Founder talk with Dan Romero, Co-founder & CEO of Farcaster

Transcript

[Music] thank you if I had to pick a pick out three themes from your talk that really stuck out to me I'd say versus getting getting feedback continuously from users iterating on it adapting on it second which you didn't State outright but I think was a theme through everything was not prematurely optimizing the docs the product the protocol even the way you engage with users everything always desperately trying to avoid premature optimization wasting resources and third cultivating a community very carefully right and thinking really really thoughtfully about what is

actually the set of people So to that end how do you let's just start with feedback you specifically talked about stated versus revealed preferences can you think of a time or give an example of how to get a little bit past the stated preference like sometime that someone give you feedback that you heard was interesting there's clearly something there except you need to do a little bit more work to figure out what they really meant you know beyond what they were explicitly saying yeah I think um it it's a bit of like you have to do it a lot like in terms of you have to be constantly talking to people and actually getting a bunch of these canned responses or these

polite responses to start to detect like okay this is where that like everyone has this piece of feedback and it's not that particularly useful yeah um so a you can you can like on a tactical level just vary the way you're asking those questions and so the example I gave before is someone's like oh I'm too busy you know sorry I'm too busy to use forecasters like well you're not too busy to use Twitter and so you can have that in a combative way which then you're probably not going to get an answer from the person because they'll be like well you know Screw you like why why like I have to tell you anything versus even just framing the question a little bit more leading at the front it was like Hey notice you're active on forecaster for a little bit and you haven't been as active recently

but I see that you're still active on Twitter is there anything specifically we could be doing better right so it's almost like you acknowledge that like you're there's a humility and I actually find that people then they kind of realize like you're you're in a mood to to hear actually the the bird yeah it sounds like you're setting the stage emotionally right because if you just kind of ask and like an almost like a blunt way someone's like oh well if I tell the person what I actually think then they're going to argue with me it's like why why am I going to I don't want to argue with you like have them be like oh this is better like and I think um I definitely have days where I just find it really frustrating like I

I have a my you know normal account and then I just have like an ALT account that I followed like 200 people on crypto Twitter right these are like people that are like okay I want to have these people on forecaster every day like that that's success people are like talking about all the stuff in ethereum and decentralization and all this and I'm just every time if I kind of catch myself at a low point I'm just like you're so full of you're like just doing this on Twitter and like you actually aren't like doing it on the you know decentralized whether it's forecaster or lens or you know pick your favorite thing and then you get a little less emotionally driven on that and then you kind of step back and say well it's like well you know who gives a like you have to figure out like to

build something compelling enough that they actually want to go switch so you can either get mad at the people for for not doing like what they should be doing or you can just deal with the reveal preference of like your isn't good enough yet so like you gotta you got to make that better and and so that's actually tends to be more productive to get back to like work yeah so those 200 people how do you select those people and also more broadly I think everybody's used forecaster agrees it has a specific style specific Community style specific Community ethos I think is really cool you immediately get the vibe right yeah how do you curate that set how to identify that set who are the best users to Target first like how how's your thinking evolved in that Community cultivation so it's a little bit of a convoluted

process in that I usually start with my main account because I'm doing the DMS anyways and anytime I see someone post something that I genuinely think is like okay they put a lot of thought into that it's not just like some you know post that's trying to get engagement or like whatever where it's you know you the person who's doing the kind of like etherscan slew thing or you know they're they're trying to like kind of construct some attempt at like maybe here's a better way to understand this part of the world um I like go to the other browser Tab and then I add that person so it's not like a I went out and tried to find 200 way it's slowly over time just because I was already using it sure but what's nice is

that account now the algo it's really good in terms of like if you if you actually are careful about who you add in there right then the algo is like actually like okay I don't have to add any sports there's there's no no distractions in that account like I have no one talking about Ukraine in that account it's like purely like people talking about whatever the the current thing in crypto is totally sounds like a algorithmic vampire attack yeah sort of and and so but but that's actually my Baseline is I checked that feed every day and then I compare it to forecaster and I say like where is the diff here and it's it's frustrating because it's like we have a lot of work to do because there's a lot of good stuff that people are putting on Twitter

right and these are people I mean like they're interested in ethereum so like we we should be able to offer something you know for them and we just haven't hit that level yet and how much of that is the dev targeting strategy because I think something I personally really like about the forecaster Community is how Builder Dev Focus they are does it come purely from that or are there other specific things yeah I think so that wasn't the original thing and actually going back to one of the questions about like the nfts the other thing that the people who stuck around tended to be people building companies and so what I founded especially in crypto and it almost became like the people who are willing to use a network when you're talking about like tens of

