The Delphi Podcast
The Delphi Podcast
Dev.Fun: Leading AI Vibe Coding Platform on Solana with 20,000+ Applications Live
Use Left/Right to seek, Home/End to jump to start or end. Hold shift to jump forward or backward.
Join Tommy Shaughnessy as he dives into the world of Dev.Fun with founders Devlord and Robi. Dev.Fun is revolutionizing app creation by enabling anyone — regardless of technical background — to build and launch applications using AI-powered vibe coding. Learn how the team is merging Web3, tokens, and consumer apps to create a decentralized ecosystem of user-generated tools, games, and communities. They cover technical architecture, distribution strategies, monetization paths, and why Solana is their chain of choice.
Explore how Dev.Fun is pushing the boundaries of app creation, developer incentives, and decentralized attention markets — and why streaming, tokens, and memes might be the next big unlock.
Dev.Fun: https://dev.fun/
🎯 Key Highlights
▸ Devlord and Robi's AI/crypto backgrounds and meeting via Pump.fun
▸ Birth of Dev.Fun and the “vibe coding” philosophy
▸ Empowering non-tech users to build token-based apps
▸ Shift from memes to tools, productivity, and full-stack apps
▸ Views on AI autonomy vs. human-led design
▸ Monetization via in-app payments, tokens, and community ownership
▸ Vision: from solo devs to tokenized micro-startups
▸ Modular backend, AI-first SDKs, and payment stack
▸ Stance on moderation, decentralization, and freedom
▸ How Dev.Fun differs from Web2 builders like Locofy
▸ Long-term incentives and plans for a Dev.Fun token
▸ Crypto-native future of streaming, distribution, and curation
▸ Why Solana powers their user-generated crypto apps
💡 Subscribe for more crypto & AI insights! 🔔
🧠 Follow the Alpha
▸ Dev.Fun's Twitter: @devfunpump
▸ Devlord's Twitter: @devlordone
▸ Robi's Twitter: @RxbiXyz
🔗 Connect with Delphi
🌐 Portal: https://delphidigital.io/
🐦 Twitter: / delphi_digital
💼 LinkedIn: / delphi-digital
🎧 Listen on
Spotify: https://open.spotify.com/show/62PR1RigLG2YN5Pelq6UY9?si=18ac7ccf36ab4753&nd=1&dlsi=50105fd66e6c4124
Apple Podcasts: https://podcasts.apple.com/us/podcast/the-delphi-podcast/id1438148082
Youtube: https://www.youtube.com/channel/UC9Yy99ZlQIX9-PdG_xHj43Q
Timestamps
00:00 – Intro: Dev.Fun, Devlord & Robi
01:00 – Meeting via Pump.fun, founding story
02:20 – Vibe coding with AI
04:15 – Empowering non-devs to launch apps
06:30 – Token mechanics vs. traditional apps
08:00 – In-app monetization strategies
10:00 – Favorite builds: memes to tools
13:30 – Viral Epstein app & other hits
15:30 – AI autonomy vs. human input
18:00 – Can vibe-coded apps scale to startups?
20:00 – Foundation models shaping UX
22:00 – Dev.Fun vs. Lovable: app store vs. SaaS
24:00 – Crypto-native distribution’s edge
26:00 – Tokens, capital, and incentives
28:00 – What’s on-chain vs. off-chain
30:00 – AI scaffolding under the hood
33:00 – Future of decentralized infra
35:00 – Curation, virality, and app caps
37:00 – Moderation and NSFW policies
40:00 – Streaming, attention, Black Mirror
42:00 – Why Solana powers Dev.Fun
44:30 – Post-launch reflections
46:30 – Building vs. distributing
48:00 – Build team spotlight
49:15 – Token rewards & retention
51:30 – Key metrics: usage & value flow
53:00 – Final thoughts & vision
Disclaimer
This podcast is strictly informational and educational and is not investment advice or a solicitation to buy or sell any tokens or securities or to make any financial decisions. Do not trade or invest in any project, tokens, or securities based upon this podcast episode.
You're now plugged into the Delphi Podcast.
SPEAKER_00Hey everyone, it's Tommy, one of the founding partners at Delphi Ventures and your host. And today I'm thrilled to have on the founders of dev.fun, which is the easiest way to turn any idea you have into a reality through vibe coding. They're the most used vibe coding platform in Web3. They have 22,000 applications launched. People visit the site, they type in their idea, they can launch an app, they can continue to iterate on it. It's really not that expensive. It's super easy to do. And they're also able to link tokens directly to uh these projects that are created. DevLord and Robbie, how's it going?
SPEAKER_01It's going great.
SPEAKER_00Hi man, how's it going?
SPEAKER_01Thanks for having us here.
SPEAKER_00Yeah, I'm super excited. I I found you guys through building a tower defense game on your site that I thought was pretty solid, but I don't think either of you were able to beat it. Is that is that true?
SPEAKER_01Yeah, I don't think anyone's been able to. Yeah, I tried my best.
SPEAKER_00It's uh it's fun. I think I went through like hundreds of iterations and then I was just sold on on the platform. It's it's so easy to use. You guys did a great job.
SPEAKER_01Yeah, I mean, uh, we're we're just thrilled to see like people building all types of games and applications and just having fun. Uh that's our our goal since day one. So I was very thrilled to see you uh do this.
SPEAKER_00That's awesome. Let's start with uh let's start with some founder stories about how Dev Fun came about. Maybe, I don't know, Devler, we could start with you and go to Robbie and mix and match.
