1 00:00:00,160 --> 00:00:03,680 Hey, I'm Kimberly with Elixir mentor Jacob recently added 2 00:00:03,680 --> 00:00:07,000 recruitment services to help companies hire pre bedded Elixir 3 00:00:07,000 --> 00:00:10,000 devs. I help with company outreach, so 4 00:00:10,000 --> 00:00:13,280 if you're hiring, I may be reaching out, or you can connect 5 00:00:13,280 --> 00:00:15,480 with us anytime at elixirmentor.com. 6 00:00:17,080 --> 00:00:19,680 Secure your Elixir apps with paraxial IO. 7 00:00:19,840 --> 00:00:22,880 From code analysis and bot defense to dependency checks, 8 00:00:22,880 --> 00:00:26,160 they've got you covered. Elixir first design with GitHub 9 00:00:26,160 --> 00:00:28,560 integration. Elevate your security at 10 00:00:28,560 --> 00:00:40,280 paraxial dot IO testing. What's up? 11 00:00:40,320 --> 00:00:42,440 I'm Jacob Litzow, your Elixir mentor. 12 00:00:42,440 --> 00:00:44,760 Welcome to another exciting episode of the Elixir Mentor 13 00:00:44,760 --> 00:00:46,560 podcast. This is where we discuss 14 00:00:46,560 --> 00:00:48,800 everything related to Elixir, from interviews with 15 00:00:48,800 --> 00:00:52,400 enthusiasts, pioneers in the community to innovative projects 16 00:00:52,400 --> 00:00:56,960 and library shaping Elixirs future Mike Energy tech founder 17 00:00:56,960 --> 00:01:01,000 and 30 year tech veteran, joins me today to talk about great 18 00:01:01,000 --> 00:01:03,400 infrastructure and the future of dev teams. 19 00:01:03,400 --> 00:01:07,800 And I'm sure we'll talk about a lot more than that, but thank 20 00:01:07,800 --> 00:01:11,720 you so much for joining me. Mike, if you could just give 21 00:01:11,720 --> 00:01:14,840 like a brief introduction of like your background and where 22 00:01:14,840 --> 00:01:16,200 you're coming from, that'd be awesome. 23 00:01:16,640 --> 00:01:18,360 Yeah, perfect. Thank you for having me on. 24 00:01:18,360 --> 00:01:21,680 I really appreciate it. So yeah, I've been a kind of a 25 00:01:21,720 --> 00:01:24,320 long and winding path in in my career. 26 00:01:24,320 --> 00:01:26,960 It's a little over 3 decades in the tech world. 27 00:01:29,280 --> 00:01:33,360 You know, I got involved in the kind of the Elixir ecosystem 28 00:01:33,520 --> 00:01:36,720 really pretty fairly early on, the kind of pre one dot O 29 00:01:36,720 --> 00:01:40,080 Phoenix types of things. I think a lot of that was a 30 00:01:40,080 --> 00:01:44,320 little bit of a reaction to some of the challenges that, you 31 00:01:44,320 --> 00:01:47,480 know, I'd experienced personally in trying to bring some larger 32 00:01:47,480 --> 00:01:52,760 scale apps in the like the Ruby on rail space into into 33 00:01:52,760 --> 00:01:56,840 existence. And, you know, back then we are 34 00:01:56,840 --> 00:01:59,960 we are actually trying to build like one of the first demand 35 00:01:59,960 --> 00:02:06,440 response applications, which is a grid service where we would, 36 00:02:07,040 --> 00:02:11,280 you know, you would pull to change their energy habits. 37 00:02:11,840 --> 00:02:15,080 And this is kind of back before like Nest thermostats and two 38 00:02:15,080 --> 00:02:16,680 way connected thermostats for the norm. 39 00:02:16,960 --> 00:02:20,840 We were trying to connect these these thermostats to to a rail 40 00:02:20,840 --> 00:02:24,400 VAP, which, you know, rail says great when you're talking about 41 00:02:25,280 --> 00:02:28,080 a kind of stateless connections. We were talking about hundreds 42 00:02:28,080 --> 00:02:30,120 of thousands of things always connected. 43 00:02:30,480 --> 00:02:32,880 It's the resource management, which wasn't there. 44 00:02:33,280 --> 00:02:36,920 We had an engineer that that found this MQTT thing that was 45 00:02:36,960 --> 00:02:38,920 built in Erlang. And at that, at that time, I 46 00:02:38,920 --> 00:02:41,120 didn't know anything about what the beam had to offer. 47 00:02:41,480 --> 00:02:45,720 And it basically just instantly took our scaling problems, you 48 00:02:45,720 --> 00:02:49,080 know, off the table on that front kind of went on the, you 49 00:02:49,080 --> 00:02:51,160 know, that one on the back burner for a while. 50 00:02:51,160 --> 00:02:56,320 And, you know, one thing led to another and, and I found found 51 00:02:56,320 --> 00:02:59,640 Elixir when we were trying to build a platform for doing 52 00:02:59,760 --> 00:03:02,440 distribution grid balancing. And I was really, really worried 53 00:03:02,440 --> 00:03:07,120 about trying to select the technology that would allow 54 00:03:07,120 --> 00:03:10,200 something to run 24/7 in the face of a lot of problems. 55 00:03:10,760 --> 00:03:14,160 And when I took a flyer on, on doing that and remembered a 56 00:03:14,160 --> 00:03:16,320 little bit about what we had done with the beam and was 57 00:03:16,440 --> 00:03:22,080 reading about that and that we, we started building that 58 00:03:22,080 --> 00:03:24,640 project. And honestly, I kind of never 59 00:03:24,640 --> 00:03:26,200 looked back. And it brought it into several 60 00:03:26,200 --> 00:03:28,960 other companies since then, really all in the grid space. 61 00:03:29,320 --> 00:03:31,600 You know, we've done things in the behind the meter, we've done 62 00:03:31,600 --> 00:03:33,760 things in the distribution grid, we've done things on the 63 00:03:33,760 --> 00:03:36,760 transmission grid. But it's just been such a 64 00:03:36,760 --> 00:03:40,160 perfect fit for this type of these types of problems. 65 00:03:40,520 --> 00:03:41,840 And I kind of haven't looked back. 66 00:03:41,880 --> 00:03:45,320 And so that's really, you know, the first half of my career is 67 00:03:45,320 --> 00:03:48,240 all about consulting and doing all kinds of different 68 00:03:48,240 --> 00:03:51,280 technology stacks. And then the last, you know, 15 69 00:03:51,280 --> 00:03:53,720 or so years, it's all been about the energy space. 70 00:03:53,720 --> 00:03:57,880 And probably the last 10, almost 10 have been, you know, beam 71 00:03:57,880 --> 00:04:00,960 related stuff. So kind of an interesting path I 72 00:04:00,960 --> 00:04:04,440 guess in some ways. It's it's funny that you're all 73 00:04:04,440 --> 00:04:07,600 over the place and then you discover the beam and elixir and 74 00:04:08,360 --> 00:04:10,560 it you stick to one tech after that. 75 00:04:11,800 --> 00:04:13,480 Yeah. I mean, when we first started 76 00:04:13,480 --> 00:04:15,040 doing it, you know, there was no live view. 77 00:04:15,040 --> 00:04:17,000 There was none, you know, none of the stuff that we kind of 78 00:04:17,000 --> 00:04:18,880 take for granted now. And it was still, you could 79 00:04:18,880 --> 00:04:21,360 still see a glimmer of like, hey, This is why this is going 80 00:04:21,360 --> 00:04:25,400 to matter, you know, and and it's just gotten better, you 81 00:04:25,400 --> 00:04:27,840 know, every project that we've that we've rolled it out on. 82 00:04:27,840 --> 00:04:31,080 And now this is five different companies that we've brought. 83 00:04:31,080 --> 00:04:35,120 It that's awesome. Yeah, I'm a I'm kind of a I'm 84 00:04:35,120 --> 00:04:36,920 late to the party. I don't think I really 85 00:04:36,920 --> 00:04:41,200 discovered Elixir until like 2020 right around there that. 86 00:04:41,480 --> 00:04:43,720 Was a really like opportune time to get into it. 87 00:04:43,720 --> 00:04:45,360 Things were really taken off at that point. 88 00:04:45,360 --> 00:04:48,200 And there was a lot of like, like live view was already like 89 00:04:48,200 --> 00:04:51,200 in a solid spot. I felt like and like everything 90 00:04:51,200 --> 00:04:56,000 was very usable at that time. Yeah, yeah, I made the mistake 91 00:04:56,160 --> 00:04:59,000 one of the in the early Elixir projects before we had live view 92 00:04:59,000 --> 00:05:03,360 that we, we connected an Elm front end to to Elixir. 93 00:05:03,400 --> 00:05:06,320 And the Elm is a really cool tech, but it just, it never 94 00:05:06,320 --> 00:05:08,080 really kind of evolved and went somewhere. 95 00:05:08,120 --> 00:05:12,960 And and I think, you know, that whole single page app thing just 96 00:05:12,960 --> 00:05:15,800 didn't really click with my with my view. 97 00:05:15,800 --> 00:05:18,440 And I'm so super glad that we've got the other options now. 98 00:05:19,160 --> 00:05:24,880 Yeah, I I'm a I'm a back end engineer at heart for sure, but 99 00:05:24,880 --> 00:05:28,800 since I've played around with live view I'm building more full 100 00:05:28,800 --> 00:05:31,800 stack projects these days, but I don't think I could do it 101 00:05:31,800 --> 00:05:37,240 without an LLM. They do a really good job with 102 00:05:37,240 --> 00:05:38,600 that. So why why not? 103 00:05:39,080 --> 00:05:44,080 That is that's super key. I think the the last, the, the 104 00:05:44,080 --> 00:05:46,720 app that we're building for this start up that I'm on, you know, 105 00:05:47,160 --> 00:05:51,400 90% of the UI, it was built very, very well with an LOL. 106 00:05:51,680 --> 00:05:55,960 It just kind of blows me away. Can can we talk about like what 107 00:05:56,000 --> 00:06:00,520 got you in like the infrastructure, like grid system 108 00:06:00,520 --> 00:06:02,600 space, like what what brought you there? 109 00:06:03,880 --> 00:06:09,320 Yeah, that was so I back when I was a consultant, there was a, a 110 00:06:09,560 --> 00:06:14,240 project that I did just a kind of a, a domain modeling kind of 111 00:06:14,240 --> 00:06:17,640 short term project trying to get them kind of look at their back 112 00:06:17,640 --> 00:06:19,960 end system and kind of reorganized how they were 113 00:06:20,040 --> 00:06:23,200 thinking about it. And and it was probably only a 114 00:06:23,200 --> 00:06:27,040 few months worth of work. But the guy that was kind of 115 00:06:27,040 --> 00:06:30,160 leading that apparently had a pretty good, you know, 116 00:06:30,160 --> 00:06:33,560 experience and cup few years later kind of called me out of 117 00:06:33,560 --> 00:06:35,800 the blue. And he's like, hey, I'm, I'm 118 00:06:35,800 --> 00:06:37,280 trying, I'm working at this company. 119 00:06:37,280 --> 00:06:38,760 It was called Converge at the time. 120 00:06:38,760 --> 00:06:40,280 It was a demand response company. 121 00:06:40,560 --> 00:06:45,240 And we really need to like build a software organization within 122 00:06:45,240 --> 00:06:47,040 this team or within this company. 123 00:06:47,240 --> 00:06:49,320 Would you be interested? And I'm like, yeah, that 124 00:06:49,320 --> 00:06:50,640 actually sounds pretty interesting. 125 00:06:50,640 --> 00:06:52,600 But I didn't have any experience in the grid. 126 00:06:52,600 --> 00:06:55,360 And so I was like, OK, well, I don't, I don't really know the 127 00:06:55,360 --> 00:06:57,760 domain very well, but I know how to build a a software 128 00:06:57,760 --> 00:06:59,080 organization. OK. 129 00:06:59,680 --> 00:07:04,680 And so we started that and it was, it was one of the more 130 00:07:04,680 --> 00:07:08,640 challenging domains that I ever had to learn because it was 131 00:07:08,640 --> 00:07:12,280 just, it grew up through these weird regulatory structures and, 132 00:07:12,280 --> 00:07:17,080 and, and, you know, technical evolutions and all these 133 00:07:17,080 --> 00:07:19,360 different, you know, weird influencing factors. 134 00:07:19,840 --> 00:07:23,280 And it was kind of like this view I had is like maybe the 135 00:07:23,360 --> 00:07:27,120 telecoms back in the 90s before they'd really modernized, there 136 00:07:27,120 --> 00:07:30,280 was just infinite problems that needed to be solved. 137 00:07:31,000 --> 00:07:34,360 And it also got me kind of excited on that, not just for 138 00:07:34,360 --> 00:07:37,720 the technical problems, but because, you know, as a species, 139 00:07:37,720 --> 00:07:40,240 we need to learn to use power in a little bit better, more 140 00:07:40,240 --> 00:07:43,200 responsible way. And so I kind of felt, started 141 00:07:43,200 --> 00:07:48,560 to feel this, I don't know this, like you were making a dent, a 142 00:07:48,560 --> 00:07:51,240 small dent in a really large important problem. 143 00:07:51,400 --> 00:07:53,960 And, you know, that was something relatively new that I 144 00:07:53,960 --> 00:07:56,360 hadn't experienced at a lot of the other types of companies or 145 00:07:56,360 --> 00:07:59,320 domains that I'd worked in. It was more just a job. 146 00:07:59,680 --> 00:08:01,800 And this ended up starting to feel a little bit more like a 147 00:08:01,800 --> 00:08:04,440 calling. And so, you know, as I poured 148 00:08:04,440 --> 00:08:06,840 myself more and more into learning the domain side of it, 149 00:08:06,920 --> 00:08:10,840 it just kind of enabled a lot more in my career and got a lot 150 00:08:10,840 --> 00:08:13,720 of opportunities to do, you know, different types of things 151 00:08:13,720 --> 00:08:16,640 and, you know, kind of climb that career ladder, for better 152 00:08:16,640 --> 00:08:20,080 or for worse, into like, you know, the CTO side of things. 153 00:08:20,080 --> 00:08:23,120 And now I kind of specialize in doing, you know, that type of 154 00:08:23,120 --> 00:08:26,880 stuff, But I really, really try hard to not be disconnected. 155 00:08:26,880 --> 00:08:31,000 CCTOI think suppose other people can judge how successful I am at 156 00:08:31,000 --> 00:08:34,400 that, But it's that's the part that really I really enjoy. 157 00:08:34,520 --> 00:08:38,919 It's why the whole elixir thing is being able to keep a hand in 158 00:08:38,919 --> 00:08:42,200 that very tightly and see what's going on with technology and be 159 00:08:42,200 --> 00:08:47,080 able to actually deploy it and and what's it makes it practical 160 00:08:47,080 --> 00:08:48,600 to do. That I was going to ask, are you 161 00:08:48,680 --> 00:08:50,720 do you still get to write code or are you too busy? 162 00:08:51,480 --> 00:08:54,400 Well, with this this start up that I'm doing, I haven't hired 163 00:08:54,800 --> 00:08:59,160 any engineers yet. So I've written the entire MVP 164 00:09:00,080 --> 00:09:02,280 and that would have never happened with the time 165 00:09:02,280 --> 00:09:05,320 commitments of starting a company without LLM. 166 00:09:05,320 --> 00:09:07,120 So that's kind of been shifting the entire. 