Dao uh it was almost like group therapy like in the sense that like you get to talk about like being a founder and working through this and actually if you just look at the people who must follow them at work like I want to say 80 of them are you know founder of a company most of them in crypto um and so that that I think contributed to the early Vibe and then I started to actually lean into explicitly and like tweet be like if you send me your GitHub as like the way to like kind of evaluate someone um but I think that the other just like tactical aspect of focusing on devs as we as we shifted towards that as a kind of a discovery is it set the network up to actually have a bunch of people who as soon as it got to be big like they

started tinkering very serendipitous the right crowd is already yeah so if like you're trying to build a protocol best group of people to have interested in it are people who can code no question yeah GPT might solve that for everybody but like I think for the time being like uh you know a public GitHub profile that has some activities usually probably a pretty good proxy or related yeah relatedly how do you think about crowding out the risk of crowding out developers right the the hope is that there will be many clients many applications many interconnections between different types of services that read and write from forecaster of course you guys started off with the warpcast client it's still the leading client how do you establish the type of

credibility to tell to let real serious Builders know like this is neutral territory you know this is fertile ground for you to build on like how do you think about that yeah so one we're figuring this out like don't have any experience doing it outside of just intuition and and kind of the experience over the last couple of years um I've been a big fan of the Apple ecosystem for a long time so there's a term in the Apple ecosystem called sherlocking This is back when finder didn't exist on Mac OS and there was an app called Sherlock it's like way to find your files and every year at WWDC for the last 10 years like some developer because Apple added some new feature uh to the OS at the OS level like got sherlocked in the sense that

like they used to have an independent business now this is being given away for free uh simple example of this is there was a flashlight app business that existed for the first few years eventually it just got added to the phone right so like overnight there are no maybe you have super flashlight I don't even know if it's not anymore but but I think um it's a balance and and the way we think about it is like so first we can do a lot of like talking I can do this I can go on podcasts we can put developer docs or whatever but it's like show don't tell like so in terms of like it's actually the actions that you take are going to build the the trust with Developers and then I think we realized especially kind of second half of last

year as people started to build and like you had the discovers of the world start to show up that okay actually now is the time to going back to the stud don't don't optimize too early right we are actually behind now like very quickly it was like oh we need to actually get the network to Credible neutrality faster than maybe we were originally planning so it's like a good problem to have but like there's a lot of work between now and then right it's like hubs need to ship we need to be on ethereum mainnet uh and then permissionless signups and this was a lot of work between now and then to do that but I think that's where um one it's helpful if your product is literally a public social network so you can communicate with people but going back to that tactic I mentioned about

telegram um so we had a an app telegram to start and what we actually did early on was we created an invite-only Dev chat and I would post on farcast or even Twitter sometimes and I'd say if you're a developer on farcast or DM me so it's like a pretty Niche audience but that channel now has I don't know three or four hundred people but it was pretty small for a long time and the other thing no offense to VCS I I didn't uh allow any VCS to come in because I didn't want the dev chat to be like someone's Alpha of like oh let me ping this person and like see if they're raising totally and and so actually that sub community so we just had two

telegram channels we had the like bug report like whole whole telegram Channel and then the dev chat and that's actually where we started to to build the trust and make mistakes and have people feel like it was actually a smaller venue even though they could have been doing this on forecaster where they can tell us like hey I think that that was like not a good decision or like you know that didn't work and then once that chat kind of started uh and I give a rune credit he really nailed the sequencing on this once it got to about 100 people he started doing a bi-weekly Dev call and his whole deal was uh I'm gonna put out the agenda the night before and uh hey I'm Gonna Cancel the meeting if there's no nothing on the agenda and no one no one has added