SPEAKER_01Yeah. Um, I mean I've been pretty much coding with AI since it's possible to do so. Like GPT had like the first kind of usable versions. Uh just trying, like, if I can recreate my school project was how I started. Uh see how much time I've wasted, you know, going to school basically. Um turns out like it was pretty good. Um at the same time, like we're like all pretty close to the pump fund team with like early investors. Uh Robbie was running their TikTok. Our view and our TCs was that pump launches like don't have to be necessarily limited to meme culture coins, and we can also maybe use it to fund like real projects, software, and consumer apps. Um we we didn't see a lot of kind of utility-based project at the time uh that's pawned there. And I mean, I I love the attention market uh as much as the next person, but uh I thought it would be nice to let people uh also iterate on consumer products and kind of have like something a bit more substantial that's also possible to be done.
SPEAKER_00Robbie, what about you?
SPEAKER_03Yeah, I mean I've been in crypto since around last cycle. Um mainly started off with trading, which wasn't super successful for me, but um got more and more successful this cycle. Um yeah, then started off doing the um social media TikTok for um Pump Top Fun, and that's basically how I met um DevLord and our uh in our other co-founder. Um and yeah, I I was pretty much on board with the idea pretty early on as well. Um I mean I have a non-technical background, um, and I always had these these ideas that I kind of wanted to go and and and get out there and create. Um, but I was always kind of limited by my ability to code, right? Um and yeah, as soon as I heard about this idea that you can basically turn any anything that you have in your mind into a working, functioning app, I was I was pretty hooked, uh pretty excited. And yeah, I was also pretty uh excited to see how how excited other people are about this. And I think that probably ties more into like how how like the product uh um launched and what we were expecting there, but I think that's something that we can probably get into um at the later stages of this podcast.
SPEAKER_00For sure. What when I was diligence to you guys for an angel check, along with my partners, the the thesis kind of came to me that you guys are helping to lower the barrier to create an app to close to zero, right? Like we saw how many DeFi apps were created in DeFi summer, and you had to code. In this world, you don't have to code, right? You just type in what you want, the apps created, and you could do capital formation and get funding and have your community right on the site. I know I'm maybe giving you a lob of a question here, but is that reality actually panning out? Like, do people come to you and say, hey, you know, now I can build in the and in the past that I couldn't? And maybe a follow-on. Is it better than just doing it without a token? What are the benefits there?
SPEAKER_01Yeah. Um we've had we've had a bunch of user interviews where like people are actually um, you know, someone working in finance or someone working like non-creative fields, just telling us like, you know, I have like all these ideas, all these things that I kind of wanted to pull out, uh, didn't really have like an outlet or a way to really do so, nor like the ability. And this really actually helps them get there. Uh they have the ideas, they have somewhat of an understanding of what's going on, but they the execution part is not really what they can do themselves. Uh, and and for these people, AI is just like a 10x catalyst, basically. They can just very quickly do a lot. Um and yeah, some of these people are using dev.fun, and it's it's pretty cool to see.
SPEAKER_03Yeah, and I also, I mean, uh, we talk to our users every day, and we always look at the most popular apps of the day and basically chat to everybody that makes a popular app on DevTon. And sometimes it's even surprising to me that these people have no technical understanding. So I'd be looking at some of these apps, and I'd I'd be like, oh, this is really cool. Like, how how come you got to integrate this or that or this? And then they're like, and then I asked them, like, do you have any sort of technical background? And they're like, no. And it kind of blows my mind sometimes because I see what these people are go ahead and create. Um, but yeah, it's it's even surprising to us sometimes that these people have no technical understanding. But that doesn't mean that we don't have people that do. Um, I think there is there's a couple successful devs that actually use dev.fun to um build a bunch of stuff um related to smart contracts or um games that involve on-chain transactions. But um, yeah, it's it's pretty cool to see that there's like a bunch of people that are actually able to turn their ideas into reality because dev fun exists.
SPEAKER_00That is really cool. And I mean, Pump Fun really comes to mind here, right? I know I'll ask you your takes on it on it later, ahead of their token launch, which we just uh tweeted out. But you know, I'd love to talk through maybe the differences, right? Like if I'm launching a token on pump.fun and I'm launching a token on dev.fun, what are what are the differences?
SPEAKER_01Um I think like fundamentally it's the same thing. Uh I think a coin is a coin. There's no like we're not like another launch pad, we're just like using Pump.fun. So technically you're launching a PumpFun token if you if you want, right? I don't think there's a need to be extremely creative on what a coin is. I think what's interesting to see is maybe the different structures that people are building around it. And what we also love to see is like how can we maybe add the utility at different ways to kind of interact with that token. Um, so it's not so much about the issuance process, it's more so like what can you build around it? Can you make an ecosystem? Can you turn it into utility coins? Uh, can you turn this project into something that actually generates revenue? Uh, maybe this can be used to like buy back the tokens and like do all those things, right? At the end of the day, it's the the devs are like kind of choosing what they want to do.
SPEAKER_00That is pretty helpful. So I guess one question for you would be you know, I create an application, you know, maybe let's take my tower defense game as an example, or or any favorite apps that you both have. Like, how does a developer think through value flows? Is it as simple as, hey, if you want to buy an upgraded skin or gun or or something, you you pay tokens and you guys capture a piece of that? Or or yeah, how does like the in-app autonomy work?
SPEAKER_03I I think that's really up to the developer to decide, right? Um, some some people want to basically monetize off of the token, some um people want to monetize off of in in-app transactions. And we've seen this model across a bunch of different apps. I mean, one app that comes to mind that necess doesn't necessarily monetize for the token, but monetizes for something in-app related is um is this ticket game where you can basically buy jackpot uh tickets for a jackpot that um basically goes ahead and um gets distributed at the end of the day. So I think what our idea here is is that uh we don't necessarily confine the developers or the users into a certain like weight for their for them to monetize, but we really give them all the freedom. And we've we've seen that these developers have come up with a bunch of like crafty, nifty, um cool ideas uh how to monetize. Um so yeah, I think there's not necessarily one specific way that they go ahead about about this uh monetization pathway, but um there's a bunch of different ways that developers can go ahead and do this.
