeah, yeah. And, and

so this is,

[01:16:15] Sina: they need to do.

[01:16:16] dwr: Well, what I would say is, it's interesting cuz I'm deeply introverted.

Um, I like to say like, I'm an introvert or you know, like I'm an extrovert. I only, but I only play it on TV or I own played on forecaster. It's, it's, uh, you can ask my wife, I'm like very much an introvert. I need my own time to kind of recharge. But, um, I think that that is actually a natural tendency and a lot of the people we've onboarded to forecaster, they fall into two camps cuz they kind of tend to be more technical and a two Ethereum builder types.

There's the extroverted type, which they get it right away. And then the introverted type is like, ah, I don't feel comfortable, or it's not perfect, or, and I think one people should just, you have to just get over that of like, just kind of put, put it out there and don't take yourself too seriously. And if you're in a, in an app or environment where, You can actually say something and then people kind of give you the benefit of the doubt or nuance actually think is more reinforcing to wanna do that.

I think what's challenging is that if you're trained to Twitter, you kind of realize if you have a bad take, like you just could be descended on with like quote tweet and then like all these other people who wanna hate on it. So I think that that's one thing that people just, when they start to try to use forecast, they're like, whoa.

Like I got like actually quality engagement and not engagement and likes. But in terms of like people actually responded to the question. So actually that's one thing I I say to everyone is one thing that Forecaster does really well right now, while it's still small and hopefully we can continue to try to scale this, is if you ask a question, you're gonna get I think generally pretty good responses.

And I think part of that is like the, the size of the community, but it's also just generally people are craving this ability to actually have like a conversation versus I think sometimes with something like Twitter, given how mature and at scale it is, people. People view that very much as kind of this broadcast platform that's where you do your shilling and all that kind of stuff.

I think that the second thing, um, is, and I, and I try to do this myself, is like, I think memes are like a pretty good icebreaker in the sense that you can either take it from somewhere else, which is the Elon strategy, right? He doesn't make his own memes. He, he gets 'em from his medium dealer. But, uh, I, I actually try to make memes on a regular basis.

I think it's, it's like people talk about Sudoku being good for your, your brain or, you know, word puzzles, but like, I actually think it's like communicating with an image template that the same thing that you would wanna say in text is, is, is a little bit more amusing. Right? To see SpongeBob doing it with squidward is, is more interesting than me writing the

[01:18:40] Sina: so so you have, you have something you wanna say and instead of trying to like craft it into a short form written thing, you're like, I have this to say, how do I, how do I put it into a meat?

[01:18:50] dwr: Yeah. And, and actually there's even more compression there, right? Because what you're importing is the point of the meme, which requires a little bit more intellectual engagement from people. But, but they actually probably find it fun, especially if it's kind of funny. And then you can actually put the, the actual message and, and you, you know, you're writing some text still.

But I think that, that like is another interesting way because I think there are certain people who are so good at Twitter, like they know exactly how to like, put everything in and it's like, you're like, I could never be this funny. But I actually always say to people is, is using a meme template, like you're basically getting someone else's creative output and then you're just putting a thin layer on top of it and you actually look, there are some accounts on Twitter that are very good at Twitter, like Turner Novak vc, not really into crypto.

He's extremely good at seeing kind of like what's happening in the Mimi sphere and then like bringing it in for the, the, the like tech Twitter take of the. And like, I think that's brilliant. And it's like, I actually think more people can do that because it's like they have an observation they wanna make, but by kind of doing it with a meme, it also kind of de uh, deescalates a bit.

Whereas like, you, you have this profound statement as like a tweet or a cast, it, it feels like you're, you're doing this versus if you kind of like my feeling when like, I see this in Ethereum and then you have like SpongeBob it, it can still make the point, but it, but it kind of by definition is like a little bit more cheeky.

So that, that would be my other piece of advice.

[01:20:10] Sina: Yeah, they're actually a deeply textured format. Like there's, there's a lot of depth to them cuz conveying something with humor and irony and like, not saying what you're exactly saying, but saying it between the lines, like there's it, it's a pretty, you know, deep. And the fact that it's like so funny and doesn't take itself so seriously only makes it more


[01:20:35] dwr: So you might call that straussian or, or something where like you, you can. Actually communicate multiple messages to multiple groups with plausible deniability, which I think with internet, that's actually important, right? Be like, Whoa, whoa. I put a picture of SpongeBob up. Like, don't, don't take that as a treatise, right?


[01:20:52] Sina: Yeah. Awesome man. Well, I think let's close here. This was a lot of fun.

[01:20:57] dwr: thanks for having me.

Hey, I'm going to make a small ask here. If you've been listening to these conversations and want to support what we're doing here, . I would really, really appreciate if you could leave a rating and review for the podcast, wherever you're listening to it. This might seem like a small thing, but it will really help other people also discover the show. Thank you. I'm grateful to be able to do this and look forward to being here together again soon.