Devtooling builders session

Transcript

Okay, I think we're live. We're good. Sorry if you're at home going, "What the heck?" Uh, this is, uh, meant to be more of like a roundt. I said round tableish. Uh, we wanted, uh, at the EF, we wanted to get tool builders together, the application layer folks building tools for people. Um funding is hard, grants are drying up. Uh people are allergic to when you say crypto, you work in crypto, you start saying AI and other things. Uh but the goal here is just to kind of get an open conversation and bring tool builders together and see uh I was just talking to the hard hat guys and we were talking about what what is different in this world of AI and how

are users even finding our tools and how are they making the decisions and one thing we were talking about is like agents are recommending whatever is the top thing. So, if you're hard hat and wondering about Foundry and Foundry verse hard hat, what does the agent recommend? And if you're in second place, you don't want to be in second place because you're not going to get recommended. Like 99% of the time, the agent's always going to just build with the other uh the the incumbent. So, that's one thing, but there's lots of different discussions around funding also and how people are staying funded in this this environment. So, basically, we have a mic. We have 15 minutes. I don't really have like I think I actually my agent

maybe asks some questions. Uh yeah, what what do you build? How is it funded? What's your biggest problem right now? That's slob. But uh if you are a tool builder and you want to ask those questions, if you want to ask other questions of folks here, if you're having problems or troubles, one of one of them was SEO, right? going going to that problem as a tool builder, distribution is still probably the hardest part now that the software is trending toward being free. How how do you get distribution? How do you get in front of builders? Uh I I was a devril for a long time and it was kind of about like time to hello world and building uh good-looking docs and now that like

humans don't read the docs at all, right? So things have really changed rapidly. Uh, I'm kind of just rambling. We could uh ask questions if you want to take some space. I want to make some space for it if anybody wants to talk or ask questions or has opinions. We're we're open to that. Uh, we got some ENS guys here. We got a Peruv are you are you wallet Chan? Is that what you are backing? But impersonator, right? Got a lot of tools from a Peruv. Uh, some other tool builders here in the space representing different standards. But um there's a microphone, there's an open space if you guys want to ask questions, if you want to talk about difficulties.

Um yeah, anything. Anybody want to take a mic and talk? I have the mic and we'll hand it to you. Anybody brave? Maybe the hard hat guys could kick us off. You You guys were We had some good discussions there. Can I bring you the mic? Will you be willing to Okay, cool. Here we go. to just reiterate the SEO question. Right. So for from our point of view, we've been we've always prided ourselves on good docs and we've been trying to keep those now egentically readable, but we're in competition with a good competitor and foundry and we are interested in any techniques for making sure that you know our tools are getting out to the agents which are now the

intermediary between ourselves and the developers that we're we're trying to reach. So if anyone has any techniques for that like we're are trying to make sure that we've got appropriate skills and the appropriate skills databases that we're writing our documentation in a way that's easily consumable by agents. Does anyone have any other take? >> Those are my first two things. Right. The typical SEO like we're writing lots of articles and then this the skill file is like the second thing where you're creating like a shelling point for here's the thing that the agent needs to know that it doesn't know yet. And so I'm sure that that stuff is used for training. But yeah, what's the next thing in terms of SEO? I think that's a great question.

Distribution, getting on Twitter and building lots of cool examples with it. >> Yeah. Yeah. Here you go, Carlos from the build guild. Um, so I think I I don't know if you remember like the SEO like early days the you know like people were doing like some like tricky stuff you know they were putting you know the background is black and the text is black something that you you couldn't see in the screen but was like Google was taking that into indexing so I'm wondering if the version of that currently is just creating a lot of a lot of content right because people are doing is oh like thinking that is like the old days and we need to create like content so that content get indexed into you know like the training but I