167 00:09:07,120 --> 00:09:10,120 I feel I feel the same way because I've been building kill 168 00:09:10,120 --> 00:09:15,320 switch which is like a encrypted document storage with a dead man 169 00:09:15,320 --> 00:09:16,480 switch. I don't know if you've seen any 170 00:09:16,480 --> 00:09:20,160 of my posts, but I would not have been able to move as quick 171 00:09:20,160 --> 00:09:25,480 as I am without, you know, without the assistance of Claude 172 00:09:25,480 --> 00:09:27,360 code. Yeah. 173 00:09:27,360 --> 00:09:30,720 And it's, it's, it's crazy in that it's also completely 174 00:09:30,720 --> 00:09:36,920 upended, like just the idea of how much capital to raise and 175 00:09:36,920 --> 00:09:40,120 what the structure should be like, you know, that was a 176 00:09:40,440 --> 00:09:42,840 fairly well known quantity just a few years ago. 177 00:09:42,840 --> 00:09:45,520 It's like, OK, if you want an application, we're going to be, 178 00:09:45,640 --> 00:09:48,200 you know, hire the these many people, break them up into these 179 00:09:48,200 --> 00:09:51,640 teams and, and go forward. And now it's like that whole 180 00:09:51,640 --> 00:09:55,080 playbook is erased. Now the ability you have as a a 181 00:09:55,120 --> 00:09:59,440 single person dev team slash like I think we're going to see 182 00:09:59,440 --> 00:10:04,640 a lot of like a lot of one man 2 man shows that are building like 183 00:10:04,800 --> 00:10:09,680 significantly important, well structured powerhouse software 184 00:10:09,680 --> 00:10:13,160 platforms now too. I think you're absolutely right. 185 00:10:13,160 --> 00:10:17,520 And for me, a lot of that comes down to the this just 186 00:10:17,520 --> 00:10:20,800 astonishment of how much noise can be taken out of the system 187 00:10:20,800 --> 00:10:24,440 when one person can, if you can get the right person that has 188 00:10:24,440 --> 00:10:28,040 domain experience, has technical experience, it has, you know, 189 00:10:28,320 --> 00:10:31,040 organizational experience, maybe some testing experience. 190 00:10:31,040 --> 00:10:34,640 And there's no more communication between 4 humans 191 00:10:34,640 --> 00:10:38,280 to get all that done. That's the multiplier even on 192 00:10:38,280 --> 00:10:42,800 top of the LLM, that it's just hard to sort of fathom how well 193 00:10:42,800 --> 00:10:44,640 this is actually working in some cases. 194 00:10:45,400 --> 00:10:48,200 Yeah, yeah, it's wild. Like it's, I have a hard time. 195 00:10:48,200 --> 00:10:51,720 Like you see a lot of noise, especially on acts where you 196 00:10:51,720 --> 00:10:55,000 know, you have your, your love for LLMS and then you have so 197 00:10:55,000 --> 00:10:58,040 much hate for it, right. I don't get the hate. 198 00:10:58,040 --> 00:11:01,840 Like if you're using a tool like in a responsible manner, it 199 00:11:01,840 --> 00:11:03,600 doesn't really produce that much slop. 200 00:11:03,600 --> 00:11:07,000 And as long as you know when it is slop, I think you're in a 201 00:11:07,000 --> 00:11:09,720 good spot. Well, it's funny you say that 202 00:11:09,720 --> 00:11:12,120 because I think I, I've experienced the same thing. 203 00:11:12,160 --> 00:11:15,800 And you know, as I talked to a bunch of people that are kind of 204 00:11:15,800 --> 00:11:19,520 on that entire spectrum, there was an, it wasn't a completely 205 00:11:19,520 --> 00:11:22,480 consistent thread, but one of the things that I noticed was 206 00:11:22,480 --> 00:11:27,360 that a lot of the engineers that would struggle to talk to other 207 00:11:27,560 --> 00:11:31,640 humans in some like natural way are struggling to talk to the 208 00:11:31,640 --> 00:11:33,520 LLMS. In the way. 209 00:11:34,240 --> 00:11:36,960 To do that. And so it's this, it's this art 210 00:11:36,960 --> 00:11:40,560 of communication that has, if you're a really, really good 211 00:11:40,560 --> 00:11:43,440 communicator and really good at organizing your thoughts and 212 00:11:43,440 --> 00:11:47,760 expressing your ideas, you, you tend to see that you can keep 213 00:11:47,760 --> 00:11:50,880 them these, these tools on the rails a little bit more. 214 00:11:51,360 --> 00:11:53,760 But if you're if you already struggle with that and your team 215 00:11:53,760 --> 00:11:56,520 was never able to understand what was going on, then you're 216 00:11:56,520 --> 00:11:58,320 probably having that same experience with Lol. 217 00:11:58,320 --> 00:12:00,640 That's wild. I never thought about that. 218 00:12:00,640 --> 00:12:02,960 Like, yeah, soft skills, they matter. 219 00:12:03,800 --> 00:12:07,400 Matter almost more now that's the thing that like talking to 220 00:12:07,400 --> 00:12:10,160 fewer humans where the soft skills almost matter more. 221 00:12:10,160 --> 00:12:12,280 And it's. I don't know that it's 100% 222 00:12:12,280 --> 00:12:14,720 true, but I I I'm feeling a thread there. 223 00:12:15,040 --> 00:12:18,520 I feel like that makes sense because I've always, like, I've 224 00:12:18,520 --> 00:12:22,480 always joked, like I don't think I'll ever be like some framework 225 00:12:22,480 --> 00:12:25,160 creator. Like I, I'm a decent engineer. 226 00:12:25,200 --> 00:12:28,800 I know where my limits are. I can build, I can build 227 00:12:28,800 --> 00:12:30,680 anything I like if I put my mind to it. 228 00:12:30,680 --> 00:12:33,120 I can build anything that comes into my mind, right? 229 00:12:33,640 --> 00:12:38,200 But I do I, I'm not going to like be inventing software, you 230 00:12:38,200 --> 00:12:41,240 know, deep level things. But I've always joked, like my 231 00:12:41,240 --> 00:12:45,360 superpower, I can communicate between those kind of people and 232 00:12:45,360 --> 00:12:48,640 like a manager and like non-technical people, I can 233 00:12:48,640 --> 00:12:51,560 bridge the gap. I think that's becoming a more 234 00:12:51,560 --> 00:12:55,360 and more valuable skill every single day as as our world is 235 00:12:55,360 --> 00:12:57,200 just kind of shifting underneath us like that. 236 00:12:57,760 --> 00:13:00,440 That's where you're the jobs are still going to be. 237 00:13:00,440 --> 00:13:05,080 And yeah, it's, it's, it's sad in some ways, but it's also 238 00:13:05,120 --> 00:13:07,440 just, it's the reality of what's happening and we can either 239 00:13:07,440 --> 00:13:09,120 fight it or we can try to go with it. 240 00:13:09,440 --> 00:13:11,400 Yeah. Like I feel like I can dig deep. 241 00:13:11,440 --> 00:13:13,960 I can like, we can have the nerdiest conversation you want 242 00:13:13,960 --> 00:13:17,480 to have, or we can have a more macro conversation at a higher 243 00:13:17,480 --> 00:13:21,920 level and still like understand what's happening at least. 244 00:13:22,520 --> 00:13:26,160 Yeah, no, that's, that is a bit definitely a superpower. 245 00:13:26,160 --> 00:13:29,400 Stick with that. I, I don't, I mean, that's all I 246 00:13:29,400 --> 00:13:36,000 got, you know? That's why, that's why I have, 247 00:13:36,160 --> 00:13:38,880 that's why I started a podcast just to talk to smarter people 248 00:13:39,120 --> 00:13:40,920 than me, you know? That's cool. 249 00:13:41,520 --> 00:13:46,160 Hoping it rubs off, you know? Well, that's, that's where we we 250 00:13:46,160 --> 00:13:49,160 all tend to learn is when we can get together and and share these 251 00:13:49,160 --> 00:13:50,480 ideas. I know that's where I get most 252 00:13:50,480 --> 00:13:53,240 of my inspiration from is listening to other folks 253 00:13:53,240 --> 00:13:55,240 pontificate about things. And even if I don't agree with 254 00:13:55,240 --> 00:13:58,400 them, because we were talking before we started like it, it's 255 00:13:58,400 --> 00:14:02,000 totally fine to to listen to somebody and go, yeah, that's 256 00:14:02,000 --> 00:14:03,520 not for me. I don't, I don't agree, but 257 00:14:04,000 --> 00:14:05,840 they'll learn something. Disagreements. 258 00:14:05,840 --> 00:14:09,040 OK, Just do it civilly, right? Yeah, exactly. 259 00:14:09,440 --> 00:14:15,560 Exactly. Yeah, I'm, I'm kind of curious 260 00:14:15,560 --> 00:14:18,360 too. So you're, you were telling me 261 00:14:18,360 --> 00:14:22,000 like your startup is more based around like data center 262 00:14:22,040 --> 00:14:24,600 infrastructure now, what does that like, what's that mean? 263 00:14:24,920 --> 00:14:30,920 Yeah, so it's a good question. So the the start up that that 264 00:14:30,920 --> 00:14:34,160 I'm working on now is called Gridvar and it's basically a 265 00:14:34,160 --> 00:14:38,480 company that is trying to attack what we call the interconnection 266 00:14:39,080 --> 00:14:41,680 Q problem. And if you think about like 267 00:14:41,680 --> 00:14:45,280 these take for example, these hyperscale data centers, it's 268 00:14:45,280 --> 00:14:49,320 not just those, but just very, very large loads when they 269 00:14:49,400 --> 00:14:52,560 attempt to connect to the grid. It's not like, you know, a 270 00:14:52,560 --> 00:14:55,800 random residential place like turning on the meter and hey, 271 00:14:55,800 --> 00:15:03,160 you got power. It's it's a process and there's 272 00:15:03,520 --> 00:15:07,400 a lot of times the real power or like, Hey, how many megawatts do 273 00:15:07,400 --> 00:15:10,320 you need is like the first and the simplest question. 274 00:15:10,320 --> 00:15:12,000 It's like, oh, it's a 500 MW data center. 275 00:15:12,000 --> 00:15:14,120 Well, we know we're going to have to do, you know, a 276 00:15:14,120 --> 00:15:16,560 substation upgrade or whatever it is in order to make that 277 00:15:16,560 --> 00:15:19,680 happen. And then so somebody writes a 278 00:15:19,680 --> 00:15:24,440 big check, they buy land, they start, you know, acquiring TP us 279 00:15:24,440 --> 00:15:27,320 and all this stuff. And then they get deep in the 280 00:15:27,320 --> 00:15:30,960 cycle and, and the utility contracts with the power systems 281 00:15:30,960 --> 00:15:33,440 engineer and says, Hey, can you make sure this is going to be 282 00:15:33,440 --> 00:15:35,560 OK? And they come back and they say, 283 00:15:35,560 --> 00:15:37,880 hey, look, you're going to have these harmonic problems. 284 00:15:37,880 --> 00:15:41,840 You have this transient issue, you've got a voltage excursion 285 00:15:41,840 --> 00:15:45,040 over here. And like all these other power 286 00:15:45,040 --> 00:15:48,360 quality issues. And then the, the people writing 287 00:15:48,360 --> 00:15:49,800 the checks are like, what, what do you mean? 288 00:15:49,800 --> 00:15:54,000 Like we, we, it's 500 megawatts. You said you had 500 megawatts. 289 00:15:54,000 --> 00:15:57,840 Why can't we connect and there's millions of dollars are just 290 00:15:57,840 --> 00:15:59,720 locked up, unable to move forward. 291 00:16:00,320 --> 00:16:03,680 And So what we're trying to do is build software that can run 292 00:16:03,680 --> 00:16:07,640 all of that deep power systems analysis that you usually had to 293 00:16:07,640 --> 00:16:11,080 contract with a power systems engineer and pay hundreds of 294 00:16:11,080 --> 00:16:13,280 thousands of dollars and wait months for that study. 295 00:16:14,320 --> 00:16:17,320 We're making it so that a non-technical user can run that 296 00:16:17,320 --> 00:16:22,120 same analysis and get all those answers upfront so they can bake 297 00:16:22,120 --> 00:16:25,200 that into their interconnection process. 298 00:16:26,360 --> 00:16:29,040 That's really cool. So you're essentially are you 299 00:16:29,040 --> 00:16:35,200 just like integrating with like like physical, like kind of IoT 300 00:16:35,680 --> 00:16:39,000 capable, like you're with sensors and meters and whatever 301 00:16:39,000 --> 00:16:41,840 it may be? Actually a little bit more 302 00:16:41,840 --> 00:16:46,720 mathematically based. And so you can, there's 303 00:16:46,720 --> 00:16:52,040 different equations of how these AC electrical systems actually 304 00:16:52,040 --> 00:16:54,440 behave under different conditions and different 305 00:16:54,440 --> 00:16:58,440 contingencies and you know, load fluctuations and generation 306 00:16:58,440 --> 00:17:00,800 fluctuations and like understanding if there's 307 00:17:00,800 --> 00:17:04,720 problems there. And so it's a lot of like deep 308 00:17:05,079 --> 00:17:07,880 numerical analysis to to run this. 309 00:17:08,200 --> 00:17:11,240 And the problem with, with today's apps that do this, which 310 00:17:11,240 --> 00:17:13,079 I was just blown away. I didn't realize this until I 311 00:17:13,079 --> 00:17:15,480 got into this. The start up is that they're all 312 00:17:15,480 --> 00:17:18,520 kind of these desktop apps that these power systems engineers 313 00:17:18,520 --> 00:17:21,880 use to like explore things. And so and they're pretty 314 00:17:22,119 --> 00:17:28,040 involved calculations that they can be and nobody has taken that 315 00:17:28,040 --> 00:17:31,160 into like data center level compute where it's massively 316 00:17:31,160 --> 00:17:35,120 parallelized and broken down. And so now you know if it, if 317 00:17:35,120 --> 00:17:37,080 they, you know if a power assistance engineer wanted to 318 00:17:37,080 --> 00:17:40,480 run like a few 100 scenarios, it might take them like several 319 00:17:40,480 --> 00:17:43,800 weeks We're now. They're doing this all manually, 320 00:17:43,920 --> 00:17:47,160 essentially. Manually some Python scripts and 321 00:17:47,160 --> 00:17:48,760 some like. Excel. 322 00:17:50,640 --> 00:17:54,160 And then, you know, so the we can run hundreds of thousands of 323 00:17:54,160 --> 00:17:56,200 scenarios, you know, within minutes. 324 00:17:57,760 --> 00:18:01,520 I wonder if that has to do with the processes and the actor 325 00:18:01,520 --> 00:18:04,400 model A. Little bit of that exactly. 326 00:18:04,520 --> 00:18:06,400 And so the ability to parallelize all these 327 00:18:06,400 --> 00:18:10,160 computations, although I got to admit we are calling out to Rust 328 00:18:10,160 --> 00:18:13,440 to do a lot of the low level. That makes sense though, just 329 00:18:13,440 --> 00:18:17,680 because you're doing a lot of computation, right? 