SPEAKER_01Yeah, and I think like what we want to do is just give some freedom, uh, observe what they do. If we see a behavior that's really dominating, where like you know 90% of people are doing this one thing, then maybe we you know had a built-in support for it. Um, but generally you want to see kind of what people do, and we just have to be positioned in a way that whatever is happening, we can just capture a fee from that. Um, and and how we do that is by basically just providing the framework, like the underlying frameworks that the apps uh are built up on, right? And this just is going to capture regardless of the way you use it, basically.
SPEAKER_00Yeah, and it giving them the freedom is pretty cool. It'll be it'll be interesting to see how developers monetize. And if if I have an app, how far is it to explore the you know, I want to monetize in any way? Do I just type in like the value flows, or is that like a coming feature? How do you how do you handle that as a developer?
SPEAKER_01Yeah, we just um recently released kind of like the on-chain and payment framework. Um, it's still like kind of a call it in beta. Um right now, receiving payment is as easy as just saying, you know, I want this feature to be you pay one sole or like 0.1 sole or whatever, um, to access this specific thing. You can just tell the AI that and it's gonna work automatically. Maybe you tell it like, oh, I want this to be like the wallet where this gets paid, like this is my treasury address or whatever. We're gonna offer more like UI-based configurations for those things that you don't even need to interact with the AI. You can just like literally click um on a few configurations. Uh, but yeah, I mean uh everything that you're doing is interacting with just text-based prompting, just like you would do on ChatGPT. That's awesome.
SPEAKER_00I want to start with Robbie and then we'll Devor go to you. I want to talk through your favorite applications to date. Like we spent the last 10 minutes talking through the platform. I want to just take the listener for a journey here on what you guys are seeing as your favorites.
SPEAKER_03Sure. Um, I mean there's a bunch of different ones in mind. I think me and Devor are sharing uh sharing a bunch of different favorites as well. Um but I mean there's there's there's different types of apps on DevOne. There's like these meme type of apps that hop on some some some sort of narrative. But then on the other hand, there's also these like small like little tools, uh, which I really like. Um one of them uh I always like talking about is called DevFuncheck. Um so basically what DevFuncheck does is it just basically gives you the uh latest coins to either receive a boost on Dexcreener or get their banner updated on Dexcreener. And it's a super simple tool. Um, it's um very, very easy. And I think it was built in uh within five to ten prompts. Um but what was funny to me is that I basically showed this to a bunch of like traders that I know, um, and they were like, yo, this is super useful, and I was like up uh X amount of sole in a day because I was just sitting on this and basically just buying the newest coins to get the deck scanner boosts or buying the newest coins to um get their deck scanner banners updated. And I thought that was uh like pretty cool to see that somebody could build something within five prompts, uh, probably in a matter of five minutes. Um, and it was actually super useful to other people. Um it wasn't necessarily super useful to an audience of millions of users, but to a couple different traders that actually went ahead and interacted with it. And I thought that was super cool to see that um an app doesn't necessarily have to be like something that serves thousands of millions of users and makes millions of dollars of revenue, but can actually be yeah, just useful to a couple people, but to those people it is super, super useful. Um but yeah, I think I think that one also has a bunch of cool apps to share.
SPEAKER_01Yeah, I mean I have like uh I think our first like kind of viral hit was this HTML related applications. We just had like new developments in the past few days, so it's uh still relevant. Like some guys had this uh unredacted version of the of the files. I don't even know if that's like true or not, but like the the the government, like the administration just released the redacted version, and this guy on the same day published like an unredacted version, and there's the they added this chat feature where like people can just kind of like scroll through the docs and just chat about it. Um and it went like super viral. That's the first app that kind of took off from dev, so it it has like a special place in my in my heart. Uh, reading the chat was just a hilarious experience. There's like this social thing happening. Um, I mean, crypto users tend to have like very interesting takes. Uh, I was just pretty heinous in general, but very funny.
SPEAKER_00Yeah, but which files were those? Like what which government files?
SPEAKER_01You know, it was like the Epstein um Black Bull. Oh wow. Like um, like it was like all these Epstein-related things, right? Um, and and so so yeah, this was pretty funny. But I think in terms of like real, like more complete apps, like we've seen recently, like Blabser, for example, is like this um Info Fi or whatever the category is called that allow people to like kind of go on Twitter and talk. And so we're starting to see like some real real applications uh that are like very solid, just kind of being built. And I think this is also super awesome to see. In general, my favorite ones are like the fun games, like can I be Gary Gensler and like shoot paintballs or like something like that, you know.
SPEAKER_00That is awesome. Uh yeah, there was like a Jerome Powell game I played when I was first looking at you guys, or you could like mess with inflation and money printing and see if you break the economy. I think, if I'm remembering correctly.
SPEAKER_01Yeah, yeah, I remember that fun. Yeah.
SPEAKER_00Yeah, that was fun. All right, so we've laid the groundwork, guys. I think everyone knows what you guys are doing, what you're targeting. I want to switch to some some potentially tougher questions on the longer term, longer term view. Um, so one is just the like the notion of vibecoding is giving everybody a software assistant, right? And I think a lot of people are debating whether AI takes over and sort of does everything for us in an AGI fashion, or if we have these tools and assistance and things to help us. And DevFun, I think, falls into that category for people. I'm curious your views on AGI just creating all the apps versus using DevFun as your assistant and your tool to achieve that goal.