don't know because I think we are going to generate like so much crap in content in in general so a lot of noise and I think um I labs are going to do something to make sure that the content is good and that might change any time because now you used to to search for in Google you know how to build on Ethereum or whatever but now it's like your agent you know and your agent is going to say I want to build this and if you don't say anything it's going to pick you know whatever is you top ranked I guess but I'm just saying that I think that rank is going to change and we don't know what is going to work and I see a lot of people that are creating like thousands of articles and in my experience I would say that that is going to be bad for

them at you know like medium term because now it seems like you are like tricking the system getting a lot of content and stuff and and LLMs are aware of that but I think at some point they will release something that is going to like the new Google algorithm into search that are going to you know create a little bit more and so I for me I think yes like you say I of course to generate content but maybe have like some kind of like curation and intention behind it that that's my take maybe maybe I'm mistaken but I think we are going to get like some new algorithms that is going to disting if it's good stuff or not >> so bearish bearish on the slop cannon it's still somewhat useful for SEO and

people are still generating but like it's so easy to generate a 100 articles now. Exactly. >> That it's probably more about like where you're placing your information and >> and some other signals that they'll use to figure out which ones are the best. Pablo did some stuff at the build guild where you were doing lots of articles too. >> Wanted to complement his his tip. I think like the basics from SEO will continue working as they always did. Of course, people are abusing of creating content now, but creating the right content, the content that makes sense for you and going back to the basics and working a lot on the architecture of your website, how all the information

along the sections are all those useful articles are structured in your website and getting links from useful and uh uh domains with authority in the um solidity and web3 domain will always work. So I will just go to the SEO basics. >> Thank you. >> My my best advice for you guys was to position yourself as number one, right? Like what would Franco say? We got to get that out to everyone or have someone else say it in the way that he says it.

Uh Wigglewag's back here. Let me get him the mic. And you you are a tool builder too. >> Yeah. Yeah. So I build to like actually a plugin for our deploy deploy plugin for that. Um and then now I'm making uh kind of a new version of it which kind of can work on other fun if you want. Well, it's mostly focused still on ADA that as but yeah, I was just wanted to say that I think in the end the thing is all about uh because AI tool AI will still look on the web for stuff. So if you target yourself something uh where you do

capabilities that don't exist elsewhere then you will still be picked up. So it's all about actually it doesn't change anything. It's like about what can you provide to what people need. And so if you if you don't if you price the same thing that other than Yeah. I mean if you are number two you stay number two but if you have probably something that the other two don't provide then you can be first on that category. >> Yeah. >> So being a differentiator in terms of functionality and you still need to slap can and talk about it. >> Yeah that's >> when you're doing something first that's no one else is doing that differentiates you.

>> Yeah. >> Cool. any you want to any >> well I one thing is that actually I sucks at the second thing which is like spreading the news about the capability for I feel my tool of probably everybody thinks that their tool have something unique uh so yeah I guess doesn't really answer the problem but in some way maybe we can be positive about that actually doesn't change maybe is not much change it's always the same thing it's like um yeah you need do something new and then make sure it's known. So, you still have to do the Yeah. the >> And you've you've scrapped for funding for like eight years like I did as a

tool builder. Any advice for someone? Let's say there's someone in the crowd who is building some some kind of independent tool like hard hat deploy that you built. It's not the hard hat of the space, but it's the hard hat deploy of the space. Uh what advice would you have for them now in this new environment 2026? Yeah, to be I I felt I was just lucky. So I don't really know how >> be lucky >> but uh yeah for me I built for myself that's it and it turned out that people find the same problem and actually that's probably likely like it's likely that if you do something that is useful for you I mean there's very likely that someone have the same problem. So yeah, maybe it's just yeah, I guess still luck

part of it, but yeah, I don't I can't give much advice. >> Anyone? >> Anybody else? I'm actually curious. Um, in the time of LLMs, we're sort of moving to writing a lot of scratch code because you're like, oh, I might as well write this thing. Whereas you probably had a need for hard hat deploy. you wrote it and then you're like, you know what, this is going to be useful for other people. Do you find that with LLMs, you have less of an inclination to actually productize an idea that you have and you could just keep it as a scratch idea locally that works for you, but you don't want to invest the time or credits in order to productize it. And this applies to everyone by the way.