330 00:18:18,320 --> 00:18:22,080 Yeah, I really wanted to to use a lot of the NX stuff to do it, 331 00:18:22,080 --> 00:18:24,720 but the particular things that we're doing aren't well covered 332 00:18:25,600 --> 00:18:27,600 at this moment in time. So I'm going to keep an eye on 333 00:18:27,600 --> 00:18:31,120 that. But you know, for the most part 334 00:18:31,120 --> 00:18:35,720 it's just these like low level numerical solvers that we're 335 00:18:35,720 --> 00:18:37,760 calling out to and and that's fine. 336 00:18:38,160 --> 00:18:40,800 It works great. You know, there's not a lot of 337 00:18:40,800 --> 00:18:43,840 REST code there, but what's there runs super fast and and 338 00:18:43,840 --> 00:18:48,880 works really really well. I'm I'm guessing then you're 339 00:18:48,880 --> 00:18:53,040 just like integrating with. Is it like like NIFS with Rust 340 00:18:53,040 --> 00:18:55,080 or how are you doing that? Yeah. 341 00:18:56,000 --> 00:18:59,840 So this is another thing where the the LMS really kind of saved 342 00:18:59,840 --> 00:19:02,120 me because I didn't have a lot of Rust experience and I 343 00:19:02,120 --> 00:19:04,560 certainly didn't have a lot of Rust experience using like 344 00:19:04,560 --> 00:19:06,480 Rustler and like things like that to. 345 00:19:08,000 --> 00:19:12,000 And you know, that was a little involved to get it work, 346 00:19:12,000 --> 00:19:13,760 especially in the deployment side of things. 347 00:19:14,840 --> 00:19:20,480 But yeah, it's, you know, they, they are just Rust Nifs and, 348 00:19:20,600 --> 00:19:24,680 and, and that actually works fairly well. 349 00:19:24,680 --> 00:19:26,960 I've been, I've been pretty impressed, although I got to 350 00:19:26,960 --> 00:19:29,960 admit I'm not a Rust expert. So there might be somebody come 351 00:19:29,960 --> 00:19:32,120 in and look at our code base and go, Oh my God, what are you 352 00:19:32,120 --> 00:19:34,920 doing? But but it does run fairly 353 00:19:34,920 --> 00:19:36,800 performantly. OK, nice. 354 00:19:37,000 --> 00:19:40,800 I've I've heard that. So I have no experience writing 355 00:19:40,800 --> 00:19:44,480 Nifs or anything, but I've heard that you kind of, you're kind of 356 00:19:44,480 --> 00:19:48,640 forced to manually handle air errors and retries and you kind 357 00:19:48,680 --> 00:19:53,240 of, you have to, you can't rely on all the great things of the 358 00:19:53,240 --> 00:19:55,520 fault tolerance that we get with Beam, right? 359 00:19:56,240 --> 00:19:57,640 Yeah. I mean, to a certain degree, I 360 00:19:57,640 --> 00:19:59,840 think if you architect it correctly, whether you've got a 361 00:19:59,840 --> 00:20:02,920 thin, you know, supervisor wrapper around it, that's pretty 362 00:20:02,920 --> 00:20:06,920 easy to recover from it because I've seen it kernel panic in the 363 00:20:06,920 --> 00:20:12,360 Rust NIF and recover OK. Appropriately so you know, you 364 00:20:12,360 --> 00:20:14,520 lose that calculation, but that's no big deal. 365 00:20:14,520 --> 00:20:16,240 I mean we've paralyzed thousands of them. 366 00:20:16,240 --> 00:20:18,360 So the fact that one of them died is no big deal. 367 00:20:19,280 --> 00:20:22,840 But yeah, it's actually been OK, at least the way we have, we 368 00:20:22,840 --> 00:20:27,120 have it built. It is the the recovery from 369 00:20:27,120 --> 00:20:30,960 those kernel panics in inside of that has has not taken anything 370 00:20:30,960 --> 00:20:32,640 else down and we've been able to keep moving. 371 00:20:33,920 --> 00:20:38,200 That's so fascinating and so like you're just building a 372 00:20:38,280 --> 00:20:42,120 fancy calculator essentially. Yeah. 373 00:20:42,120 --> 00:20:44,920 I mean if you think about it like a lot of those, you know 374 00:20:44,920 --> 00:20:48,440 those math routines are, are pretty low down in there. 375 00:20:48,440 --> 00:20:51,080 And then there's another layer of the system that's a bunch of 376 00:20:51,280 --> 00:20:55,600 AI stuff for kind of exploring different optimization solutions 377 00:20:55,600 --> 00:20:58,720 like hey, should we put a capacitor bank here or statcom 378 00:20:58,720 --> 00:21:00,720 there or battery system over here? 379 00:21:01,000 --> 00:21:04,760 And like trying to find the minimum configuration of cost 380 00:21:04,760 --> 00:21:07,800 and lead time and all these other things in order to solve 381 00:21:07,800 --> 00:21:09,960 all, all the contingencies that you find. 382 00:21:10,240 --> 00:21:12,320 And that's a little bit more interesting problem and stays 383 00:21:12,320 --> 00:21:14,280 within the Elixir ecosystem more. 384 00:21:14,280 --> 00:21:19,160 But yeah, that's kind of another layer to the whole problem set 385 00:21:19,160 --> 00:21:21,920 there. Like so is there anything like 386 00:21:21,920 --> 00:21:27,040 dynamic that you need to feed in that varies data center to data 387 00:21:27,040 --> 00:21:31,560 center other than like power consumption and things like 388 00:21:31,560 --> 00:21:33,880 that? Yeah, if you think like one of 389 00:21:33,880 --> 00:21:38,720 the interesting things is it's just like the, the load profile, 390 00:21:39,000 --> 00:21:43,920 like how quickly does something shift 'cause that changes a lot 391 00:21:43,920 --> 00:21:47,120 of the dynamic nature of how you create different harmonics and 392 00:21:47,120 --> 00:21:51,800 stuff. And so, you know, being able to 393 00:21:51,800 --> 00:21:56,160 like, and I had something I, I had to learn like understanding 394 00:21:56,160 --> 00:22:00,480 when an AI data center is doing inference versus when it's doing 395 00:22:00,480 --> 00:22:04,360 training and how those load profiles look completely 396 00:22:04,360 --> 00:22:07,560 different and why. And I thought I had it 397 00:22:07,560 --> 00:22:10,000 completely reversed in my mind when I first approached it. 398 00:22:10,000 --> 00:22:12,440 I was like, oh, well, inference would just be all over the board 399 00:22:12,760 --> 00:22:14,840 and then the training would be much more consistent. 400 00:22:16,760 --> 00:22:20,680 And it's actually the opposite. The, the, the training has these 401 00:22:20,680 --> 00:22:24,080 weird spikes 'cause when you go through and like have to reload 402 00:22:24,080 --> 00:22:27,120 all the GPUs with more data, like the whole thing comes 403 00:22:27,120 --> 00:22:30,000 crashing down for like a few seconds and then ramps right 404 00:22:30,000 --> 00:22:31,920 back up. And that is like the worst thing 405 00:22:31,920 --> 00:22:34,880 in the world you can do the grid is to have like this 500 MW 406 00:22:34,880 --> 00:22:37,600 swing back and forth, back and forth, back and forth. 407 00:22:37,600 --> 00:22:39,240 And it scares utilities to no end. 408 00:22:39,240 --> 00:22:43,360 And it's why they Nix these connections all the time. 409 00:22:44,040 --> 00:22:46,800 Of interest in like these are like, yeah, you don't realize 410 00:22:46,800 --> 00:22:50,160 how many problems exist at this level, like these are we just 411 00:22:50,160 --> 00:22:53,200 take the like, I take it for granted that things just work, 412 00:22:53,240 --> 00:22:56,600 right. I can't believe, like for me 413 00:22:56,600 --> 00:23:00,240 now, having been in the in the energy space for almost 15 414 00:23:00,240 --> 00:23:03,800 years, I'm stunned when a light comes on and I flip the switch. 415 00:23:04,080 --> 00:23:07,480 There's just so much weirdness going on behind the scenes that 416 00:23:07,480 --> 00:23:09,400 it's just, it's amazing that anything works. 417 00:23:09,760 --> 00:23:13,800 And you always hear like, it's like how fragile some of these 418 00:23:13,800 --> 00:23:16,800 systems can be, but yet they're still very resilient. 419 00:23:16,800 --> 00:23:21,360 And like, you know, I rarely lose power or even like Internet 420 00:23:21,360 --> 00:23:23,560 connection so. Yeah. 421 00:23:23,560 --> 00:23:26,440 And it it, there's some really heroic work being done by some 422 00:23:26,440 --> 00:23:29,240 really talented engineers to kind of to keep things 423 00:23:29,320 --> 00:23:32,280 operating. And that is there's a lot that 424 00:23:32,280 --> 00:23:33,920 goes on. I have tremendous amount of 425 00:23:33,920 --> 00:23:36,040 respect for a lot of the power systems engineers that are 426 00:23:36,040 --> 00:23:36,640 sweating. Out. 427 00:23:37,880 --> 00:23:40,360 Those details, it's, it's important. 428 00:23:40,680 --> 00:23:44,280 I feel like it, like it's such it's an industry that's been 429 00:23:44,280 --> 00:23:48,200 around for so long. Do you find that they they're 430 00:23:48,200 --> 00:23:56,800 slow on adopting technology? Yes, utilities, especially 431 00:23:56,800 --> 00:24:00,160 utilities are very slow on all fronts. 432 00:24:00,760 --> 00:24:02,360 They're, you know, they're regulated entities. 433 00:24:02,360 --> 00:24:07,520 They're and they're, they're very scared of different new 434 00:24:07,520 --> 00:24:09,960 technologies. And that's a little bit of a 435 00:24:09,960 --> 00:24:12,400 blanket statement, but you do see that there's different 436 00:24:12,400 --> 00:24:14,640 geographies, maybe a little bit more forward thinking than 437 00:24:14,640 --> 00:24:17,840 others. But I think it's a pragmatic 438 00:24:17,960 --> 00:24:19,880 place to be. When you have something as 439 00:24:19,960 --> 00:24:23,320 important as the power grid, you don't just decide to go out and 440 00:24:23,600 --> 00:24:26,440 like you can't. Update everything to the latest 441 00:24:26,440 --> 00:24:27,960 and greatest because you could break a lot. 442 00:24:28,720 --> 00:24:29,880 Exactly. And that's kind of the one of 443 00:24:29,880 --> 00:24:32,200 the reasons why I'm a little bit excited about this project 444 00:24:32,200 --> 00:24:36,920 because I'm not doing like control systems on the grid like 445 00:24:36,920 --> 00:24:40,280 we have before. And so it's almost like a breath 446 00:24:40,280 --> 00:24:43,000 of fresh air where hey, I can, I can maybe actually have an error 447 00:24:43,000 --> 00:24:44,280 and, and it will be the end of the world. 448 00:24:44,280 --> 00:24:46,600 There won't be like a major blackout or something like that. 449 00:24:47,520 --> 00:24:52,200 Do you implement LLMS or anything in your analysis with 450 00:24:52,200 --> 00:25:00,640 your current project? We, we don't, there's some what 451 00:25:00,640 --> 00:25:03,480 I would more consider like machine learning type of things 452 00:25:03,480 --> 00:25:09,280 going on, but we don't have like a, an LLM interface today, 453 00:25:09,280 --> 00:25:11,480 although we've, we've been scratching our heads about like, 454 00:25:11,480 --> 00:25:14,440 is there a, is there a good way to have a conversational, you 455 00:25:14,440 --> 00:25:16,600 know, approach to interacting with this? 456 00:25:16,600 --> 00:25:18,960 Because this is really a back and forth conversation about 457 00:25:18,960 --> 00:25:22,320 where should we buy land or how should we connect and like this. 458 00:25:22,600 --> 00:25:25,440 So it could lend itself to that type of interface. 459 00:25:27,040 --> 00:25:30,480 But it, you know, I'm kind of trying to hold back and get the, 460 00:25:30,680 --> 00:25:33,240 the pure mathematical infrastructure in place. 461 00:25:33,560 --> 00:25:37,000 And I think what we may end up having is being able to extract 462 00:25:37,000 --> 00:25:43,400 a lot of these core mathematical concepts into a kind of a fleet 463 00:25:43,400 --> 00:25:47,880 of agents that can collaborate to find interesting solutions on 464 00:25:47,880 --> 00:25:49,840 their own. And so I think that's kind of 465 00:25:49,840 --> 00:25:52,440 the direction we'll end up going Version 2. 466 00:25:53,880 --> 00:25:56,640 And but I you know, we need each one of those things to be 467 00:25:56,640 --> 00:26:00,280 proven, like how they solve individual little problems 468 00:26:00,280 --> 00:26:02,680 before we start turning them loose trying to collaborate. 469 00:26:03,120 --> 00:26:06,200 Yeah, I didn't, I, I didn't have time, but I was supposed to 470 00:26:06,200 --> 00:26:11,920 create a sponsor video for for Jeto Michael Stettler's project. 471 00:26:13,520 --> 00:26:18,880 He's been making some videos for my YouTube channel and in 472 00:26:18,880 --> 00:26:23,680 exchange he gets sponsored and I don't know if you're familiar 473 00:26:23,680 --> 00:26:27,000 with that project at all, but. Absolutely familiar, as a matter 474 00:26:27,000 --> 00:26:29,400 of fact, that you know, I've been doing a little bit of a 475 00:26:29,400 --> 00:26:32,360 technology spike on the side trying to explore that exact 476 00:26:32,360 --> 00:26:37,200 idea using Cheeto. So I'd be surprised if this 477 00:26:37,200 --> 00:26:39,240 comes to bear if it ends up using that framework. 478 00:26:39,960 --> 00:26:42,720 Yeah, very awesome. Yeah, MM is doing some awesome 479 00:26:42,760 --> 00:26:45,160 things here. Let me give him a let me give 480 00:26:45,160 --> 00:26:51,240 him a plug since I I own that Let's see coding Elixir with AI 481 00:26:51,240 --> 00:26:53,520 or building AI into your Elixir apps. 482 00:26:53,800 --> 00:26:58,120 The Elixir AI collective discord is your community for both, and 483 00:26:58,120 --> 00:27:03,640 you can find that link at agent gito dot XYZ forward slash 484 00:27:03,640 --> 00:27:06,720 discord. So join that discord guys. 485 00:27:07,040 --> 00:27:12,000 It's a awesome group of people all nerding out on LMS and a 486 00:27:12,000 --> 00:27:15,480 bunch of projects all based around it and feel like Elixir 487 00:27:15,480 --> 00:27:17,400 is just the perfect fit for this stuff. 