SPEAKER_01Yeah, I'm um extremely skeptical, at least on the short to midterm, that the AI is just gonna build the whole thing itself and just like kind of make millions of iterations. Um, I just don't see it. I've been coding since it's possible to do so uh using AI. I I barely write code manually um nowadays, but uh I would never just let the thing build itself by AI. And I actually think it's been getting worse um over time. Like as the smarter the models are getting, uh the worse they are at kind of just building something autonomously almost. Because I don't know, it's like they kind of get lost in the source. Uh they try to come up with like these very smart solutions, um, and they have like way larger context windows now and way larger like possibility to output, and they just create monsters if you let them. Um, it is completely unmaintainable, basically. Um I I think like it's it actually got worse over time. That's that's what I feel. But if you kind of uh be able to like guide them very properly, like if you're able to think with good structure beforehand, if you're able to plan correctly, um, then they are like just incredibly incredibly powerful. Um, so I just don't see us being just completely replaced. Um I think Microsoft was saying like 30% of the code is written by AI or something in in the company nowadays. I don't know how true the the figure is. It's probably not completely false. But I I would not imagine that to be like 100% anytime soon. Uh humans are like maybe the code is written by AI, but like you're gonna have like software architects still kind of chiming in and like guiding the whole process, basically, right?
SPEAKER_00Yeah, it is it is hard to think through if it does sound like most of the time letting them go off on their own leads to issues. I'm just worried, I'm wondering like the velocity of their improvement, like if they improve to a degree where they AGI just creates any app at will. But yeah, it it's a little confusing.
SPEAKER_01Yeah, I just I just don't see it as someone who's written so much code, I just don't see it happening anytime soon. But you know, we're like maybe just one breakthrough away, so it's also pretty hard to predict, right? I would say unless like there's a significant breakthrough in in the next like two or three years, or like we're not we're not replacing humans.
SPEAKER_00Yeah. And I mean the other question I have for you is I think the idea is to create super viral apps on Death Sub, right? And there becomes a time when these live-coded applications need to graduate to full-blown companies, right? They have full backends and tech stacks and you know, full team. How are you thinking through that? Like, is there a pathway for me to create a TikTok, a Snapchat, a Telegram, or is that just outside the scope?
SPEAKER_01Um I would say it's outside the scope of the current iteration of the product. Um however, like that's been a topic of discussion with with Robbie internally for a while. Is like, how do we scale this to be not just one app, but maybe like a more like a project slash micro, like small company entity where you don't just have maybe a single app or maybe just don't have a single dev. Uh, how do we think about these entities? How do we think about scaling them? It's it's the current iteration of the product is is apps, is one app that you're building. Um, I think that's that's what we can do very well for now. Uh in the future, I think is gonna be companies.
SPEAKER_00Nice. Yeah, I'm excited to see what it turns into on that side. And then a different question I had for you is every time I open Twitter, I realize how good Cursor is, how good Claude Code is, just one-shotting like highly technical project fixes and enhancements, right? I'm curious if the advancement of the underlying models helps or hurts dev.fun. Like those foundational models improving. Does that help the experience of the user? Does that hurt your competitive ability? How does that work?
SPEAKER_03I would certainly say it helps. Um, I think every time we've seen one of these underlying models kind of improve, um, all our users come back to us and mention like, yo, did you guys just update DevFun? We can build so much more cool stuff with this. Um, and we've seen the quality of apps increase since since we kind of launched in January, um, just because also the underlying models are kind of improving. So I would certainly say it definitely is add adds to adds to DevFun, and it definitely helps your users.
SPEAKER_01Yeah. The um there's the way these things get better. The model is improving, uh, which is kind of like a passive buff we're getting basically. Um but also like the scaffolding and all like the backend and internal tools that we're building and giving the AI uh just gets used better, just gets like leveraged better, basically. So there's like this double like improvement almost. So I think we highly benefit from models getting better. I just hope they get better faster, you know.
SPEAKER_00And then on the the web 2 side, outside the foundational models, there's some big players, right? Like there's lovable, there's bolt. Um, I think I saw the lovable founder say that they created 10% of the websites last month. I don't think those numbers totally make sense, but but directionally maybe. How do you guys view DevFund competing with the web two giants that aren't crypto native, not on chain, don't have this capital formation? But to be honest, they are really good, really good application generators in their own right.
SPEAKER_01Yeah, um, I I don't see us as exactly trying to conquer the same markets. I mean, I I don't know their like whole plan, obviously, but uh from what I'm seeing, they're like really positioned as tools to just build something. Uh which I mean obviously we do that, but I think we really position ourselves as like kind of the more like an app store model almost, uh, where distribution is what's the most important. Like the numbers of websites lovable claims they're like building. I'm I mean I'm willing to kind of buy into it. I actually think maybe it's possible, but you're never seeing that this is built on Lovable, basically, they don't really get credit for it. Um, what they're after is just like their MRR, like, can they get users to subscribe to build those things? Um, I think we really look at Roblox, Steam, and App Store uh as kind of like what we aspire to be more so than lovable. Where essentially the distribution is key and like everyone that touches content there uh kind of create this flywheel where like the winners will bring more customers to you, will like increase the distribution, which will make um more people want to kind of deploy there. So at the end of the day, we just want that's a place where people discover like new companies, new apps, new new consumer products, and just interact with them, you know.
SPEAKER_03Yeah, and I think as all of these kind of models get better and as all of these tools survival get better and better and better, the main focus will not necessarily be like, okay, what can you build? Because I think at some point everybody will be kind of on par there. Um, but it will become like I think the key component there will be like, okay, how can you distribute? Because we'll see everybody um building their own little tools, building their own projects, building their own um websites and companies. But um, as everybody is basically given this opportunity, like it's really going to be difficult to kind of stand out. And I think that's where Web3 and Crypto is really the niche to kind of go for, um, just because of access to CT, which I think is an incredible way to basically go out and and spread the word, um, just because you can basically just DM every user and people are pretty responsive, but also like this token aspect is pretty, pretty important, I would say, because it's like an incredible way to basically evangelize early supporters, early backers. Um, and I do think this is some some some way we're gonna be seeing uh it's gonna be pretty normalized in the future, just because it's gonna be like the best way to kind of go to market, um, and also just the best way to kind of grasp that attention that's gonna be so scarce and and so heavily fought over um as soon as everybody just basically gets the ability to build whatever they kind of want and and and limit it to no time.
