1 00:00:00,040 --> 00:00:02,200 This episode is brought to you by Beam OPS. 2 00:00:02,400 --> 00:00:05,520 If your Elixir deployments are painful, or your infrastructure 3 00:00:05,520 --> 00:00:07,920 is a mess, or you just need experience hands to help you 4 00:00:07,920 --> 00:00:11,840 scale, these are your people. Beam OPS specializes in scalable 5 00:00:11,840 --> 00:00:14,800 Elixir development, automated deployment pipelines, and 6 00:00:14,800 --> 00:00:18,000 infrastructure migrations. They will help you work smarter 7 00:00:18,040 --> 00:00:21,480 and ship faster. As authors of the Pragprog book 8 00:00:21,560 --> 00:00:24,520 Engineering Elixir Applications, they know their stuff. 9 00:00:24,720 --> 00:00:30,560 Check them out at beamops.co.uk. Secure your Elixir apps with 10 00:00:30,560 --> 00:00:33,280 paraxial IO. From code analysis and bot 11 00:00:33,280 --> 00:00:35,880 defense to dependency checks, they've got you covered. 12 00:00:36,040 --> 00:00:38,840 Elixir first design with GitHub integration. 13 00:00:38,960 --> 00:00:46,280 Elevate your security at paraxial dot IO testing. 14 00:00:49,880 --> 00:00:52,320 What's up? I'm Jacob Litzow, your Elixir 15 00:00:52,320 --> 00:00:55,000 mentor, and welcome to another exciting episode of the Elixir 16 00:00:55,000 --> 00:00:57,080 Mentor Podcast. This is where we discuss 17 00:00:57,080 --> 00:00:59,160 everything related to Elixir, from interviews with 18 00:00:59,160 --> 00:01:01,680 enthusiasts, pioneers in the community, to innovative 19 00:01:01,680 --> 00:01:04,000 projects and libraries shaping Elixir's future. 20 00:01:04,400 --> 00:01:07,800 Today I'm joined by Vasilis. He's head of software 21 00:01:07,800 --> 00:01:11,320 development at Teacher Space and he's here we're going to talk 22 00:01:11,320 --> 00:01:15,960 LLMS, Ash, you know, just agentic software and Elixir in 23 00:01:15,960 --> 00:01:18,400 general. And I'm sure we'll dive into a 24 00:01:18,400 --> 00:01:21,440 ton more. Thanks so much for joining me 25 00:01:21,440 --> 00:01:24,200 today, Vassilis. I really appreciate it and it's 26 00:01:24,200 --> 00:01:27,800 great to meet you. Thank you and nice to be here. 27 00:01:28,640 --> 00:01:31,480 And sorry if I butchered your name, I feel like I I'm not 28 00:01:31,480 --> 00:01:35,520 doing it right but I'm trying. It was good, It was good. 29 00:01:36,360 --> 00:01:40,400 All right. You want to for our audience 30 00:01:40,400 --> 00:01:43,120 here, do you want to just kind of give us a brief introduction 31 00:01:43,120 --> 00:01:46,080 about yourself, like where are you based? 32 00:01:46,080 --> 00:01:51,400 What is your background? Kind of where did you start to 33 00:01:51,400 --> 00:01:54,200 before you got into Elixir? Sure. 34 00:01:54,320 --> 00:01:57,520 Yeah. So I'm I'm based in Prague and 35 00:01:57,880 --> 00:02:00,760 basically so I'm it's where my parents grew. 36 00:02:00,760 --> 00:02:03,640 That's a funny story and come from my agrees that both my 37 00:02:03,640 --> 00:02:06,760 parents can like go up here and yeah. 38 00:02:06,760 --> 00:02:11,200 So I've been working with Ellixi for like around 9 years now. 39 00:02:11,640 --> 00:02:17,760 And yeah, so basically I guess like I started with with Ruby, 40 00:02:18,280 --> 00:02:20,000 like most of Ellixi programmers do. 41 00:02:20,800 --> 00:02:29,160 It was like around like I think 2000, like 14 that I started and 42 00:02:29,160 --> 00:02:31,840 basically it was in college. So I wanted to start building 43 00:02:31,840 --> 00:02:37,480 some, something like a product that I had in mind and started 44 00:02:37,480 --> 00:02:40,280 looking around what can be used. And I found like Ruby and Rails. 45 00:02:40,720 --> 00:02:46,840 So I kind of like got into doing Ruby and I kind of like really, 46 00:02:47,160 --> 00:02:50,240 I really kind of started loving the language because like in the 47 00:02:50,240 --> 00:02:52,080 school we were doing only C, Java. 48 00:02:52,280 --> 00:02:55,800 So kind of like these, like I would say like not very 49 00:02:55,800 --> 00:02:57,360 exciting, kind of like languages. 50 00:02:58,240 --> 00:03:00,800 So it was really interesting to me to kind of like find Ruby and 51 00:03:00,800 --> 00:03:02,760 it was very practical. He could do things. 52 00:03:03,880 --> 00:03:08,280 I remember I was like working with us, the Rails book, I think 53 00:03:08,280 --> 00:03:12,440 it's called, basically it's you're creating this like online 54 00:03:12,440 --> 00:03:15,040 store. And yeah. 55 00:03:15,040 --> 00:03:16,880 And I was like really excited about the language. 56 00:03:17,320 --> 00:03:23,440 So I wanted to like learn more. I went to some meet ups in, in 57 00:03:23,440 --> 00:03:25,440 Athens. So like it's a fine. 58 00:03:25,440 --> 00:03:27,400 So I was kind of basically like studying. 59 00:03:27,640 --> 00:03:31,400 I would go like I would take the train travel like 4 hours to 60 00:03:31,400 --> 00:03:35,320 Athens and go on a meet up. And yeah, I'd come back 61 00:03:35,320 --> 00:03:38,960 afterwards. And yeah, in, in one of the meet 62 00:03:38,960 --> 00:03:44,400 ups, there was a guy, Jimmy, he's a good friend now. 63 00:03:45,840 --> 00:03:48,520 And he started like talking about Elixir basically. 64 00:03:50,240 --> 00:03:53,280 So, yeah. So that's kind of like my first 65 00:03:54,320 --> 00:03:57,000 introduction to Alex here. A couple of years later, I 66 00:03:57,000 --> 00:03:59,800 wanted to learn something functional and, and thought 67 00:03:59,800 --> 00:04:03,080 Alexia was a good fit to like start to learn that. 68 00:04:03,760 --> 00:04:07,000 And since then, I guess I was like an Alexia advocate. 69 00:04:07,000 --> 00:04:12,680 So I, I really kind of started like language pushed it in. 70 00:04:13,560 --> 00:04:17,640 At the time I, I was working in a small consultancy here in 71 00:04:17,640 --> 00:04:20,320 Prague, or like it's, it's actually fairly big now. 72 00:04:20,320 --> 00:04:24,040 And at the time it was small. And yeah, since then I've been 73 00:04:24,040 --> 00:04:25,480 kind of like working with the Lex here. 74 00:04:25,640 --> 00:04:30,560 I've been working mostly with startups going from yeah, from 75 00:04:30,560 --> 00:04:34,440 like fintech to supply chain, like, yeah. 76 00:04:34,680 --> 00:04:38,000 And now I'm working in a tech. That's awesome. 77 00:04:38,640 --> 00:04:42,440 Very cool. So you you got like you have a 78 00:04:42,440 --> 00:04:44,760 computer science degree. That was your background. 79 00:04:45,880 --> 00:04:52,640 Well, not really. So basically so I dropped off 80 00:04:52,640 --> 00:04:56,480 the university because like I started learning Ruby and it was 81 00:04:57,280 --> 00:05:04,480 quite, I would say like I didn't feel I was learning enough in 82 00:05:04,480 --> 00:05:07,200 the university. It was like very technical. 83 00:05:07,200 --> 00:05:10,800 So we were learning things like electrical engineering as well 84 00:05:10,800 --> 00:05:16,600 and, and kind of like, yeah, which is interesting, but I kind 85 00:05:16,600 --> 00:05:19,240 of like wanted to like do things like imagine we didn't even 86 00:05:19,240 --> 00:05:22,040 learn git in in school. So like it is something that I 87 00:05:22,040 --> 00:05:24,320 discovered by. Myself, people still don't learn 88 00:05:24,320 --> 00:05:26,920 git in school, which is wild. Yeah that's great. 89 00:05:27,040 --> 00:05:30,920 You should be starting like group projects where you have to 90 00:05:30,920 --> 00:05:34,520 use git and do pull requests and code reviews, right? 91 00:05:34,520 --> 00:05:37,200 And they don't. Yeah, that's that's very 92 00:05:37,200 --> 00:05:39,920 interesting to me, right. Like it was great. 93 00:05:39,920 --> 00:05:44,080 Like when I discovered git have it was like, you know, like like 94 00:05:44,080 --> 00:05:46,200 get in the candy store they say right. 95 00:05:46,840 --> 00:05:52,040 So like it's like projects. I was like, yeah, so so I've 96 00:05:52,040 --> 00:05:54,400 never finished the university. I just like went to work. 97 00:05:54,400 --> 00:05:59,440 So when I started like doing the kind of like Ruby, I came here 98 00:05:59,440 --> 00:06:03,400 to Prague for like 1 summer and I met some people. 99 00:06:03,400 --> 00:06:06,720 They were very kind of like we started talking about Toby and 100 00:06:06,720 --> 00:06:08,640 stuff. It was very exciting and and I 101 00:06:08,640 --> 00:06:11,560 just started to work for them. Nice. 102 00:06:11,920 --> 00:06:14,040 I, I kind of have a similar background. 103 00:06:14,040 --> 00:06:18,160 I, I got my like undergrad degree in civil engineering, 104 00:06:18,160 --> 00:06:20,360 which is like, you know, construction and building, 105 00:06:20,600 --> 00:06:25,800 whatever. And I then got the smart idea a 106 00:06:25,800 --> 00:06:29,600 few years later, I was going to go and get my master's degree in 107 00:06:29,600 --> 00:06:33,280 software engineering, but all they were teaching was Java. 108 00:06:33,400 --> 00:06:35,680 And I actually dropped out because of Java. 109 00:06:35,840 --> 00:06:40,960 I like, I hated Java and I could just teach myself and move so 110 00:06:40,960 --> 00:06:43,160 much quicker, kind of like being self-taught. 111 00:06:43,160 --> 00:06:46,920 So I just went that route. And so software kind of the same 112 00:06:46,920 --> 00:06:50,560 path. I I quit university and I've 113 00:06:50,560 --> 00:06:53,760 taught myself, yeah. Yeah, I think like you can learn 114 00:06:53,760 --> 00:06:56,560 a lot, a lot like in a in a company, like much more than 115 00:06:56,560 --> 00:06:58,920 like studying. But I guess it depends. 116 00:06:59,040 --> 00:07:01,280 Yeah. Like if you want to do something 117 00:07:01,280 --> 00:07:03,760 very focused, like like it, it depends, right. 118 00:07:03,760 --> 00:07:06,360 Like if you, for example, like want to do AI and you go to the 119 00:07:06,360 --> 00:07:08,920 company doing AI, even there you're going to learn more. 120 00:07:08,920 --> 00:07:12,480 But there's some things in the academic world that's I feel 121 00:07:12,480 --> 00:07:15,080 like interesting, right. And, you know, definitely 122 00:07:15,080 --> 00:07:18,000 there's also like many parts you can take from an academic 123 00:07:18,000 --> 00:07:20,640 perspective and, and do a lot. Yeah. 124 00:07:21,040 --> 00:07:25,000 I do, I do agree. Like you learn a lot of theory 125 00:07:25,040 --> 00:07:29,080 and like computer science and like algorithms and like why you 126 00:07:29,120 --> 00:07:31,320 do certain things, which is important. 127 00:07:31,880 --> 00:07:34,160 But I think as long as you're self-taught and you're willing 128 00:07:34,160 --> 00:07:37,320 to go and learn those things too, you're good. 129 00:07:37,320 --> 00:07:39,520 Like go build. But then like when you kind of 130 00:07:39,960 --> 00:07:43,400 hit the bump in the road where you have to use, I don't know, 131 00:07:43,640 --> 00:07:45,480 big O notation or something, right? 132 00:07:45,480 --> 00:07:49,480 Dive in and actually understand why, and then I think you'll be 133 00:07:49,480 --> 00:07:52,280 good 'cause you don't really know what you don't know until 134 00:07:52,280 --> 00:07:59,200 you run into it, right? Yeah, I, I remember when I first 135 00:07:59,200 --> 00:08:04,080 started just, I could just build like an iOS app, right? 136 00:08:04,120 --> 00:08:07,120 And I could get a couple screens, push some buttons, do 137 00:08:07,120 --> 00:08:09,560 everything client side. But I thought I was like on top 138 00:08:09,560 --> 00:08:12,440 of the world, 'cause I can, I can build software now. 139 00:08:13,840 --> 00:08:16,480 Yeah. Yeah, and then the longer you're 140 00:08:16,480 --> 00:08:19,480 in the game, you're like, dude, there's so much I don't know. 141 00:08:22,120 --> 00:08:24,480 Yeah, it's a classic, the handout in Kruger effect, right? 142 00:08:24,480 --> 00:08:27,280 Yeah, exactly. Yeah, it's funny. 143 00:08:28,520 --> 00:08:33,200 So now you work for teacher space or is it? 144 00:08:33,360 --> 00:08:37,120 Is it your company? No, I, I, I work for them and 145 00:08:37,280 --> 00:08:38,919 basically it's, we just put a choir. 146 00:08:38,960 --> 00:08:40,520 So it's like. Awesome. 147 00:08:41,159 --> 00:08:46,000 Weeks ago and yeah, so we were a very small team. 148 00:08:46,000 --> 00:08:50,680 It was like 3 people. So yeah, so I was the only 149 00:08:50,680 --> 00:08:53,800 engineer there basically taking care of the whole system. 150 00:08:54,200 --> 00:08:57,600 And yeah, we now go. To DevOps and everything. 151 00:08:58,920 --> 00:09:03,440 Yeah, yeah, everything. And it's actually there's a lot 152 00:09:03,440 --> 00:09:05,640 because there's a lot of integrations with like third 153 00:09:05,640 --> 00:09:08,920 parties. So it's in the Danish 154 00:09:08,920 --> 00:09:14,000 educational system and you have a lot of like basically I think 155 00:09:14,280 --> 00:09:16,800 they are private companies that like work with the government 156 00:09:16,800 --> 00:09:21,080 and they develop these like very large applications that like the 157 00:09:21,080 --> 00:09:22,800 whole educational system is using. 158 00:09:22,880 --> 00:09:26,480 So it's like big contractors and yeah, you have to integrate with 159 00:09:26,480 --> 00:09:28,160 all of them, right? So like you have like 1 160 00:09:28,360 --> 00:09:33,960 contractor that is like for basically the, the user 161 00:09:33,960 --> 00:09:35,680 management kind of like the identity. 162 00:09:35,880 --> 00:09:38,560 Then you have like different for like the schedules, others for 163 00:09:38,560 --> 00:09:42,000 like grading. And so so a lot of integrations. 164 00:09:42,000 --> 00:09:44,720 And I feel that's kind of like the hardest part because you 165 00:09:44,720 --> 00:09:49,840 have some really modern partners that like use, you know, like 166 00:09:51,320 --> 00:09:54,120 modern technology like maybe graph QL or something, or kind 167 00:09:54,120 --> 00:09:58,480 of like they're using like view plug insurance and some others 168 00:09:58,480 --> 00:10:02,480 that are on very old technology. So, so for example. 169 00:10:02,760 --> 00:10:06,200 And yeah, so you have to like, balance it out, yeah. 170 00:10:07,560 --> 00:10:09,840 Did it? When you integrate a ton of like 171 00:10:09,840 --> 00:10:13,360 third party APIs, I feel like it's so hard to like, keep your 172 00:10:13,600 --> 00:10:17,160 like UX peaceful and flowing nicely. 173 00:10:17,160 --> 00:10:20,200 It's a difficult problem 'cause you have to like not only deal 174 00:10:20,200 --> 00:10:23,160 with your errors, you have to also deal with errors you can't 175 00:10:23,160 --> 00:10:27,160 control. Yeah, yeah. 176 00:10:27,200 --> 00:10:31,760 What just like kind of like a a macro description? 177 00:10:31,760 --> 00:10:33,280 What? What is teacher space? 178 00:10:34,680 --> 00:10:38,400 So we, we basically build software to help teachers. 179 00:10:40,120 --> 00:10:43,600 Like we have like various different like sub products I 180 00:10:43,600 --> 00:10:47,400 would say. So we have one where it's helps 181 00:10:47,400 --> 00:10:49,720 like with the video. So basically teachers record 182 00:10:49,720 --> 00:10:53,280 their classes and then they can like receive feedback. 183 00:10:53,640 --> 00:10:57,640 Customers are using it for actually different like us like 184 00:10:58,120 --> 00:11:00,040 in different ways. So we have, for example, like 185 00:11:00,080 --> 00:11:04,440 one of our biggest customers is actually a university that will 186 00:11:04,440 --> 00:11:06,960 record that. Like it's not teachers who 187 00:11:06,960 --> 00:11:09,280 record the videos, it's their students. 188 00:11:09,640 --> 00:11:13,760 And then they are getting graded on the on those videos. 189 00:11:13,760 --> 00:11:16,680 So basically it's students that are becoming teachers. 190 00:11:17,480 --> 00:11:21,120 So they're students, but yeah, they're on on the path to become 191 00:11:21,120 --> 00:11:23,560 a teacher. And then they have like as part 192 00:11:23,560 --> 00:11:25,960 of the assignment, like they have to like take some take 193 00:11:25,960 --> 00:11:29,240 videos and then there is a professor that like rates them. 194 00:11:29,240 --> 00:11:35,720 So for example, that's one use case and then we have like a 195 00:11:35,720 --> 00:11:42,200 chat with a with a RAC system. And this is more for like most 196 00:11:42,200 --> 00:11:46,320 of our customers are like primary school is called like 197 00:11:46,360 --> 00:11:49,280 right, like that the first school you go like the primary I 198 00:11:49,280 --> 00:11:49,640 guess. Yeah. 199 00:11:51,640 --> 00:11:57,680 And so, yeah, so it's, it's a chat with a, with a rack system. 200 00:11:57,840 --> 00:12:02,560 We have like like a product to where they can like correct text 201 00:12:02,840 --> 00:12:06,080 and yeah. And like you generate images. 202 00:12:06,080 --> 00:12:11,840 So basically it has the element degrees kind of like the 203 00:12:11,840 --> 00:12:15,040 generative, kind of like a type of functionality. 