488 00:27:18,880 --> 00:27:22,160 It is I I couldn't agree more. And so at this point, I can't 489 00:27:22,160 --> 00:27:24,520 really imagine what type of technology stack is going to 490 00:27:24,520 --> 00:27:27,520 have to come along to displace me to take me down a different 491 00:27:27,520 --> 00:27:28,720 path. I'll probably be long dead by 492 00:27:28,720 --> 00:27:31,520 the time that happens. Yeah, yeah. 493 00:27:31,560 --> 00:27:35,120 I don't know what would have to change because just writing an 494 00:27:35,120 --> 00:27:37,440 Elixir is just so enjoyable for me too. 495 00:27:37,520 --> 00:27:42,880 And then just having like message passing and processes 496 00:27:42,880 --> 00:27:46,520 and you can just leverage so much and do so much cool stuff 497 00:27:46,520 --> 00:27:48,480 in parallel. Yeah. 498 00:27:48,920 --> 00:27:52,360 What I think my experience has also been that it's just the 499 00:27:52,360 --> 00:27:55,760 quality of the code that you get out of LLMS is so much better in 500 00:27:55,760 --> 00:27:58,520 the Elixir world. Maybe that's some of that, 501 00:27:58,520 --> 00:28:00,800 because I know how to correct it a little bit better than with 502 00:28:00,800 --> 00:28:04,760 other languages, but it still just seems to be a little more 503 00:28:04,840 --> 00:28:05,760 on top of things. Oh. 504 00:28:08,720 --> 00:28:15,320 Dang it, it's so I use restream to do these and everything in 505 00:28:15,320 --> 00:28:20,520 OBSI can't use keyboard shortcuts for copying and 506 00:28:20,520 --> 00:28:22,720 pasting. I had to right click and paste. 507 00:28:22,720 --> 00:28:26,000 It's the most annoying thing but I just shared. 508 00:28:26,000 --> 00:28:28,880 I just shared the link for everyone in the in the chat to 509 00:28:29,280 --> 00:28:31,840 get on the judo discord. Cool. 510 00:28:32,480 --> 00:28:37,000 But but yeah, cool project. And right when you said like, 511 00:28:37,000 --> 00:28:39,960 you know, I want a bunch of agents that can all message pass 512 00:28:39,960 --> 00:28:43,040 and communicate, I was like, I know a guy. 513 00:28:43,840 --> 00:28:46,680 Yeah, exactly. Well, we'll, we'll see. 514 00:28:46,720 --> 00:28:49,280 I, I have AI have a pretty strong feeling that it won't get 515 00:28:49,280 --> 00:28:51,480 accepted. But there's a, there's a talk 516 00:28:51,480 --> 00:28:57,000 that I, I put a proposal in for Elixir comp EU and it, it kind 517 00:28:57,000 --> 00:28:59,960 of talked, it's about this complexity and like how 518 00:28:59,960 --> 00:29:02,600 complexity can be a great thing if it's, if it's emergent 519 00:29:02,600 --> 00:29:04,800 complexity and and kind of that, that type of thing. 520 00:29:04,800 --> 00:29:08,240 And it, and it all kind of builds up to showing a 521 00:29:08,280 --> 00:29:12,880 simulation of a grid being of a like a 10,000 point grid being 522 00:29:12,880 --> 00:29:15,800 balanced through a bunch of agents just communicating about 523 00:29:15,880 --> 00:29:19,000 their state back and forth. And like hitting these like 524 00:29:19,280 --> 00:29:23,520 dropping, you know, kind of like a, a, a rock in the water and 525 00:29:23,520 --> 00:29:27,680 just seeing the wave go through the, the grid and you know, like 526 00:29:27,680 --> 00:29:29,600 a voltage disturbance. And then how they all 527 00:29:29,600 --> 00:29:31,320 collaborate and bring it back to normal. 528 00:29:31,320 --> 00:29:33,280 And it's just, it's just fascinating how you can write 529 00:29:33,280 --> 00:29:37,200 these really, really simple things that have very simple 530 00:29:37,200 --> 00:29:40,960 rules and you get this beautiful emergent behavior out of it just 531 00:29:40,960 --> 00:29:43,680 by the simple collaboration. And so when I was playing with 532 00:29:43,680 --> 00:29:45,280 that, I'm like, this is just amazing. 533 00:29:45,560 --> 00:29:48,440 And you know, maybe I can kind of turn this into a talk so. 534 00:29:48,800 --> 00:29:54,440 That would be awesome and just like the like the like the idea 535 00:29:54,440 --> 00:29:58,520 behind like this distributed nodes and things. 536 00:29:58,520 --> 00:30:03,400 And then sorry, excuse me, I've had a freaking cough for a 537 00:30:03,400 --> 00:30:08,440 month. I hate it but the idea behind 538 00:30:08,440 --> 00:30:12,600 just how easy it is to message, pass and communicate between 539 00:30:13,120 --> 00:30:16,240 distributed nodes. Or I guess like you would have. 540 00:30:18,280 --> 00:30:20,960 I guess you could just spin up agents as well, but you would 541 00:30:20,960 --> 00:30:23,560 probably still have a node at each location. 542 00:30:24,600 --> 00:30:25,280 Yeah. I mean, I think there's a 543 00:30:25,280 --> 00:30:26,960 million ways you could you could solve that. 544 00:30:27,840 --> 00:30:30,680 Yeah, I was, I was playing more like just experimenting with the 545 00:30:30,680 --> 00:30:32,000 Asian thing. And so it was. 546 00:30:32,600 --> 00:30:34,920 It's kind of natural there. Yeah, that's cool. 547 00:30:35,040 --> 00:30:39,760 I like that. Let's see. 548 00:30:45,400 --> 00:30:50,080 I, I kind of find it fascinating whenever people do like more 549 00:30:50,360 --> 00:30:52,760 like physical. So like before I got fully into 550 00:30:52,760 --> 00:30:55,760 the software world, I was involved in manufacturing. 551 00:30:55,760 --> 00:31:02,320 I did a lot of like automation and like PLC's and like linear 552 00:31:02,320 --> 00:31:04,080 drives, all that, all that stuff. 553 00:31:04,080 --> 00:31:08,560 And it's cool to, you know, do the timing and program machinery 554 00:31:08,560 --> 00:31:14,040 that works together. And I, I saw that in your, in, 555 00:31:14,160 --> 00:31:18,920 in your in a past gig, you did a lot of like controllers with 556 00:31:18,920 --> 00:31:21,000 batteries and wind farms and things. 557 00:31:21,000 --> 00:31:27,040 And I'm curious that was, that was all elixir and everything 558 00:31:27,040 --> 00:31:29,200 right for controlling that. And was it? 559 00:31:29,200 --> 00:31:31,600 Did you use nerves? Yeah. 560 00:31:31,600 --> 00:31:35,560 So I mean, there's been a evolution of my experience in 561 00:31:35,560 --> 00:31:39,080 the hardware world because I was not a hardware guy from like 562 00:31:39,080 --> 00:31:42,200 that. I, I originally like took over a 563 00:31:42,200 --> 00:31:44,920 hardware team back in my converse days and I had know 564 00:31:44,960 --> 00:31:49,480 nothing about it and hardware and firmware. 565 00:31:49,480 --> 00:31:52,400 And so we've, it's kind of come in and out of my career in a 566 00:31:52,400 --> 00:31:56,280 couple different ways. I would say that they're like 567 00:31:56,280 --> 00:31:59,320 the, the big batteries and wind farms, like interfacing with 568 00:31:59,320 --> 00:32:01,240 those is a little bit more hands off. 569 00:32:01,920 --> 00:32:03,960 The part that has been more interesting though is that when 570 00:32:03,960 --> 00:32:09,280 I was at grid point and we, we actually did bring nerves in and 571 00:32:09,280 --> 00:32:12,760 that was that was a fascinating thing is I I was a little bit of 572 00:32:12,760 --> 00:32:15,400 a flyer that we that we took because I didn't have any 573 00:32:15,640 --> 00:32:19,760 personal experience in nerves, But we were re architecting that 574 00:32:19,960 --> 00:32:23,840 entire system in Elixir and there was a lot of challenges 575 00:32:23,840 --> 00:32:27,520 with the hardware and firmware that were that was being 576 00:32:27,520 --> 00:32:30,240 deployed. Just a lot of there was a lot of 577 00:32:30,640 --> 00:32:33,320 bugs and, and other things that were going on and some, some 578 00:32:33,320 --> 00:32:34,840 things that were just hard to explain. 579 00:32:35,400 --> 00:32:40,160 And so we, I kind of put, I kind of put it out there, it's like, 580 00:32:40,160 --> 00:32:43,000 hey, you know, is there like maybe a way that you guys can 581 00:32:43,200 --> 00:32:45,480 start off not even nerves. It's like, can you guys like 582 00:32:45,480 --> 00:32:50,760 maybe compile into your stack a way that you can attach this 583 00:32:50,840 --> 00:32:56,160 thermostat or, or controller to one of our like live view 584 00:32:56,160 --> 00:32:59,200 channels? And one of the engineers like 585 00:32:59,200 --> 00:33:01,760 took it on and then like within, you know, like a day, he had the 586 00:33:01,760 --> 00:33:04,760 hard he had compiled in a a little library inside of that 587 00:33:04,760 --> 00:33:07,840 and, and we were had bi directional, you know, just web 588 00:33:07,840 --> 00:33:09,640 socket communication down to there. 589 00:33:09,960 --> 00:33:13,400 And they were just like, so they were like super excited about 590 00:33:13,400 --> 00:33:14,520 it. And so I kind of used that. 591 00:33:14,520 --> 00:33:17,240 I was like, well, you know, you know, we can kind of take this a 592 00:33:17,240 --> 00:33:21,560 step further and you know, this this whole nerves thing. 593 00:33:21,560 --> 00:33:24,760 And so we brought in, you know, a couple of a pretty senior well 594 00:33:24,760 --> 00:33:28,280 known nerves guys into the team to kind of really kind of kick 595 00:33:28,280 --> 00:33:32,520 that off. And then a few, we had a couple 596 00:33:32,520 --> 00:33:34,840 of the engineers that were not nerves engineers or, you know, 597 00:33:34,840 --> 00:33:38,960 more traditional seed type of developers on the firmware side 598 00:33:39,240 --> 00:33:42,360 really take it up. And, and they, they saw like 599 00:33:42,360 --> 00:33:46,160 what the productivity that you could get out of not bit tooling 600 00:33:46,160 --> 00:33:50,400 down at the almost at the metal. And so watching them kind of get 601 00:33:50,400 --> 00:33:53,000 excited about it. And so the, my whole goal with 602 00:33:53,000 --> 00:33:57,760 that was how can we bring the software and the firmware teams 603 00:33:57,760 --> 00:34:00,440 closer so that they're speaking more, closer to the same 604 00:34:00,440 --> 00:34:02,840 language and can maybe even share resources. 605 00:34:04,720 --> 00:34:07,240 And really that last piece didn't start to come together 606 00:34:07,240 --> 00:34:10,480 until we've actually made a step where we took one of our team 607 00:34:10,480 --> 00:34:14,880 leads that was running the the back end side, like the, the 608 00:34:14,880 --> 00:34:20,520 software on the Phoenix and Elixir side and put him over the 609 00:34:21,360 --> 00:34:25,040 firmware nerds team. And that was when some really 610 00:34:25,040 --> 00:34:27,320 interesting magic happened. Because now it's not, it wasn't 611 00:34:27,320 --> 00:34:29,560 just about like the tools and the technology. 612 00:34:29,560 --> 00:34:33,320 It was more about, Hey, how do you think about this software 613 00:34:33,320 --> 00:34:37,159 from like the deep experience he had in architecting things? 614 00:34:38,120 --> 00:34:40,480 And it just completely started to unlock a lot more 615 00:34:40,480 --> 00:34:43,120 productivity and quality and and throughput. 616 00:34:43,679 --> 00:34:46,320 And that was, you know, so we didn't completely change 617 00:34:46,320 --> 00:34:49,560 everything over there were still some devices that were too low 618 00:34:49,560 --> 00:34:52,920 power to really run nerves at that time. 619 00:34:52,920 --> 00:34:55,360 And I know some changes are happening in the ecosystem now 620 00:34:55,360 --> 00:34:58,840 that maybe it would enable that. But I think it's actually, it 621 00:34:58,840 --> 00:35:01,160 was actually turned out to be a really great thing. 622 00:35:01,440 --> 00:35:06,080 And if nothing else, I slept better in doing our firmware 623 00:35:06,080 --> 00:35:10,160 updates when we had the the nerve system in place because it 624 00:35:10,160 --> 00:35:12,880 was just a lot more reliable. When things went bad, it would 625 00:35:12,880 --> 00:35:16,360 just kind of auto roll back. And you know, the idea of 626 00:35:16,360 --> 00:35:19,600 pushing the button and updating 10s of thousands of devices and 627 00:35:19,840 --> 00:35:22,160 possibly breaking them and costing your company millions of 628 00:35:22,160 --> 00:35:24,080 dollars, it's just too stressful. 629 00:35:24,560 --> 00:35:27,280 That that sounds stressful. Yeah. 630 00:35:28,200 --> 00:35:30,640 But yeah, nerves was a great way of of getting around that. 631 00:35:30,640 --> 00:35:36,080 And so I think that, you know, if I ever did another hardware 632 00:35:36,400 --> 00:35:40,200 oriented company or startup, I would seriously consider using 633 00:35:40,200 --> 00:35:41,160 Nerds again. It was. 634 00:35:41,480 --> 00:35:44,720 It's awesome. I have a hardware startup that I 635 00:35:44,760 --> 00:35:47,200 don't have. I want someone else to build my 636 00:35:47,200 --> 00:35:51,640 idea because I think it's a good idea, but I I want to. 637 00:35:51,680 --> 00:35:56,000 I think it would be cool to like for large properties, at least I 638 00:35:56,000 --> 00:35:59,040 have in mind where it's like hard, like hard to place cameras 639 00:35:59,040 --> 00:36:02,680 and security all over the place. I think it'd be really cool to 640 00:36:02,680 --> 00:36:07,080 use like a mesh Wi-Fi network And you lay it out wherever you 641 00:36:07,080 --> 00:36:10,560 want, you know, spots that you can monitor, right? 642 00:36:11,120 --> 00:36:15,600 And then you can use like AI or some form of it to determine if 643 00:36:15,600 --> 00:36:17,960 movement is an animal or a person. 644 00:36:18,240 --> 00:36:21,760 And then you can send a swarm of drones with cameras. 645 00:36:21,920 --> 00:36:24,160 And you don't have to place cameras all over your property. 646 00:36:24,160 --> 00:36:29,120 You can fly them to it and you can you can see any spot on your 647 00:36:29,120 --> 00:36:32,080 property with a camera. That's super interesting. 648 00:36:32,400 --> 00:36:35,080 We're probably going to need all of that, as you know, the the 649 00:36:35,080 --> 00:36:39,320 world starts to develop and just having your own swarm of this 650 00:36:39,320 --> 00:36:42,080 may be a great defense. I mean, if I end up buying 651 00:36:42,080 --> 00:36:45,920 property, which would be my end goal, I will probably have to 652 00:36:45,920 --> 00:36:48,160 build this system because I just think it'd be awesome. 653 00:36:48,760 --> 00:36:52,120 That would be a lot of fun. I I bet I know a few people that 654 00:36:52,440 --> 00:36:54,400 that are really really good nearest developers. 655 00:36:54,440 --> 00:36:57,120 Maybe we should talk offline? Wait, all right, I'm down for 656 00:36:57,120 --> 00:37:01,040 that because I think it would be a fun project and like I would 657 00:37:01,080 --> 00:37:04,960 like that product I would buy like even at a residential 658 00:37:04,960 --> 00:37:07,720 level, if you could just have them like come out and if 659 00:37:07,720 --> 00:37:10,720 someone's trying to get in your back door, you can pretty much 660 00:37:10,720 --> 00:37:14,240 like get a good facial recognition on them boom and 661 00:37:14,240 --> 00:37:16,480 then. I think you could you could 662 00:37:16,480 --> 00:37:18,840 absolutely sell that to the entire prepper community. 