>> Yeah, it's for me. No, actually I with I can make it product like very polished. That's what I spend the last like now that I fully using AI it's I go polish actually I was part of the like GitHub fund and for security and so is AI could do all the job to make like uh make it like more secure like the way I distribute or all this the things so uh I would have never touched that before I will never care so I think I mean but maybe it's not the same for everybody Yeah.

>> Yeah. >> I just wanted to ask a follow-up question on that. So, uh if someone is I mean if some tool is uh simple enough and uh I'm creating a project and previously I would not build that tool myself but my LM would just uh build that tool from scratch myself instead of installing someone's tool. >> Yeah. So uh do you think it's worth it to just keep on making packages uh like that? >> You know I think it's I mean I think for me it's true like I will rather build it again myself for but I think it's all about if it's simple or if it's yeah I don't know I think it depend again of what what you I don't what my take.

Yeah, I think distribution is the hardest part. Like once you've built some little product, you still have to get it out there and see if other people want to use that product. And that's I you should and like Wigglewag said, like a lot of people will just recreate their own product that does the same thing, but I don't think it should be a chilling effect that anyone could just oneshot the product you built because sometimes you have a really good idea and it kind of works, right? And someone else sees it and they're like, "Oh, that's what I need." So, I think you should still build it and ship it and see if people use it. But that's your your your kind of trial by fire is once you put it online and tweet about it. If no one uses it, then it's probably not

worth supporting. Did you want to jump in there? Another builder in the space, what are you building? What do you Oh sorry, you already had a mic. Sorry. >> Yes. So, I had an idea about the distribution aspect of it. So I think if we have more like trusted directories and stuff so those already rank better on SEO websites. So like for L2s we have L2 beat and for wallets there's wallet beat. So if we have a trusted source where the LLM already knows to query that instead of having our own blog post on some random domains. So if you can get there there also the LLM can actually analytically analyze the various options that is there. So like in case of foundry versus hardhead if it can see and compare the numeric values

like the compilation takes this much seconds or those so then it can make like educated guess to what would be the better of the two options. So I think if we can get better domains in such of those things that can be useful >> and is ethereum.org basically ethereum toolbe like is there better ethereum toolbe than ethereum.org tools orbuilders or whatever you find there. Yeah, I think that's the best domain for Ethereum developers. If there's I think comparative tables so the LLM can have like data to back the claim up then that would be ideal.

>> Not just hey this one's the best but this one's the best and here's why and go through that >> because the users can have separate use cases. So when they're chatting with LLM if they have a specific use case in mind a tool can be better in one aspect but not better in the other one. So if they can clarify that, that would be I >> Greg's prompting his bot to go build Ethereum tool beat. >> No. So I had uh kind of like two points that are actually counter to each other now. Um but one is like I have recently uh I recently built like a little indexer thing for that runs in the front end that works differently than something that was out there before for a very niche use case. And I thought it was cool and I when I was building it I

was like in a past life I would have just built this as a library from the start but now it's like I'll just kind of embed it in this thing. And then um two days ago I was rebuilding the same or I wanted the same exact thing again and I was like okay I could either pause now and go turn it into a library and then use the library in both or I could just say build this the way it is in this other repo. And I was like yeah I'll just build it the way it is in the other repo. And so I have very very similar code in two places now. >> That's like >> which feels really bad as a coder. I know a programmer you hated this but now >> right it's like um it was quick and easy and it worked on the first try and um

it's like if I need it a third time will I then feel the need to turn it into a library? Maybe. Um but that was like a kind of counter to the point or maybe not a counter an example to the point of like you keep things as scrap instead of publishing as a package. Um but then also yesterday um just to kind of like I I had a I was doing a safe transaction um on a whatever and there was a bunch of call data because now they don't decode it for you unless you're on the safe pro or something and instead of dumping the call data into an agent I googled decode call data just for fun and um the first two results are or the second result at