204 00:12:17,320 --> 00:12:19,920 So yeah. But like the main goal is to 205 00:12:20,440 --> 00:12:27,160 help teachers kind of like find like, like when they're starting 206 00:12:27,160 --> 00:12:30,800 their journey and they don't kind of like know a lot to kind 207 00:12:30,800 --> 00:12:34,880 of like boost them so they have access to the material that they 208 00:12:35,080 --> 00:12:38,200 that is they can ask questions and, and yeah, things like that. 209 00:12:38,640 --> 00:12:41,040 OK. So it's like kind of it networks 210 00:12:41,040 --> 00:12:44,680 teachers a little bit. And yeah. 211 00:12:45,160 --> 00:12:48,800 And teachers in like university or whatever that are becoming 212 00:12:48,800 --> 00:12:50,760 teachers. Yeah. 213 00:12:50,760 --> 00:12:53,920 So that's the, that's The thing is that like it's still a very 214 00:12:53,920 --> 00:12:55,840 small company. Like now that we've we've been 215 00:12:55,840 --> 00:12:58,680 acquired, we, there's going to be like much more use cases. 216 00:12:58,920 --> 00:13:02,760 But up till now, it was like very small, like we have just a 217 00:13:02,760 --> 00:13:06,680 few customers and overall like the whole operation you can say 218 00:13:06,680 --> 00:13:10,320 was, was quite small and we're still trying to find the part of 219 00:13:10,320 --> 00:13:11,760 it. So that was like, I think the 220 00:13:11,760 --> 00:13:15,040 hardest part because the, the company did like a pivot like 3 221 00:13:15,040 --> 00:13:18,280 times where they started with an idea of something that could 222 00:13:18,280 --> 00:13:20,200 work. It would start like selling to 223 00:13:20,200 --> 00:13:22,440 some customers and then they saw that kind of like, OK, this, 224 00:13:22,440 --> 00:13:25,560 this wouldn't scale. Like the initial idea even 225 00:13:25,560 --> 00:13:29,840 before I joined was to record videos so that the teachers 226 00:13:29,840 --> 00:13:32,400 themselves, they can kind of like give feedback to each 227 00:13:32,400 --> 00:13:35,480 other. And that kind of like didn't 228 00:13:35,480 --> 00:13:39,760 work out because it's like it just didn't scale. 229 00:13:39,760 --> 00:13:43,360 The teachers didn't have like time for it and it didn't make 230 00:13:43,360 --> 00:13:46,720 sense for for like most of the schools. 231 00:13:46,880 --> 00:13:50,160 So then we have to start thinking, what else can we do? 232 00:13:50,360 --> 00:13:53,760 We created like this automated feedback where basically we 233 00:13:53,760 --> 00:13:57,840 would like read and make a transcript from, from the, from 234 00:13:57,840 --> 00:14:02,680 the video and then kind of like analyse it with an LLM and, and 235 00:14:02,680 --> 00:14:04,720 provide feedback to the teachers, teachers. 236 00:14:04,720 --> 00:14:09,600 But then again, that didn't work out because they either they 237 00:14:09,600 --> 00:14:13,480 didn't have time to kind of like record the videos that come to 238 00:14:13,480 --> 00:14:16,440 the first problem. But also they didn't like the 239 00:14:16,440 --> 00:14:18,720 LLMS at the time. They weren't at the level of 240 00:14:18,800 --> 00:14:22,720 giving good feedback. It was very hard to transfer the 241 00:14:22,800 --> 00:14:29,600 audio correctly because you have and basically like Danish and 242 00:14:29,680 --> 00:14:33,320 like bad quality audio. So the models weren't as good. 243 00:14:33,520 --> 00:14:36,600 So we had to have like layers to kind of like improve the 244 00:14:36,600 --> 00:14:41,960 transcription and and then kind of like it was very hard to 245 00:14:42,080 --> 00:14:47,560 create the set, like figuring out how to create good feedback, 246 00:14:47,560 --> 00:14:49,800 right? Like something that's kind of 247 00:14:49,800 --> 00:14:52,480 like constructive and really helps the teachers. 248 00:14:52,720 --> 00:14:55,800 So, yeah, so we can like spend a lot of time on creating this 249 00:14:55,800 --> 00:14:57,800 product, but then it just like didn't work out. 250 00:14:57,800 --> 00:14:59,240 So we have to do something different. 251 00:15:00,240 --> 00:15:03,920 The fun, the fun of pivoting to find product market fit right. 252 00:15:05,600 --> 00:15:08,200 I wonder if any companies nail it at the start because it's 253 00:15:08,200 --> 00:15:11,720 always funny like the the hear what the MVP product is versus 254 00:15:11,720 --> 00:15:16,200 what you have to do to actually make it whatever market ready. 255 00:15:17,440 --> 00:15:19,040 Yeah. I think like if you're building 256 00:15:19,040 --> 00:15:23,840 a SAS product and you can already have a customer from the 257 00:15:23,840 --> 00:15:27,040 beginning, right, like you can kind of like have somebody with 258 00:15:27,040 --> 00:15:28,960 a need. And I think that's the easiest 259 00:15:29,280 --> 00:15:32,720 way to find product market fit is where you actually start 260 00:15:32,720 --> 00:15:36,000 building for like a problem. Let that not, not just like you 261 00:15:36,000 --> 00:15:38,680 know exists, but you have like the person like that that needs 262 00:15:38,680 --> 00:15:40,960 this, you know, like the the company that needs it and you 263 00:15:40,960 --> 00:15:42,360 start building for them directly. 264 00:15:42,360 --> 00:15:43,640 Yeah. Yeah. 265 00:15:44,160 --> 00:15:46,920 So you're a solo developer there? 266 00:15:46,920 --> 00:15:49,400 Or not anymore, but you started off that way. 267 00:15:50,600 --> 00:15:52,360 Yeah. Well, yeah. 268 00:15:52,400 --> 00:15:56,120 And then so I'm I'm going to go down the like, what is your full 269 00:15:56,120 --> 00:15:58,400 stack? Like you have Elixir on the back 270 00:15:58,400 --> 00:16:01,000 end. You're using Ash I think, right? 271 00:16:02,120 --> 00:16:03,800 We actually don't in teacher space. 272 00:16:03,800 --> 00:16:09,120 So I'm using yeah, so in teacher space it's kind of like a a 273 00:16:09,680 --> 00:16:15,280 classical, like a Phoenix with live view and yeah, like a ecto 274 00:16:15,320 --> 00:16:18,080 and kind of like, I would say like the traditional Phoenix 275 00:16:18,080 --> 00:16:22,320 app. And yeah, where I'm using Ash is 276 00:16:23,720 --> 00:16:27,160 in in the startup that I'm working on. 277 00:16:27,720 --> 00:16:28,440 Yeah. OK. 278 00:16:28,520 --> 00:16:30,560 And can we talk about that one? What? 279 00:16:30,560 --> 00:16:32,320 What is that start up and what are you doing? 280 00:16:33,240 --> 00:16:35,400 Yeah. So that's like all communities. 281 00:16:36,000 --> 00:16:37,960 It's basically, how would I describe it? 282 00:16:38,320 --> 00:16:47,920 It's, it's a platform for, it's like a local first platform for 283 00:16:48,040 --> 00:16:53,160 like social network, right? Where it focuses on, on making 284 00:16:53,160 --> 00:16:56,680 the connection to a community. So through groups, like how one 285 00:16:56,680 --> 00:17:01,520 could think about it is like if you put together like Google 286 00:17:01,720 --> 00:17:05,960 Facebook groups, plus Google Maps and Craigslist, kind of 287 00:17:05,960 --> 00:17:09,839 like all together. That's kind of like what what 288 00:17:09,839 --> 00:17:13,200 you would get is so, so basically how it works is that 289 00:17:13,599 --> 00:17:19,000 you can like register and you can see things that are 290 00:17:19,000 --> 00:17:23,760 happening around you and you can join kind of like local groups, 291 00:17:23,960 --> 00:17:26,960 see kind of like events, places that are happening around you. 292 00:17:27,200 --> 00:17:32,360 People can like post items. So for example, they might want 293 00:17:32,360 --> 00:17:35,200 to share some some things with others so they can kind of like 294 00:17:35,880 --> 00:17:40,640 put it there. And yeah, so it's a social 295 00:17:40,640 --> 00:17:45,800 network that is like focused on kind of like that what's 296 00:17:45,800 --> 00:17:48,720 happening around you. And the main idea is that 297 00:17:48,720 --> 00:17:52,280 everything revolves around kind of like the communities, which 298 00:17:52,280 --> 00:17:55,160 are basically groups of people. OK. 299 00:17:55,440 --> 00:18:00,400 Are the communities like based off regions or more like topics 300 00:18:00,400 --> 00:18:04,440 and interests? So anybody can create like a 301 00:18:05,120 --> 00:18:08,560 community that like what they would like to right now. 302 00:18:09,480 --> 00:18:11,720 We we have like a very few communities at the moment. 303 00:18:11,720 --> 00:18:14,800 So basically it's been really, really hard to kind of like get 304 00:18:14,800 --> 00:18:18,800 people on the platform. And I think that's like that's 305 00:18:18,800 --> 00:18:22,240 the biggest challenge, right? Like it's especially on on, I 306 00:18:22,240 --> 00:18:26,960 think on consumer facing apps, like having the thought that 307 00:18:28,120 --> 00:18:30,960 building something is usually the easiest part. 308 00:18:31,320 --> 00:18:32,920 And especially that's what. I said. 309 00:18:35,800 --> 00:18:38,360 So, so we have very few communities and, and most of 310 00:18:38,360 --> 00:18:40,840 them kind of like are semi active, right? 311 00:18:40,840 --> 00:18:43,800 Like people will log in maybe like once every couple of weeks 312 00:18:43,800 --> 00:18:44,960 to kind of like see what's happening. 313 00:18:44,960 --> 00:18:47,800 But it's not like like activity you can rely on. 314 00:18:48,240 --> 00:18:53,000 And yeah, so we have like like one that's like around the 315 00:18:53,040 --> 00:18:58,840 philosophy there was like a now it's like not active because 316 00:18:58,840 --> 00:19:01,760 they only do it like part time of the year, but they do like a 317 00:19:01,760 --> 00:19:08,440 Film Festival and and yeah, so. OK, what, what is like, what is 318 00:19:08,440 --> 00:19:14,000 your ideal like use case? So like you have one that's a 319 00:19:14,000 --> 00:19:16,920 Film Festival, but they're probably only active during the 320 00:19:16,920 --> 00:19:19,600 festival and maybe a couple weeks before. 321 00:19:19,960 --> 00:19:22,240 And then like philosophy, obviously, like anyone 322 00:19:22,240 --> 00:19:24,840 interested would probably want to go there and chat about 323 00:19:24,840 --> 00:19:28,000 philosophy, right? Is your goal to kind of support 324 00:19:28,000 --> 00:19:31,480 both or 'cause I feel like they're, they're kind of 325 00:19:31,480 --> 00:19:34,400 different use cases, right? Yeah. 326 00:19:34,400 --> 00:19:38,880 And that's also like a challenge to like try to figure out what's 327 00:19:39,240 --> 00:19:45,200 like to focus on like what I would think. 328 00:19:45,880 --> 00:19:48,440 Like I'm still like figuring things out as I go, to be 329 00:19:48,440 --> 00:19:53,840 honest. And but like 1 like how I'm 330 00:19:53,840 --> 00:19:56,480 thinking about it is like, if you want to do an in person 331 00:19:56,480 --> 00:20:00,000 community that you like meet with people and you kind of like 332 00:20:00,240 --> 00:20:05,840 want to be able to track physical locations, track like 333 00:20:06,240 --> 00:20:10,760 maybe items that the community like has access to, also like 334 00:20:10,760 --> 00:20:13,880 contacts. So basically one challenge that 335 00:20:13,880 --> 00:20:16,480 a lot of communities have is that they have like a concept of 336 00:20:16,480 --> 00:20:18,520 an actor and basically an actor is. 337 00:20:19,320 --> 00:20:22,360 A person or an organization that can help them with some goals 338 00:20:22,680 --> 00:20:27,480 and yeah, they have like different ways of like keeping 339 00:20:27,480 --> 00:20:33,840 all that information and yeah, I think it this can like this can 340 00:20:33,840 --> 00:20:37,200 help organize it in some way, right. 341 00:20:37,800 --> 00:20:40,200 Another thing is that, for example, like taking decisions. 342 00:20:41,240 --> 00:20:44,800 So a lot of yeah, I can have like a like they want to take 343 00:20:44,800 --> 00:20:47,120 decisions and like have everybody contribute. 344 00:20:47,120 --> 00:20:51,400 So you either have to like meet in person or of course you can 345 00:20:51,400 --> 00:20:53,560 have like a signal chat or something. 346 00:20:53,560 --> 00:20:58,440 But yeah, it's there's kind of like a, a few problems from like 347 00:20:58,440 --> 00:21:01,160 talking with people that, that this can help. 348 00:21:01,160 --> 00:21:05,000 But my main focus is like to be for in person communities 349 00:21:05,400 --> 00:21:09,360 because like I think there's already things for online like 350 00:21:09,360 --> 00:21:11,360 platforms that are for online communities. 351 00:21:11,840 --> 00:21:15,880 And I haven't found like any like good platform that's kind 352 00:21:15,880 --> 00:21:18,920 of truly help solve the the problems of of in person 353 00:21:18,920 --> 00:21:20,920 communities. Yeah, fair. 354 00:21:21,240 --> 00:21:23,360 Do you? I don't know if you know what 355 00:21:23,360 --> 00:21:24,920 meet up is. I don't know if that's a whole. 356 00:21:26,040 --> 00:21:28,280 World. Thing or just AUS based thing? 357 00:21:29,080 --> 00:21:31,200 No, no, like meet up is really big, yeah. 358 00:21:31,440 --> 00:21:35,920 And yeah. So it's like, yeah, I would say 359 00:21:35,920 --> 00:21:38,320 it's actually the closest to to what I'm trying to do. 360 00:21:38,560 --> 00:21:43,320 But the issues there is that like it's not really helping 361 00:21:43,640 --> 00:21:45,720 organize the community. It's more kind of like you have 362 00:21:46,120 --> 00:21:50,480 an A host who host events and and then you have like the 363 00:21:50,480 --> 00:21:53,760 members and this sense of kind of like doing things together. 364 00:21:53,760 --> 00:21:56,120 It's it's not really there. So kind of like that's that's 365 00:21:56,160 --> 00:21:59,000 part of the, I would say like the differences. 366 00:21:59,000 --> 00:22:00,960 Yeah. I feel like meet up destroyed 367 00:22:00,960 --> 00:22:04,240 their like monetization strategy during COVID. 368 00:22:04,400 --> 00:22:07,520 Like they, they like spiked their rates. 369 00:22:07,520 --> 00:22:12,720 They like tried to force every host now to like pay money and 370 00:22:12,760 --> 00:22:15,080 like it used to be free. I don't know what they were 371 00:22:15,080 --> 00:22:16,480 doing. It's probably just venture 372 00:22:16,480 --> 00:22:20,120 backed or VC backed, right? But I don't know, I feel like 373 00:22:20,240 --> 00:22:24,360 they kind of went downhill. Yeah, they changed ownership 374 00:22:24,360 --> 00:22:27,800 actually a few times. So I think that's about right. 375 00:22:27,800 --> 00:22:31,480 I think they they when doing as well financially as they wanted 376 00:22:31,480 --> 00:22:34,440 to and kind of like it could sold the whole company a couple 377 00:22:34,440 --> 00:22:37,960 of times and you know, like new pressures and and all that, you 378 00:22:37,960 --> 00:22:40,040 know. That's the problem with too many 379 00:22:40,040 --> 00:22:43,560 board members, right or investors. 380 00:22:44,400 --> 00:22:50,280 I'm curious, what do you have like a monetization strategy or 381 00:22:50,280 --> 00:22:53,120 what is your goal? Are you going to, are you making 382 00:22:53,120 --> 00:22:55,720 money or not yet? Is it going to be free for a 383 00:22:55,720 --> 00:23:00,040 while? So I'm not making money yet from 384 00:23:00,040 --> 00:23:03,880 from it. The the idea is to basically 385 00:23:03,880 --> 00:23:06,600 have memberships, but there's overall like many ways to 386 00:23:06,600 --> 00:23:11,200 monetize it actually. So what I want to focus first is 387 00:23:11,200 --> 00:23:15,880 to do like government contracts. So, so there is a business model 388 00:23:15,880 --> 00:23:19,720 that you can like. So basically the idea is to also 389 00:23:19,720 --> 00:23:23,080 help people become connected with their neighbors, with their 390 00:23:23,080 --> 00:23:25,560 locality, right? Like support local businesses. 391 00:23:26,320 --> 00:23:32,640 So the idea is to try to like, like, you know, basically work 392 00:23:32,640 --> 00:23:38,960 with the government to find regions and try to basically do 393 00:23:38,960 --> 00:23:43,440 pro social work. So basically help people engage 394 00:23:43,440 --> 00:23:46,680 socially. Yeah. 395 00:23:46,680 --> 00:23:49,080 Through through virus activities, organizing kind of 396 00:23:49,080 --> 00:23:50,840 events and all. Yeah. 397 00:23:51,120 --> 00:23:55,640 So this is something that seems to be working in Scandinavian 398 00:23:55,640 --> 00:23:57,040 countries. There's like a few companies 399 00:23:57,040 --> 00:24:00,600 that do that, but I haven't seen it happening here. 400 00:24:00,600 --> 00:24:03,160 So I want to see if that's possible. 401 00:24:03,480 --> 00:24:08,040 Of course, lately it hasn't been like easy because you know, like 402 00:24:08,040 --> 00:24:10,320 a little bit of right leaning governments and everything. 403 00:24:10,320 --> 00:24:15,840 So not the best time right now. And but I've talked with some 404 00:24:15,840 --> 00:24:20,200 some people that are kind of like in a position to like help 405 00:24:20,200 --> 00:24:22,520 me with that and they think it's a good idea. 406 00:24:22,520 --> 00:24:24,880 But they told me there's that right now it's not the best time 407 00:24:24,880 --> 00:24:27,200 to kind of like for projects like this. 408 00:24:28,120 --> 00:24:33,160 So yeah, trying to kind of like, yeah, like with, I want to see 409 00:24:33,160 --> 00:24:35,600 with time how it evolves because like, yeah, time change. 