SPEAKER_01Yeah, exactly. And I I think also like the way we want to monetize and we approach like um how we're making money is also very different because basically we just want to have it's a volume game, so we just want as much UGC as possible, we just want as much apps as possible, as much companies deploying there as possible. Um, we're gonna kind of win with the winners. Um, like the ones that are like really popping off uh will bring most of our like kind of revenue and financial upside will come from those guys instead of just yeah, having people build um tons of websites and they're just like paying a subscription. Um, and I think like this approach you can tell by the output and what people create on dev.fun versus what they would create on like lovable. You'll notice that it's actually very, very different because we really want to have like winners. We have to you know max out like their viral potential, we have to max out like how fun and good it can be to use one of those apps. And you see the creative output of our devs is just way superior uh in in general. Like I've been looking a lot at their like competitions that they organize, and they end up with a lot of like web front page, um, you know, more like boring stuff, um, maybe useful to a different type of audience. But uh yeah, I think we're just very different products at the end of the day.
SPEAKER_00Yeah, it's you guys have clearly thought a lot about it. One avenue I do want to double-click on is the nuance between dev.fun, like requiring tokens for apps, and and like a lovable and a bolt and cloud code, etc. not requiring them. So the there's obvious positives, right? Capital formation, building a community, potentially even a forcing function on on if this thing's gonna be successful or not, and to move on. But I'm curious, like, have you thought through not requiring a token? Or is that something that's core to the core to dev.fun forever?
SPEAKER_03Yeah, so so basically when we went ahead and we launched, um, we required every app to kind of add and link a token. Um and a couple users asked us um to be able or to be given the option to go ahead and let them add a token later since they kind of wanted to build out a little project and build up some hype for it, traction, whatever, to go ahead and launch a token later. So that's basically what we did. So um yeah, at this at this point of time, you can either um add a token immediately or you can add a token later, and it's really up for the user to decide. Um so yeah, we we did give that uh some some sort of some thought. Um, but we we also have seen that most users can still choose to kind of launch a token. Um but yeah, we have we have given people the opportunity to go ahead and um yeah, add a token later.
SPEAKER_00That is pretty cool. I didn't know that. I thought you had to, I thought you had to launch with one immediately.
SPEAKER_01Yeah, that's something we recently kind of changed. We're like always kind of trying to refine and see see what's possible to do, um see what's needed to change. Like this has been a lot of internal debate about whether the token is mandatory or not. But we may come back to it being mandatory at some point. Like uh, you know, we're just like always kind of seeing what works and what doesn't as well.
SPEAKER_00Yeah, I guess I mean you know, the whole essentialism playbook, you you wanna you can't be everything to everyone. I I do see the the value though in letting some developers just monetize with with in-app transactions first launching a token, but I guess it's yeah, it's up to you guys at the end of the day. Yeah, exactly. I see we're just experimenting a lot as well. Cool. A different avenue of question I wanted to ask you guys was on the technical architecture of of DevPun, right? Like, I'm not sure the best questions to ask you here. The first ones that come to mind are, you know, what's on-chain, what's off-chain. And I think the second most important question um would be, you know, what AI models are being used under the hood? Like, is this a prop proprietary model? Is this a bunch of different models? Do you not want to share that information? Just curious to get the low down there.
SPEAKER_01No, I mean, I think we can we can share. Um basically, in terms of what's on-chain and are off-chain, um, pretty much most is off-chain. Obviously, we're allowing people to do Solana apps and we have like a Solana framework. Uh, but in terms of like accepting payments, for example, this is not obviously there's a transaction involved, uh, but a lot of the payment verification, for example, is like how do we make sure um user cannot buy PASI payment? How can we make sure the amount is correct? Uh, this is all like kind of back-end services. I think the idea is to have the least amount of on-chain possible because I mean, even Solana is is is pretty fast, pretty great. I I think like having just back-end verification is still the most like kind of efficient, the best way to go about it. Um, and works really great with AI. So most of it is off-chain, but we obviously provide RPCs, we provide uh Helios RPCs to all our users. So by default, you can create an app that will have a wallet connection, that will have uh ability to connect to any smart contract uh that you desire, basically, right? So, yeah, that's for like kind of the on-chain buses, off-chain part. In terms of the models, uh, we're not like at the stage where like we want or it's worth it for us to create our own like prop models. Uh, we're just using combinations of uh whatever is state of the art. Uh nowadays it's I mean, obviously, Cloud is doing a great job uh in terms of coding model. I think uh Gemini is also great. We're using both, depending on like what exactly is happening, uh Gemini being way cheaper and kind of a good way for like doing debugging, for example, because it's it's cheap, quite fast. And so basically, what we're doing is we're taking those models and we're like scaffolding, we're building like a whole scaffolding around those. And most of the props thing that we do comes from all the services that we build and provide that are designed to be used by AI, basically. So we have like uh database services, people can prompt and have an app that has like uh permanent states with the database, which um is pretty unique feature. Um, like if you go and like love a ball of these other, like you have to kind of connect to your own uh back-end service. Uh, we kind of provide a lot of those primitives by default and we design those to be usable by AI. Uh we create like AI first SDKs that we just plug into like our uh app generation framework. So like the Solana SDK is the same and all the other things we provide are the same.
SPEAKER_00What do you think would be necessary to move on-chain in the future that isn't on-chain now? Like, I don't know if it would be the the models or storage or yeah, not uh maybe using decentralized or distributed compute providers to to run the apps. Like what do you think would be the most valuable thing to decentralize moving forward?