anything to the agenda and then keeping it really really short and efficient and I think that that also started to build a group of people and now I think about 30 people and people have like full-time jobs they're doing other things but they they attend the death call because now they feel bought into this inner sanctum of the community so it was like a community within a community that you're actually building that's super interesting that's super interesting have you thought much at all about how these developers and later clients will monetize you have a theory there and it's obviously there's some important prerequisites so so we um are planning to do subscriptions we're gonna try that because I think if you think about forecaster that

on-chain element in some ways is like a domain like you need almost like a registrar who can like manage that for you and I have this belief that like uh if you can just get Apple pay in front of people having been at coinbase and know like what the drop off on like Fiat rails and all that kind of stuff is like basically if the sign up experience for forecaster is you open up the app you sign up for a subscription we do a meta transaction behind the scenes on chain to register you and you're good to go yeah um you get two things out of that one uh I think just even having subscription like the low effort trolls just never show up Bots don't show up uh at some amount of scale like I actually I'm a big believer in a lot of what Elon is

doing with Twitter like like if you ask me it's like what is the biggest pain in the ass thing to like growing forecaster it's like elon's running Twitter and I know a lot of people think he's like running it into the ground but like the guy can land two rockets at the same time on the ground I think he's gonna be able to figure out like having run a version of Twitter albeit much smaller like not rocket science um so I think that the idea of subscriptions as a a increasing Baseline quality is is something that I'm excited about that said I think the revealed preference at scale on the Internet is people like free yep um and so I think that maybe at some point you know Spotify actually is is a good example and Netflix is basically getting there

too where you have the ad supported tier and the Primitive tier and I think that that like the way I want to do it is build the premium tier first and then add a free tier at some amount of scale later but I think um I don't know I think I think it worked pretty well for like you know Netflix and Spotify is like and I think if Twitter could go back to the beginning assuming you had like stripe and the internet infrastructure that you have now Apple pay I think Twitter as a subscription product from the beginning what have actually been an amazing business assuming you don't hire 8 000 people like that's right and so so I'm actually relatively bullish on on what Elon is doing there and I think we can actually apply that to decentralized social at least in our client and I think one

thing people forget is like Elon has already built a subscription business that's like a 1.5 billion dollar run right in starlink that that will probably be a 10 billion business in the next 18 months and so it's like he he actually understands subscriptions and so I I think consumers like subscriptions assuming they they get like a better user experience yeah there's just so little there's so little game film on subscription social right but yeah there's still a lot of interesting counter cutting evidence that suggests certain things are broken there should be certain experiments proof of personhood is definitely one of the most interesting angles for it I think maybe in the early days of social networks and they were still growing I mean needed the initial user base the initial

value proposition a little harder but now we've reached a special critical mass where we can try to experiment with different ways to isolate to interesting groups yeah well one thing I would say is like you're you have an American college degree call it about 200 000 worth of Education expense yeah that's its own social network and if you just amortize that over the entire course of your adult life probably something like eight bucks a month so like I do think people like have these things it just slightly priced differently yeah and so I think it's it's it all comes out of the user experience and it's like is this is this worth it and I think people showed that like as soon as Netflix showed up it's like why would I ever watch television where I'm watching

a you know 30 minute show 20 23 minutes and then I have to watch seven minutes of like you know pill ads and car ads like like people just don't want that at a certain level and so so I think that consumers are primed for this and it's just like it all comes down to execution that's super interesting changing gears a little bit since you didn't come from a social network background how did sufficiently decentralized develop like what was the thought process to where you isolated you know platform risk and the minimal components of protocols that need to be put into the neutral layer to make them credible credible a credible building Foundation yeah so I I um I I basically really enjoyed Ben