663 00:37:19,000 --> 00:37:21,920 Like without a. Well, unless they don't trust 664 00:37:21,920 --> 00:37:23,880 technology. Well, they're that. 665 00:37:23,880 --> 00:37:24,960 You have a very good point there. 666 00:37:26,440 --> 00:37:29,240 Some of those guys want to just stick to strings and bells, you 667 00:37:29,240 --> 00:37:30,840 know? There's somebody said for that 668 00:37:30,840 --> 00:37:32,600 too. Technology can't be going. 669 00:37:33,120 --> 00:37:36,320 I get it. Well, and like one thing, I 670 00:37:36,320 --> 00:37:38,640 think it would have to be a closed system too because you 671 00:37:38,640 --> 00:37:41,800 don't want it to be easily hackable. 672 00:37:41,800 --> 00:37:45,160 I wouldn't say it's not going to be hackable anymore, but if it's 673 00:37:45,160 --> 00:37:47,320 a closed system at least it's a little harder. 674 00:37:47,680 --> 00:37:49,720 Yeah, I can see that. That'd be interesting. 675 00:37:50,800 --> 00:37:53,520 Yeah, you should do that. Someday, someday. 676 00:37:54,000 --> 00:37:56,480 It's on my list of Oh my good ideas. 677 00:37:56,480 --> 00:38:00,600 What's funny is Kill Switch was an idea I had for probably over 678 00:38:00,600 --> 00:38:04,720 five years now, and I just started building it this over 679 00:38:04,720 --> 00:38:05,640 the last few months. Yeah. 680 00:38:06,920 --> 00:38:09,480 I think that's the that's the world where you live in now that 681 00:38:09,480 --> 00:38:12,520 a lot of things that maybe weren't economically viable or 682 00:38:12,520 --> 00:38:15,640 just seem too daunting or it's just like, well, why not? 683 00:38:15,920 --> 00:38:18,200 You know, I've got projects, they're just kind of being built 684 00:38:18,880 --> 00:38:21,080 out of band because they're just, they're just happening, 685 00:38:21,080 --> 00:38:22,640 you know, in another window over there. 686 00:38:22,640 --> 00:38:25,680 And it's like. Maybe fruitions, maybe they 687 00:38:25,720 --> 00:38:27,880 won't. Well, like honestly too, when I 688 00:38:27,880 --> 00:38:32,480 first had the idea, it was before Elixir was in my life and 689 00:38:32,640 --> 00:38:35,040 I just didn't know how to fully solve it properly. 690 00:38:35,040 --> 00:38:39,680 And then it was just a very clear once, once elixir was in 691 00:38:39,680 --> 00:38:42,440 my life. So that's good too, yeah. 692 00:38:42,840 --> 00:38:44,400 That's really cool. Yeah. 693 00:38:45,280 --> 00:38:47,040 Yeah, that you have a lot of that going on. 694 00:38:47,880 --> 00:38:51,360 And Oban or Oban, I always say it differently every time I see 695 00:38:51,360 --> 00:38:56,040 it is a lifesaver. Yeah, yeah, that's, that's 696 00:38:56,120 --> 00:38:58,960 definitely been a piece of of almost any of the projects. 697 00:38:58,960 --> 00:39:00,120 So I, I can't remember what it was. 698 00:39:00,120 --> 00:39:02,440 It seemed like there was somebody building a another 699 00:39:02,440 --> 00:39:05,040 alternative to that processing I was just reading about. 700 00:39:05,040 --> 00:39:06,360 I don't remember the name of it now. 701 00:39:07,320 --> 00:39:10,720 Yeah, I'm not sure. Same kind of same kind of thing 702 00:39:10,720 --> 00:39:12,960 like scheduled workers and. Yeah. 703 00:39:12,960 --> 00:39:15,960 But I think it's it's difference was that it was very like 704 00:39:15,960 --> 00:39:17,720 declarative and how you configured it. 705 00:39:18,080 --> 00:39:21,080 So, you know, a la ash type of of things. 706 00:39:21,800 --> 00:39:24,000 So I just had a different kind of DSL to it. 707 00:39:24,360 --> 00:39:26,880 I can't remember what the name of the words didn't actually end 708 00:39:26,880 --> 00:39:29,680 up playing with it yet. Well, if anyone listening knows, 709 00:39:30,000 --> 00:39:32,960 throw it in the chat and then we'll we'll give it a shout out. 710 00:39:33,920 --> 00:39:36,720 But yeah, I've been kill switches. 711 00:39:36,720 --> 00:39:40,600 Actually, I have two projects now that are using Ash and I'm a 712 00:39:40,600 --> 00:39:42,360 big fan. Like I feel like it helps me 713 00:39:42,360 --> 00:39:46,760 move quickly and I enjoy it. Took me a long time to pick up 714 00:39:46,760 --> 00:39:50,800 the DSL though, Like it didn't click in my brain until the book 715 00:39:50,800 --> 00:39:54,240 came out and I could just really stare down examples. 716 00:39:55,160 --> 00:39:59,600 Yes, I have interesting background with Ash, so this 717 00:40:00,040 --> 00:40:03,920 startup is the first time I'm using Ash and Anger for real 718 00:40:03,920 --> 00:40:04,720 production. OK. 719 00:40:05,560 --> 00:40:07,160 Which has been, it's been fantastic. 720 00:40:07,160 --> 00:40:11,080 I no complaints, but I also remember, like, I remember so 721 00:40:11,080 --> 00:40:13,720 many years going to Elixir comp and seeing the ash table there 722 00:40:13,720 --> 00:40:15,560 and be like, I don't want to, I don't want to go to learn about 723 00:40:15,560 --> 00:40:16,600 that. I don't even know the framework. 724 00:40:16,600 --> 00:40:20,160 And it was just out of sheer stupidity and like ignorance 725 00:40:20,240 --> 00:40:22,200 that I didn't really understand what it was all about. 726 00:40:22,200 --> 00:40:25,640 And I guess in some ways I kind of lucked out that I, I didn't 727 00:40:25,640 --> 00:40:29,440 get involved early on until it like hit this really interesting 728 00:40:29,440 --> 00:40:31,160 inflection point just a couple years ago. 729 00:40:31,640 --> 00:40:34,000 And that's when, you know, kind of started to get a little more 730 00:40:34,000 --> 00:40:36,880 interested in it. And I probably had the battle 731 00:40:36,880 --> 00:40:43,960 scars of like seeing several big Phoenix projects kind of plateau 732 00:40:43,960 --> 00:40:48,000 out in terms of productivity or even drop off because that layer 733 00:40:48,000 --> 00:40:50,760 got just got so difficult to deal with. 734 00:40:52,200 --> 00:40:55,240 You just don't see that when you go and you like do the initial 735 00:40:55,440 --> 00:40:58,480 work with Phoenix and everything works great and maybe you got a 736 00:40:58,480 --> 00:41:00,760 really tiny team, you can keep that going for a little while 737 00:41:00,760 --> 00:41:03,360 longer. Like I had one team that was, 738 00:41:03,360 --> 00:41:07,720 you know, 30 engineers working on a platform and that, you 739 00:41:07,720 --> 00:41:11,120 know, you just don't stand a chance of of keeping that under 740 00:41:11,120 --> 00:41:14,080 control. And I think Ash would probably 741 00:41:14,080 --> 00:41:17,720 have made a difference in keeping that coordinated a 742 00:41:17,720 --> 00:41:19,200 little bit better. Yeah, that's interesting. 743 00:41:19,200 --> 00:41:23,240 Like where, where do you think? Like I, I think I agree with 744 00:41:23,240 --> 00:41:25,320 you. Like as projects grow they get 745 00:41:25,320 --> 00:41:30,000 hard to maintain, especially if your architecture isn't right. 746 00:41:30,440 --> 00:41:32,560 But where do you think the biggest problem comes in? 747 00:41:32,560 --> 00:41:35,920 Like at your like schema level or like the context like the 748 00:41:35,920 --> 00:41:37,360 CRUD layer? What do you think? 749 00:41:38,240 --> 00:41:42,760 Yeah, I think that. I think it's just easy to let 750 00:41:42,760 --> 00:41:46,920 the contexts get a little bit. Bloated. 751 00:41:47,240 --> 00:41:49,800 Yeah, bloated and messy and duplicative. 752 00:41:50,480 --> 00:41:53,200 You know, nobody can keep in their mind what's available and 753 00:41:53,240 --> 00:41:56,720 you know, those types of things. And so I think that's where. 754 00:41:56,720 --> 00:41:59,200 Yeah, you start rewriting things, probably a lot of 755 00:41:59,200 --> 00:42:01,840 duplicated code doing similar things. 756 00:42:01,840 --> 00:42:04,000 And yeah, I get that. Yeah, ash. 757 00:42:04,240 --> 00:42:05,840 Ash definitely helps prevent that. 758 00:42:06,440 --> 00:42:09,400 It does and you know, nothing's perfect and but I think it is 759 00:42:09,400 --> 00:42:10,800 also. It's just about. 760 00:42:11,000 --> 00:42:14,480 Sometimes it's just about a few percentage points one way or the 761 00:42:14,480 --> 00:42:18,520 other will create a massive difference in your trajectory 762 00:42:18,520 --> 00:42:20,240 over the course of two or three years. 763 00:42:20,600 --> 00:42:23,480 And I, I think that's, that's a little bit of, of my hope, even 764 00:42:23,480 --> 00:42:26,920 though honestly, I think that the team size thing has been 765 00:42:26,920 --> 00:42:31,960 taken off the table 'cause I will not ever hire another large 766 00:42:31,960 --> 00:42:35,440 team again like that. It's just a little, if you can't 767 00:42:35,440 --> 00:42:38,600 build it with, you know, three or four people, I'm not sure. 768 00:42:38,760 --> 00:42:40,640 Like that's not something I'm interested in anymore. 769 00:42:41,000 --> 00:42:46,840 Yeah, I'm curious like with like you've gone from managing or 770 00:42:46,840 --> 00:42:50,160 working with like large engineering teams like 20 Plus 771 00:42:50,160 --> 00:42:53,520 to now your startup, you're the single engineer currently. 772 00:42:54,120 --> 00:42:58,200 What, what, what's your, and this is mostly like a question 773 00:42:58,200 --> 00:43:02,160 coming from me because like, what do you think like your 774 00:43:02,160 --> 00:43:04,560 first hire is going to be? Are you going to step away from 775 00:43:04,560 --> 00:43:08,440 development or how is how? What does that look like? 776 00:43:08,960 --> 00:43:12,160 Man, I've been thinking about this a lot and I think it is 777 00:43:13,160 --> 00:43:17,320 like the to answer the question, I'm going to have to probably 778 00:43:17,320 --> 00:43:22,240 just talk about what I think a development organization might 779 00:43:22,240 --> 00:43:25,000 look like going forward. And we touched on it a little 780 00:43:25,000 --> 00:43:27,680 bit earlier, but I'll just kind of flesh it out a little bit. 781 00:43:28,440 --> 00:43:32,400 I don't think that the role of a product owner and a developer 782 00:43:32,400 --> 00:43:36,400 and a tester, I don't think that those should be or can be 783 00:43:36,400 --> 00:43:41,760 independent roles anymore. And so I'm, I will be, you know, 784 00:43:41,760 --> 00:43:46,960 hiring a, a really interesting engineer here coming very soon. 785 00:43:46,960 --> 00:43:49,960 But it will put that whole thing will be somebody that knows the 786 00:43:49,960 --> 00:43:54,840 domain, somebody that knows Elixir inside and out, somebody 787 00:43:54,840 --> 00:43:59,880 that can communicate and a really, really adept way and 788 00:43:59,880 --> 00:44:04,520 somebody that has a very strong skeptical testing mentality all 789 00:44:04,520 --> 00:44:08,240 in one head. That's a really rare skill like, 790 00:44:08,240 --> 00:44:09,880 oh, by the way, if anybody's listening and you have that, 791 00:44:09,880 --> 00:44:12,400 please reach out to me because we want you. 792 00:44:13,600 --> 00:44:16,600 But I the days of me hiring somebody that want us, that 793 00:44:16,600 --> 00:44:19,160 wants to sling code, that's that's over. 794 00:44:19,440 --> 00:44:21,200 Yeah, that's interesting. Again. 795 00:44:21,640 --> 00:44:23,840 Yeah, I agree. You're going to just have your 796 00:44:24,000 --> 00:44:27,000 dedicated rock stars. Yeah. 797 00:44:27,000 --> 00:44:30,760 And you put him in charge of one like application or one major 798 00:44:30,760 --> 00:44:32,240 area. And it's like it's your 799 00:44:32,240 --> 00:44:34,080 responsibility to make this happen. 800 00:44:34,160 --> 00:44:37,200 Like you've got to have that, that entrepreneurial experience 801 00:44:37,200 --> 00:44:40,080 of like, no, you've got the latitude, like you know what's 802 00:44:40,080 --> 00:44:42,000 going on, you know, everybody's here to help. 803 00:44:42,440 --> 00:44:46,080 But it's also like figure out what, what, how, how you have to 804 00:44:46,080 --> 00:44:48,520 move forward on this. And then how do you see your 805 00:44:48,520 --> 00:44:51,320 role changing? Would you always be slinging 806 00:44:51,320 --> 00:44:56,760 code or like working on the next project or step more into a 807 00:44:57,040 --> 00:45:01,360 founder CEO role? I think what I'm wondering about 808 00:45:01,360 --> 00:45:04,640 myself is I there for a while, I wanted to go down the CEO path 809 00:45:04,640 --> 00:45:11,040 and I don't think I think I dodged a bullet in taking ACEO 810 00:45:11,040 --> 00:45:14,720 role in this, this go round because I kind of wanted to and 811 00:45:16,560 --> 00:45:18,400 I'm just much happier on the tech side. 812 00:45:18,440 --> 00:45:20,880 Like it, it's, it's, that's where I should be. 813 00:45:20,880 --> 00:45:23,240 It's it's where I'm productive, it's where I have something to 814 00:45:23,240 --> 00:45:27,560 offer. But I believe like I'm gonna try 815 00:45:27,560 --> 00:45:30,960 to build this company like the the CEO and I work together a 816 00:45:30,960 --> 00:45:34,000 lot very closely. And we talk a lot about like 817 00:45:35,080 --> 00:45:37,880 it's, it's all about profit per head per head count. 818 00:45:38,120 --> 00:45:40,840 Like that's one of the major metrics is like, how do we 819 00:45:41,040 --> 00:45:46,560 create a, a massively pop profitable company with very, 820 00:45:46,560 --> 00:45:50,560 very few people? And, you know, I think that that 821 00:45:50,560 --> 00:45:53,880 is, that goes back to what I said about the type of people 822 00:45:53,880 --> 00:45:56,360 we're going to hire. And it also means that my 823 00:45:56,360 --> 00:46:00,280 responsibility, I think as I can start to hand off major parts of 824 00:46:00,280 --> 00:46:04,240 the core application are going to be about, Hey, how do we, how 825 00:46:04,240 --> 00:46:08,080 do we build the, the tech infrastructure in place so that 826 00:46:08,080 --> 00:46:13,040 we don't have to hire, you know, sales execs and things like 827 00:46:13,040 --> 00:46:15,000 that. I mean, you may have like one 828 00:46:15,000 --> 00:46:19,360 really high end, you know, like chief revenue officer type of 829 00:46:19,360 --> 00:46:23,560 person, but there's a lot of things you can do to make the 830 00:46:23,560 --> 00:46:27,280 entire sales pipeline that was always dealt with with account 831 00:46:27,280 --> 00:46:30,720 executives and, you know, inside sales and all those other roles 832 00:46:30,720 --> 00:46:32,600 like that shouldn't happen anymore. 