410 00:24:35,600 --> 00:24:40,520 And also it's a lot about trust when you work with the 411 00:24:40,520 --> 00:24:42,160 government, right? So they want to see something 412 00:24:42,160 --> 00:24:43,960 happening and it having already impact. 413 00:24:44,200 --> 00:24:46,880 And once you can like show that and they're much more like 414 00:24:46,880 --> 00:24:52,280 willing to work with you. It sounds like a very painful 415 00:24:52,280 --> 00:24:56,520 experience working on contracts with the government well. 416 00:24:57,640 --> 00:24:59,520 It doesn't have to be actually like, it depends. 417 00:24:59,520 --> 00:25:02,040 Yeah, of course, it's a developed cult, but it can also 418 00:25:02,040 --> 00:25:06,120 be kind of like provide stability because like, yeah, 419 00:25:06,120 --> 00:25:08,760 you can, you don't have like as much pressure. 420 00:25:09,000 --> 00:25:15,160 And it sounds like what attracts me there is that of course, if 421 00:25:15,160 --> 00:25:17,080 the government is not corrupt and you have like a good 422 00:25:17,080 --> 00:25:19,720 government, which I think you're in Czech Republic, that the 423 00:25:19,720 --> 00:25:24,960 government's pretty good, then you can do like pro social work, 424 00:25:24,960 --> 00:25:27,080 you can do like something that's like really beneficial for the 425 00:25:27,080 --> 00:25:29,440 public. And I don't know, I just like, 426 00:25:29,720 --> 00:25:32,440 feel really good about that. That's kind of like what I want 427 00:25:32,440 --> 00:25:36,640 to do is build things that like, help society in a in a 428 00:25:36,640 --> 00:25:39,000 meaningful way. Yeah, I like that. 429 00:25:39,040 --> 00:25:40,920 That's always, it's always a good goal. 430 00:25:40,920 --> 00:25:45,120 Right now this this project is using Ash right? 431 00:25:45,640 --> 00:25:47,520 Yeah. All right, cool. 432 00:25:47,960 --> 00:25:52,040 Now what, what are like some cool features that you've been 433 00:25:52,040 --> 00:25:55,160 working on? Because I know you're, you're 434 00:25:55,680 --> 00:26:00,920 like, you're a big advocate of like Ash and agentic coding and 435 00:26:00,920 --> 00:26:03,560 like it all kind of works well together. 436 00:26:03,560 --> 00:26:07,080 And I, I know you contribute to a lot of open source projects. 437 00:26:07,440 --> 00:26:14,600 And what has your experience been using LLMS and using the 438 00:26:14,600 --> 00:26:19,600 Ash framework? So I think Ash and LMS work 439 00:26:19,600 --> 00:26:22,080 really, really well. I've been actually very 440 00:26:22,080 --> 00:26:24,880 surprised like in the beginning I thought because you know, like 441 00:26:25,120 --> 00:26:29,360 Ash is like a niche like project in a niche language, right? 442 00:26:29,680 --> 00:26:35,080 So I thought kind of it would be very hard to for LMS to kind of 443 00:26:35,080 --> 00:26:39,320 like learn it and work with it. But it turns out it works 444 00:26:39,320 --> 00:26:45,760 really, really well. And I think the main reason for 445 00:26:45,760 --> 00:26:52,560 that is so like if you think about it, LMS are recognition 446 00:26:52,640 --> 00:26:57,560 like recognized patterns, right? And then because Ash is kind of 447 00:26:57,560 --> 00:26:59,200 like unique in a way it's written. 448 00:26:59,840 --> 00:27:03,320 I think even if there is just a small amount of like data it's 449 00:27:03,320 --> 00:27:06,200 been trained on it like knows that data very well, if that 450 00:27:06,200 --> 00:27:09,640 makes sense. So you can think about it like, 451 00:27:09,640 --> 00:27:11,840 let's say you have like JavaScript and there's like a 452 00:27:11,840 --> 00:27:13,400 lot of JavaScript, different versions. 453 00:27:13,400 --> 00:27:15,520 You have like Coffeescript pure script. 454 00:27:15,800 --> 00:27:19,320 So even if there's like a lot of data, in theory, it could get 455 00:27:19,320 --> 00:27:21,760 kind of like mixed up because it's like things that look 456 00:27:21,760 --> 00:27:24,920 similar, but like they behave differently, right? 457 00:27:25,640 --> 00:27:29,680 And I think Ash being like a very unique syntax, even with 458 00:27:29,680 --> 00:27:32,960 just a small amount of data, like the LMS, they can kind of 459 00:27:32,960 --> 00:27:36,040 like learn that and they get pointed to the right direction 460 00:27:36,040 --> 00:27:38,600 and they know, OK, this is Ash and I, I, you know, it's kind of 461 00:27:38,600 --> 00:27:41,160 like they recognize it and they know how to work with it. 462 00:27:41,160 --> 00:27:45,120 So I think that's like one of the reasons why they might be 463 00:27:45,120 --> 00:27:47,600 good. Also, like, yeah, so we talked 464 00:27:47,600 --> 00:27:50,520 about it a little bit before is that it's, it's declarative and 465 00:27:50,520 --> 00:27:53,280 you can like introspect how things work, right? 466 00:27:53,280 --> 00:27:56,360 So like if you're using Tidewave, which I really 467 00:27:56,360 --> 00:28:00,480 recommend using Tidewave with any, any Elixir project, then 468 00:28:00,680 --> 00:28:05,560 basically it can look up how things are and it can get so 469 00:28:05,560 --> 00:28:09,000 much information, right. So it's in LMS, I feel it's all 470 00:28:09,000 --> 00:28:12,680 about getting the right context. And when they can get the right 471 00:28:12,680 --> 00:28:15,960 context of how things are structured, it's much more easy 472 00:28:15,960 --> 00:28:20,160 for them to to work with things. Yeah, regarding kind of like 473 00:28:20,160 --> 00:28:23,800 interesting, like features. What I was like really 474 00:28:23,800 --> 00:28:27,600 surprised. So basically in Ash you can or 475 00:28:27,840 --> 00:28:30,840 there is like a project called like Spark that you can declare 476 00:28:30,840 --> 00:28:34,880 your own DSLS. And I kind of like didn't feel 477 00:28:34,880 --> 00:28:37,120 very comfortable kind of like using it, but there were some 478 00:28:37,120 --> 00:28:41,720 use cases I wanted to write. So more specifically, like we 479 00:28:41,720 --> 00:28:46,480 have this notion of and of like a community resource. 480 00:28:46,720 --> 00:28:50,920 And, and what that is, is basically like a different type 481 00:28:50,920 --> 00:28:53,120 of data that like a community can have. 482 00:28:53,120 --> 00:29:00,120 So can imagine like a, a post or like an item, an event, a place. 483 00:29:00,240 --> 00:29:03,960 So like there's like a like right now we have like 9 484 00:29:04,400 --> 00:29:09,080 different types of like resources and I wanted kind of 485 00:29:09,080 --> 00:29:12,760 like something to combine to like make it easier to create 486 00:29:12,760 --> 00:29:14,800 more resources. But also like, do you have some 487 00:29:14,800 --> 00:29:17,680 shared code, right? Some some shared like 488 00:29:17,680 --> 00:29:21,560 foundations for, for all of it. And that is also configurable. 489 00:29:21,560 --> 00:29:24,520 So like each of them can have like different settings, right? 490 00:29:25,960 --> 00:29:29,160 So it was something that I knew I wanted to build like early on, 491 00:29:29,520 --> 00:29:34,960 but I didn't like feel that comfortable like starting Spark 492 00:29:35,160 --> 00:29:37,800 kind of like driving my own. I don't know, I just like felt a 493 00:29:37,800 --> 00:29:40,800 little bit intimidated, I think, to be honest. 494 00:29:42,640 --> 00:29:47,840 But I decided to actually just try to do something simple with 495 00:29:47,840 --> 00:29:54,000 Spark through the 11:00. And basically it was to like try 496 00:29:54,000 --> 00:29:57,280 sluggable, kind of like functionality where it can say 497 00:29:58,080 --> 00:30:01,840 you sluggable and and then it kind of like makes things 498 00:30:01,840 --> 00:30:05,040 sluggable. And like I say used because, 499 00:30:06,160 --> 00:30:07,400 yeah, we're getting used to the use. 500 00:30:07,520 --> 00:30:10,600 But basically you just say slackable in the in the 501 00:30:10,600 --> 00:30:15,920 resource. And yeah, so I tried to kind of 502 00:30:15,920 --> 00:30:20,280 like prompt that I explained to the what I wanted and I was like 503 00:30:20,400 --> 00:30:24,600 really surprised that it made the Spark DSL. 504 00:30:24,840 --> 00:30:27,160 It kind of like implemented the feature and it just like all 505 00:30:27,160 --> 00:30:32,200 works like right off the bat. So that kind of was like my, my 506 00:30:32,200 --> 00:30:37,520 start and think that OK, that's like like it can do even like 507 00:30:37,800 --> 00:30:40,920 like the ones can do kind of like even things that you 508 00:30:40,920 --> 00:30:44,840 wouldn't expect them to. And maybe it was like six to 509 00:30:44,840 --> 00:30:48,080 seven months ago. It was like maybe longer right 510 00:30:48,080 --> 00:30:49,000 now. Yeah. 511 00:30:49,160 --> 00:30:54,240 So yeah, it's a while ago. And yeah, what was really 512 00:30:54,240 --> 00:30:56,640 interesting is that then I kind of just like went through what 513 00:30:56,640 --> 00:31:01,400 it did and I understood kind of like how like the Spark DSL 514 00:31:01,400 --> 00:31:04,240 works, how kind of like you can configure it. 515 00:31:04,520 --> 00:31:08,080 And, and kind of like, it was an interesting experience for me 516 00:31:08,080 --> 00:31:15,360 because I learned that you, you basically, yeah, you can learn a 517 00:31:15,360 --> 00:31:17,640 lot as well through kind of like LMS. You don't have to just like 518 00:31:17,640 --> 00:31:19,760 generate code and kind of like, just like use it. 519 00:31:19,960 --> 00:31:23,160 You can then like create it and understand it. 520 00:31:23,160 --> 00:31:26,320 And it can like boost your, your, your learning curve, let's 521 00:31:26,320 --> 00:31:28,960 say. Yeah, I would. 522 00:31:30,440 --> 00:31:32,720 I got buzzers going off in my ear. 523 00:31:33,600 --> 00:31:38,560 I would say I like realized like the upbeat of ash being really 524 00:31:38,560 --> 00:31:43,280 usable by like Claude Coder. It was usage rules and I feel 525 00:31:43,280 --> 00:31:47,120 like ever since that came out, it just was like night and day 526 00:31:47,120 --> 00:31:50,080 difference too, 'cause I feel like once you like, like what 527 00:31:50,080 --> 00:31:52,480 you're saying, you have like such few examples, but there's 528 00:31:52,480 --> 00:31:56,600 really only like one way to do it and to approach problems. 529 00:31:56,600 --> 00:32:00,000 Like once you have your resources and actions and 530 00:32:01,800 --> 00:32:04,600 whatever the like authentications behaviors, 531 00:32:04,600 --> 00:32:08,280 whatever you're doing, they're all, it's all done the same way. 532 00:32:08,280 --> 00:32:11,600 So I feel like it's like such a repeatable pattern to keep just 533 00:32:12,040 --> 00:32:14,560 producing more resources or domains. 534 00:32:14,560 --> 00:32:19,480 And it's it's not I like so kill switch my little side project, 535 00:32:19,480 --> 00:32:23,160 my SAS project, which by the way, is very hard to get users 536 00:32:23,400 --> 00:32:26,240 on board. So I'm right there with you. 537 00:32:26,520 --> 00:32:32,160 But it's been a lot of fun. But yeah, Quad code's amazing 538 00:32:32,160 --> 00:32:34,680 with like helping me knockout features now. 539 00:32:35,760 --> 00:32:38,120 Yeah, Yeah. They've become really good. 540 00:32:38,960 --> 00:32:41,800 It's it's surprising, Yeah. I'm, I'm sometimes kind of like 541 00:32:41,800 --> 00:32:46,320 feel uncomfortable because I, I can feel that you're delegating 542 00:32:46,320 --> 00:32:50,440 too much and yeah, it's like, I'm just like afraid where this 543 00:32:50,440 --> 00:32:53,000 goes, like if you continue on that route. 544 00:32:54,040 --> 00:32:57,200 But the interesting thing is that you can start focusing on 545 00:32:57,240 --> 00:33:00,040 on other areas, right? Like you can start thinking 546 00:33:00,280 --> 00:33:04,560 bigger about the system. I don't think I would ever be 547 00:33:04,560 --> 00:33:07,080 able to achieve so much and kind of like start thinking about. 548 00:33:07,920 --> 00:33:12,240 I think that I don't, right. Like when I started working on 549 00:33:12,240 --> 00:33:14,560 it, I was like very much, OK, I'm building like this product 550 00:33:14,880 --> 00:33:19,360 and all my focus was on that. But now I'm starting to think 551 00:33:19,360 --> 00:33:22,160 about like how to market it, like how to build a system of 552 00:33:22,160 --> 00:33:25,200 kind of like creating leads and, and all these like things that 553 00:33:26,200 --> 00:33:28,680 you just wouldn't have the bandwidth as a single developer. 554 00:33:28,680 --> 00:33:32,880 So I think it's, you're gonna use this kind of like like you 555 00:33:32,880 --> 00:33:35,200 delegate something like you start thinking more of as a 556 00:33:35,200 --> 00:33:37,760 manager, I would say, right? Like where a manager doesn't 557 00:33:37,760 --> 00:33:40,920 think about the, the degree of how something works, but like 558 00:33:40,920 --> 00:33:44,240 how to make a project successful. 559 00:33:44,600 --> 00:33:47,880 And I think that's kind of like what like the skills that enable 560 00:33:47,880 --> 00:33:50,840 you to train like when you basically are work with the OMS, 561 00:33:51,640 --> 00:33:54,000 you train your manager skills, you know? 562 00:33:55,440 --> 00:33:58,080 Yeah, like I just had Luca on last week. 563 00:33:58,080 --> 00:34:02,880 He he's ACTO or Co founder of a fintech and I'll he said it 564 00:34:02,880 --> 00:34:05,320 perfectly. He's like, I never thought like 565 00:34:05,360 --> 00:34:08,560 I'm a software engineer, right? I was never I never thought 566 00:34:08,560 --> 00:34:11,440 coding was my job, like writing code. 567 00:34:11,719 --> 00:34:16,120 He said my job was creating solutions and in it like, you 568 00:34:16,120 --> 00:34:21,159 know, producing a product that, you know, solves a problem, 569 00:34:21,159 --> 00:34:23,360 right? Not, not how you get there. 570 00:34:23,600 --> 00:34:27,400 And I kind of like, I like that because we're just, we still 571 00:34:27,400 --> 00:34:29,560 have to solve the problem, know how to solve it. 572 00:34:29,880 --> 00:34:33,840 And now we just get a lot of assistance and a lot of the 573 00:34:33,840 --> 00:34:37,120 manual labor that we used to do. Yeah. 574 00:34:38,159 --> 00:34:40,520 Because like, like you're saying, it gives us more time. 575 00:34:40,520 --> 00:34:43,719 Like I, I hate it, but I need to, but you have to spend more 576 00:34:43,719 --> 00:34:45,639 time marketing and selling your product. 577 00:34:45,960 --> 00:34:50,320 And now you can spend like it's a lot easier to do like 5050 578 00:34:50,320 --> 00:34:55,320 split with your time when you can kind of just like, like if I 579 00:34:55,320 --> 00:34:58,960 organize a feature, I keep, I, I try to keep my features like, 580 00:34:58,960 --> 00:35:03,320 you know, at least like consumable, like not like just 581 00:35:03,320 --> 00:35:05,320 go build this whole platform, right? 582 00:35:05,680 --> 00:35:07,880 But like when I create a feature, I try to keep it like 583 00:35:07,880 --> 00:35:13,280 very small and to the point. And I plan it out as detailed as 584 00:35:13,280 --> 00:35:17,440 possible, But then I can just let Claude go do its thing. 585 00:35:17,440 --> 00:35:19,920 Let Tidewave. I've been using Tidewave quite a 586 00:35:19,920 --> 00:35:22,320 bit. It's a lot of fun, especially 587 00:35:22,320 --> 00:35:25,280 with like UI, just like the ability to click on components 588 00:35:25,280 --> 00:35:27,680 and be like, why doesn't this work like it should? 589 00:35:28,120 --> 00:35:30,880 And it just solves problems. But like, you can kind of let 590 00:35:30,880 --> 00:35:32,680 those things work in the background. 591 00:35:32,720 --> 00:35:37,960 And I can be wasting my time trying to post on Reddit or, you 592 00:35:37,960 --> 00:35:42,640 know, market my product, write blog posts, whatever it may be. 593 00:35:44,480 --> 00:35:47,160 I'm not a huge fan. I'd rather be building features, 594 00:35:47,160 --> 00:35:49,720 but I also want to be a product owner. 595 00:35:49,720 --> 00:35:52,760 And that's, that's kind of the double edged sword, right? 596 00:35:52,760 --> 00:35:55,000 You have to do a lot of things you don't want to do. 597 00:35:56,120 --> 00:36:02,000 Yeah, I think what's important is like not to stop doing good 598 00:36:02,000 --> 00:36:03,560 plans. And I don't. 599 00:36:03,680 --> 00:36:06,520 And like saying the word plan now is kind of like you. 600 00:36:06,520 --> 00:36:08,920 It makes you think about like the the plan you give to the 601 00:36:08,920 --> 00:36:11,000 Kojic agent. But I mean more about 602 00:36:11,000 --> 00:36:13,680 understanding the system and kind of like constructing the 603 00:36:13,680 --> 00:36:16,400 system and having a good idea how the system is like wired 604 00:36:16,400 --> 00:36:19,400 together. Because like I, I feel like when 605 00:36:19,400 --> 00:36:24,960 I've, when it didn't work well, it's kind of like like when I 606 00:36:24,960 --> 00:36:27,800 tried to use like LMS and it didn't work well was always when 607 00:36:27,800 --> 00:36:30,680 I kind of like delegated too much of the plan, right? 