Thompson's aggregation Theory during the 2010s and like reading strategory as a Blog it's like I didn't go to business school like I worked at companies and then kind of read stuff on the internet Peter Thiel Chris Dixon um Paul Graham and so you kind of like take all that stuff and just think about like how did people build successful products and kind of all those Frameworks together and then just for me I'm like an infowar so I've been using Twitter as like a power user for 15 plus years RSS is still used um and so the original idea for forecaster was kind of like could you make RSS competitive with Twitter right and the kind of I I think approached from a social standpoint was you have to get to kind of some shelling

point of like an identity space that that was one of the core things that I think Twitter beat RSS like the reality is domains are not user friendly like we the email today yeah DNS domains like we use email despite the fact like that yeah they should have never been exposed to normal people and the you know counter to like yeah we use email I actually think the only reason we still use email today is because it's it's a function of work whereas I actually think like if companies didn't use email consumers definitely wouldn't use email it would all be like WhatsApp phone number Instagram like username and so I think if you're thinking about social like the world has moved to this like you're represented by a name right

um so the theory there is that the most valuable Network effects are accumulating to different parts of the system yes and if the network affects maybe one of the most valuable ones maybe the most valuable one I think we're going to learn that is to the names yes and so if you dislocate that and put that into the neutral layer then that's enough that the system will remain decentralized yeah and I think that we [Music] um when would that hypothesis in V1 and then what we realized is so first of all people like upgrading their usernames so on Twitter you can upgrade your username if you tie your entire identity to a single username and a user control Network you're not going to be able to update the follows if you change that so

right V2 we have a dual part identity system so we have an FID which is an integer so I am number three uh Bruno's number two uh and and so just increments each time someone signs up and then the F name the username that's the human readable layer that sits on top that's optional and in theory like because I actually very much believe in like you should be driving Things based on their value not rent extraction in in protocol World an fname could be interchangeable with ens like you you could use the NS or build a client that just respects CNS and the beauty is everything that lives at the Hub level doesn't need a name right it's not an FID so we've felt like that was more Intel actually honest that was a mistake we had made in the

original I think that's right I totally agree and and so we're bullish unlike forecast or F names just because people are gonna who spend a lot of time in the ecosystem or have a better name there and so they're going to describe more value to that um that said I think what we realized is like and and this crystallized it's always controversial because triggering for people is when Trump got kicked off Twitter it like was so clear to me what the sufficiently decentralized social network should be right it doesn't care about his old tweets doesn't care about his likes doesn't care about his DMS the only thing he cared about about being kicked off Twitter and this is anyone with a large audience getting kicked off that's why probably Elon bought Twitter he didn't want to lose his big audience

is you have a a public pointer on Twitter your username that 80 hundred million people have opted into getting a totally an update from and then you from your phone can just boom hit 100 million people that's an insanely powerful like if you were to just go back to show like kings and Emperors throughout history it's like yeah I have this little device that I can just like beam a thought to 100 million people instantly like that that's a lot of power and so our point of view is if you look at email as a contrast you don't hit that scale with email in terms of I don't know what the biggest email newsletter is but it's it's probably not it's not billions yeah yeah but with email if I'm using sub stack as my

newsletter right and I have a custom domain on it and subsack decides they don't want to work with me or I don't want to work with substack I can move to MailChimp or on my own server and I can continue to email all those people and like that for us was like okay how do you actually take Twitter and move it to that level of decentralization when the reality is no one wants to run their own server and so like you view running your own servers even more decentralized like that that was the moment we like kind of like were like okay this is the level of the centralization we really need to focus on and this is the layer that we actually need to get out of the hands of any individual company or country so you let people own the thing that the

network effect accumulates to and then these other components which are incredibly important obviously the server where's the data store the schema the model those things are really important need to be built for the application to work but as long as those things all ultimately serve the controller of the username then uh maybe that's decentralized enough yeah and I think that like the other thing you have to just from a first principle standpoint is so forecaster hubs there's a consensus algorithm right so it's starting to sound like a blockchain it's like you've got a dedicated server you have a consensus algorithm but first principles there is no uh double spend problem with social there's a double spend problem with identity that's the thing that lives on a blockchain but if