SPEAKER_01We're uh we're not like the most um we're not the biggest people on like decentralization. Um I think what is important uh would be I mean essentially kind of like the hosting. I mean, I think whatever makes um an app harder to like be taken down by you know, we we we love like the freedom, obviously, that people can deploy kind of whatever they want and have like some sort of like you know, freedom of speech and all and all those things. So making this resilient is important. There's no uh like best kind of ways to go about it right now, is how we feel.
SPEAKER_00Yeah, I I guess a different, maybe a different question on on the protocol itself and an improvement and opening it up. There's obviously that the platform exists for anyone to come to build their app. Do you ever envision a world where people come to Dev.fun to enhance the the platform itself, like incentivizing AI developers to come and build on like the core, the core protocol itself?
SPEAKER_01Yeah, 100%. I mean, we're already exploring those things, uh, whether it comes from, you know, I mean, we've explored different things and like obviously all those MCP servers, like all those types of infrastructure uh that you can kind of use to plug and play and extend capabilities of your engines. That's something we kind of had in mind since the very first day, where like, how can we leverage like, you know, in crypto, it's very common to kind of leverage your community to work for you. Well, like kind of like this idea. So instead of doing all the integrations and like AI first SDKs on like extending the capabilities of our engine ourselves and like taking this burden, uh, we definitely want to like open it up eventually, where like people can start contributing their own like uh extensions um and SDKs basically, and then uh it becomes like you know usable by everyone, or maybe they can even monetize that. Uh, these are definitely things we're like thinking of.
SPEAKER_00That's awesome. It's always nice when other people come and help to build and things like that. I know we're jumping around a bit, but another topic I want to make sure we cover on this is just relation, right? Like there's 22,000 apps. I think there's hundreds of developers. I'll I'll let you share the the stats so I don't mess them up. But how do you guys can't just filter and sort and and cut through the noise for for users to access the best apps?
SPEAKER_03Yeah, I I think that's also pretty interesting. It is it it follows like I think apps, like obviously people go on dev fun and check out their favorite ones. Um, but I think dev fun kind of leverages a similar flywheel that, or like flywheel, but like a similar core mechanism that kind of makes PumpFun successful, where a dev will actually go ahead and create the app on DevFun, um, obviously get some attention from it on the platform itself, but will actually go ahead and basically show it on on Twitter, um, share it with people and get other people excited. Um and I think Twitter is really like the best way, best way to kind of do this. Um and these type of people actually come back to DevOne, check out the app and play around with it. So I think that's a really great way these people actually go ahead and promote their apps. Um, and also how a bunch of people actually go ahead and and and find out about DevOne, I think.
SPEAKER_00I guess the token, like do the token market caps per app do you think that's a good form of curation, or could that be a bad, a bad signal? I'm trying to figure out if that helps or hurts on the curation side.
SPEAKER_03I I think it does help to some extent because I mean obviously tokens are just like reflective of the attention that flows towards a given app. So it just basically shows you, okay, this is an app that's hot right now, this is something that has a lot of mind share. I mean, it's it's just like a simple way of how a token works, right? Um, but I wouldn't necessarily say that it's like super important for a really cool or useful app to have a really strong token. Um we see a bunch of different um apps on our platform that people use um that don't necessarily have a token and that's performing incredibly well, um, but are still apps that are pretty cool or useful or um yeah, have some some kind of users. Um but I I do think it is it is a decent way of kind of basically allowing users to like or like allowing apps to curate um like what what's working or what's not working. Um so yeah, I I I I don't think it's like necessary, necessary, but it definitely is like a good signal.
SPEAKER_00So you ever had to step in and like turn an app off or just restrict an app or or can you or should you? Uh it it's obviously a really hard question to ask founders because people take extreme views of this, and but yeah, curious if that's happened.
SPEAKER_03I don't uh we never had the uh situation where we really had to go ahead and step in. Um because like from from from there's like I guess two issues at hand where it would require us to step in. One would be if there's any sort of malicious smart contract um stuff going on. Um with the basic dev base that we have that allows users to build these applications uh using smart contracts, we have we have a pretty secure framework that basically um like um only allows users to build apps that I guess don't include any like malicious, um malicious um malicious content or malicious features. Um but on the other hand, when it comes to like um yeah, I guess like this NS NSFW type of content um where people uh create these apps that are like I guess like racist or or whatever, um, I think that also kind of gets filtered out a lot by um the underlying models that we use. Um, because I mean you can't really go to like any sort of like LLM and um come up with like the most racist concoction uh known to man because it will block you from doing that. So I I think there is some sort of like uh um safeguards there that basically filter that sort of stuff out for us. Um so yeah, there was never really like a situation where we had to go in and um yeah, address somebody that built an app. And I don't think it's something that we would would want to be doing, right? Um let the market let the market, yeah, let the market decide, let people decide what they kind of want to bet and speculate on. Um yeah, we don't really want to be the juror there.
SPEAKER_01Yeah, I think our stance is basically that we don't really want to intervene um unless it's like blatantly illegal. Um we we let kind of users comment, like you can comment on an app, you can kind of do like a review. Uh there's a report button, we have like disclaimers, so obviously we're mindful that you know some things can happen. Like it's it's crypto, people will try to maybe you know extract or like do some do some weird things. Unless it's illegal, we're not gonna touch. I think we just want to like kind of let people also like some freedom to do whatever they want. Uh, we're not really positioning ourselves as the judge uh of what is you know considered like good or bad. I think um it's really hard to kind of define those things, especially in crypto, like everything's a scan now uh because price is going down. But is it really? Um it's kind of hard to tell. Um I think if we start having this burden and if we start kind of having full responsibility for this, um it's just like an impossible, it becomes like an impossible task, basically. So, you know, there's like laws, and we're gonna like follow the laws. Uh, and basically it's like you know, child porn, like this kind of stuff is like obviously uh in the instant remove, right? Um, everything else is just up to kind of it's you know user-generated content, basically.