608 00:36:30,680 --> 00:36:34,160 Like I had like some like half baked idea and kind of like was 609 00:36:34,160 --> 00:36:35,920 OK, like, let's see what it does with that. 610 00:36:36,080 --> 00:36:38,600 And then I run with it for a little more than I should. 611 00:36:38,760 --> 00:36:41,840 So I think that's great to kind of like get an idea of what you 612 00:36:41,840 --> 00:36:44,120 want to build, but don't rely on that. 613 00:36:44,120 --> 00:36:46,640 So like I think it's very important to have a good idea of 614 00:36:46,640 --> 00:36:50,320 what you're building and yeah, and design the system. 615 00:36:51,280 --> 00:36:53,280 Yeah, I'm not not delegate the design of the system. 616 00:36:53,520 --> 00:36:56,320 Yeah, 100%. I'm going to just pause this for 617 00:36:56,320 --> 00:36:58,560 about 30 seconds and we'll be right back as. 618 00:36:58,960 --> 00:37:02,480 Hey, I'm Kimberly with Elixir mentor Jacob recently added 619 00:37:02,480 --> 00:37:05,800 recruitment services to help companies hire pre bedded Elixir 620 00:37:05,800 --> 00:37:08,800 devs. I help with company outreach, so 621 00:37:08,800 --> 00:37:12,080 if you're hiring, I may be reaching out, or you can connect 622 00:37:12,080 --> 00:37:14,280 with us anytime at elixirmentor.com. 623 00:37:15,600 --> 00:37:18,720 Coding Elixir with AI or building AI into your elixir 624 00:37:18,720 --> 00:37:21,440 apps. The Elixir AI collective discord 625 00:37:21,440 --> 00:37:24,320 is your community for both link in the description below. 626 00:37:25,840 --> 00:37:28,280 And we're back. All right, we got some we got 627 00:37:28,280 --> 00:37:31,880 some good chat in the in the comments here. 628 00:37:32,960 --> 00:37:36,440 Dharma, sorry, I probably butchered that, but he he just 629 00:37:36,440 --> 00:37:39,720 said that it's personality driven development, like founder 630 00:37:39,720 --> 00:37:44,600 mindset. The ability, I think like AI is 631 00:37:44,600 --> 00:37:48,440 helping developers that are more entrepreneurial kind of own 632 00:37:48,440 --> 00:37:51,960 their ideas, right? Like kill switch for me was an 633 00:37:51,960 --> 00:37:56,440 idea I wanted to build for. I don't know, I want maybe 634 00:37:56,440 --> 00:37:59,600 almost like 8 years. Like I came up with it eight 635 00:37:59,600 --> 00:38:01,280 years ago. I wanted to build it, but I 636 00:38:01,280 --> 00:38:03,840 always put it off because I just I didn't have time. 637 00:38:04,160 --> 00:38:08,080 I was like what one like I, I'm not a front end developer. 638 00:38:08,080 --> 00:38:09,880 I actually hate front end development. 639 00:38:11,200 --> 00:38:14,640 It's not great. I hack my way through it, but 640 00:38:14,920 --> 00:38:19,960 like Claude gives me the ability to now deal with tailwind and 641 00:38:19,960 --> 00:38:23,800 all the things that I hate. And so now I can actually just 642 00:38:24,040 --> 00:38:29,480 make kill switch in my free time and like make relatively nice UI 643 00:38:29,480 --> 00:38:32,960 designs because I can be like, hey, Claude, look at this 644 00:38:32,960 --> 00:38:36,400 website, look at this website. I want kind of like a mixture of 645 00:38:36,400 --> 00:38:40,360 these components. Help me, you know, make my, my 646 00:38:40,360 --> 00:38:42,560 branding or my, my color palette. 647 00:38:42,560 --> 00:38:45,680 And it's been awesome. So now I'm a full stack 648 00:38:45,680 --> 00:38:49,680 developer, even though I hated front end development and I can 649 00:38:49,680 --> 00:38:53,680 build a product and like, I think like the cool thing is 650 00:38:53,680 --> 00:38:57,360 like I'm a back end dev and so like security has always been a 651 00:38:57,360 --> 00:38:59,600 priority. So I'm less concerned. 652 00:38:59,880 --> 00:39:01,360 Like everyone's always concerned. 653 00:39:01,360 --> 00:39:03,880 Like if you're vibe coding too much, are you going to have like 654 00:39:03,880 --> 00:39:07,960 huge security vulnerabilities? And I would say probably yes, 655 00:39:08,280 --> 00:39:11,600 but at least like I still review all the code going in. 656 00:39:11,600 --> 00:39:15,520 I understand my DevOps to a degree and I know that my back 657 00:39:15,520 --> 00:39:19,280 end work is solid. And so that helps me kind of 658 00:39:19,280 --> 00:39:23,840 sleep at night. And I think I think like we're 659 00:39:23,840 --> 00:39:27,280 going to see a lot more. I don't know, like I see all 660 00:39:27,280 --> 00:39:29,920 these doom and gloom articles about like the end of SAS, 661 00:39:29,920 --> 00:39:31,280 right? Because now everyone can just 662 00:39:31,280 --> 00:39:34,520 kind of make their own products for to fill their needs. 663 00:39:34,920 --> 00:39:38,280 But I still you still are going to have like well established 664 00:39:38,920 --> 00:39:42,080 solutions that people will be willing to pay for because it's 665 00:39:42,080 --> 00:39:46,560 not always that easy just to vibe code your own solution. 666 00:39:47,160 --> 00:39:48,480 So. Definitely, yeah. 667 00:39:49,040 --> 00:39:51,840 And it's a lot of the time it's also like the ecosystem, right? 668 00:39:51,840 --> 00:39:55,160 It's not just like having a single particle, but it's how it 669 00:39:55,160 --> 00:39:58,840 integrates with everything else. And yeah, that can be like a 670 00:39:58,840 --> 00:40:03,360 harder problem, right? Like, let's say, right, like. 671 00:40:04,320 --> 00:40:07,280 Anybody could like build like a a huge platform, but like making 672 00:40:07,280 --> 00:40:10,200 it work at scale and and kind of like having the the people 673 00:40:10,200 --> 00:40:11,560 there. That's the hard part. 674 00:40:11,560 --> 00:40:16,880 So I think yeah, consumer products are definitely kind of 675 00:40:16,880 --> 00:40:18,880 like, it's still like the mold there is kind of like having 676 00:40:18,880 --> 00:40:21,480 this like network effect and having kind of like the people 677 00:40:21,720 --> 00:40:24,160 there. And even for, for kind of like 678 00:40:24,160 --> 00:40:26,400 salsa, it's kind of like business, like pure business to 679 00:40:26,400 --> 00:40:29,480 business. I think like a lot of the times 680 00:40:29,480 --> 00:40:32,000 we think we understand a problem, but we don't really. 681 00:40:32,000 --> 00:40:34,800 And it's only when you get into that, the problem that you kind 682 00:40:34,800 --> 00:40:39,280 of like see all the nuances. And yeah, I think if you're 683 00:40:39,280 --> 00:40:42,000 naive to think that kind of like you already understand it fully, 684 00:40:42,000 --> 00:40:43,560 then you're probably wrong, right? 685 00:40:43,560 --> 00:40:45,960 Like, yeah. And if you do understand it 686 00:40:45,960 --> 00:40:49,680 fully, you probably see that, you know, like all these years 687 00:40:49,720 --> 00:40:52,160 people like, you know, like these companies have put into 688 00:40:52,160 --> 00:40:55,400 building these products, you probably cannot like, you know, 689 00:40:55,400 --> 00:40:58,920 like compacted to kind of like a few weeks of bad development. 690 00:40:58,920 --> 00:41:01,880 It's just, it's just impossible. Yeah, yeah. 691 00:41:01,880 --> 00:41:06,400 And another good message in the chat is like SAS is no longer a 692 00:41:06,400 --> 00:41:07,720 Moat. And I agree. 693 00:41:07,720 --> 00:41:10,280 So I think we're going to just see a lot more products hit the 694 00:41:10,280 --> 00:41:13,400 market, but we're also going to see a lot more garbage. 695 00:41:13,400 --> 00:41:17,040 But I think that's kind of like that's kind of where competition 696 00:41:17,040 --> 00:41:19,680 will just like kind of even things out, right? 697 00:41:19,680 --> 00:41:23,080 Like the garbage is going to die and the good products are going 698 00:41:23,080 --> 00:41:26,360 to survive. Or maybe the better marketers 699 00:41:26,360 --> 00:41:28,760 will survive, who knows. Yeah. 700 00:41:28,880 --> 00:41:31,880 But I do think like software is going to change a lot because if 701 00:41:31,880 --> 00:41:34,440 you think about software, it has been about like managing 702 00:41:34,440 --> 00:41:38,200 information, right? And we got a new way to manage 703 00:41:38,200 --> 00:41:40,520 information now that is like very, very flexible. 704 00:41:40,840 --> 00:41:45,800 So, yeah, I think a lot of the like the pipelines we had, you 705 00:41:45,800 --> 00:41:49,400 had like a person that had like this flexibility and then you 706 00:41:49,400 --> 00:41:52,400 have like these rigid systems that like were very kind of like 707 00:41:52,400 --> 00:41:53,920 deterministics and they were doing something. 708 00:41:53,920 --> 00:41:56,600 And then the person would like make these choices, right? 709 00:41:56,600 --> 00:42:01,000 They would have like that. It's like like human 710 00:42:01,000 --> 00:42:03,240 understanding and kind of like that the ability to kind of like 711 00:42:03,240 --> 00:42:07,160 decide what's going to happen. And now we've built machines 712 00:42:07,160 --> 00:42:09,240 that can't make decisions, right? 713 00:42:09,520 --> 00:42:13,920 And that fundamentally changes how some software works, right? 714 00:42:13,920 --> 00:42:19,200 Like so I think we're gonna see a lot like I do agree that like 715 00:42:19,200 --> 00:42:22,680 a lot of the suspects are probably gonna you know, like 716 00:42:22,760 --> 00:42:24,880 die out or kind of like they're gonna like change. 717 00:42:25,280 --> 00:42:27,920 But it's not necessarily because somebody can like buy product, 718 00:42:28,640 --> 00:42:32,080 you know, like a vibe coded product, but it's more about 719 00:42:32,080 --> 00:42:34,840 like the needs are going to change, right, Because we're 720 00:42:34,840 --> 00:42:38,040 going to be able to, we are able already to manage information in 721 00:42:38,040 --> 00:42:42,560 a different way. And that's kind of like depend 722 00:42:42,560 --> 00:42:46,680 on this human making a decision and that's no longer be going to 723 00:42:46,680 --> 00:42:49,160 be like a requirement maybe in the future for some. 724 00:42:49,160 --> 00:42:52,680 Have you have you or I? I guess I'm curious on your 725 00:42:52,680 --> 00:42:54,600 take. I've heard a couple people talk 726 00:42:54,600 --> 00:42:58,920 about like stop building products, just build like APIs. 727 00:42:59,320 --> 00:43:04,960 What like and I I'm assuming like the angle on that is we're 728 00:43:04,960 --> 00:43:08,040 going to probably vibe code our own personal dashboards to 729 00:43:08,040 --> 00:43:10,320 consume AP is the way we want to consume them. 730 00:43:10,680 --> 00:43:12,560 Like what do you think? Do you think they're? 731 00:43:12,680 --> 00:43:16,760 Do you think that is going to be a direction things go like more 732 00:43:16,760 --> 00:43:21,160 integrating clawed into our lives to work with Salute like 733 00:43:21,160 --> 00:43:25,440 products we have out there? I think definitely there's going 734 00:43:25,440 --> 00:43:30,080 to be I can see kind of like this idea of like having an API 735 00:43:30,080 --> 00:43:34,640 having like a system that is easy to kind of like that the 736 00:43:34,640 --> 00:43:37,320 machine can integrate with directly and it doesn't need to 737 00:43:37,320 --> 00:43:40,280 to use the UI. So the fact that there's there's 738 00:43:40,280 --> 00:43:43,120 a merit there. Yeah. 739 00:43:44,680 --> 00:43:47,880 So I do think so. But I still think that there's 740 00:43:47,880 --> 00:43:51,480 going to be like some need for for some UI. 741 00:43:51,480 --> 00:43:58,400 But yeah, I definitely think UI is less important today than it 742 00:43:58,400 --> 00:44:01,400 used to be because there's going to be so much like you're going 743 00:44:01,400 --> 00:44:03,520 to have like your agent who's going to consume the API is 744 00:44:03,520 --> 00:44:06,040 going to show you the information exactly the way, 745 00:44:06,240 --> 00:44:09,840 well, hopefully exactly the way you need it, right? 746 00:44:09,840 --> 00:44:11,960 But it can be a double edged sword, right? 747 00:44:11,960 --> 00:44:17,920 Because a lot of these companies have put like time to understand 748 00:44:17,920 --> 00:44:23,240 how a person wants to view the information and you don't always 749 00:44:23,440 --> 00:44:25,320 know how you want to see the information, right? 750 00:44:25,320 --> 00:44:29,640 So like you have like a product like linear, right, That like 751 00:44:29,640 --> 00:44:35,760 managed to like there was been like like task management apps 752 00:44:35,760 --> 00:44:38,400 for like so long. And then lunar comes and it kind 753 00:44:38,400 --> 00:44:41,400 of like makes things a little bit faster, smoother, makes it 754 00:44:41,400 --> 00:44:44,280 easier to like navigate around and and all of a sudden kind of 755 00:44:44,280 --> 00:44:46,960 like everybody loves it, right, because it's makes the things 756 00:44:46,960 --> 00:44:50,880 faster for them. And it it kind of like they 757 00:44:50,880 --> 00:44:54,280 didn't even know how they wanted things to be represented, but 758 00:44:54,280 --> 00:44:56,520 the moment they use it, they understand, Oh, this is the way 759 00:44:56,760 --> 00:45:00,640 right. So I think like good design and 760 00:45:00,760 --> 00:45:05,120 kind of like thinking through something and understanding how, 761 00:45:05,440 --> 00:45:07,560 how somebody might want to see it. 762 00:45:09,600 --> 00:45:11,840 Like I, I still, it's still a hard problem. 763 00:45:11,840 --> 00:45:15,880 And I don't think LMS are at least for now, particularly good 764 00:45:15,880 --> 00:45:18,080 at it, right. Yeah. 765 00:45:18,080 --> 00:45:21,440 So, but they, they might be, I don't know. 766 00:45:21,440 --> 00:45:24,000 It's kind of like it depends. And I, I, I think there's 767 00:45:24,000 --> 00:45:26,360 definitely the definitely going to become better. 768 00:45:26,360 --> 00:45:31,360 And already you can see that like, like a common thing is to 769 00:45:31,760 --> 00:45:34,800 tell an LM to explain to you like a repo and, and make an 770 00:45:34,800 --> 00:45:37,920 HTML page out of it. And that's been like really 771 00:45:37,920 --> 00:45:40,960 useful, right? So like you can even like, let's 772 00:45:40,960 --> 00:45:43,280 say like you write APR and then at the end you say, hey, can, 773 00:45:43,280 --> 00:45:47,800 can you kind of like write me an HTML page that explains kind of 774 00:45:47,800 --> 00:45:50,560 like what we did in this PR, what were the changes, how it 775 00:45:50,560 --> 00:45:53,880 impacts the system? And it can easily give you like 776 00:45:53,880 --> 00:45:55,720 an overview. That's, that's very helpful. 777 00:45:56,080 --> 00:46:00,040 So yeah, I, I guess it, I guess it depends. 778 00:46:00,440 --> 00:46:04,000 I'm honestly not sure, but I feel it's going to be less 779 00:46:04,000 --> 00:46:08,880 important like the UI. Yeah, I don't even think we we 780 00:46:08,880 --> 00:46:11,600 know where this is going either. Like I think the way we probably 781 00:46:11,600 --> 00:46:15,240 consume software is going to be drastically different than we 782 00:46:15,240 --> 00:46:17,800 are today. And I don't have the answer. 783 00:46:17,800 --> 00:46:20,440 I just, I, I just feel like we're not doing it right. 784 00:46:22,880 --> 00:46:26,720 But like back to Ash, I'm not sure if you saw it, but like 785 00:46:26,720 --> 00:46:32,120 recently there have been a new release of like Ash TypeScript, 786 00:46:32,120 --> 00:46:34,760 it's called and basically it's a project. 787 00:46:34,760 --> 00:46:36,600 I have seen. That yeah. 788 00:46:38,680 --> 00:46:41,680 So I started. So basically I had this idea of 789 00:46:42,200 --> 00:46:47,160 nature orchestrator basically. So the idea is that you have. 790 00:46:48,520 --> 00:46:51,400 So it's interesting because like I had this idea and then I saw 791 00:46:51,520 --> 00:46:55,200 an exact product on on GitHub called paper clip. 792 00:46:55,480 --> 00:47:00,600 And basically the idea is that you have like you have the 793 00:47:00,600 --> 00:47:03,280 project, you have like issues of, you can create like issues 794 00:47:03,280 --> 00:47:05,000 of the project. And also you can create goals 795 00:47:05,280 --> 00:47:08,240 and then you can like create agents and you can assign issues 796 00:47:08,240 --> 00:47:10,400 and goals to the agents and they're gonna like work on it. 797 00:47:10,760 --> 00:47:14,320 And the interesting idea is that like when you have like a goal, 798 00:47:14,520 --> 00:47:17,440 the agent can like read that goal and it could create its own 799 00:47:17,440 --> 00:47:19,960 issues and kind of like, you know, like work of that. 800 00:47:20,160 --> 00:47:23,840 So so basically you can like define project in terms of like 801 00:47:25,960 --> 00:47:29,600 goals and issues and have agents like work on them and you kind 802 00:47:29,600 --> 00:47:31,400 of like create this organization structure. 