you think about social if I type hello world twice three different messages uh that doesn't cause a catastrophic loss of faith in the value of the underlying currency whereas if you obviously spend an ethereum transaction or Bitcoin transaction twice everyone is impacted by that and so like not needing that level of you know Byzantine fault tolerance or like you know allowed us to use crdts and and build a consensus algorithm so hubs are very fast they don't have any cost effectively relative to like an on-chain cost but we realize that like identity collisions double spends are very like expensive for the network because they do lose faith in the network and therefore they have effect so that's why we opted to

put it on chain even though there's cost makes sense how do you think about authenticity and projecting authenticity is something you mentioned in your talk you know trying to going from you know fake BS chill that's an uncomfortable behavior for a lot of people yeah certainly is for me but on the other hand very true that people like updates people like to be in the loop people like to take a little bit of ownership of the thing know about it be a part of it that's a very authentic Behavior what does that mean specifically like and even for the founders in this room right how do they how can they be both pushing the project inviting people to participate with them encouraging their uh you know people they work with to take ownership but in

an authentic way yeah so it goes back to the dog fooding to the brush test Point um I I don't even know what it is now but I think it's like I'm double digit percentages of like all activity ever on the forecaster Network and it's not me like writing scripts to do that like sure when I'm doing GM and welcome that's literally me typing like GM and welcome to like the people who are joining the network or like responding on the Twitter and so I think there's a bit of proof of work component to that of if people feel like you're trying to like you know scale your time and get leverage like people are really sophisticated yeah and they're gonna they're gonna quickly see that um whereas if you're willing to kind of

like put the work in and and you know pay attention to feedback respond to people the stuff that they don't see publicly helps build that brand and then I think publicly the the other thing is like you you all know when like when someone has an update that they're not overly marketing or they're not adding a bunch of like you know syntactic sugar around that that like talking talking about the protocol it's like trying to keep things narrow to a milestone or a screenshot yeah like a like a LinkedIn post Vibe versus a right and and so so I think the simplest thing is just like keep the post short and just like you know and show not telling in the sense that like anytime you have a screenshot anytime

you have a link like always more interesting for people like people aren't on Twitter to like basically have a LinkedIn post but they're on Twitter if they want to discover something interesting so when something is actually interesting happening in your protocol try to just let the interesting thing do the work rather than you know putting a bunch of additional Shilling on top of it that's great for so some of the founders here are less Technical and I'm personally very generous with how I think about technical right like just because someone's not a software engineer day to day it doesn't mean they aren't a technical mind I know your co-founder of Varun is deep in the weeds how do you advise uh to Founders who are a little bit less

technical or who are focusing on the non-technical components to get familiar and eloquent enough and and you know deep enough into things so they can be you know competent in a space like crypto that's incredibly technical not just in a strategic way but even like you were talking about you know ecdsa curves earlier today like what's what's the right framework for the founder in that position yeah so I feel pretty passionate about this I was an English major so like if I'm if I if I can talk about ecdsa curves I can promise you you can get up the curve on that um but I think the it goes back to authenticity it's like if you're interested in this stuff like it's important not to get sucked in yeah

and it won't feel like work um if you're not interested in it then that's a little harder because then it that it'll like basically feel like work or homework and you kind of don't want to do it um that said like you know can I explain to you like all of the nuances of the evm or no and like even when I was at coinbase like Bitcoin utxos still hurt my head like trying to think about that yeah that said I think just trying to surround yourself with other people who are really passionate about this stuff and then just kind of let them talk or ask them questions like the the beauty of crypto people like if if you kind of can surround yourself a little bit with them

and they can also be like over telegram or whatever is they love to just like talk about crypto like forever and and it I think that they get a bad rap like generally as a group because it's just like the people who want to talk about like price and whatever token versus the people who literally we could just talk about the technology like forever totally and just just listen to those people I'm I remember just at the coinbase lunch table when I first joined in 2014 um I don't know if you guys know who Olaf is he's a he's a bit of a character uh but he was one of the first employees at coinbase and he he was talking about ethereum 2015 at coinbase and like just if you just pay attention to like what other smart people are doing like your