SPEAKER_00Yeah, that is fair. It it I guess there comes a time when you'll have your polymarket is Zelensky wearing a suit uh uh concern, but but but sometimes having that much concern and issue is great because it just shows that hypergrowth, I guess.
SPEAKER_01Yeah, there's obviously gonna be some growing pains. Uh I think like one of I followed pretty closely when uh you know Pam Fun was having this whole debate about the streaming back when they had streaming the first time, before they shut it down. Um like people were doing a lot of crazy things um for attention. And I think it was extremely entertaining personally, but a lot of people were not really liking it. And um, I think it was just very interesting to observe this this debate. Nothing illegal was actually happening. Uh I think like you would see like some fake kidnappings, whatever, but it was obviously all like performative. Uh none of this was actually real, right? None of this was actually illegal, but it's just like kind of people just didn't like it, so they tried to like shut it down by by canceling it and being just loud. Like who was right? You know, like it's it's kind of like uh there's nothing illegal happened, you know. So it's just like very interesting. Uh people like when you associate kind of money, uh, coins, and attention, somehow it starts becoming like this very weird relationship where they're like um you know, like it's like the next black mirror or something. Yeah, I think we're just gonna follow the laws for for when it comes to this.
SPEAKER_00Yeah, it it is interesting, especially because it appears live streaming is like one of the core theses of Pump Fun given their tweet this morning, right? Like killing Instagram, TikTok, etc. So they they may have launched it early, but they're definitely leaning back into it.
SPEAKER_01Well, obviously, I think like when you saw what was happening there, I'm I I was convinced this is the future of entertainment. Like literally, I'm like, okay, this is insane. Like, this is the craziest thing I've ever seen. To me, it really doesn't look like a Black Mirror episode. It's just like people do these kind of things all the time, but this was like a crazy platform to do so with a lot of attention back then.
SPEAKER_00Yeah, I mean I could be totally wrong, but if if they're able to target like influencers and creators and target all the people not in crypto that use those web 2 platforms, like it could be a massive onboarding event.
SPEAKER_01Yeah, 100%. I mean streaming is the future, especially like with all I mean, we're in the age of uh obviously so much, especially when it comes to video. I think the AI slopification of content. Like if you look at Twitter, uh basically everything is AI slop right now. Uh when I tweet, like some of the replies are like just bots. There's like barely any human content, it feels like. Um, having this authentic experience of just streaming is gonna become increasingly valid more valuable, I think. Uh so yeah, it's just like a huge future for streaming in general, whether it's web two, web three, like yeah.
SPEAKER_00Yeah, no, I'm with you. Um back to back to you guys though for the final 15 minutes we have. I I wanna, I mean, maybe a couple disparate questions here, but why Solana, right? Like we see uh, you know, there's there's really successful L2s on Ethereum. You could you could launch your own chain. Like, why did you decide to build here and why have you decided to stay here?
SPEAKER_03I I I genuinely think there's no other option. I think everything that is is kind of interesting in crypto, like consumer facing happens on Solana. And I think it just has like the best support rails, um, it has the users, and is just like the easiest, uh, easiest way to also onboard new users, easiest UI, all this sort of stuff. So I think it's just the most consumer-facing um chain there is. Um, and I wouldn't really see it another option, to be honest. Um, I mean, even when I talk to friends, um, I I I then I talked to them. I I don't know when the last time was that I kind of interacted with Ethereum or any of our other people that are active in crypto. Um really actually we went on and used an app on Eve. Um so yeah, I I think that's just like the only option to be honest.
SPEAKER_01Yeah, we didn't really think about it too much, yeah. We like we we just knew like it was gonna be Solana. Uh we're also like, I mean, obviously we wanted to like the coins to launch to like PAM.phone as well. It was just an obvious choice. Uh, I think if we were to like if you ask us, as is there like actually another option, I think the only kind of viable one we've seen would be base, um, which has some merits. Uh, I think in terms of like distribution and their coinbase integrations, like the Coinbase wallets, all those things. Uh, but yeah, it was Solana was the obvious choice uh to to to to get the product out there.
SPEAKER_00Yeah, no, that is maybe it being the default no-brainer is the is the signal there, right?
SPEAKER_01Uh yeah, it was just the no-brainer, exactly. Like we didn't have like a debate about it. We just like it just was that from the beginning.
SPEAKER_00Yeah, that is that is pretty cool. And I think one of the cool things about having hosting you guys and other founders, I get to ask like the before and after questions along the way. When you first started, you know, no apps on the platform, and now we we bring up the site today, and there's 22,000, right? So everything you were thinking about then, the growing pains, the features to build, the headaches, the the fun experience have have changed, right? What are you both thinking about now that differs from when you started? Is there different problems to overcome, different features you want to launch? Like, how has your mentality changed as founders from now until then?
SPEAKER_01Yeah, there's there's many things. I think we we both have things to say here. Um the my first assumption, and especially as a software engineer, was we have to make the most um like uh you know feature complete, you know, like software engineers all want to like build like very complex like things, like we want complex systems, you want like I don't know, there's like this kind of uh fetishism, like the software complexity almost. Then we launched, and then people just want to make the funniest, I think the funniest meme. And we're like, okay, like I guess market is always correct, you know. Um they actually don't care about like very complex features. What they want is just make uh funny interactive experiences. Uh that was like a very interesting thing I I learned from going to market. Um I think in general, it's also like, and that's something we want to expand more into. We assume that the people who are building the apps are building the software, the companies, are also gonna be completely taking care of the distribution. But actually, the people who are like good at doing the selling are not exactly the same profile as the people who are good at doing the building. And so I think we're thinking into how can we structure this in like more of a complete like project company approach where you have like maybe people are taking good software and distributing it, and you know, maybe people are just like building the software and like waiting for someone to distribute, and maybe they get like some reward for that, and like they get paid. Because having like the same person do both the building and both the selling. Uh, the people who are really good at doing the selling generally just don't really maintain their ads, they just try to do like some hype stuff and basically farm the users. Um, and the people who are really good at doing the building, they can't really get traction. Uh, so how do we kind of solve this? Is like a next challenge.