803 00:47:31,640 --> 00:47:35,560 And so like paper clips specifically tries to do this 804 00:47:35,560 --> 00:47:38,000 thing called like autonomous companies. 805 00:47:39,400 --> 00:47:43,000 But yeah, have like this a similar idea and I was like, OK, 806 00:47:43,000 --> 00:47:47,840 I'm going to try to kind of like work on that. 807 00:47:48,320 --> 00:47:51,080 I just saw like that our stride sprit was came out. 808 00:47:51,080 --> 00:47:58,280 So I wanted to see how it works. And so I basically kind of 809 00:47:58,280 --> 00:48:04,720 created this like plan to where I kind of like told it like how 810 00:48:04,720 --> 00:48:07,600 we want to integrate, kind of like what are the tools that we 811 00:48:07,600 --> 00:48:11,520 want to use? And, and yeah, and I started 812 00:48:11,520 --> 00:48:15,040 like building it and it's been working really well. 813 00:48:15,040 --> 00:48:20,640 So like highly recommend. So basically what it does is 814 00:48:20,640 --> 00:48:24,240 that it takes all you can define which resources you want to be 815 00:48:24,240 --> 00:48:28,640 mapped into like JavaScript. And then it creates like this 816 00:48:28,640 --> 00:48:32,960 RPC interface so that you can basically work with your 817 00:48:32,960 --> 00:48:38,120 resources directly in, in React and create like, yeah, the front 818 00:48:38,120 --> 00:48:39,880 end. So that has been working really 819 00:48:39,880 --> 00:48:41,480 well. That's pretty cool. 820 00:48:43,360 --> 00:48:47,800 Like that's one thing I think kind of like like with live is 821 00:48:47,800 --> 00:48:53,400 with live view that like live view is great for for like back 822 00:48:53,400 --> 00:48:56,000 office internal thing, even like I think like SAS. 823 00:48:56,400 --> 00:49:01,680 But when you like, The thing is that the ecosystem, like the 824 00:49:01,680 --> 00:49:03,680 JavaScript ecosystem is like so huge. 825 00:49:04,000 --> 00:49:07,120 So there's always kind of like these things that you will want 826 00:49:07,120 --> 00:49:09,680 to through kind of like the JavaScript ecosystem. 827 00:49:10,120 --> 00:49:15,000 And then you have to kind of like like basically combine live 828 00:49:15,000 --> 00:49:20,160 view with the JavaScript And honestly, hooks are working 829 00:49:20,160 --> 00:49:21,920 really well. It's like a very good system and 830 00:49:21,920 --> 00:49:27,320 kind of like how you can like use hooks to use JavaScript 831 00:49:27,520 --> 00:49:31,000 with, with live view. But I always feel like you have 832 00:49:31,000 --> 00:49:34,320 like this, like this connection and it's still like a lot of 833 00:49:34,320 --> 00:49:38,240 like libraries and like there's a lot in the JavaScript 834 00:49:38,240 --> 00:49:40,760 ecosystem. And I sometimes feel that we are 835 00:49:40,760 --> 00:49:43,320 trying to fight it too much when embracing it. 836 00:49:43,400 --> 00:49:48,040 And kind of like just like using it would just like make some, 837 00:49:48,040 --> 00:49:49,920 some stuff like much easier, right? 838 00:49:50,000 --> 00:49:55,760 I don't know, like there's always been kind of like this, 839 00:49:59,160 --> 00:50:01,960 like this gap to bridge. Yeah. 840 00:50:02,320 --> 00:50:04,960 And it was interesting with like, I feel like it's, it's 841 00:50:04,960 --> 00:50:06,000 just like integrate. It's very nice. 842 00:50:06,000 --> 00:50:09,520 So we can, you can integrate the hard parts like, so you don't 843 00:50:09,520 --> 00:50:12,080 have to like build like APIs. You don't have to kind of like 844 00:50:12,600 --> 00:50:15,200 make clients on the front end. That's like all handled for you. 845 00:50:15,680 --> 00:50:19,640 And you can just like use your resources and and use kind of 846 00:50:19,640 --> 00:50:24,840 like established libraries and, and on the front end like React. 847 00:50:24,840 --> 00:50:30,160 And I don't like massage CM for like styles and components and 848 00:50:30,160 --> 00:50:31,280 just like works. So. 849 00:50:31,280 --> 00:50:33,000 So that's been a really nice experience. 850 00:50:34,600 --> 00:50:37,720 Yeah, that I got to probably look more into it. 851 00:50:38,560 --> 00:50:42,040 What is your current? What is your front end for your 852 00:50:42,040 --> 00:50:44,840 communities app? So that's live view. 853 00:50:45,240 --> 00:50:48,440 Oh, it is live view. OK, OK. 854 00:50:48,520 --> 00:50:52,120 And then do you have a mobile app or no? 855 00:50:53,120 --> 00:50:56,080 No. So I have like the web 856 00:50:56,080 --> 00:50:59,760 progressive app. So I'm like using Wow, but it's 857 00:50:59,760 --> 00:51:02,120 not a mobile app. And that's kind of like one of 858 00:51:02,120 --> 00:51:06,560 the like the challenges is that like almost like it is essential 859 00:51:06,560 --> 00:51:09,560 to have like a mobile app. And I'm trying to figure out how 860 00:51:09,560 --> 00:51:13,560 to do it like effectively. And it's been, it's kind of like 861 00:51:13,560 --> 00:51:18,960 hard to get like an Apple, like, AW, like a web progressive app 862 00:51:19,200 --> 00:51:23,200 published on the App Store. And it's a little bit easier on 863 00:51:23,200 --> 00:51:25,720 Android. But yeah, so then I kind of like 864 00:51:25,720 --> 00:51:27,720 want to go into that. It's kind of like in the back. 865 00:51:27,720 --> 00:51:31,240 Look, I have almost everything ready, but just like, yeah, have 866 00:51:31,280 --> 00:51:36,760 to go through that, Yeah. Apple's such a pain in the butt. 867 00:51:37,400 --> 00:51:40,640 I've been trying. So like I kind of have kill 868 00:51:40,640 --> 00:51:44,880 switch set up as like a PWA right now and it works great on 869 00:51:44,880 --> 00:51:50,760 a desktop and on on iOS mobile though, for whatever reason and 870 00:51:50,760 --> 00:51:56,680 if anyone's listening and knows how to fix my configs, basically 871 00:51:56,680 --> 00:51:58,920 I don't know if it's like in my manifest or where I'm setting 872 00:51:58,920 --> 00:52:02,800 up, but like where I load on my dashboard looks perfect. 873 00:52:02,800 --> 00:52:04,960 It's all the UIS exactly how I want it. 874 00:52:05,160 --> 00:52:09,160 The second I navigate to another view, the browser like 875 00:52:09,240 --> 00:52:12,560 navigation pops up and it's like no longer. 876 00:52:13,280 --> 00:52:17,600 It no longer looks like a PWA and I can't figure it out for 877 00:52:17,600 --> 00:52:19,280 the life of me. I don't know what I'm doing 878 00:52:19,280 --> 00:52:25,560 wrong but it's super annoying. Maybe I can, like help you with 879 00:52:25,560 --> 00:52:27,200 that, Like they can look at it, yeah. 880 00:52:27,440 --> 00:52:30,920 I'll share my code, feel like what am I doing because I know 881 00:52:30,920 --> 00:52:34,480 it's probably something silly. It's probably just because it 882 00:52:34,480 --> 00:52:37,480 shouldn't be difficult to set up a PWA. 883 00:52:38,560 --> 00:52:41,240 Yeah, like it it, I got it working on kind of like the 884 00:52:41,240 --> 00:52:42,720 browser. You can like audit everything 885 00:52:42,720 --> 00:52:47,000 like works, but it's still not like a, like it wouldn't be an 886 00:52:47,000 --> 00:52:49,280 native app anyway, but it's still not on the App Store. 887 00:52:49,280 --> 00:52:51,720 And for some people, it's kind of like they just want to have 888 00:52:51,720 --> 00:52:54,120 it on the App Store, you know? Yeah, there's like a level of 889 00:52:54,120 --> 00:52:56,080 trust, I think having it on the App Store. 890 00:52:56,480 --> 00:53:01,720 And I don't know, I don't Apple might have changed this, but I I 891 00:53:01,720 --> 00:53:05,360 don't think you have access to any kind of notifications 892 00:53:05,360 --> 00:53:07,040 without going through the App Store. 893 00:53:07,720 --> 00:53:09,760 You do, but you have to open your app. 894 00:53:09,760 --> 00:53:12,440 So basically you only get the notification on opening the app, 895 00:53:12,440 --> 00:53:15,200 which is. Kind of pointless, yeah. 896 00:53:15,640 --> 00:53:18,280 Then you open it and you just get flooded with a bunch of push 897 00:53:18,280 --> 00:53:19,400 notifications. Yeah. 898 00:53:20,120 --> 00:53:23,360 Yeah, that's so dumb. I feel like Apple has gone so 899 00:53:23,360 --> 00:53:28,360 far out of their way to stop Pwas because they want to just 900 00:53:28,560 --> 00:53:31,160 be greedy and force everyone through the App Store and get 901 00:53:31,160 --> 00:53:38,400 their 30% cut or 20% cut. Yeah, yeah, all. 902 00:53:40,400 --> 00:53:43,960 Right, yeah, let's see what like I I don't know if I fully 903 00:53:43,960 --> 00:53:48,680 understand like when would I want to use ash TypeScript like 904 00:53:49,080 --> 00:53:53,080 I would use it in my live view apps instead of JS hooks or how 905 00:53:53,080 --> 00:53:57,960 would I use it? So basically what if you want 906 00:53:58,360 --> 00:54:02,200 like, yeah, if you don't want to use like live view and you want 907 00:54:02,200 --> 00:54:06,880 to be fully on, on like the JavaScript framework, then yeah. 908 00:54:07,160 --> 00:54:09,160 So that's that's that's the ideal use case for it. 909 00:54:09,480 --> 00:54:16,080 And yeah, it So what it does is it automates all the API kind of 910 00:54:16,080 --> 00:54:20,720 like interactions. So you so you get basically it 911 00:54:20,720 --> 00:54:23,720 generates for you a client and you can use that client right 912 00:54:23,720 --> 00:54:25,080 away. So you don't have to like write 913 00:54:25,080 --> 00:54:27,240 the client code. It's generated for you and it's 914 00:54:27,240 --> 00:54:29,360 generated with types. So you can use basically 915 00:54:29,400 --> 00:54:32,200 TypeScript, right? And it's all kind of like typed 916 00:54:32,200 --> 00:54:34,320 for you and all that kind of like work. 917 00:54:34,320 --> 00:54:38,320 So you can think about it as like generating the client for 918 00:54:38,320 --> 00:54:44,240 you and you can directly use a client and there's basically no, 919 00:54:44,320 --> 00:54:45,680 no effort. You don't have to create the 920 00:54:45,680 --> 00:54:48,480 API. You just use the the HTSL to 921 00:54:48,480 --> 00:54:51,920 kind of like define the which resources you want to use and 922 00:54:51,920 --> 00:54:55,840 kind of like any configurations you want to to to give and and 923 00:54:56,040 --> 00:54:58,880 it just works. I see, OK, I love that dude. 924 00:54:59,040 --> 00:55:02,720 Like that's actually it needs to be talked about more is how 925 00:55:02,720 --> 00:55:07,520 awesome ash generators are for like the for your like your 926 00:55:07,520 --> 00:55:11,600 migrations and all. Like have you ever had to go and 927 00:55:11,600 --> 00:55:15,240 fix one of your migrations like add an index to it or any of 928 00:55:15,240 --> 00:55:16,960 that using hash? Because I haven't. 929 00:55:17,760 --> 00:55:21,320 Like, yeah, I can also like, think of all the time, like, 930 00:55:22,120 --> 00:55:25,800 like the only time I remember, like there were sometimes I did 931 00:55:25,800 --> 00:55:29,800 indexes and but then I realized that actually I wasn't 932 00:55:30,440 --> 00:55:32,600 configuring the resources correctly to do the indexes. 933 00:55:32,880 --> 00:55:36,200 And I kind of like, yeah. And if you wanna do extensions 934 00:55:36,200 --> 00:55:38,120 or something like that, obviously like that, you're 935 00:55:38,120 --> 00:55:40,640 gonna write the migration. Yeah. 936 00:55:41,040 --> 00:55:43,080 But everything else like yeah, you can just. 937 00:55:43,480 --> 00:55:47,280 But for like simple CRUD operations and like you, it's 938 00:55:47,280 --> 00:55:49,880 just magical and you don't even have to think about it. 939 00:55:50,880 --> 00:55:52,320 Yeah. And even like things that I 940 00:55:52,320 --> 00:55:54,120 wouldn't expect. So for example, there is like 941 00:55:54,120 --> 00:55:58,920 these like complex indexes that you can do on the vector and 942 00:56:00,680 --> 00:56:04,880 vector embeddings and you can define that in ASH as well. 943 00:56:07,320 --> 00:56:10,440 So basically it gives you the ability to basically, yeah, you 944 00:56:10,440 --> 00:56:13,800 can configure almost anything. I haven't accounted really 945 00:56:13,800 --> 00:56:16,960 something that you can, but I'm sure there are cases like that. 946 00:56:18,640 --> 00:56:23,800 And overall, I would say the ecosystem around ASH is so good, 947 00:56:24,680 --> 00:56:28,560 I'm surprised that, you know, it's, it's not like more 948 00:56:28,560 --> 00:56:31,520 popular. And yeah, because like you have 949 00:56:31,520 --> 00:56:36,240 things like Igniter, I think it really like changes the game in 950 00:56:36,240 --> 00:56:42,120 many ways, right? And so I think in general, like 951 00:56:42,120 --> 00:56:45,400 the, the libraries around Ash are amazing. 952 00:56:45,400 --> 00:56:47,680 And if somebody hasn't like looked into that, I highly 953 00:56:47,680 --> 00:56:51,080 recommend it. Igniter's really nice and like, 954 00:56:51,080 --> 00:56:54,520 if you have any library out there to integrate it or set it 955 00:56:54,520 --> 00:56:58,640 up with Igniter because it makes using your library so easy. 956 00:57:01,440 --> 00:57:03,560 True. And and it's just the same with 957 00:57:03,560 --> 00:57:05,560 usage rules. Add that to your libraries 958 00:57:05,560 --> 00:57:09,000 because I feel like like instantly unlocks like a 959 00:57:09,000 --> 00:57:13,520 superpower that you know, all of a sudden your your agent MD file 960 00:57:13,520 --> 00:57:15,640 just knows how to use everything in your project. 961 00:57:18,440 --> 00:57:22,120 What what does your current like workflow look like? 962 00:57:22,600 --> 00:57:25,120 Like using claw? Do you use clawed code? 963 00:57:25,120 --> 00:57:27,840 And how do you how do you plan a feature and write 1? 964 00:57:28,880 --> 00:57:33,200 So mostly cloud code and I would say it, it depends. 965 00:57:33,280 --> 00:57:35,400 Now I've been doing like the past couple of weeks, I've been 966 00:57:35,400 --> 00:57:37,760 doing like a lot of prototyping since we kind of like got 967 00:57:37,760 --> 00:57:40,120 acquired because we're launching a new product. 968 00:57:40,120 --> 00:57:42,040 So I'm kind of like just testing things out. 969 00:57:42,040 --> 00:57:46,920 So it's been kind of like like raw Proms and, and kind of like 970 00:57:46,920 --> 00:57:52,640 reading code, planning basically what we're going to build and 971 00:57:52,640 --> 00:57:54,040 how kind of like designing the system. 972 00:57:55,720 --> 00:57:57,000 Yeah. So that's been kind of like the 973 00:57:57,000 --> 00:57:59,720 past couple of weeks. But otherwise I've developed 974 00:57:59,720 --> 00:58:05,160 this system where we make these plans and then implement these 975 00:58:05,160 --> 00:58:09,880 plans. So, so the idea is it's kind of 976 00:58:09,880 --> 00:58:15,120 like interesting, like concept, I call it like easy development 977 00:58:15,840 --> 00:58:19,000 wrote like A blog about it. And basically you go through the 978 00:58:19,000 --> 00:58:21,320 cycles of like you have like an idea. 979 00:58:21,320 --> 00:58:25,920 So you have like 5 cycles, ideation, exploration, 980 00:58:27,480 --> 00:58:31,520 specification, then kind of like implementation, and then finally 981 00:58:31,520 --> 00:58:34,400 evaluation. And the idea is that you start 982 00:58:34,400 --> 00:58:37,320 with a very rough idea, then you kind of like explore that idea 983 00:58:37,320 --> 00:58:40,560 and kind of like understand what do you want out of it. 984 00:58:40,800 --> 00:58:43,520 And then there's the planning phase where you basically like 985 00:58:43,520 --> 00:58:47,320 make a plan and you kind of like the specification phase. 986 00:58:47,920 --> 00:58:51,680 The specification phase, you specify what you want to build 987 00:58:52,200 --> 00:58:58,440 and and that's kind of like of course where product is involved 988 00:58:58,440 --> 00:59:02,480 in general, kind of like people who have input in in how it 989 00:59:02,480 --> 00:59:04,960 should be. And then you go to the 990 00:59:05,080 --> 00:59:07,800 implementation. And then finally you have like 991 00:59:07,800 --> 00:59:11,120 an open phase where like maybe after a few weeks or after a 992 00:59:11,120 --> 00:59:14,160 month, you kind of like assess of how the product was built. 993 00:59:14,360 --> 00:59:18,280 And if kind of like there's like you can do like probably 994 00:59:18,280 --> 00:59:25,640 iterations right after you ship. And so usually kind of like how 995 00:59:25,640 --> 00:59:29,960 it works. Is that like either like help or 996 00:59:29,960 --> 00:59:32,240 try to understand kind of like what the plan is. 997 00:59:32,760 --> 00:59:40,000 And then I like I have actually a skill that I've prepared, 998 00:59:40,000 --> 00:59:44,800 which is like more of a command where I roughly kind of like 999 00:59:44,800 --> 00:59:48,200 tell it the plan. And sometimes we have like an, A 1000 00:59:48,200 --> 00:59:49,760 more official spec that we create. 