least is E thesh um and that is like that's a tool by a small >> builder who's in the room >> right here in the room >> right um and that is like proof that SEO still works and then that um was like the best called data decoder and I've used it before by going to it um but like that was the best one and then there are also articles from like chain link and alchemy on how you decode call data and all that But the second result is Eth. Um, and that's great. And so like distribution is important and it is important that people still build tools even if you could just kind of slop your own in 2 minutes. Um, but also

like I myself sometimes still just slop stuff in two minutes. So um, it's not like a useful insight here. It's just kind of brain dumping. >> That's what we're here for, I think. Yeah. So the traditional models are still working a lot as things change a little bit. Yep. Okay. M maybe appar like what's the hardest problem you have right now? What's the thing that like pisses you off the most? >> I think I haven't monetized my tools yet because money.

>> Yeah. >> Yeah. >> Like right now building my tools like I built so is kn.xyz then the v2 version is the 8.sh message but right now the tools don't need any funding so I just do it in my spare time to solve my problems and once it's live it's just the hosting cost for me but right now as more and more users are coming to the space I think like I'm have to explore the monetization aspects of it because someone has already reached out like they want to place some ads on it because they see that like the SEO is working for it so I want to explore like what's the best approach for those things like this is a hosted platform so it's different from the other sort of tools where the user has to do the SDK and uh install the packages. So yeah, I

haven't I've got the incoming request but I haven't pushed anything there because once you have a ad then it looks like paid and then you have to trust the entity that's posting that ad there. I saw like roti wallet or like the roadk platform they have a tiers list. So each month or like each week they have tiers of the sponsors that they want. So like gold, silver and bronze those sort of tiers. So maybe I can explore that. So I wanted to know views from other people here. >> We should have left Tarius here. I don't think he is.

>> He's the expert grant funer. He's been around doing it for like 15 years it feels like. Yeah. So uh on the topic of funding you uh like in this space there's kind of like the people who say we're builders and we don't talk about price and they look down on the trenchers and the token people but like you did do a token on uh banker I think and that did help fund you somewhat. But I think there's an interesting tension here where a tool builder who's building useful tools that are showing up that we're using isn't getting like the funding that they need to survive and they are using some of these other methods like a token to

survive and it was sort of useful to you. But there's definitely like a tension here between what meme coins can do for builders in the space and how we fund builders in a good way. And I don't know if I even have a question around this but I don't know how do you think about that tension it was helpful for you right? >> Yeah so like the token was helpful so I built wallet chan which was earlier called banker wallet so I saw that banker released their APIs and so I quickly built a wallet on top of that and the community was token first so someone like used the banker tokens and like deployed my own version of it. So it was good for the short term but people were more the users of the token instead of the actual product that was

behind it. So more were traders and more wanted features related to that. But in the short term I could sustain based on that. But in the long term you have to keep on building more features and you don't have to like stay away from your original like vision that you had for the product because once you have a token then more efforts need to be gone down that path instead of like in my case I have to build like various like hardware support or some other things but the traders or like the users wanted to have more functionality to open like staking and all those things. So it can be useful for like the funding aspect of it but you get diverted from the actual use case that you started with initially.