SPEAKER_00Robbie, what about what about you? Devloor, that was awesome. I want to get Robbie's thoughts too before we move on.
SPEAKER_03Yeah, I think I genuinely agree with Devloir there, to be fair.
SPEAKER_00Yeah, that is the way to go. And I mean, thinking ahead, like it is hard to build an ecosystem from scratch, but I think I've noticed this, you know, the the buildle team is is really helpful for you guys, and I think you're you know closely in touch and it kind of demonstrates an early community around you guys. Can can we maybe talk a little bit about what them, what they've built, and how they're helping Dev Fun?
SPEAKER_03Yeah, for sure. Um, I mean, it's pretty funny because they're actually um, I mean, not really associated with us uh directly, but uh basically they were just like one of some of the earliest users of DevFun that were super excited about it and were one of the first people to kind of join our whitelist when we launched. Um and since then um they've been super bullish on this vibe coding thesis and have also thought to like add to kind of the DevTon ecosystem. And basically what they've been doing is they've been trying to build this like um accelerator for vibe coded projects, which I think they've done pretty successfully so far. Um so yeah, we we we we we are in touch with them. Um we try to help them where we can. Um obviously they're uh super beneficial to us and um also uh great guys, a really cool, uh really cool team. And and yeah, they they support um projects on Devon, um support builders on Devon, and yeah, I've just been really kind of a really good asset to have since since we launched. And and yeah, the the project is cool, the team is nice, um really, really bullish on them.
SPEAKER_00It's pretty cool to have an organic accelerator out of nowhere.
SPEAKER_03It's definitely been helpful.
SPEAKER_00All right, now one of the the hardest questions for you guys. Um if you were to eventually launch a token, you you could go fine if you do or don't, or it's undecided at this time, which is also fine. Figuring out how to reward all of your creators is tough, right? Because a lot of people are here quickly, they leave quickly, some stick around and and they stick around for a long time, like just normal power law distro. How do you how do you go about incentivizing people and you know keeping them? And it just seems really difficult.
SPEAKER_01Yeah, it is really difficult. I think ideally, like the ideal world is like the incentive speaks for itself. Basically, we don't have to do anything. What we just have to do is let them get to money as fast and as efficiently as possible. And and this is enough for them to be like, you know, to see the value. I think we we we've seen people kind of act this way already, where they're like, oh, you know, like I would benefit from uh going from dev.phone instead of just doing something on my own because there's like better distribution, maybe because like I'm gonna get more attention, or uh it's gonna take care of the tech aspect for me. I have the idea, I know how to distribute, but I cannot build the thing. Like we've seen many cases where people kind of see the intrinsic value. Um, I don't think it's like like like jobs not done, basically, I would say. Like we still have work to do to provide even more value and make it even clearer that you know this is the place to be where like uh if if you don't go there, like basically you're you're not gonna make it. Uh, we're not at that stage yet, but that's the goal, right? It's like um think of it like a game launching on on Steam versus creating their own like launcher. Like almost no one does that because the benefit is so obvious uh to just go on Steam and pay them the fee, right? Uh you just get like best distribution and it's just easier uh for everyone. So we we just want to position ourselves this way. Uh I don't think tokens uh necessarily help. I think it's more like a vanity thing in a lot of cases where I can see the value of doing it, but uh doing it in the wrong way will be like just a bunch of vanity metrics where basically you're bribing like user for like short-term um growth, but it's unclear whether it's gonna stick around or not uh once it's over. So I think our approach has always been like can we see them sticking around before we start like bribing them basically? Because the token incentivization is just like kind of boring against the future. You know, you need to just pay for user acquisition, basically, right? Uh so we're just careful about this.
SPEAKER_00Yeah, it kind of feels like a good metric to track would just be that in-game spend, you know, seeing when uh seeing when people are are spending a lot of money in the apps. I guess that shows and then the creators make money outside of token anyway, so it could work well.
SPEAKER_01I'm actually curious, like as a as a where you're obviously an investor in the space for a long time. I'm actually curious, like if you're looking at you know, investing in like, let's say you're like a trench VC now and you're looking at projects you want to you know gamble a little bit, like buy some really low caps, like invest in very early stage companies, like what would you look at? Like, what's the criteria that would interest you the most?
SPEAKER_00Yeah, it's it's I try to not wear too many hats at the same time because it's just the the mental capacity is is obviously limited and the context switching really really eats into things. So I'm I'm trying to stay a little bit out of the trenches and more on the VC side. But if I were to envision dev fun a year from now, and I'm an investor and I want to spend money it and and invest in in these applications, I mean just having a dev fund dashboard of users per app and the longevity of the creators and the money spent within their applications, I think would be and then just like being able to see like videos and MAs and stuff of of the creators and and their vision, I think would be a really cool, cool site to have. So you could get people to invest into these things. And I think that could be a really cool French site too to have.
SPEAKER_01Yeah, that's that's they're insightful. That's actually exactly uh what we're thinking. Yeah, cool.
SPEAKER_00Well, guys, I want to thank you so much for coming on um and having me and my partners as as investors, and want to thank uh Zero X LTR of our on-chain data sleuth at the research company Delphi Digital for introducing us. And uh and thanks so much, guys. I'm really excited for uh for Defun.
SPEAKER_03Yeah, thanks for having us. Yeah, man, what's great at shopping?