1001 00:59:50,320 --> 00:59:55,040 And then I create like a initiative plan. 1002 00:59:55,480 --> 00:59:57,440 And then I go and I read the plan carefully. 1003 00:59:57,440 --> 00:59:59,240 I make any edits, I kind of like update it. 1004 00:59:59,640 --> 01:00:03,680 And it has like this very kind of like defined structure with 1005 01:00:04,640 --> 01:00:06,040 kind of like a definition of done. 1006 01:00:06,440 --> 01:00:09,800 And then I work on the definition of done items either 1007 01:00:09,800 --> 01:00:12,760 kind of like through like prompting nowadays. 1008 01:00:12,760 --> 01:00:19,080 And to be honest, like I prompt using cloud code mostly. 1009 01:00:19,920 --> 01:00:25,360 I will like look at the code and I have I have like this, this 1010 01:00:25,360 --> 01:00:28,120 flow like I call like interactive review. 1011 01:00:28,520 --> 01:00:31,760 And it's it's again like a dislike command that is like a 1012 01:00:31,760 --> 01:00:34,600 seven step command. But basically you tell you tell 1013 01:00:34,600 --> 01:00:41,320 the phone agent to read the read the changes that we've done in 1014 01:00:41,320 --> 01:00:46,840 this like branch present in a product oriented way the whole 1015 01:00:46,840 --> 01:00:49,440 features and how it impacts the system, how the system looks 1016 01:00:49,440 --> 01:00:53,040 now. And you basically say to include 1017 01:00:53,040 --> 01:00:56,680 like diagrams. And then the third step is you 1018 01:00:56,800 --> 01:01:01,400 have to create diagrams and present them and then wait for 1019 01:01:01,400 --> 01:01:05,640 user instructions. And, and then the other steps is 1020 01:01:05,640 --> 01:01:10,120 kind of like a like a after you get this, like towards the end, 1021 01:01:10,120 --> 01:01:12,000 it's basically kind of like to ensure that there's like good 1022 01:01:12,000 --> 01:01:15,920 documentation and, and tests and, and kind of like to do QA. 1023 01:01:16,400 --> 01:01:22,480 And so after I kind of like run the first kind of like prompt 1024 01:01:22,480 --> 01:01:25,840 where it's kind of like the initial specifications, then I 1025 01:01:25,840 --> 01:01:28,800 run through this interactive review where I kind of like look 1026 01:01:28,800 --> 01:01:32,320 at what it did basically from its own output, right? 1027 01:01:32,320 --> 01:01:36,080 And then I kind of like see if it makes sense, make any 1028 01:01:36,080 --> 01:01:37,400 changes. So I basically can tell it to 1029 01:01:37,400 --> 01:01:43,320 like change things. There's like this other like 1030 01:01:43,320 --> 01:01:47,200 command that I use that kind of like says like look at things 1031 01:01:47,200 --> 01:01:51,520 that you've like implemented and kind of like now that you've 1032 01:01:51,520 --> 01:01:54,480 implemented it, basically is there anything that you would 1033 01:01:54,480 --> 01:01:57,520 change or have done differently? And then kind of like you said, 1034 01:01:57,800 --> 01:02:02,640 these are we should ensure that there is kind of like that it's 1035 01:02:02,760 --> 01:02:06,040 architecturally sound and some other kind of things. 1036 01:02:06,520 --> 01:02:11,320 And yeah, so kind of like almost like sculpt it, right? 1037 01:02:11,320 --> 01:02:13,840 So basically you get like this like rough thing and then I 1038 01:02:13,840 --> 01:02:15,840 sculpt it. And then I finally kind of like 1039 01:02:15,840 --> 01:02:20,600 look at the like the code, like put create a bra. 1040 01:02:20,600 --> 01:02:24,040 I create a PR like review it from there, like the final pass. 1041 01:02:24,040 --> 01:02:28,320 And then like I merge it because I'm the only developer I self 1042 01:02:28,320 --> 01:02:31,400 merge for now. So that's dangerous, no? 1043 01:02:31,960 --> 01:02:36,720 Yeah. So I also self merge at least 1044 01:02:36,720 --> 01:02:40,360 with kill switch right? But I'm curious, do you find 1045 01:02:40,360 --> 01:02:44,640 like that whole setup and like kind of workflow can are is 1046 01:02:44,640 --> 01:02:48,320 clawed like one shot in things for you because you take that 1047 01:02:48,320 --> 01:02:52,680 time? I wouldn't say it's almost never 1048 01:02:52,920 --> 01:02:56,000 like one shots things. What I would say is kind of like 1049 01:02:56,000 --> 01:03:01,320 it goes like 80% of the way or like 70% of the way. 1050 01:03:01,560 --> 01:03:06,400 But then like the last 30 can be like the like, takes almost as 1051 01:03:06,400 --> 01:03:09,640 much time or, or like, well, basically the prompt, the first 1052 01:03:09,640 --> 01:03:12,280 prompt kind of like is, is like the easiest, right? 1053 01:03:13,560 --> 01:03:18,520 But like, I, I would say it's gonna, there's gonna be like 1054 01:03:18,560 --> 01:03:20,040 issues. And, and usually it's kind of 1055 01:03:20,040 --> 01:03:23,040 like more architectural or kind of like more like, not 1056 01:03:23,080 --> 01:03:25,920 necessarily like nitpicks, but like how I would like to see the 1057 01:03:25,920 --> 01:03:31,840 system like be right. So there's definitely kind of 1058 01:03:32,800 --> 01:03:38,880 things that I, I do afterwards and yeah, but, but it's, it's 1059 01:03:38,880 --> 01:03:42,160 rare for me to actually delve into the code and change things 1060 01:03:42,160 --> 01:03:43,840 myself. Like I've, I've seen that 1061 01:03:43,840 --> 01:03:48,360 especially like like lately that I'm able to describe what I want 1062 01:03:48,680 --> 01:03:53,960 and, and it basically gets it. Yeah, for the most part. 1063 01:03:54,120 --> 01:04:00,560 Like, yeah, usually where I go and like actually write stuff is 1064 01:04:00,560 --> 01:04:04,520 usually when I have like API keys and and configurations 1065 01:04:04,920 --> 01:04:07,760 basically because you don't want to just, you know, like tell you 1066 01:04:07,760 --> 01:04:09,520 this is the API keys. Can you configure it? 1067 01:04:09,520 --> 01:04:12,760 So basically I configure the API keys and then I say hey you got 1068 01:04:13,240 --> 01:04:16,200 there's like this. You don't trust Anthropic with 1069 01:04:16,200 --> 01:04:21,960 your API keys, do you? Think? 1070 01:04:22,480 --> 01:04:28,440 Like so I keep my production keys off off my obviously my 1071 01:04:28,440 --> 01:04:33,040 project and I manually put them on my server and then I let 1072 01:04:33,120 --> 01:04:37,520 Anthropic touch my dev keys because I care less about that. 1073 01:04:37,520 --> 01:04:41,360 But I am, I am like always paranoid because you know, 1074 01:04:41,360 --> 01:04:45,320 they'll like act like they can't go and look at your dot env 1075 01:04:45,400 --> 01:04:49,080 files, but you know damn well that it's scanning all your 1076 01:04:49,080 --> 01:04:51,000 files and reading those constantly. 1077 01:04:52,080 --> 01:04:54,400 Yeah. Like I know for a fact that it 1078 01:04:54,400 --> 01:04:57,440 definitely has like read through kind of like API keys that they 1079 01:04:57,440 --> 01:05:00,440 have on development. Like I guess now we just have to 1080 01:05:00,440 --> 01:05:04,680 rotate the keys more often and the best our engine kind of like 1081 01:05:04,680 --> 01:05:07,320 have keys that are not as impactful. 1082 01:05:08,040 --> 01:05:09,880 Like me working with the LMS a lot. 1083 01:05:10,480 --> 01:05:13,840 I do actually have a lot of the LM like we have like different 1084 01:05:13,840 --> 01:05:21,240 keys for for like Devon staging and for production and and but 1085 01:05:21,240 --> 01:05:23,800 like basically it's using the same model. 1086 01:05:24,120 --> 01:05:27,560 Like basically we host them on Asia. 1087 01:05:27,840 --> 01:05:34,800 So so if that he were where to be leaked, it's actually yeah, 1088 01:05:35,280 --> 01:05:38,560 they could kind of like charge us a bunch by using a key, 1089 01:05:38,560 --> 01:05:38,920 right. Yeah. 1090 01:05:39,880 --> 01:05:43,360 Yeah, like, I think about that all the time because like I, I 1091 01:05:43,360 --> 01:05:48,240 like have like Twilio set up for, you know, SMS auth and 1092 01:05:48,240 --> 01:05:52,520 messaging and notifications and yeah, if someone got that key, 1093 01:05:52,520 --> 01:05:55,960 they could just send text messages to everyone. 1094 01:05:57,360 --> 01:05:59,560 I'm trying to think like another damaging one. 1095 01:05:59,560 --> 01:06:05,280 I guess it could be like my S3 buckets, Stripe, Stripe wouldn't 1096 01:06:05,280 --> 01:06:10,400 be a good API key to get. Yeah, yeah. 1097 01:06:11,320 --> 01:06:13,120 But yeah, I've been extra paranoid. 1098 01:06:13,120 --> 01:06:18,320 But I, I am pretty LAX because like my dev keys, they're always 1099 01:06:18,320 --> 01:06:22,960 like obviously just for dev and like local hosts that 1100 01:06:23,600 --> 01:06:28,600 environments. But I feel like I would just, if 1101 01:06:28,600 --> 01:06:33,200 one of them started having like rate limits or weird, weird 1102 01:06:33,200 --> 01:06:37,000 access, I would just disable them all like instantly and 1103 01:06:37,280 --> 01:06:43,920 regenerate them, yeah. I think that's where like, like 1104 01:06:43,920 --> 01:06:46,440 there's gonna be like, we should focus on kind of like thinking 1105 01:06:46,440 --> 01:06:50,920 about how do we make systems that are resilient in that way? 1106 01:06:50,920 --> 01:06:54,520 That's kind of like, I can have like a kill switch, right? 1107 01:06:54,520 --> 01:06:58,520 Like or something. But you can kind of like, you 1108 01:06:58,520 --> 01:07:03,440 can fast kind of like rotate keys and things like that, 1109 01:07:03,440 --> 01:07:04,720 right? So because like if it takes you 1110 01:07:04,720 --> 01:07:08,040 like a few days to, to notice or rotate something, then that 1111 01:07:08,040 --> 01:07:11,440 could be an issue, right? So, and it can totally happen 1112 01:07:11,440 --> 01:07:14,920 like in a larger organization, especially I guess unless people 1113 01:07:14,920 --> 01:07:18,640 are taking up more things like the system becomes very complex 1114 01:07:18,640 --> 01:07:21,720 and and if you are like down kind of like focusing on a 1115 01:07:21,720 --> 01:07:24,320 feature, like remembering exactly where to go to kind of 1116 01:07:24,320 --> 01:07:26,680 like change that key of that service, it might not be that 1117 01:07:26,680 --> 01:07:30,280 straightforward, right. So yeah, having. 1118 01:07:30,320 --> 01:07:33,000 I just thought of a product. Well, I think you thought of a 1119 01:07:33,000 --> 01:07:35,520 product. I think we need like A1 password 1120 01:07:35,520 --> 01:07:40,680 or bit warden like a password manager for API keys and you can 1121 01:07:40,840 --> 01:07:46,040 enter like do API integrations, you can generate new keys and 1122 01:07:46,440 --> 01:07:49,080 revoke keys all from 1 dashboard. 1123 01:07:49,080 --> 01:07:50,800 That would be a really cool product. 1124 01:07:51,600 --> 01:07:54,120 Someone should build that. Whoever's listening, someone 1125 01:07:54,120 --> 01:07:57,080 built that of. Course you can create like this 1126 01:07:57,080 --> 01:07:59,480 whole security hole as well, right? 1127 01:07:59,480 --> 01:08:00,640 Because like you. Right. 1128 01:08:00,640 --> 01:08:03,040 You have to really trust. Yeah, you got to trust the 1129 01:08:03,040 --> 01:08:07,920 platform. But actually you can't. 1130 01:08:09,520 --> 01:08:11,200 Yeah, I don't know a good way you could do that. 1131 01:08:12,320 --> 01:08:16,800 You'd have to have API access kind of all through the client 1132 01:08:16,800 --> 01:08:19,479 side. And then you could, you could 1133 01:08:19,479 --> 01:08:21,520 encrypt everything client side as well. 1134 01:08:21,520 --> 01:08:23,399 If you store anything on the database. 1135 01:08:23,840 --> 01:08:27,279 Yeah, it would be. It's messy though, like 'cause 1136 01:08:27,279 --> 01:08:30,319 there's you want your client apps to be dumb for the most 1137 01:08:30,319 --> 01:08:33,840 part, right? They're they, they view data, 1138 01:08:34,319 --> 01:08:37,399 that's all you wanted to do. But then like since I've done 1139 01:08:37,399 --> 01:08:40,439 Killswitch and it's zero knowledge, I have to do a lot of 1140 01:08:40,439 --> 01:08:42,359 encryption and decryption client side. 1141 01:08:42,359 --> 01:08:46,240 And it kind of goes against the paradigm of what I think front 1142 01:08:46,240 --> 01:08:50,520 end should be. But yeah, there's like a a 1143 01:08:50,520 --> 01:08:53,319 balance I think you got kind of have to hold there. 1144 01:08:54,520 --> 01:08:58,240 Yeah, it's always, you have to expect that there's like 1145 01:08:58,359 --> 01:09:01,000 different types of users and some are gonna be like more 1146 01:09:01,000 --> 01:09:03,359 technical like like feel comfortable with kind of like 1147 01:09:03,359 --> 01:09:06,000 owning keys and and owning responsibility. 1148 01:09:06,000 --> 01:09:10,920 But yeah, like that's like why crypto hasn't like really 1149 01:09:11,120 --> 01:09:15,200 worked, I think is that like most people, they, they're not 1150 01:09:15,760 --> 01:09:17,920 ready to take responsibility or kind of like they're not as 1151 01:09:17,920 --> 01:09:19,720 technical, right. So yeah, it's too. 1152 01:09:19,720 --> 01:09:24,160 Complicated to get set up like what like to open a crypto 1153 01:09:24,160 --> 01:09:29,600 wallet, you have your your your key and then you have your 1154 01:09:29,600 --> 01:09:33,000 private key and you have to kind of understand how encryption 1155 01:09:33,000 --> 01:09:38,800 works a little bit, yeah. Or if you lose that or give it 1156 01:09:38,800 --> 01:09:43,160 to anyone, you lose all access or give away all access, right? 1157 01:09:44,439 --> 01:09:47,000 Yeah. And it's either like you, Yeah, 1158 01:09:47,680 --> 01:09:50,880 it's because for example, it's easy today to kind of like get a 1159 01:09:50,880 --> 01:09:52,760 crypto word, right. You can go on your phone, 1160 01:09:52,760 --> 01:09:55,160 there's like many apps that kind of like are crypto words, but 1161 01:09:55,400 --> 01:10:00,480 and either like allow them to have the keys and they can't 1162 01:10:00,480 --> 01:10:02,040 actually have access to them, right? 1163 01:10:02,040 --> 01:10:04,680 Like many work like there's some that work through like Google 1164 01:10:04,680 --> 01:10:06,000 login. So basically even if you lose 1165 01:10:06,000 --> 01:10:08,240 your phone or something, you still have access. 1166 01:10:08,400 --> 01:10:12,120 But the truth is that somebody has your keys, you know, like 1167 01:10:12,120 --> 01:10:16,280 and hopefully they're not gonna like do anything malicious with 1168 01:10:16,280 --> 01:10:19,680 it. But yeah, you do kind of like 1169 01:10:20,880 --> 01:10:24,360 somebody has to kind of like have that knowledge and and if 1170 01:10:24,360 --> 01:10:26,840 you're not willing to have the kind of like their own knowledge 1171 01:10:26,840 --> 01:10:28,920 and responsibility. And if you're not willing, 1172 01:10:28,920 --> 01:10:32,040 somebody else has to, right? Yeah, yeah. 1173 01:10:32,520 --> 01:10:36,320 I will say I have leaked some API keys on some live streams 1174 01:10:36,320 --> 01:10:40,000 and the Elixir community is are are pretty honest folks. 1175 01:10:41,600 --> 01:10:44,760 So no one, no one grabbed it and went and used it. 1176 01:10:44,760 --> 01:10:47,440 So that was good. And then I obviously revoked it 1177 01:10:47,440 --> 01:10:52,080 after the live stream, but it was it was active the whole time 1178 01:10:52,080 --> 01:10:54,080 and no one stole it. So that was always, that was 1179 01:10:54,080 --> 01:10:56,080 good to see. Nice. 1180 01:10:56,080 --> 01:10:58,800 Yeah, I think Alex is one of the, like, the sweetest 1181 01:10:58,800 --> 01:11:02,000 communities of encounters. Definitely. 1182 01:11:04,000 --> 01:11:08,360 Oh, that's kind of a cool idea. So Dharma in the chat here 1183 01:11:08,760 --> 01:11:14,040 saying that, you know, the whole API key manager is just 1184 01:11:14,040 --> 01:11:18,720 essentially a password manager, but I like keys like family keys 1185 01:11:18,720 --> 01:11:20,600 would make sense. That would be kind of cool. 1186 01:11:20,880 --> 01:11:26,080 Like you could actually have more you could actually generate 1187 01:11:26,080 --> 01:11:31,880 keys that point at a key and actually like never give access 1188 01:11:31,880 --> 01:11:35,280 to your real key too. So you could kind of kind of 1189 01:11:35,280 --> 01:11:39,840 like a a URL shortener or like a a redirect. 1190 01:11:39,840 --> 01:11:41,760 You could actually, that would be it guys. 1191 01:11:41,760 --> 01:11:46,960 This could be a cool product you could read like you could share 1192 01:11:46,960 --> 01:11:53,160 keys that go to your products API and then it grabs the actual 1193 01:11:53,160 --> 01:11:58,720 API key and returns that. And that could be encrypted or 1194 01:11:59,560 --> 01:12:01,360 handled however you want. I don't know. 1195 01:12:01,360 --> 01:12:03,840 I think we just really overcomplicated this, but I like 1196 01:12:03,920 --> 01:12:08,880 it. It's, it's also like, yeah, 1197 01:12:08,920 --> 01:12:12,760 thinking about if it has access to revoke because it's also can 1198 01:12:12,760 --> 01:12:15,480 be a security risk, right? Like if you can revoke the keys 1199 01:12:15,480 --> 01:12:17,760 and all of a sudden somebody's getting errors in production. 