by osmosis you're going to be picking that stuff up um I mean that like ethereum was not consensus in in 2015. so I think uh true I think you just have to have an open mind and an ability to listen and and sometimes the interface of the person that you're talking to might not be like the most polished but the content might be great and so that would be my my recommendation the other thing is like there's an infinite amount of content on the internet and maybe use I don't know chat GPT now to summarize it for you but I don't know it's easier than ever to go up the climb up the hill really yeah I I generally find YouTube is is really good if you can kind of find the right things I will plug the uh Canon reading list from a16z it's very

good I have a competing reading list which I think is also really good so if you type down remember reading no but yeah that's great any parting words for the founders in here I know everybody's been really appreciative of your time so thank you Dan yeah look I feel a little weird because it's like forecasts are as a really long way to go but I I did have a good run at coinbase and um the thing I'd say is like crypto is pretty brutal in terms of the ups and downs uh and it's interesting that this Market is actually timed with like the first down market for Tech generally like it was always really weird at

coinbase especially in the first crypto winner where basically you know mid 2010s in San Francisco it's like everything's up and to the right totally and like crypto is just sideways and it was like two years of basically being like Oh you're still working a Bitcoin company if a Bitcoin diet or like whatever yeah and then it's a very weird thing to have like stuck that out and there are moments I I almost left coinbase in you know 2016 but like 2017 it was like a fever dream because all of a sudden like we didn't do anything differently to just Market shifted and then crypto it was just like every single person all they wanted to do was talk about crypto and it's like

you work at coinbase like whereas a year before it would have been like what's that versus and so I think it's not to say that like oh just wait for the market to correct but but what I would say is with crypto especially like it just by virtue of the financialization of the category like just you know Highs are higher and the lows are lower and so Equanimity and and a little bit of stoicism and I think this is what Brian Armstrong's superpower is and why coindesk has been successful as it is is like you got into it for a reason you you have a vision for what you actually want to build there are so many different Temptations on like getting distracted and not actually going ahead and just keep building on that and so and I think we

were as mentioning to you before is like time preference is a superpower in crypto in that that if you are really excited about what you're building and you can actually you know keep burn low don't get distracted just keep trying to make week over week progress get the feedback from The Limited number of users you you have like that that cockroach mentality that is I think yeah um it's way harder to do than you know for me to say but like having seen it at coinbase like that that's like The Guiding Light for for us when we're working on podcaster like we've raised much money we're a 10 person team we're holding at 10 like it's like the goal is like keep making progress every week and

at some point we might find product Market fit and it might have nothing to do with actually the market changing and at that point then I think you can accelerate it crypto has a really intense Financial cycle right all markets have all products and technology has product development Cycles Financial Cycles that's true universities but in crypto the financial cycle is over the top because these are assets that can be freely traded yeah and that's super weird yeah it was like so I went through the first cycle and then 2017 you know it was just insane year it's like 100x growth and like every metric going from 10 million in Revenue to a billion dollars in revenue and it was just like so we go and hire all these Executives in 2018. then the market corrects and

all these people who are supposed to be like the people I'm supposed to be learning from from like they're just like freaking out they're like what like I've never been at a company that's like just had a 90 correction in Revenue in a year and all of the the board members slash people had been there just like oh well it's not as bad as the last correction yeah yeah and and so I think it's like so that that was the point where I realized like oh the superpower in crypto is you have to be able to to tune out the noise yeah it's like the the crypto Twitter haters like you know web 3 is is going great BS like your friends giving you a hard time like that is the those are all the the things that are going to distract you away from what

you're actually focusing on yep so can you describe for us how you think about the competitive landscape for forecaster and then what if any advantages you think you might have to creating a sustainable social network yeah so the competition's Twitter like I don't think about any decentralized social networks this competition like all of the the highest status people the people that everyone wants to pay attention to the current thing the like the the thing like when there's a big news event you're on Twitter right like FTX blows up you're on Twitter like svb blows up you're on Twitter it's too much stuff blowing up that are financial institutions but like like just like anything that's happening you're on