833 00:46:34,200 --> 00:46:36,560 I mean, I, I was listening, there's a podcast that I listen 834 00:46:36,560 --> 00:46:39,400 to in Lenny's podcast that is a lot of amazing interviews with 835 00:46:39,400 --> 00:46:42,320 people that are doing some really cool stuff in scaling 836 00:46:42,320 --> 00:46:43,640 companies, especially in the start up. 837 00:46:43,640 --> 00:46:46,160 And there was one about this very thing about how this 838 00:46:46,160 --> 00:46:50,840 company, you know, basically replaced 90% of their sales team 839 00:46:50,840 --> 00:46:53,160 with a, with a bunch of agents and they kind of kept their 840 00:46:53,160 --> 00:46:56,240 productivity flat, didn't go better, but it didn't really get 841 00:46:56,240 --> 00:46:59,360 worse either. And you know, I think there's, 842 00:46:59,360 --> 00:47:02,280 there's a lot of opportunities to, to figure out how to apply 843 00:47:02,280 --> 00:47:06,560 this technology to keep a company's growth manageable. 844 00:47:08,080 --> 00:47:09,320 That's what I'm kind of excited about. 845 00:47:09,320 --> 00:47:11,960 That's what gets me out of bed in the morning is hopefully I 846 00:47:11,960 --> 00:47:15,840 can hand off this core product to some to one or two really 847 00:47:15,840 --> 00:47:18,800 talented engineers and I can start looking at the other parts 848 00:47:18,800 --> 00:47:20,400 of the company and how we're going to scale that. 849 00:47:21,360 --> 00:47:23,840 Lenny's podcast is really cool indeed. 850 00:47:23,840 --> 00:47:26,320 Daniel Bergholtz jumped in the comments. 851 00:47:26,320 --> 00:47:28,720 I'm going to have to write a note down and check that out 852 00:47:28,720 --> 00:47:32,320 because. It's you could do nothing better 853 00:47:32,320 --> 00:47:34,600 than to just go and listen to a bunch of the historical ones 854 00:47:34,600 --> 00:47:37,280 because I've just blown away what I've learned and some of 855 00:47:37,280 --> 00:47:38,440 the people that he's had on there. 856 00:47:38,760 --> 00:47:42,600 I'm like, I'm kind of in the I feel like you and I are very 857 00:47:42,600 --> 00:47:45,280 similar. Like I always have that idea. 858 00:47:45,320 --> 00:47:48,760 I want to be like aceo role. But like, when it comes down to 859 00:47:48,760 --> 00:47:53,280 it, I like to build and solve problems And, and I don't mean 860 00:47:53,280 --> 00:47:59,920 solve business problems because like I kind of go ADHD and lose 861 00:47:59,920 --> 00:48:03,240 focus on what I should be doing with, you know, whatever, 862 00:48:03,320 --> 00:48:05,720 whatever I can distract, like I'm gonna implement a new 863 00:48:05,720 --> 00:48:09,120 feature. But I, I enjoy writing code and 864 00:48:09,120 --> 00:48:13,720 solving like a core issue. It's an enjoyable part. 865 00:48:14,400 --> 00:48:17,120 And there's like completion mentality about it, like you 866 00:48:17,120 --> 00:48:19,640 when you push something out and you're like, ha, I got something 867 00:48:19,640 --> 00:48:21,960 done. Like that's, that's a big part 868 00:48:21,960 --> 00:48:24,720 of my personality is just having that a little bit of a dopamine 869 00:48:24,720 --> 00:48:29,400 hit of like done, done, done. That cadence is really 870 00:48:29,400 --> 00:48:32,160 important. Well, I just I find the simplest 871 00:48:32,160 --> 00:48:36,480 things really cool, like I the the other night, like, you know, 872 00:48:36,920 --> 00:48:41,040 when I wrap up my day job, I'll I'll either I'll like be 873 00:48:41,040 --> 00:48:43,920 listening to audiobooks and just pumping out a feature on kill 874 00:48:43,920 --> 00:48:47,120 switch. And all I did was integrate a 875 00:48:47,120 --> 00:48:50,640 third party affiliate program. Right. 876 00:48:50,640 --> 00:48:55,800 But like just a simple like redirect with, you know, a URL 877 00:48:55,800 --> 00:48:59,480 parameter in there. And then I set it into a, into a 878 00:48:59,480 --> 00:49:03,440 cookie and just watch it work into my like whole registration 879 00:49:03,440 --> 00:49:06,720 flow and automatically apply discounts to my user. 880 00:49:06,920 --> 00:49:08,720 I just signed up like little things like that. 881 00:49:08,720 --> 00:49:11,760 I'm just like, that's so cool. It's not even an impressive 882 00:49:11,760 --> 00:49:15,400 thing, but. Well, I think that's where a lot 883 00:49:15,400 --> 00:49:17,360 of people get. That's why a lot of people get 884 00:49:17,360 --> 00:49:19,880 kind of frustrated with the LM stuff because they, they think 885 00:49:19,880 --> 00:49:23,560 it's like, you know, degrading development down to like 886 00:49:23,560 --> 00:49:25,880 connecting pipes and whatnot. And I, I don't agree. 887 00:49:25,880 --> 00:49:31,720 Like I think that's, I think that for me, I get the, the 888 00:49:31,800 --> 00:49:35,520 reward and the dopamine hit from progress, like solving a 889 00:49:35,520 --> 00:49:38,240 problem, not fighting with a problem. 890 00:49:38,240 --> 00:49:40,320 And I know that not everyone thinks about it that way. 891 00:49:40,320 --> 00:49:44,600 Like some people truly enjoy the, you know, the mental 892 00:49:44,600 --> 00:49:47,320 gymnastics of actually having to go through and solve that. 893 00:49:47,320 --> 00:49:50,520 And that's great. It's not, it's not gonna be 894 00:49:50,520 --> 00:49:54,480 profitable anymore, but it's, but it the, the reality is like 895 00:49:54,600 --> 00:49:57,360 the people that are, that are all about like, hey, how do I 896 00:49:57,360 --> 00:49:59,560 solve problems? And they get excited about 897 00:49:59,560 --> 00:50:02,800 seeing the solution as opposed to fighting with the problem. 898 00:50:04,200 --> 00:50:06,120 It's your. It's your time to Yeah. 899 00:50:06,200 --> 00:50:10,120 I see like, look, I'm a developer at heart and I, I do 900 00:50:10,120 --> 00:50:15,120 love writing code, but I, I see a lot of people are like, well, 901 00:50:15,120 --> 00:50:18,920 if I use an LLM too much, I actually don't have the muscle 902 00:50:18,920 --> 00:50:20,600 memory to write a function anymore. 903 00:50:20,640 --> 00:50:24,880 Like I feel like it's slipping and I agree with that, but I 904 00:50:24,880 --> 00:50:27,560 also don't care. It doesn't bother me. 905 00:50:27,560 --> 00:50:30,520 I'd I see the end result, I can read the code, I see that it's 906 00:50:30,520 --> 00:50:33,720 good. Me writing it or not doesn't 907 00:50:34,320 --> 00:50:36,680 make me feel good or bad about myself. 908 00:50:36,680 --> 00:50:39,800 It's just makes me feel good because my product's moving 909 00:50:39,800 --> 00:50:42,560 along. Yeah, that makes a lot of sense. 910 00:50:42,560 --> 00:50:44,240 I tend to agree with that a little bit more. 911 00:50:44,400 --> 00:50:46,600 And I would even actually take it a step further that I don't 912 00:50:46,600 --> 00:50:50,200 think that, at least for me, I don't think my muscle memory has 913 00:50:50,200 --> 00:50:51,920 gotten worse. Maybe that's just because I 914 00:50:51,920 --> 00:50:54,160 wasn't that great to begin with, but I think there's a lot of 915 00:50:54,160 --> 00:50:57,320 times I learned something where it's like I have a mental idea 916 00:50:57,320 --> 00:50:59,840 of how it should be solved, and then the LLM goes off and does 917 00:50:59,840 --> 00:51:01,640 something. I'm like, Oh my God, I would 918 00:51:01,640 --> 00:51:03,080 have never thought to do it that way. 919 00:51:03,480 --> 00:51:06,360 What's interesting there too, like with kill switch, I'm doing 920 00:51:06,360 --> 00:51:09,080 a lot of like it's zero knowledge encryption. 921 00:51:09,160 --> 00:51:13,280 And like I, I knew I know roughly how it works, right? 922 00:51:14,080 --> 00:51:18,640 But I also used Claude to like make a really like in depth plan 923 00:51:19,320 --> 00:51:23,560 and reading through that plan like made me level up. 924 00:51:23,640 --> 00:51:28,400 Like I understand so much more about how zero and like 0 925 00:51:28,400 --> 00:51:30,480 knowledge encryption works now because of that. 926 00:51:31,320 --> 00:51:35,480 I I'm, yeah, I couldn't agree with you more like. 927 00:51:35,520 --> 00:51:38,120 And I think it's so important that you brought up that 928 00:51:38,760 --> 00:51:44,360 planning process like I have, I always used to hate writing 929 00:51:44,360 --> 00:51:48,120 specs because it just helpful. Like it took so long to get the 930 00:51:48,120 --> 00:51:52,280 nuance out there that was needed to have a a meaningful document 931 00:51:52,880 --> 00:51:55,440 that it I just got bored And you know, like, no, that's not for 932 00:51:55,440 --> 00:51:58,520 me. But working with, you know, 933 00:51:58,520 --> 00:52:03,120 Claude or something and describing what I want and 934 00:52:03,120 --> 00:52:08,880 having like so much more come back and so and challenging me. 935 00:52:08,880 --> 00:52:12,120 It's like, you know, this, this example of what we're what I'm 936 00:52:12,120 --> 00:52:15,840 building today in the, in the this start up, I said it was a 937 00:52:15,840 --> 00:52:19,120 lot of hardcore power systems calculations and things that are 938 00:52:19,120 --> 00:52:21,120 like well understood. There's a lot of, you know, PhD 939 00:52:21,120 --> 00:52:22,720 level papers and stuff out on the web. 940 00:52:23,160 --> 00:52:27,400 And I can, I can have it come back with all of that context 941 00:52:27,400 --> 00:52:30,520 and push back on me in a way that I'm just blown away. 942 00:52:30,520 --> 00:52:33,480 And I can build out, you know, a 20 or 30 page spec over the 943 00:52:33,480 --> 00:52:36,640 course of a few hours that is so much better than I could have 944 00:52:36,640 --> 00:52:38,960 ever written on my own. Just hands. 945 00:52:39,360 --> 00:52:43,200 Just better. Empirically better than I could 946 00:52:43,200 --> 00:52:45,640 have ever done. Yeah, I I agree. 947 00:52:45,720 --> 00:52:50,440 And I my HTML emails look a lot better with Claude doing them, 948 00:52:50,440 --> 00:52:53,120 let me tell you, I bet. Yeah. 949 00:52:53,840 --> 00:52:55,360 No, I think it's great. And then when you look at the 950 00:52:55,360 --> 00:52:58,840 quality of what the code comes out, if you take this time to 951 00:52:58,840 --> 00:53:03,600 really get a good specification and then turn that specification 952 00:53:03,600 --> 00:53:06,040 into, you know, a set of acceptance tests that you're 953 00:53:06,040 --> 00:53:09,360 still keeping an eye on. Like when you when you invest 954 00:53:09,360 --> 00:53:12,520 that a little bit of time up front, you don't get slop. 955 00:53:12,560 --> 00:53:14,040 I come out of these things, you can see I. 956 00:53:14,040 --> 00:53:17,560 Just actually that reminds me, I saw a really interesting ex post 957 00:53:17,560 --> 00:53:22,840 about how someone is moving forward with interviewing 958 00:53:23,200 --> 00:53:27,280 software engineers and it's less about putting them on the spot 959 00:53:27,280 --> 00:53:29,680 writing code. They give them a problem to 960 00:53:29,680 --> 00:53:32,960 solve and they want to see how they solve it using LLMS and the 961 00:53:32,960 --> 00:53:36,440 tools at hand. And if they're not good, like I 962 00:53:36,440 --> 00:53:40,400 think it's like very important, like the results that you get 963 00:53:40,600 --> 00:53:43,960 out of LLMS is directly correlated to the prompts you 964 00:53:43,960 --> 00:53:46,840 put in. And I found that really 965 00:53:46,840 --> 00:53:51,440 interesting because a lot of these really good developers 966 00:53:51,440 --> 00:53:54,360 that refuse to touch this stuff want to do everything manually. 967 00:53:54,360 --> 00:53:58,400 But I think you have to accept the changes that are happening 968 00:53:58,400 --> 00:53:59,880 in our space. Yeah. 969 00:54:01,040 --> 00:54:02,800 Because we're going. To see a lot more of that, like 970 00:54:03,320 --> 00:54:07,080 what you're saying, you want one Rockstar developer that knows 971 00:54:07,080 --> 00:54:12,120 how to do everything. And I like same here. 972 00:54:12,120 --> 00:54:15,720 Like if I hire a developer at Killswitch, I'm I want that 973 00:54:15,720 --> 00:54:17,360 Rockstar, I want to be able to trust him. 974 00:54:17,360 --> 00:54:20,920 I want to trust his architectural decisions, his 975 00:54:20,920 --> 00:54:24,200 code smell ability, and his ability to communicate. 976 00:54:24,920 --> 00:54:26,920 Yep. Yeah, that's exactly it. 977 00:54:27,200 --> 00:54:32,200 I I agree. I think it's yeah. 978 00:54:32,200 --> 00:54:33,920 I'm glad you brought up the interview thing because I'm 979 00:54:33,920 --> 00:54:37,040 actually, that's one of the things I don't know what an 980 00:54:37,040 --> 00:54:39,000 interview should look like today. 981 00:54:39,000 --> 00:54:40,880 I mean, I've interviewed hundreds and hundreds of 982 00:54:40,880 --> 00:54:44,000 developers over the past few years, and I think everything 983 00:54:44,000 --> 00:54:46,440 about the way I've done in the past is probably wrong now. 984 00:54:47,280 --> 00:54:50,480 I think, I think everything that I've done a month ago is wrong 985 00:54:50,480 --> 00:54:52,560 now. Like ever since I saw that post, 986 00:54:52,560 --> 00:54:55,320 it made me rethink because so like I've been doing a little 987 00:54:55,320 --> 00:54:57,720 like pre vetted recruitment stuff on the side where I was 988 00:54:57,720 --> 00:55:01,240 like like 1. I don't believe in white 989 00:55:01,240 --> 00:55:02,720 boarding. I don't believe in putting 990 00:55:02,720 --> 00:55:05,760 people on the spot with algorithms and sorting. 