1200 01:12:18,400 --> 01:12:23,200 So it would also like the keys, it has to have the ability to 1201 01:12:23,200 --> 01:12:25,600 regenerate. Like how how is that match? 1202 01:12:25,600 --> 01:12:30,440 Is that on the client, right? So yeah, which could be on on 1203 01:12:30,440 --> 01:12:32,800 the client, but then like how does the client know it in the 1204 01:12:32,800 --> 01:12:35,920 first place if it is not signed into the platform? 1205 01:12:35,920 --> 01:12:40,800 So yeah. It's a cool product, but there's 1206 01:12:40,800 --> 01:12:43,760 some complicated architectural decisions to make. 1207 01:12:45,040 --> 01:12:48,440 Yeah. Yeah, that's a tough one because 1208 01:12:48,440 --> 01:12:52,760 like, imagine if anyone gets access to your platform, let's 1209 01:12:52,760 --> 01:12:56,240 pretend we created this manager of API keys. 1210 01:12:56,600 --> 01:13:00,640 And yeah, you could be like Community. 1211 01:13:00,640 --> 01:13:02,000 What is the name of your Community app? 1212 01:13:02,000 --> 01:13:07,800 Is it just Community? OK so communities I could just 1213 01:13:07,800 --> 01:13:13,440 be like I hate that guy and then I hack his API key manager and 1214 01:13:13,440 --> 01:13:15,400 then just revoke all his keys in production. 1215 01:13:16,280 --> 01:13:20,480 Yeah, exactly. And then your system blows up. 1216 01:13:20,480 --> 01:13:23,520 That would actually be a nightmare, but at least you have 1217 01:13:23,520 --> 01:13:26,600 the past the API key manager and it's really easy to just 1218 01:13:26,600 --> 01:13:33,640 regenerate. Yeah, but yeah, it's true. 1219 01:13:33,640 --> 01:13:34,400 It's true. Yeah. 1220 01:13:34,760 --> 01:13:38,320 So then you can recover relatively fast. 1221 01:13:41,760 --> 01:13:44,840 Great. The cool thing would be like if 1222 01:13:44,840 --> 01:13:46,960 you. That would be an interesting 1223 01:13:46,960 --> 01:13:49,040 thing. So what if it was? 1224 01:13:49,040 --> 01:13:53,400 If it regenerated, it would apply the changes directly in 1225 01:13:53,400 --> 01:13:55,320 your app. So basically the worst case 1226 01:13:55,320 --> 01:13:58,600 would be that your app restarts to like, apply the new keys, but 1227 01:13:58,640 --> 01:14:00,000 basically every time you regenerate. 1228 01:14:00,480 --> 01:14:02,880 But of course then you could do like a denial of service or 1229 01:14:02,880 --> 01:14:05,920 something. But imagine like when you 1230 01:14:05,920 --> 01:14:09,080 regenerate the keys, it automatically kind of like sends 1231 01:14:09,080 --> 01:14:11,320 the message to your app to kind of like apply the new key. 1232 01:14:12,840 --> 01:14:14,240 Yeah. So yeah. 1233 01:14:14,680 --> 01:14:19,120 What would be cool is then all your API communications all go. 1234 01:14:19,360 --> 01:14:22,600 It would have to all kind of, well I guess just retrieval of 1235 01:14:22,600 --> 01:14:26,960 keys would all go through your API instead of your environment 1236 01:14:26,960 --> 01:14:29,440 variable file, which I don't know. 1237 01:14:29,440 --> 01:14:33,280 I kind of like this idea. It might slow things down though 1238 01:14:33,280 --> 01:14:38,640 if speed matters on requests, but I don't know, I think 1239 01:14:38,640 --> 01:14:40,520 someone should someone should build this. 1240 01:14:40,680 --> 01:14:45,960 I don't want to, it sounds hard. A lot of nuances, like when you 1241 01:14:45,960 --> 01:14:48,080 start, if you get into it, you're going to discover a lot 1242 01:14:48,080 --> 01:14:51,040 of things. Maybe we should start an Elixir 1243 01:14:51,040 --> 01:14:53,480 mentor open source project and build this thing. 1244 01:14:53,880 --> 01:15:00,080 That could be fun, but someone should take take the the reins. 1245 01:15:00,200 --> 01:15:05,200 I'm too busy. Back to LMS, I'm kind of like 1246 01:15:05,200 --> 01:15:08,160 curious about, I've been thinking about this concept and 1247 01:15:08,160 --> 01:15:10,360 I'm trying to see what other people also like think about it. 1248 01:15:10,360 --> 01:15:16,640 But do you think that LMS understand or have the ability 1249 01:15:16,640 --> 01:15:21,840 to understand? I don't know, I think, I don't 1250 01:15:21,840 --> 01:15:23,400 think they have the ability to understand. 1251 01:15:23,400 --> 01:15:26,480 I think they have the ability to see patterns and copy patterns 1252 01:15:27,960 --> 01:15:30,880 and they're obviously really good, like search algorithms, 1253 01:15:30,880 --> 01:15:33,840 right, for finding data in a mass amount of data. 1254 01:15:35,080 --> 01:15:39,800 So I think it's just, I, I mean, I'm, I'm a complete idiot, so I 1255 01:15:39,800 --> 01:15:42,480 don't listen to anything I'm saying, but like, I just feel 1256 01:15:42,480 --> 01:15:45,160 like they're really good search algorithms, like, you know, next 1257 01:15:45,160 --> 01:15:49,720 level Google search, right? And then I think you're just 1258 01:15:49,720 --> 01:15:54,720 kind of using it as a pattern recognition and it sees it. 1259 01:15:54,920 --> 01:15:57,680 That's why it's like so good at functional programming and 1260 01:15:57,680 --> 01:16:01,120 immutable languages, right? Yeah, Cuz like a pattern is just 1261 01:16:01,120 --> 01:16:03,000 so much cleaner and easier to see. 1262 01:16:03,320 --> 01:16:05,680 And then it can just regurgitate and write code. 1263 01:16:06,200 --> 01:16:08,520 Cuz that's like one of the things that I always do. 1264 01:16:08,520 --> 01:16:11,440 Like you were talking about, you kind of have like a review 1265 01:16:11,440 --> 01:16:14,960 skill. I'll always ask, I'll always ask 1266 01:16:14,960 --> 01:16:18,600 it before like APR or even if I have APR open. 1267 01:16:18,800 --> 01:16:21,400 But like, are the changes in this branch? 1268 01:16:21,680 --> 01:16:25,120 Do they follow like architectural decisions and like 1269 01:16:25,120 --> 01:16:27,480 patterns that are already in the app? 1270 01:16:28,000 --> 01:16:31,200 And if it says that like a part of it sticks out and doesn't 1271 01:16:31,200 --> 01:16:35,240 look right, it's like it helps you find nasty code right it all 1272 01:16:37,000 --> 01:16:40,200 like I yeah, I don't know. I always ask it to compare 1273 01:16:40,200 --> 01:16:43,480 against itself, like do a get difference here and. 1274 01:16:43,720 --> 01:16:46,080 Yeah, yeah. That's the that's really 1275 01:16:46,080 --> 01:16:49,960 important I think for it to be able to like reflect and like 1276 01:16:49,960 --> 01:16:53,480 most do it naturally like most of them I would say. 1277 01:16:54,440 --> 01:16:57,960 Yeah, all the LLMS I have like thinking they basically reflect. 1278 01:16:58,160 --> 01:17:02,000 Well, LLMS reflect by default because that they're make that 1279 01:17:02,000 --> 01:17:05,040 tension mechanism is to look what you did before, right. 1280 01:17:05,040 --> 01:17:10,240 So they have like a yeah. But yeah, they like if you tell 1281 01:17:10,240 --> 01:17:12,640 them that they kind of like should like black critical 1282 01:17:12,640 --> 01:17:14,200 thinking. If they can't even do that, that 1283 01:17:14,200 --> 01:17:17,440 like, it definitely helps them kind of like looking at the 1284 01:17:17,440 --> 01:17:20,040 patterns and and and try to see if there's something like wrong 1285 01:17:20,080 --> 01:17:24,400 with them. Yeah, Daharma wants to know if 1286 01:17:24,400 --> 01:17:29,840 he'll see you at Elixir Conf EU. Yes, I will definitely be there 1287 01:17:29,840 --> 01:17:31,600 actually giving out the training. 1288 01:17:31,600 --> 01:17:38,720 So I'm going to be a trainer teaching about basically how to 1289 01:17:38,720 --> 01:17:42,880 create all enabled applications. It's an interesting program. 1290 01:17:43,480 --> 01:17:47,840 We're going to building this like to do app and that's the 1291 01:17:47,840 --> 01:17:49,200 foundation. So you have like A to do app 1292 01:17:49,200 --> 01:17:55,680 that's all done for you and then you start putting like a smart 1293 01:17:55,680 --> 01:17:59,720 functionality on top of that. And it's basically there's a 1294 01:17:59,720 --> 01:18:02,520 command palette and you can like type things into it and it will 1295 01:18:02,520 --> 01:18:03,880 do. So they can like tell it to 1296 01:18:03,880 --> 01:18:08,520 create tickets to move tickets to do like research on the web 1297 01:18:08,520 --> 01:18:10,920 and kind of like prepare some things for you. 1298 01:18:10,920 --> 01:18:12,840 So, so I'm looking forward to it. 1299 01:18:12,840 --> 01:18:15,600 Actually, it's my first time I'm doing, it's the first time I'm 1300 01:18:15,600 --> 01:18:18,720 doing training in Elixir conf. I've done one small training 1301 01:18:19,240 --> 01:18:24,280 about Elixir here in Prague some time ago, some like years ago 1302 01:18:24,280 --> 01:18:26,480 right now. But yeah, it's going to be the 1303 01:18:26,480 --> 01:18:30,160 first time in Elixir call. So I'm actually, yeah, I'm 1304 01:18:30,160 --> 01:18:32,240 looking forward to. That's awesome. 1305 01:18:32,520 --> 01:18:34,520 You'll do awesome. I'm sure it'll be great. 1306 01:18:36,680 --> 01:18:40,160 I I wish I could go. I, I tried to go to the the 1307 01:18:40,160 --> 01:18:44,160 Lisbon 1 and I got sick the the morning of my flight and had to 1308 01:18:44,160 --> 01:18:48,280 cancel the whole trip and then life's just been too crazy. 1309 01:18:48,280 --> 01:18:52,800 I haven't even had time to really travel, but I want to get 1310 01:18:52,800 --> 01:18:57,200 to Chicago this year. It it might not happen. 1311 01:18:58,600 --> 01:19:02,320 We'll see. Yeah, just life. 1312 01:19:02,320 --> 01:19:06,560 Life is crazy. Yeah, it sounds like a long, a 1313 01:19:07,040 --> 01:19:10,080 long travel, right Like to go to Europe, but I would imagine that 1314 01:19:10,080 --> 01:19:12,680 you would like to combine it with some vacation or? 1315 01:19:12,680 --> 01:19:15,480 It would have to, yeah, it would have to be vacation for sure. 1316 01:19:15,480 --> 01:19:17,200 But like, there's so much to see. 1317 01:19:17,240 --> 01:19:22,200 Like, yeah, like twist my arm. Where is where is the EU 1318 01:19:22,200 --> 01:19:25,000 conference this year? It's in Malaga, Spain. 1319 01:19:25,440 --> 01:19:26,800 Oh, OK, nice. Yeah. 1320 01:19:26,800 --> 01:19:28,440 So like, there's so much to see in Spain. 1321 01:19:29,160 --> 01:19:31,680 Yeah, it's my first time actually to Spain, so. 1322 01:19:32,160 --> 01:19:35,520 Oh nice. So yeah, like, I'm nice. 1323 01:19:37,640 --> 01:19:39,920 Oh, Daharma's giving his first talk too. 1324 01:19:40,520 --> 01:19:41,840 That is awesome. What? 1325 01:19:42,800 --> 01:19:46,920 Let us know what your talk is and do you need to be a guest on 1326 01:19:46,920 --> 01:19:52,760 my podcast? Maybe you do see, I always got a 1327 01:19:52,760 --> 01:19:56,720 recruit. I have to catch back up and I 1328 01:19:56,720 --> 01:19:59,960 have to. The worst part about my podcast, 1329 01:20:00,040 --> 01:20:03,080 or I should say the part that I procrastinate the most, is 1330 01:20:03,160 --> 01:20:06,120 scheduling the events to show up on my website. 1331 01:20:06,920 --> 01:20:10,440 I don't know why, but something about creating thumbnails I just 1332 01:20:10,440 --> 01:20:12,080 don't want to do until I have to. 1333 01:20:12,480 --> 01:20:15,840 It's so much work. I need to maybe get Claude to 1334 01:20:15,840 --> 01:20:18,360 start doing it for me, like integrate it with Canvas 1335 01:20:18,360 --> 01:20:19,720 somehow. I don't know if I can. 1336 01:20:21,080 --> 01:20:25,760 There's so basic to do to do what to. 1337 01:20:26,080 --> 01:20:28,760 To create like the thumbnail, like, you know how like, I have 1338 01:20:28,760 --> 01:20:32,880 your headshot and then I just have like, you know, I have like 1339 01:20:33,400 --> 01:20:35,800 the title of this, the podcast episode. 1340 01:20:35,800 --> 01:20:37,880 But then like, a lot of it is the same. 1341 01:20:37,960 --> 01:20:41,120 It stays static. But yeah, just creating those 1342 01:20:41,120 --> 01:20:46,400 takes a long time. So I always procrastinate it not 1343 01:20:46,400 --> 01:20:49,800 a long time, but more time than I'd rather write a feature for 1344 01:20:49,800 --> 01:20:58,640 kill switch, you know? So I'm curious, solo 1345 01:20:58,640 --> 01:21:01,600 entrepreneur communities, right? What? 1346 01:21:01,600 --> 01:21:03,960 What do you do to get the word out? 1347 01:21:03,960 --> 01:21:07,400 How do you onboard people? Is it just a lot of hand holding 1348 01:21:07,400 --> 01:21:10,920 right now and like manual work or what's your strategy? 1349 01:21:12,160 --> 01:21:17,640 So in, in the beginning, I was kind of like trying to find the, 1350 01:21:17,760 --> 01:21:20,720 the kind of like who would be kind of like irrelevant to the, 1351 01:21:20,800 --> 01:21:23,760 to the project. And now I'm doing like a lot of 1352 01:21:23,760 --> 01:21:28,120 networking. So I basically, I'm, I'm 1353 01:21:28,200 --> 01:21:32,600 becoming familiar with like, like community builders here in 1354 01:21:32,600 --> 01:21:36,240 Prague and mostly people who are like focused on like social 1355 01:21:37,480 --> 01:21:40,080 impact. So there's for example, we have 1356 01:21:40,080 --> 01:21:43,520 a partnership now with the organization called Big Act and 1357 01:21:43,520 --> 01:21:48,720 they are doing basically they're helping immigrants integrate 1358 01:21:48,840 --> 01:21:53,720 into like Czech society. So like Prague is around like 1359 01:21:53,720 --> 01:21:59,640 maybe like 20 to 30% immigrants. So there's a lot of immigrants 1360 01:21:59,640 --> 01:22:02,840 living here. And yeah, part of their mission 1361 01:22:02,840 --> 01:22:07,080 is to kind of like help the immigrants integrate and kind of 1362 01:22:07,080 --> 01:22:11,040 like be part of, of everyday life. 1363 01:22:12,800 --> 01:22:16,760 So, yeah, so kind of like working with them and just 1364 01:22:16,760 --> 01:22:20,560 trying to find people who are socially active and and try to 1365 01:22:20,560 --> 01:22:24,640 do like these initiatives. What I want to get into, but I'm 1366 01:22:24,640 --> 01:22:29,480 kind of like still haven't like gotten into it is to work with 1367 01:22:29,520 --> 01:22:30,960 people who are doing volunteering. 1368 01:22:31,680 --> 01:22:35,480 So organizations who are for example, like maybe they are 1369 01:22:35,480 --> 01:22:42,160 doing programs that like help like like homeless people or 1370 01:22:42,160 --> 01:22:46,760 somebody to like find food or shelter and basically kind of 1371 01:22:46,760 --> 01:22:52,520 like try to work with them. And because I really want this 1372 01:22:52,520 --> 01:22:56,680 to be kind of like about making your region like a better place. 1373 01:22:57,240 --> 01:23:01,960 And yeah, I think volunteering is a really good way to like 1374 01:23:02,440 --> 01:23:06,000 connect with people, socialize and, and also kind of like help 1375 01:23:06,000 --> 01:23:09,560 somebody and, and it's really rewarding like when when you do 1376 01:23:09,560 --> 01:23:12,280 some, some volunteering. So that's that's kind of like a 1377 01:23:12,280 --> 01:23:15,480 space I want to like see and find people and, and, and see 1378 01:23:15,480 --> 01:23:18,640 also kind of like what problems do they have and how kind of 1379 01:23:18,640 --> 01:23:21,960 like I can help them maybe like integrate there. 1380 01:23:22,480 --> 01:23:26,720 And yeah, so kind of like right now I'm more focused on kind of 1381 01:23:26,720 --> 01:23:29,200 like networking, trying to find like key people that could like 1382 01:23:29,200 --> 01:23:36,760 help me kind of like because I feel like what I would like to 1383 01:23:36,760 --> 01:23:39,240 do is kind of get into these organizations that are already 1384 01:23:39,240 --> 01:23:40,640 existing and already functioning. 1385 01:23:40,920 --> 01:23:43,600 And it's not always easy to kind of breakthrough, especially if 1386 01:23:44,080 --> 01:23:47,120 they are not very always welcoming to outsiders trying to 1387 01:23:47,800 --> 01:23:52,880 quote UN quote sell them stuff. So I want to kind of get 1388 01:23:52,880 --> 01:23:56,200 involved because my goal is to like help them really. 1389 01:23:56,200 --> 01:24:00,360 So I kind of like when I first show up as a human, you know, 1390 01:24:00,360 --> 01:24:05,280 like help them kind of like be be involved and kind of like see 1391 01:24:05,480 --> 01:24:08,520 through that kind of like how I can actually also build a 1392 01:24:08,520 --> 01:24:11,640 product that that that helps these goals. 1393 01:24:12,040 --> 01:24:13,680 Yeah, I think that's a good strategy. 1394 01:24:13,680 --> 01:24:17,920 I think people underestimate just knowing people, right? 1395 01:24:17,920 --> 01:24:21,280 Like if people know you and they like you, they're more likely 1396 01:24:21,280 --> 01:24:26,480 just to like trust your product and give you a try. 1397 01:24:26,480 --> 01:24:29,480 So I think that's a good strategy for sure, especially 1398 01:24:29,480 --> 01:24:32,240 for like a platform, like communities. 