Twitter you're not on those other networks and so that is the that's the hard thing is like how do you chip away at that and so our approach is actually to just go narrow and we're we're specifically focused on ethereum because I think there's a lot of interesting stuff in crypto but having been here for 10 years I think the the highest concentration of people building interesting stuff tends to be in ethereum or like evm put it that way um and that's not a judgment on other things it's just like like it's an opinionated point of view um the the so so with that focus on ethereum I think just have a lot of work just like how do you get like naturally people who are interested in in the evm to actually

be wanting to use something like forecaster and that's that's the exploration and effort we're doing right now hey can you talk a bit about your resource allocation strategy between building a protocol and a client in parallel yeah and how you've been thinking about that how that might have evolved from the early days to now and the future 80 90 on the initial desktop client that was a mistake we should have built mobile everyone if you're building a consumer app you should be a building mobile I don't care if the App Store is going to be like a hard thing to get through it's like that's the hard thing like mobile user experience like it it doubled down overnight like 80 of usage of forecaster is on mobile

um so we started with the desktop app because it was easier I wish I had just like gone back in time and knocked myself in the head to like go go build mobile um we really focused on the the app for year and a half and that was that got us to the point where I think in orders of magnitude you're the first user figure out some friends get to 10. and once we got into hundreds of Dao not not any fake metrics then I think we started to actually have enough usage that we had two things one we started to have developers be interested and two we knew all the problems with the uh the V1 of the system and so V2 was informed by

early developer usage as well as like just our own experience building our own client on top of the protocol so that shifted 50 50 to the protocol versus client and then more recently it's probably been like 60 or 70 percent protocol we were really really pushing a lot of resources in the other interesting thing is we didn't have any open source contributions for the first two years and now we're up to like 40 individuals total and so just like now all of a sudden you just have random people who like find the technical problems really interesting Varun has done an excellent job of finding issues that are just like small things that the team could go do but he

takes the time to spec it out a little bit and then specifically tags it on GitHub issues of like need help and their digestible pieces of work so I think sometimes with open source projects uh you know like everything just is like too big of a chunk of work with person who doesn't necessarily have enough context um but by by taking a little bit of time to like maintain the garden of potential issues you find people and we've actually now uh someone full-time came in just as a result of uh they ran across the repo they liked how well documented everything was remember zero documentation on the website but the GitHub was like really really well organized and their full-time higher now uh which which was really awesome

um hiring is really hard and like so to have people just like randomly show up off the internet with a PR and then you like start working with them and then they become a full-time employee this is like pretty cool it's cool thank you and then Marat again from uh primav can you talk a little bit about the white paper versus build process like for our case you know there's a white paper for uh um a car that we're kind of working on but we're also building a skateboard in the meantime and trying to get a sense of how you allocated uh resources maybe to those things in the early days or maybe things things that you would do differently knowing your knowledge today I I think that the exercise of this is like to just even broaden it like the

classic should you write a business plan for your business or like build a product I think there's a useful intellectual exercise of like can you think through the system like beforehand of like what you're trying to go build what that car looks like and then extract the essence of that to figure out like what what is your skateboard from here and there are always going to be exceptions they're going to be certain things where it's just like sorry like no like we're building a car like we can't have a skateboard first right like obviously Tesla itself like they built a car they didn't build a skateboard but I think in terms of like what is the the core of whatever you're building and like really really focus on that even trying to get that in the hands of users developers your own product to start and then and then kind

of work your way there I think if you think of something like unit swap where it's like and I'm not an expert but like if you just look at the difference between V1 V2 and V3 like obviously how much more sophisticated V3 is um but like V1 was vitalik wrote the white paper you wrote it in a Reddit post and then Hayden just like implemented it so like amazingly successful protocol but like what the implementation plus like that really just simple website like that that is enough to kind of like start getting those Loops kicked off thanks that helps yep cool all right well thanks again Dan thank you

[Music]