>> I'm also confused in that regards because the token runs in hypes. So at that time there was like AI and the AI matter was running on base. So more and more people were using that but once the hype dies down so there's a token and other things. So you always need to keep innovating but you can't like keep on launching tokens because that's also not sustainable. >> Not even the tension of uh this like whole meme coin and builders hate meme coins and memecoins give crypto a bad name but even the tension of once you have a coin and you're surviving off of it, you have to do suboptimal things to like keep driving the hype cycle rather than like build the tool that people want. So I mean

>> what's great for distribution like it's sort of >> you do get tons of distribution like immediately userenerated content. >> Yeah. >> Yeah. People share screenshots of it. They share it in their telegram channels. So I don't discreard that aspect of it because it's great for distribution. So like even the existing like deck screener it's like a marketing page for a token. If it's ranking top there more and more eyes would go there and from there the people can go to your main website and learn about your project. So that's another aspect of it. So there's more good and bad. >> Good and bad. I think for other possible tool builders in the room though, uh what a Puru does well is like pay attention to the trends and build things

quickly in like trendy places and a lot of interesting things happen out at the edge. And I think uh someone on lock computer said something like you uh it was Adam who did um FWA, but he said uh you can't be a great writer if you don't read a lot. And I think you use a lot of tools. You use Ethereum all the time. You same thing with Wigawa. you you end up needing a tool because it's some weird case that is brand new or at the frontier and you just build it and show it off and ship it to people rather than keeping it quiet and sometimes you have a banger and it it gets used a lot by a lot of people right so I think for other

tool builders in the space like take that as a lesson like be tinkering at the edge be shipping be showing the things off on crypto Twitter that's kind of the place to to do it and get lucky Okay, we're probably getting close to the end, but really like kind of open open mic if you have questions or you want to Yeah, let me send a mic your way. Here you go. >> So, hey, this is Ravi and I'm a founding engineer in a company called asa. So uh we are building an open source tool called as Raven which is a private information retrieval system which is a PI and uh currently it helps you to do private it's a closer blockchain data

and currently it's on Ethereum mainet and yeah we are thinking to uh integrate with rail gun and uh the recent Kohaku wallet as well. So if there are any views of you people or what do you want to think about it as well as we also building a mixet called as NOX NO X so where uh you could uh route your brackets uh without uh finding your uh without giving your IP address. So you have the specific uh IP privacy metadata privacy. So our main aim is to have a full stack privacy over Ethereum where institutions and high network individuals can have a uh a better market experience as in uh regular web

web to market with some regulations and stuff. So yeah, if anyone want to know anything or else if you have any ideas on this, I would really like to know about it. that that for me that brings up the topic of we're talking about putting your tool out there and getting distribution but at the EF we're thinking about like this deep adoption term where it's not necessarily about the numbers it's more about like the type of people if you're providing privacy to you know if let's say there's a hundred users out there in the world your tool in your mind might be useful if 40 of them or 60 of them are using your tool but uh In terms of deep adoption, there may be

like a specific three people in that hundred that need to use your tool. And if you build a tool that works well for them, that I still consider that a success. It's just you you're not measuring it in terms of numbers anymore. You're measuring it in terms of impact for uh certain uh high use case individuals that that need self- sovereign money. And that's kind of what we're here building. >> I don't know if we're getting kicked out. Do is there someone else that needs to go next? Yeah, I just had one point to share >> that like we should invest more of our time in the retention system of our products.

>> So like uh retention system like as in like whatever loop we are building like earlier in the preAI era whenever like we like used to build something uh like the cost of experimentation was quite high but uh right now uh the cost of experimentation is quite low. So we can probably experiment more. So that's good in one sense but uh also wrong if we just keep experimenting. So yeah just wanted to add that point that uh we should uh invest more of our time in uh creating products that have a good retention and probably like uh get away from the inner loop to the outer loop like inner loop basically uh means short-term transactions on the product.

But uh when we go to the outer loop uh we can return the users uh for like a long period of time. I think that's good advice. Anybody else questions? Hard things in the space. There we go. I was curious for us like the the target audience for hard hat has always been developers and with the transition to I presume most people here using a gentic engineering at this point there's sort of a middleman in between

for us where our user is becoming claude or whatever Gemini um and has that shifted people's requirements for the tools that they're using to develop do you find that you used to have specific values that you had as a software for a developer that you really cared about, but as your workflow has dramatically shifted, do you no longer care about those and now you care about different things? And what would those things be in your experience? I mean, at this point, it's about getting the job done, right? Like, does the thing do what it needs to do? That's a great question, though, as to as things change. Anybody have any good answers to that?