991 00:55:05,800 --> 00:55:10,080 You know, just the, the stupid, I forget the name of it right 992 00:55:10,080 --> 00:55:14,000 now, but all the whiteboarding equations like you can look this 993 00:55:14,000 --> 00:55:17,480 crap up, It doesn't matter. Like you can understand how to 994 00:55:17,480 --> 00:55:22,800 do, you know, Big O and all that things referencing Google like 995 00:55:22,800 --> 00:55:26,480 it doesn't matter. But so I was, I would just have 996 00:55:26,840 --> 00:55:29,680 conversations with them. I would give people a technical 997 00:55:29,680 --> 00:55:32,400 take home assignment depending on the level they're applying at 998 00:55:33,000 --> 00:55:36,320 and the job. And then what I do is I look 999 00:55:36,320 --> 00:55:41,040 through the code before a a call and I have like, I have 1000 00:55:41,120 --> 00:55:44,120 questions, obviously from just looking at their code and we 1001 00:55:44,120 --> 00:55:46,360 have a conversation around it because I want to just make sure 1002 00:55:46,360 --> 00:55:50,600 they understand what they wrote. And like, as far as them using 1003 00:55:50,600 --> 00:55:54,160 Claude, I don't care. But if I see something that 1004 00:55:54,160 --> 00:55:58,480 looks overly complicated and complex and I start digging into 1005 00:55:58,480 --> 00:56:01,920 it and you don't know what it's doing, then I know you, you 1006 00:56:01,960 --> 00:56:04,120 know, it doesn't have any right in your code base. 1007 00:56:04,240 --> 00:56:07,360 And that's my opinion with it. Yeah, no, I think that's fair. 1008 00:56:07,680 --> 00:56:09,760 And I think, you know, like the other thing that I'm trying to 1009 00:56:09,760 --> 00:56:15,480 figure out is like, how do you like uncover like the, the 1010 00:56:15,480 --> 00:56:17,680 blood, sweat and tears that somebody's gone through to 1011 00:56:17,720 --> 00:56:21,720 understand how to develop their own workflow with these tools? 1012 00:56:21,720 --> 00:56:24,400 Because I, I can tell you, you know, from my own personal 1013 00:56:24,400 --> 00:56:29,120 experience, you know, just nine months ago, how backwards and 1014 00:56:29,120 --> 00:56:35,040 slow and just wrong my way of interacting with this stuff was. 1015 00:56:35,040 --> 00:56:38,640 And I think that, you know, we're in that time where it's 1016 00:56:38,640 --> 00:56:41,760 things are changing quickly enough that if somebody can't 1017 00:56:41,960 --> 00:56:46,400 explain the hardships and like why they've made decisions about 1018 00:56:46,400 --> 00:56:48,600 how they're interacting with their tools this way. 1019 00:56:48,600 --> 00:56:50,680 And it's because of this. And even if it's like no longer 1020 00:56:50,680 --> 00:56:53,360 relevant, I just want to, I feel like I kind of want to 1021 00:56:53,360 --> 00:56:55,960 understand that path to see like, you know, have you been 1022 00:56:55,960 --> 00:57:00,280 living this? You like spent 8 or 10 hours a 1023 00:57:00,280 --> 00:57:04,440 day trying to figure out how to make this stuff work because it 1024 00:57:04,440 --> 00:57:06,200 doesn't work out-of-the-box today. 1025 00:57:06,680 --> 00:57:12,080 And if you speak in a certain way, you do have to like have 1026 00:57:12,080 --> 00:57:14,520 certain guardrails up. And if you don't have to do 1027 00:57:14,520 --> 00:57:17,880 those things maybe a year from now or maybe five years from 1028 00:57:17,880 --> 00:57:20,120 now, that'll just be out-of-the-box for everybody. 1029 00:57:20,120 --> 00:57:21,960 But it's not the box we have today. 1030 00:57:22,520 --> 00:57:27,080 Yeah, I'm like with all the shift in like the the 1031 00:57:27,120 --> 00:57:31,320 superpowers, I feel like it unlocks like I think it makes a 1032 00:57:31,320 --> 00:57:36,040 great engineer like even greater, if that makes sense. 1033 00:57:36,080 --> 00:57:41,360 I don't if that's even a word. What is your like ideal team 1034 00:57:41,360 --> 00:57:43,840 size for like a successful company now? 1035 00:57:43,840 --> 00:57:45,360 Like what do you think that looks like? 1036 00:57:46,360 --> 00:57:51,040 What I really want to say 1, even though I'm not sure that 1037 00:57:51,080 --> 00:57:56,000 like it is the right answer, but I think what I'm going to try is 1038 00:57:56,760 --> 00:58:00,640 to have something that's very close to teams of one. 1039 00:58:00,640 --> 00:58:03,040 Now they of course, you know, there's going to be, you know, a 1040 00:58:03,040 --> 00:58:06,880 lot of support around that, you know, like we've got a APHD 1041 00:58:06,880 --> 00:58:09,000 level power systems engineer that we just hired a couple 1042 00:58:09,000 --> 00:58:12,160 weeks ago. Because you know, you need like 1043 00:58:12,960 --> 00:58:15,600 when you, when something doesn't make sense or you have a, you 1044 00:58:15,600 --> 00:58:17,800 know, a judgement call to make about how this should actually 1045 00:58:17,800 --> 00:58:20,520 work, You need somebody that has gone through that level of 1046 00:58:20,520 --> 00:58:23,680 training and can and can provide that level of guidance. 1047 00:58:25,280 --> 00:58:28,560 But that should be a resource that, you know, spans across, 1048 00:58:28,600 --> 00:58:31,640 you know, our entire, you know, company of, of engineers. 1049 00:58:31,640 --> 00:58:36,320 And so I think that if, if I'm lucky that, you know, we've got 1050 00:58:36,320 --> 00:58:41,040 like 3 or 4 different core product ideas that all relate 1051 00:58:41,040 --> 00:58:44,240 together very tightly. So I would imagine that it'll be 1052 00:58:44,240 --> 00:58:48,520 like 3 or 4 killer engineers that really know how to use 1053 00:58:48,520 --> 00:58:51,000 these tools and then have an entrepreneurial mindset. 1054 00:58:51,000 --> 00:58:52,800 They're like, now I want to own this, I want to brick. 1055 00:58:52,960 --> 00:58:57,200 I want the success or failure of this application to be on my 1056 00:58:57,200 --> 00:58:59,600 shoulders. Like those are the people that I 1057 00:58:59,600 --> 00:59:02,280 think are going to be moving the needle for companies now. 1058 00:59:02,600 --> 00:59:07,160 And I think if you've got that, I cannot imagine, I think you 1059 00:59:07,160 --> 00:59:11,000 have to beat people away with a stick for like opportunities. 1060 00:59:11,000 --> 00:59:15,240 Because I, I tell you, the, the, the company that came in and 1061 00:59:15,240 --> 00:59:19,400 provided our seat funding invests in companies like this. 1062 00:59:19,880 --> 00:59:24,800 And you know, I like they're watching and they know and 1063 00:59:24,800 --> 00:59:28,960 they're calibrating their expectations around the output 1064 00:59:28,960 --> 00:59:31,200 that they're seeing from the companies that they're investing 1065 00:59:31,200 --> 00:59:35,840 in. And that is setting a bar for 1066 00:59:35,840 --> 00:59:39,120 what the throughput is. And so you cannot, you just 1067 00:59:39,120 --> 00:59:42,280 cannot do handcrafted code anymore at that bar. 1068 00:59:42,640 --> 00:59:45,880 It just doesn't. Yeah, yeah, it's either a 1069 00:59:45,880 --> 00:59:49,000 fortunate or unfortunate truth, right, Depending on your. 1070 00:59:49,440 --> 00:59:51,800 Holding judgement. I'm just saying it's what it is, 1071 00:59:51,800 --> 00:59:54,480 like good battery different, it's what it is. 1072 00:59:54,840 --> 00:59:59,040 Yeah, what? What does a customer support 1073 00:59:59,040 --> 01:00:01,000 look like in the future do you think man? 1074 01:00:03,000 --> 01:00:07,960 You know, it's funny, I think I mean there's been a lot of work 1075 01:00:07,960 --> 01:00:10,640 done on trying to automate customer support. 1076 01:00:10,640 --> 01:00:13,320 And I was just actually listening to somebody talk about 1077 01:00:13,320 --> 01:00:16,720 this just a few days ago. And I think there's a lot of 1078 01:00:16,720 --> 01:00:19,520 potential to automate that. But where most people get it 1079 01:00:19,520 --> 01:00:23,480 wrong is they forget that there is a definite part of that 1080 01:00:23,480 --> 01:00:29,120 process of learning if you've got it right before you actually 1081 01:00:29,600 --> 01:00:32,600 allow very much agency on these systems. 1082 01:00:35,040 --> 01:00:37,160 And I've, I mean, I don't know, you've probably done like my 1083 01:00:37,160 --> 01:00:40,960 experience has been that 90% of the customer support agents that 1084 01:00:40,960 --> 01:00:43,200 I interact with are trash. And I'm just trying to. 1085 01:00:43,360 --> 01:00:45,240 Just. Trying to figure out how to get 1086 01:00:45,240 --> 01:00:49,200 out of that cycle and get to a human because like they don't, 1087 01:00:49,200 --> 01:00:53,880 they don't solve the problem. And anybody that tried to use it 1088 01:00:54,080 --> 01:00:55,960 could see that within a few minutes. 1089 01:00:56,480 --> 01:01:01,080 So clearly who is ever doing a lot of these projects is not 1090 01:01:02,120 --> 01:01:04,600 eating their own dog food and like going in through those 1091 01:01:04,600 --> 01:01:06,720 channels and trying to solve real problems. 1092 01:01:07,440 --> 01:01:10,800 So I, I don't know what it's going to take to to get customer 1093 01:01:10,800 --> 01:01:13,520 support to actually work on this. 1094 01:01:13,920 --> 01:01:17,520 Do you like this is just a random idea that I have and it 1095 01:01:17,520 --> 01:01:23,760 kind of goes into your your your idea of just knowledge, your 1096 01:01:23,760 --> 01:01:27,880 experts, you know, in whatever Ave. parts of your company that 1097 01:01:27,880 --> 01:01:31,680 you have them right, your your engineers, whatever it may be. 1098 01:01:32,160 --> 01:01:35,800 Do you think you might just have like a Rockstar customer support 1099 01:01:35,800 --> 01:01:39,720 person that could train an in house LLM for real low level 1100 01:01:39,960 --> 01:01:42,920 tasks? But then if it does, if it if it 1101 01:01:42,920 --> 01:01:47,240 escalates high enough, they get that person in real life like. 1102 01:01:49,040 --> 01:01:53,240 I think it's, I think it's possible, you know, come back 1103 01:01:53,240 --> 01:01:55,640 and talk to me in like a year or two when we've actually 1104 01:01:55,640 --> 01:01:57,520 launched. And I've probably failed at this 1105 01:01:57,520 --> 01:02:00,920 a few times, but I've never done the like the customer support 1106 01:02:00,920 --> 01:02:02,800 thing. You know, at Grid Point, you 1107 01:02:02,800 --> 01:02:05,760 know, we had a huge customer support organization and it was, 1108 01:02:05,880 --> 01:02:08,080 it was tough. And we had somebody that came in 1109 01:02:08,080 --> 01:02:11,200 to try to revamp it. They took it, you know, off, 1110 01:02:11,320 --> 01:02:15,240 they took it down to Mexico and that actually improved some 1111 01:02:15,240 --> 01:02:17,600 things. And there were some ideas of 1112 01:02:17,600 --> 01:02:21,720 trying to bring more and more of this, of, of the automation side 1113 01:02:21,720 --> 01:02:26,080 of things into it. But man, like you just also have 1114 01:02:26,080 --> 01:02:29,360 to remember like, like and ask yourself who are your customers 1115 01:02:29,360 --> 01:02:31,400 and what like, where are they coming from? 1116 01:02:32,680 --> 01:02:35,320 Like a grid point. Our customers were like a lot of 1117 01:02:35,320 --> 01:02:38,160 quick serve restaurants people so like Chipotle and whatnot. 1118 01:02:38,160 --> 01:02:41,320 And they're like, Hey, if my, my HVAC is not working and like I'm 1119 01:02:41,320 --> 01:02:45,680 responsible for, for the getting through this lunch rush, I'm not 1120 01:02:45,680 --> 01:02:50,480 going to be patient talking with some random LOM that's asking me 1121 01:02:50,480 --> 01:02:52,320 basic questions like. That's true. 1122 01:02:52,920 --> 01:02:55,400 And so you just have to like understand, like, where are your 1123 01:02:55,400 --> 01:02:57,960 customers coming from and what are what are their capabilities? 1124 01:02:57,960 --> 01:03:01,280 What are their skill sets? So maybe there's some that'll 1125 01:03:01,280 --> 01:03:03,560 that that'll work on, but then there's others where it's just 1126 01:03:03,560 --> 01:03:05,360 like you're you're berk at the wrong tree. 1127 01:03:05,800 --> 01:03:08,720 The person. I have like very little patience 1128 01:03:08,720 --> 01:03:12,600 on customer support. It's a problem but at least I 1129 01:03:12,640 --> 01:03:14,360 know it's a weakness of mine. Right? 1130 01:03:15,040 --> 01:03:18,400 But I've been thinking about this a lot and as soon as I can 1131 01:03:18,400 --> 01:03:21,200 afford it with kill switch, like as soon as income is coming 1132 01:03:21,200 --> 01:03:25,920 through the door, I think my first hire will be some you 1133 01:03:25,920 --> 01:03:29,600 know, $20.00 an hour customer support person to help me like 1134 01:03:29,600 --> 01:03:33,520 put things in place. But they're the the main point 1135 01:03:33,640 --> 01:03:36,480 of contact for everyone using my app. 1136 01:03:38,360 --> 01:03:41,160 And because there is something to be said about a human. 1137 01:03:42,200 --> 01:03:45,440 There is, and I think, you know, the people don't call into 1138 01:03:45,440 --> 01:03:47,280 support when things are going well. 1139 01:03:47,400 --> 01:03:51,080 And so you have to like understand that that is a 1140 01:03:51,080 --> 01:03:53,800 definitely a moment where you're going to be losing customers. 1141 01:03:53,960 --> 01:03:57,400 Like like that's the default when somebody's calling in. 1142 01:03:57,960 --> 01:04:01,960 Like you got to like view that as this is a customer loss and 1143 01:04:01,960 --> 01:04:03,920 I've got to figure out how to get it back on track. 1144 01:04:04,760 --> 01:04:10,720 And you know, while I'm a little down on the the whole tech side 1145 01:04:10,720 --> 01:04:14,040 of this, I also have to be very honest and say that I haven't 1146 01:04:14,040 --> 01:04:16,560 had a great experience with the human side of most tech support 1147 01:04:16,600 --> 01:04:18,760 either. So maybe that's true pretty low 1148 01:04:19,640 --> 01:04:23,200 on that sense. I guess that what that means is 1149 01:04:23,200 --> 01:04:25,160 just build better products so that you don't have to call into 1150 01:04:25,160 --> 01:04:26,400 tech support. Maybe there's that's. 1151 01:04:26,760 --> 01:04:30,000 That's my goal, but then you always like the problem. 1152 01:04:30,240 --> 01:04:35,080 Also, QA in your own code, it doesn't work. 1153 01:04:35,240 --> 01:04:36,760 You can't talk. It's because. 