1399 01:24:32,800 --> 01:24:36,800 You kind of, you know, you get a couple, like you know, big 1400 01:24:36,960 --> 01:24:39,800 people or agencies on there, right? 1401 01:24:40,160 --> 01:24:45,320 And you'd probably see just an influx of use after that too, 1402 01:24:45,320 --> 01:24:48,840 because then word of mouth starts traveling and people talk 1403 01:24:48,840 --> 01:24:54,480 about liking your your product. Yeah, Yeah, it's true. 1404 01:24:54,480 --> 01:24:56,120 Yeah. But it's, it's kind of like 1405 01:24:56,120 --> 01:24:59,520 difficult what I'm now kind of like getting into. 1406 01:24:59,520 --> 01:25:03,200 And I really like the idea is I, I mentioned earlier that I'm 1407 01:25:03,200 --> 01:25:05,960 kind of like thinking about this or like building this product 1408 01:25:06,320 --> 01:25:10,040 that basically you have like agents and you assign them tasks 1409 01:25:10,040 --> 01:25:14,600 and and all that. And I want to basically what 1410 01:25:14,600 --> 01:25:19,120 I've like done is you've I've integrated with like ACMS. So 1411 01:25:19,120 --> 01:25:22,080 there's like the CMS code like 20 where it's like an open 1412 01:25:22,080 --> 01:25:26,160 source CMS is self hostable. So I hosted that. 1413 01:25:26,160 --> 01:25:29,560 There's also like a product called Affine that's also self 1414 01:25:29,560 --> 01:25:32,880 hostable and it's basically an open source notion. 1415 01:25:33,320 --> 01:25:37,960 And now the idea is kind of like you put agents to do like 1416 01:25:37,960 --> 01:25:41,480 research and try to find organizations and contacts and 1417 01:25:41,480 --> 01:25:44,560 basically like fill up your like CMS with kind of like relevant 1418 01:25:44,560 --> 01:25:46,080 contacts. And then you go through that 1419 01:25:46,400 --> 01:25:51,280 and, and kind of like have people to kind of like reach out 1420 01:25:51,320 --> 01:25:54,160 to. It's been actually hard because 1421 01:25:54,160 --> 01:25:57,680 like I, I found that it works best when you go through an 1422 01:25:57,680 --> 01:26:00,560 introduction like called, like called messaging or called 1423 01:26:00,560 --> 01:26:02,600 emailing. Like these people have got like 1424 01:26:02,600 --> 01:26:07,520 very few responses. And yeah, but like what I also 1425 01:26:07,560 --> 01:26:10,200 kind of like now want to do is I've seen that there's a lot of 1426 01:26:10,200 --> 01:26:15,960 podcasts that like people are talking about like relevant. 1427 01:26:16,400 --> 01:26:22,320 Like there's like a podcast here called like a chess plot and 1428 01:26:22,400 --> 01:26:26,160 check. And I want to write something 1429 01:26:26,160 --> 01:26:31,000 that like downloads the podcast, creates a transcript, kind of 1430 01:26:31,000 --> 01:26:34,080 like, and then like tries to find relevant people and kind of 1431 01:26:34,080 --> 01:26:36,880 like gives me information about them, tells me what their 1432 01:26:36,880 --> 01:26:39,920 projects are working on. And then kind of like creating 1433 01:26:39,920 --> 01:26:42,320 like this pipeline. And then you can like direct it 1434 01:26:42,320 --> 01:26:44,520 to like various, right? Like, yeah. 1435 01:26:44,520 --> 01:26:48,200 So kind of like I want to some more, more thinking about things 1436 01:26:48,200 --> 01:26:52,640 like that now, which, yeah, it's not like features, but it's 1437 01:26:52,640 --> 01:26:54,080 definitely things that need to get done. 1438 01:26:54,080 --> 01:26:57,040 It's actually kind of fun because like it's a lot of 1439 01:26:57,120 --> 01:27:00,080 authentic work and kind of like thinking about these like new 1440 01:27:00,080 --> 01:27:03,000 types of systems where they make decisions for you, right. 1441 01:27:03,000 --> 01:27:07,720 And, and yeah, and then just like unleash them. 1442 01:27:07,720 --> 01:27:11,240 Of course, I don't have a setup so they can actually send emails 1443 01:27:11,240 --> 01:27:14,640 or anything, but like generate leads and get information for 1444 01:27:14,640 --> 01:27:15,960 you. Do that like research part. 1445 01:27:17,800 --> 01:27:21,320 Yeah, we we have a funny, a funny comment from Star Mount 1446 01:27:21,320 --> 01:27:23,760 Pictures. He has the ultimate Save the 1447 01:27:23,760 --> 01:27:25,560 World show. Bankman Friedman. 1448 01:27:26,240 --> 01:27:28,680 I'm you're talking about Sam Bankman Fried. 1449 01:27:28,680 --> 01:27:33,840 I think the problem is, I think everyone starts off with good 1450 01:27:33,840 --> 01:27:38,200 intentions, right? And then like, money changes a 1451 01:27:38,200 --> 01:27:43,120 lot of things and then drugs start and it becomes more fun to 1452 01:27:43,120 --> 01:27:45,160 do drugs than actually save the world. 1453 01:27:46,920 --> 01:27:50,520 Yeah, the Sabre Complex, a lot of people have it nowadays, it 1454 01:27:50,520 --> 01:27:54,560 seems like. Yeah, I mean, like throughout 1455 01:27:54,560 --> 01:27:57,720 history there's always been like people with Sabre Complex. 1456 01:27:57,720 --> 01:28:03,200 But yeah, it's, it's hard. Like I feel that the pressure 1457 01:28:03,200 --> 01:28:05,240 on, on kind of like succeeding when you're like in these 1458 01:28:05,240 --> 01:28:10,120 circles, it becomes so big that you forget how to live and how 1459 01:28:10,120 --> 01:28:12,080 to enjoy your life, you know, like, and you're all you're 1460 01:28:12,080 --> 01:28:15,640 chasing is kind of like this external validation and kind of 1461 01:28:15,640 --> 01:28:16,280 get back. Yeah. 1462 01:28:16,920 --> 01:28:22,680 Yeah, yeah, slippery slope. I I've been thinking about my 1463 01:28:22,680 --> 01:28:27,280 own like networking strategies too, because 11 mistake that I 1464 01:28:27,280 --> 01:28:29,880 made was like, oh, I have an audience, right? 1465 01:28:29,880 --> 01:28:33,320 Elixir mentor like this is a product that I want to use. 1466 01:28:33,320 --> 01:28:35,680 So I feel like people listening to me would want to use it too 1467 01:28:36,000 --> 01:28:38,920 and not true, right? Like my audience, my current 1468 01:28:38,920 --> 01:28:43,520 audience wants to build things in elixir and learn Elixir, not 1469 01:28:43,520 --> 01:28:47,560 buy some random SAS product. So like I've been working on 1470 01:28:47,560 --> 01:28:52,240 like I need to start hanging out in like proper organization, 1471 01:28:52,240 --> 01:28:55,440 like where people are kind of like security paranoid, but they 1472 01:28:55,440 --> 01:28:58,680 think about planning for the worst case scenario. 1473 01:28:59,000 --> 01:29:02,200 I had to like get on podcasts that are outside of the 1474 01:29:02,200 --> 01:29:07,000 developer community and talk about my product and do that. 1475 01:29:07,360 --> 01:29:12,200 And it sounds exhausting, but I think it's like one of those 1476 01:29:12,200 --> 01:29:15,040 things that you just have to kind of you have to do, you have 1477 01:29:15,040 --> 01:29:19,000 to eat the glass and do things you don't want to do because 1478 01:29:19,760 --> 01:29:21,600 Killswitch is a cool product, I think. 1479 01:29:21,640 --> 01:29:25,040 And that like the best part that makes it the most fun for me is 1480 01:29:25,040 --> 01:29:27,480 like, I'm building it because I like when I add a feature, it's 1481 01:29:27,480 --> 01:29:31,040 like something I'm using. So like everything that I build, 1482 01:29:31,040 --> 01:29:35,360 I'm like already using. Like I use Killswitch for all my 1483 01:29:35,360 --> 01:29:39,640 stuff. So all my loved ones, friends 1484 01:29:39,640 --> 01:29:43,040 and family are going to get Deadman switch triggers if I 1485 01:29:43,040 --> 01:29:47,760 stop checking in and they're they're going to get videos from 1486 01:29:47,760 --> 01:29:50,520 me and I'm just going to make sure they they can't forget 1487 01:29:50,520 --> 01:29:54,400 about me when I'm gone. That would be fine and you. 1488 01:29:57,840 --> 01:30:03,240 Infowars, Oh, if I could get on Infowars, you know, get what is 1489 01:30:03,240 --> 01:30:07,680 Alex Jones, that would be funny. There we go. 1490 01:30:08,840 --> 01:30:11,440 I'll get on Infowars. I don't know if I have like the 1491 01:30:12,040 --> 01:30:14,480 personality for that show, honestly. 1492 01:30:16,400 --> 01:30:19,440 Do you know who Alex Jones is? Yeah, of course, of course, 1493 01:30:19,440 --> 01:30:21,480 yeah. He's he's. 1494 01:30:21,680 --> 01:30:23,040 A. He's an. 1495 01:30:23,040 --> 01:30:25,840 Interesting guy. What's that? 1496 01:30:26,120 --> 01:30:29,840 Oh, the frog's kid. That's hilarious. 1497 01:30:30,000 --> 01:30:32,680 Yeah, The funny thing is that like. 1498 01:30:32,720 --> 01:30:36,000 You actually like he has like some points in everything, 1499 01:30:36,000 --> 01:30:36,960 right? Like, there's like, it's not 1500 01:30:36,960 --> 01:30:39,800 like, it sounds absurd, but then you find out that actually, 1501 01:30:39,800 --> 01:30:42,480 yeah, there's been like some environmental changes and like 1502 01:30:42,480 --> 01:30:46,360 the fact actually, like, there are gays, there are frogs that 1503 01:30:46,360 --> 01:30:50,880 can change sex, you know? The problem, like I, I think why 1504 01:30:50,880 --> 01:30:54,040 he's like so entertaining, right, is like, I think he's got 1505 01:30:54,040 --> 01:30:57,520 some sort of autism or like Asperger's, right? 1506 01:30:57,800 --> 01:31:02,120 And he he has the like the right idea, but like the way he 1507 01:31:02,120 --> 01:31:05,520 delivers it, like you're turning the frogs gay instead of like 1508 01:31:05,520 --> 01:31:10,400 explaining what that means, like, you know what I mean? 1509 01:31:10,400 --> 01:31:14,560 Like he just jumps, he jumps, he jumps the gun and like delivers 1510 01:31:14,560 --> 01:31:18,360 it in like a really autistic manner instead of just like 1511 01:31:18,360 --> 01:31:22,880 communicating the problem. But he had he gets headlines 1512 01:31:22,880 --> 01:31:26,920 because of it. Strong, strong. 1513 01:31:28,560 --> 01:31:31,320 But like, yeah, what was, like the whole thing behind that? 1514 01:31:31,320 --> 01:31:34,640 I think it was like some pesticide that was being sprayed 1515 01:31:34,640 --> 01:31:39,920 on crops was like changing the genders of frogs, which is wild, 1516 01:31:39,920 --> 01:31:42,880 right? But like saying it that way 1517 01:31:42,880 --> 01:31:45,160 might like make people pay attention. 1518 01:31:46,280 --> 01:31:47,920 Yeah, yeah. Yeah. 1519 01:31:47,920 --> 01:31:53,720 Well, I think it doesn't always because like it's well, I don't 1520 01:31:53,720 --> 01:31:56,840 know much about it, just like from from, from things I've like 1521 01:31:57,320 --> 01:31:58,960 heard from podcast and stuff like that. 1522 01:31:59,240 --> 01:32:03,520 But it kind of like consumes a lot of information and it's like 1523 01:32:03,560 --> 01:32:04,880 you're getting lost in that, right. 1524 01:32:04,880 --> 01:32:08,040 So like you partly understand why you don't get the full 1525 01:32:08,040 --> 01:32:12,160 picture and you just kind of like know the the end numbers, 1526 01:32:12,160 --> 01:32:14,680 but you don't know the story. And I'm not sure he always like 1527 01:32:14,680 --> 01:32:17,400 knows the stories to kind of like, you know, like gets out to 1528 01:32:17,400 --> 01:32:19,920 the like the the numbers and it it just sounds absurd. 1529 01:32:19,920 --> 01:32:24,120 But then I just yeah. Yeah, yeah, I'm sure that's the 1530 01:32:24,120 --> 01:32:26,320 problem. Like you can only consume so 1531 01:32:26,320 --> 01:32:30,320 much data and retain it, right? And then like, it's like, I'll 1532 01:32:30,320 --> 01:32:36,400 randomly remember minor, like, weird, you know, stats or some 1533 01:32:36,400 --> 01:32:38,520 weird story. But I don't know how I got 1534 01:32:38,520 --> 01:32:40,360 there. And if I just like, randomly 1535 01:32:40,360 --> 01:32:42,880 blurted it out one day, no one would know what I was talking 1536 01:32:42,880 --> 01:32:48,960 about. But yeah, this has been a lot of 1537 01:32:48,960 --> 01:32:49,880 fun. I think. 1538 01:32:49,960 --> 01:32:52,000 I think we can, we can wrap this up. 1539 01:32:53,000 --> 01:32:55,320 Thank you so much for joining me today. 1540 01:32:55,680 --> 01:32:58,400 I really, it was great chatting, getting to know you. 1541 01:33:00,080 --> 01:33:03,760 Is there any words of advice or anything that you want to leave 1542 01:33:03,760 --> 01:33:08,360 our audience before we sign off? Like definitely check out Ash, 1543 01:33:08,960 --> 01:33:12,760 even if like it's, it's a little bit of a of a jump, but it's 1544 01:33:12,760 --> 01:33:14,880 it's it's so, so amazing. I would say, like I said, as a 1545 01:33:14,880 --> 01:33:21,400 project and yeah, right self right, your skills don't don't 1546 01:33:21,400 --> 01:33:25,240 go up coding your skills and commands like kind of like be 1547 01:33:25,240 --> 01:33:27,360 mindful, like it's really interesting. 1548 01:33:27,360 --> 01:33:31,360 Like words matter, right? So try to find the correct words 1549 01:33:31,360 --> 01:33:35,080 to to describe things and explore. 1550 01:33:35,960 --> 01:33:40,280 Yeah, I one thing I like, I want to add to that is I find myself 1551 01:33:40,280 --> 01:33:42,840 kind of like tweaking my skills a lot. 1552 01:33:43,640 --> 01:33:47,600 But like as you see things that don't work and like, you know, 1553 01:33:47,600 --> 01:33:52,080 if Claude or an LLM just kind of like goes off the board and 1554 01:33:52,080 --> 01:33:56,880 hallucinates in some crazy direction, stop it and like ask 1555 01:33:57,000 --> 01:34:01,440 what made it go that direction. And like a lot of times you can 1556 01:34:01,440 --> 01:34:05,120 like solve the problem and it'll be, yeah, it'll be like just 1557 01:34:05,400 --> 01:34:10,640 simple word in your skill MD that just this minor tweak could 1558 01:34:10,640 --> 01:34:13,640 probably solve the problem. Yeah. 1559 01:34:15,440 --> 01:34:19,040 So, yeah, if I'm like, for example, like it's, it's been 1560 01:34:21,080 --> 01:34:23,640 like there's been like papers that if you use words like 1561 01:34:23,640 --> 01:34:29,120 analyze it kind of like really can change the kind of like how 1562 01:34:29,120 --> 01:34:31,440 it approaches the problem, right. 1563 01:34:31,440 --> 01:34:33,200 So kind of like you really kind of like some, some, some 1564 01:34:33,200 --> 01:34:38,280 details, yeah, can matter in terms of how you phrase your, 1565 01:34:38,720 --> 01:34:41,840 your problems basically, yeah. Yeah, Yep, I agree. 1566 01:34:42,400 --> 01:34:47,360 Awesome. Well, Vassilis, thank you so 1567 01:34:47,360 --> 01:34:49,560 much for joining me. This was fun. 1568 01:34:49,880 --> 01:34:56,760 I hope your your elixir conf. What is the it's like before the 1569 01:34:56,760 --> 01:34:59,800 conference the the smaller workshop stuff? 1570 01:34:59,800 --> 01:35:02,920 I hope that goes well. I feel like that would be more 1571 01:35:02,920 --> 01:35:05,760 fun than like a big talk, at least less stressful. 1572 01:35:05,760 --> 01:35:07,400 Like you got smaller groups right? 1573 01:35:08,920 --> 01:35:12,600 I think so, yeah, like it's my first time kind of like doing 1574 01:35:12,600 --> 01:35:15,680 it. So yeah, it's longer, right? 1575 01:35:15,720 --> 01:35:20,800 It's like 8 hours. I feel really anxious like when 1576 01:35:20,800 --> 01:35:24,240 it's because I've also given a talk and in the past it's really 1577 01:35:24,240 --> 01:35:26,160 kind of like anxious when you know that it's just like we'll 1578 01:35:26,160 --> 01:35:29,040 have one hour and you have to kind of like do perfectly while 1579 01:35:29,040 --> 01:35:30,520 there's. I feel it's more relaxed. 1580 01:35:30,520 --> 01:35:34,040 You can like people are going to be doing work and you can kind 1581 01:35:34,040 --> 01:35:36,040 of like just observe what they're doing, right. 1582 01:35:36,040 --> 01:35:39,480 So it's definitely, I feel this kind of thing to be more 1583 01:35:39,480 --> 01:35:43,120 relaxed, yeah. Yeah, that'll be fun. 1584 01:35:43,480 --> 01:35:46,240 Well, enjoy. And it sounds like you'll get to 1585 01:35:46,240 --> 01:35:49,160 meet Deharma there, so that's cool. 1586 01:35:49,600 --> 01:35:54,360 And yeah, thanks so much. And for everyone that tuned in, 1587 01:35:54,360 --> 01:35:58,000 thank you for watching. And I'm taking off next week, 1588 01:35:58,080 --> 01:36:02,640 but then I'll be back at it with a guest to actually someone that 1589 01:36:02,640 --> 01:36:06,000 I work with that collects is coming on and it's going to be a 1590 01:36:06,000 --> 01:36:09,560 cool conversation. We're kind of changing the way 1591 01:36:09,560 --> 01:36:13,520 that we're writing code, which is kind of cool. 1592 01:36:13,520 --> 01:36:16,480 Like I think ID ES are dying because we're going to be 1593 01:36:16,480 --> 01:36:19,800 writing less and less code and it's more about reviewing code, 1594 01:36:19,800 --> 01:36:23,120 right? So little teaser there. 1595 01:36:23,120 --> 01:36:24,680 I think it'll be a fun conversation. 1596 01:36:24,680 --> 01:36:29,240 But yeah, I hope everyone has a good weekend and see you in two 1597 01:36:29,240 --> 01:36:33,560 weeks testing.