It's a like it's a hard question, right? Like it's not >> I think I I can speak from my own experience. The thing that we've very much noticed is >> um more polish which I think aligns with what you were saying uh because you can polish more at no extra cost of time like that has become a free commodity that you can use. So like just going 100% on your engineering prowess and really delivering on all the things, better security, better testing off the thing. Um the other thing is more experimentation that I think was mentioned. You can just experiment very easily um at no cost. Um and you don't care if it's a throwaway prototype.

I think Kenzang Kenzang, is that right? Am I saying it right? You had a question. He's a builder from Bhutan. He built the NDI system. They use it for their identity. >> Uh hi everyone. Uh I am not a dev tool builder but uh I would love to be one. So I see that there are lots of experts and builders in the room. So my question would be is there any advice from you for a beginner that would love to be a dev tool builder? Thank you.

I kind of shredded through some of my advice and that was using lots of tools in the first place, using lots of apps and building lots of things and getting them out there. But anybody else have any advice for, you know, other young tool builders or tool builders just getting into the space that want to be successful, launch a tokens. Yeah, my advice is to solve your own issues first and if you can share it with the community. So the community might live in a discord or some specific niche community but if you can solve their issues first then you're have a chance to gain traction. So I started with a very basic tool just earlier

there was some like long ago there was some airdrop that happened but there wasn't like a UI for the users to easily check the thing. So I just shared in a discord and then I shared it on Twitter and people started following from that. So if you can build your own following via that and solving specific niche issues, that's always a plus thing. >> That's great. Yeah. So it's not just like using lots of of things and tools, but using tools and then finding frustrating or missing pieces and then building those things and showing them up on Twitter and then you like build your following. So the next time you launch a tool, you get 10 users automatically and then the next time you launch a tool, you get a 100 users automatically. That's That's good

advice. Yeah. Okay, we're probably toward the end. Anybody have any last words? Go use clear signing. Is anybody still looking at the code? That was a question we had uh last Yeah, I bet the hard hat guys look at the code. I bet they're some of the only guys in the room that still look at the code. We got We got another hand back there. >> Rarely. >> Rarely. Yeah, I don't Yeah, I don't look at the code.

I haven't looked at the code for years, though. I'm a bad case. Yeah. Can we get him a mic? Yeah. Oh, sorry. Were you going somewhere else? >> Oh, yes. >> We can go there next. Oh, here we go. >> Has anyone used a debugger in the last six months? >> Oh, man. Like stepped through code. >> Yeah. Like, has anyone used a debugger? Like, it's a genuine question because that is on our road map and >> my agent might, right? Like if it has a good skill file and I can be like, "What the heck is going on here?" It'll be like, "It shows in the skill file. I can use the debugger. I'm gonna go run the debugger real quick. Here's the problem.

I'm gonna go solve it." But me with my fingers no but maybe someone in here. I don't know. Just for fun, for for nostalgia. Yeah. to try it. >> Yeah. >> Waiting too long for it. >> Yeah, I would say it's more like uh uh these things is more like um when you want to do it for fun. Like I learned coding before because I thought it was fun. So sometimes when I have to have fun coding, I just uh do it manually and then I will just like write uh I don't like huff code or use like a debugger or something or write like assembly and

it's like a like a good practice for the brain hijen you know trying to not be dumb and uh just trust the machine but you get to rewire your brain to think again about some stuff. But uh yeah, I feel like it's kind of like I don't like sudoku. It's like hard to do it. So sometimes just for fun you're like okay I want to write like this algorithm from scratch. And I mean some people are more attached by the end product but uh personally for me I prefer like the process starting with an idea and figure out how to build it. So if you prefer this aspect then sometimes you prefer to

still do things manually today because you have more um uh reward in the end than just prompting but it's like personal matters in the end >> and and your role is changing from a developer to more of like an artist. We were talking about making clay pottery like you don't make clay pottery to have the dishes in your house anymore. like you're doing clay pottery for the like the not just the aesthetics but the enjoyment of making the pot. You enjoy the syntax of the code and you enjoy the art of of building and it's a different thing altogether now and it used to be a means to an end to get the product in someone's hand and now there are certain

people who like to just build for the art of it. Probably won't make a ton of money doing that but maybe you'll learn a lot and enjoy it. Yeah. What? Yeah. Yeah. Uh hi folks. So I uh I was flutter developer. I was because I don't write flutter code anymore. Uh so I wanted to uh so in flutter there are two things. One is that you test the actual devices because it builds mobile app. Uh other thing is like checking memory, checking leaks that we used to do with debugger. Uh so what I found out that AI agents were good in summarizing

stuffs but as soon as I asked it to find out exactly a code which was leaking something or which would break something. So it was not able to do this and it is about uh 2025 Q4. So I built a tool for myself and I think that this would be a good advice for all of us. So I built a tool myself to do exactly what I was doing. So for example, I used to start simulator. So the tool will start simulator and I used to connect the app through dev tools uh flutter's dev tools. So it the tool will again find out URL and connect to dev tools and then do sets like okay scroll the app check for memory scroll the app check

for memory. So now the tool is uh good enough to do automated development but in the end it gives me a whole report how did it did the debugging and how did it uh did all the things. uh my confidence level with that tool is so high that whenever I have a feature I just give that tool to agent and it is able to build complete feature uh including UI uh including back end of the app uh and yeah so I just have to spend like 5 minutes before I publish it to play store or app store.

Uh hey guys sorry uh sorry to interrupt uh I mean we are already past the uh time now but if there's any uh closing or last question uh you can do that. >> Yeah let me get your mic. >> Oh there we go. >> Thank you. >> Hi folks. So uh we just talked about how nobody's writing code anymore and uh I'm kind of like uh in a difficult place with new generation of builders that you kind of build some things using different kind of tools and over a period of time you make mistakes. You take various architecture choices and then you realize what is good architecture and what is good system

design and what isn't. But since nobody is writing code anymore and uh this new generation of builders needs to uh like I need to get them up to speed very quickly. If anybody has any suggestions on how do we uh you know make that happen I would be very grateful. A very simple example is that there used uh there was a hack two three years ago where uh the smart contract code developer wrote uh continue instead of break and because of that a million of hack something like that happened. So just curious if anybody has any advice on getting the new generation of builders up to speed.

>> Yeah. Uh I mean this is not advice but I found out that gap uh when I was working for income tax of India and I face a bank the biggest bank in India. Uh so to tackle that because I am a community guy. I run lot of community in Mumbai. I started a community known as mobile architectural lab. Uh so we did four events out of it. Uh a lot of people were students and some working professionals. Uh this Sunday we did hackathon for mobile architectural lab just to focus on that problem. So we did not build any app into that hackathon.

Instead I asked every attendee to start with one pager problem statement given to them uh then work on HLD LLD and then come up with EDRs architectural decision records and then by end of the day we chosen a winner who was able to explain all of this that to us. uh a winner who uh the winner uh three people who who are today here who also got free tickets to uh uh today's event were able to explain each and every stuff that they did or they had in their documents.

What I felt what I really liked was throughout this so before this we had three events where we had lot of speakers talking only on architecture not anything else. These guys attended that and they were able to confidently talk about lot of principles, a lot lot of tradeoffs and but yeah, so this is what I'm doing. Maybe it may be useful. Uh but yeah, >> we're we're out of time. Also for builders, we have like speedrun Ethereum and a lot of tools and stuff. So if you have someone who's technically good at building or an idea guy, send them to speedrun Ethereum also. Okay, I think that's it. Thank you guys.