1154 01:04:39,160 --> 01:04:41,680 I learned this the hard way. Well, not the hard way, I guess, 1155 01:04:41,680 --> 01:04:45,200 just living right. But when I was younger and doing 1156 01:04:45,200 --> 01:04:50,080 automation and things like that, you know, you'd have these touch 1157 01:04:50,080 --> 01:04:53,000 screens and you know you'd implement everything that you're 1158 01:04:53,000 --> 01:04:57,200 doing and this is how you operate this piece of machinery. 1159 01:04:57,760 --> 01:05:00,800 And the second I let someone else on that thing, they did 1160 01:05:00,800 --> 01:05:05,520 everything in different order and every way I didn't program 1161 01:05:05,520 --> 01:05:08,200 the damn thing to work right. And so things break. 1162 01:05:08,240 --> 01:05:12,160 And then you're like, OK, so now I have to put safeguards in for 1163 01:05:12,360 --> 01:05:16,360 all these other things that people think they want to do, or 1164 01:05:16,360 --> 01:05:19,160 I have to change the UX because something's not clear. 1165 01:05:19,840 --> 01:05:24,160 And so, yeah, I always, you always think you're building 1166 01:05:24,280 --> 01:05:28,320 solid systems, but you never test them in any other ways than 1167 01:05:28,320 --> 01:05:31,080 the way you built them from the engineering standpoint. 1168 01:05:32,000 --> 01:05:34,200 I think that's really true, but I, I think one of the things you 1169 01:05:34,200 --> 01:05:37,280 said in there was maybe even more critical and important, 1170 01:05:37,280 --> 01:05:39,520 which is about the UX side of things. 1171 01:05:39,520 --> 01:05:43,680 And I don't view UX as like how something looks, you know, I 1172 01:05:43,680 --> 01:05:48,200 view it like the, the part, the UX that matters is how do I, how 1173 01:05:48,200 --> 01:05:52,640 do I reduce the surface area of the application to the essence 1174 01:05:52,640 --> 01:05:55,200 of what matters to keep somebody moving forward? 1175 01:05:56,920 --> 01:06:02,680 You know, as engineers, we have a tendency to, I don't know. 1176 01:06:03,120 --> 01:06:05,920 Overcomplicate things. Yeah, it's like, oh, wouldn't it 1177 01:06:05,920 --> 01:06:08,320 be cool if, and it's like every time you have an O wouldn't 1178 01:06:08,320 --> 01:06:11,680 wouldn't it be cool if that's not a one, that's not a + 1 1179 01:06:11,680 --> 01:06:16,400 feature, It's a like it's a quadratic or it's a it's a power 1180 01:06:16,400 --> 01:06:22,960 function of a solution space or bug space that you've created in 1181 01:06:22,960 --> 01:06:26,160 this weird combinatoric effect that you can't possibly test. 1182 01:06:26,680 --> 01:06:31,360 And so, you know, I, I'm a, I fight this everyday. 1183 01:06:31,360 --> 01:06:35,560 So that's why I'm kind of saying this out loud of trying to 1184 01:06:35,560 --> 01:06:42,080 simplify and reduce and, and, and most users want things that 1185 01:06:42,080 --> 01:06:44,920 just work more than more features. 1186 01:06:45,520 --> 01:06:48,680 I think we'll do a little bit more service to our community by 1187 01:06:49,120 --> 01:06:51,520 trying to adhere to that amount. True, just a little bit more. 1188 01:06:52,120 --> 01:06:54,320 Yeah, when I first started building Kill Switch, I had all 1189 01:06:54,320 --> 01:06:59,000 these grand plans of like I wanted like very specific use 1190 01:06:59,000 --> 01:07:05,200 cases for storing like crypto wallets and and like OTP code, 1191 01:07:05,240 --> 01:07:09,280 rotating codes for passwords. And I overcomplicated everything 1192 01:07:09,280 --> 01:07:12,400 and I ended up stripping it down to just look you can create 1193 01:07:13,080 --> 01:07:16,760 encrypted notes and you can upload any document or file that 1194 01:07:16,760 --> 01:07:23,560 you want and that that's it. And I stopped and I, I think as 1195 01:07:23,560 --> 01:07:28,280 more people start using it, I'll probably get some good requests 1196 01:07:28,280 --> 01:07:31,160 that I want to add in the future, But I really feel like 1197 01:07:31,160 --> 01:07:34,240 keeping it simple. Like I want people to be able to 1198 01:07:34,240 --> 01:07:37,160 open any app that I build and not need an instruction guide to 1199 01:07:37,160 --> 01:07:40,480 use it. I think it's a very, that's 1200 01:07:40,600 --> 01:07:44,280 perfect, like the guiding principle to strive towards 1201 01:07:44,280 --> 01:07:46,600 because I think if you can do that, then all those things 1202 01:07:46,600 --> 01:07:49,640 you're talking about how people use things incorrectly and, and 1203 01:07:49,640 --> 01:07:53,240 put things your system in weird state and all that just becomes 1204 01:07:53,760 --> 01:07:56,640 a decreasing function. And I'm a big believer that 1205 01:07:56,640 --> 01:07:59,000 nothing stays static. It's either like things are 1206 01:07:59,000 --> 01:08:02,240 either going up or they're going down and it's all trying to stay 1207 01:08:02,240 --> 01:08:04,800 on the right side of that. And it's a million little 1208 01:08:04,800 --> 01:08:07,880 decisions you make that you wake up 1 morning and you're either 1209 01:08:07,880 --> 01:08:10,120 in this death spiral or you're like, oh wow. 1210 01:08:10,120 --> 01:08:13,160 Like this is something that that's actually growing and like 1211 01:08:13,160 --> 01:08:17,520 creating like a synergistic effect instead of a avalanche of 1212 01:08:17,520 --> 01:08:21,520 doom. Yeah, one of my like, simplest 1213 01:08:21,560 --> 01:08:25,600 features. It's literally, you know, hard 1214 01:08:25,600 --> 01:08:30,720 coded, kind of like an onboarding checklist. 1215 01:08:31,240 --> 01:08:34,760 And because so many people are like, well, what kinds of 1216 01:08:34,760 --> 01:08:37,240 documents should I be uploading to kill switch? 1217 01:08:37,680 --> 01:08:43,399 And I was like, OK, I need like a drip campaign e-mail or like a 1218 01:08:43,399 --> 01:08:47,160 checklist to track these things. So like now all I have is a 1219 01:08:47,160 --> 01:08:51,319 checklist of like you. OK, you should upload a will, 1220 01:08:51,319 --> 01:08:54,479 you should upload power of attorney, you should use crypto 1221 01:08:54,479 --> 01:08:56,319 wallets. And I have all these categories 1222 01:08:56,319 --> 01:08:58,760 of different things you want to add to kill switch. 1223 01:08:59,160 --> 01:09:01,160 And then I have a drip e-mail system. 1224 01:09:01,160 --> 01:09:03,200 You get like 8 emails over two months. 1225 01:09:03,200 --> 01:09:07,160 So it doesn't annoy people, but it goes into each specific 1226 01:09:07,160 --> 01:09:11,760 category into detail and then helps you like, OK, what type of 1227 01:09:11,760 --> 01:09:14,800 person or relationship should be the beneficiary of this 1228 01:09:14,800 --> 01:09:17,120 document? And it kind of just like guides 1229 01:09:17,120 --> 01:09:20,920 them through this process. And so by the end, like so 60 1230 01:09:20,920 --> 01:09:24,200 days later, they shouldn't be overwhelmed, but then everything 1231 01:09:24,240 --> 01:09:26,600 should be set up in kill switch form and then they should never 1232 01:09:26,600 --> 01:09:30,439 have to log in again. Yeah, that's really cool. 1233 01:09:30,479 --> 01:09:33,880 I like the fact that you people are building software that is, I 1234 01:09:33,880 --> 01:09:36,399 never have to log in again, but it just does something for me 1235 01:09:36,399 --> 01:09:38,399 and I'm, I don't have to think about it because everybody's 1236 01:09:38,399 --> 01:09:41,800 wanting goals and, and like engagement and stuff and like 1237 01:09:41,800 --> 01:09:44,000 that's, that's important for certain types of things, but 1238 01:09:44,000 --> 01:09:47,200 there's also like stuff out there that's, that's exactly 1239 01:09:47,200 --> 01:09:48,680 like what you're doing, which is just like. 1240 01:09:48,960 --> 01:09:52,240 My goal is once it's set up, you just make sure things are up to 1241 01:09:52,240 --> 01:09:56,520 date maybe yearly, twice a year, whatever it may be, right? 1242 01:09:56,960 --> 01:09:59,120 And what's crazy is I didn't even think about this. 1243 01:09:59,120 --> 01:10:02,440 I have an awesome beta tester. He's like, look, I really like 1244 01:10:02,440 --> 01:10:06,720 the the check in link, but I hate that I have to auth in and 1245 01:10:06,800 --> 01:10:09,680 then it's like it turned it was a multi step thing. 1246 01:10:09,960 --> 01:10:12,920 And This is why I'm saying like QA in your own shit doesn't work 1247 01:10:13,760 --> 01:10:17,520 because I, I didn't realize how annoying it was because you had 1248 01:10:17,520 --> 01:10:19,640 an auth in and then you had to check in. 1249 01:10:19,640 --> 01:10:21,920 It was like 2-3 steps you had to take. 1250 01:10:22,440 --> 01:10:27,720 And now I send an auth link and all it does is say OK, your dead 1251 01:10:27,720 --> 01:10:30,000 man switch has been checked in, you're good. 1252 01:10:30,520 --> 01:10:33,320 And like if you want to go further than that, you can sign 1253 01:10:33,320 --> 01:10:36,280 into your account, but it only gives you access to check into 1254 01:10:36,280 --> 01:10:39,560 that single dead man switch. Yeah, so you don't. 1255 01:10:39,560 --> 01:10:40,920 Have to, you don't have to log in. 1256 01:10:40,960 --> 01:10:45,440 It's just one, one click and it comes into your texts and then 1257 01:10:45,440 --> 01:10:48,520 you're done. Yeah. 1258 01:10:48,560 --> 01:10:51,960 My my goal is like, I don't want to see people logging into my 1259 01:10:51,960 --> 01:10:54,160 app, you know? Yeah. 1260 01:10:55,160 --> 01:10:57,000 And that's, that's nice. It's good to hear. 1261 01:10:58,320 --> 01:11:00,640 Yeah, because like what you're saying, everyone was like, how 1262 01:11:00,640 --> 01:11:03,000 do I up engagement? How do I make people addicted to 1263 01:11:03,000 --> 01:11:04,640 this? Like, no, it's a tool. 1264 01:11:04,640 --> 01:11:06,640 It's just there. It needs to work when you don't 1265 01:11:06,640 --> 01:11:10,080 want to think about it. Yeah, yeah, yeah, I'm there. 1266 01:11:14,840 --> 01:11:20,880 But yeah, I, I, I, I am excited of where LLMS are taking us 1267 01:11:20,880 --> 01:11:22,680 because I think it opens the door. 1268 01:11:22,680 --> 01:11:26,040 Like people like me, I don't want to run giant companies. 1269 01:11:26,400 --> 01:11:31,360 I want to run small. Like I'm a big like I've been 1270 01:11:31,360 --> 01:11:35,240 talking about this for a while now, but like micro products and 1271 01:11:35,240 --> 01:11:39,040 if you're making a few $1000 off of each product every month, 1272 01:11:39,040 --> 01:11:42,880 like you can make a pretty good life without a lot of hard work. 1273 01:11:42,880 --> 01:11:46,280 Like if you just have a couple products with a few 1000 users. 1274 01:11:47,440 --> 01:11:50,000 Like you don't. You don't need multiple nodes, 1275 01:11:50,000 --> 01:11:52,200 you don't need a crazy dev OPS solution. 1276 01:11:52,280 --> 01:11:54,280 You just have a product that works. 1277 01:11:55,520 --> 01:12:00,520 Yeah, I think that's that's like an amazing path to to having a 1278 01:12:00,520 --> 01:12:04,120 very sane level of success. And I think there's actually 1279 01:12:04,120 --> 01:12:07,680 another path opening up as well, which is like building really, 1280 01:12:07,680 --> 01:12:10,080 really big companies that just don't have a lot of people. 1281 01:12:11,360 --> 01:12:14,040 And I think that, you know, that's becoming more and more 1282 01:12:14,040 --> 01:12:15,640 practical. And I know a lot of people will 1283 01:12:15,640 --> 01:12:17,480 have a visceral negative reaction to that. 1284 01:12:17,480 --> 01:12:19,880 They're like, well, you know, it's taking jobs away and 1285 01:12:19,880 --> 01:12:21,680 whatnot. And maybe it is. 1286 01:12:21,680 --> 01:12:24,880 I, I, I, I don't like, I'm not here to judge that as good or 1287 01:12:24,880 --> 01:12:25,960 bad. I'm just saying it's now 1288 01:12:25,960 --> 01:12:28,600 possible. So when did when did we get the 1289 01:12:28,600 --> 01:12:31,520 first Unicorn? You know, one person Unicorn 1290 01:12:31,520 --> 01:12:33,240 company like. That'll be wild. 1291 01:12:33,280 --> 01:12:34,840 Actually. That's exciting to think about. 1292 01:12:35,200 --> 01:12:39,440 And like as far as jobs go, I think that's BS. 1293 01:12:39,440 --> 01:12:43,240 I think we're going to just see a lot more creative solutions at 1294 01:12:44,000 --> 01:12:47,520 larger quantity. So we're going to see all these 1295 01:12:47,640 --> 01:12:52,800 awesome products and companies out there by the thousands and 1296 01:12:52,800 --> 01:12:55,160 they're it'll employ the same amount of people. 1297 01:12:55,520 --> 01:12:58,360 It'll just be more solutions. Yep. 1298 01:12:58,640 --> 01:13:02,680 And I think you can extend that same argument to like the just 1299 01:13:02,680 --> 01:13:06,280 the software development career path, like it's changed, there's 1300 01:13:06,280 --> 01:13:10,160 no doubt about it. But we're also in a path where 1301 01:13:10,320 --> 01:13:14,240 people that could not afford bespoke software can now 1302 01:13:14,480 --> 01:13:17,720 possibly afford it. And so the, even though the 1303 01:13:17,720 --> 01:13:23,480 hourly output equation is a little bit different, the target 1304 01:13:23,480 --> 01:13:27,600 environment has just gone from here to here in terms of 1305 01:13:27,840 --> 01:13:32,040 potential people that could use bespoke software. 1306 01:13:32,040 --> 01:13:35,120 And that's really important to me because I, I hate most SAS 1307 01:13:35,120 --> 01:13:39,160 products for the fair and that they're just bloated with all 1308 01:13:39,160 --> 01:13:41,360 kinds of crap that you don't actually need. 1309 01:13:41,640 --> 01:13:43,720 And they're difficult. You have to end up bending your 1310 01:13:43,720 --> 01:13:47,280 own company to try to fit into their, their way of doing 1311 01:13:47,280 --> 01:13:50,920 things. And it's like never again like. 1312 01:13:51,040 --> 01:13:53,400 And you need a user guide just to get started. 1313 01:13:53,880 --> 01:13:59,280 And you need a user guide. All of that is I, I think is, is 1314 01:13:59,360 --> 01:14:01,480 being changed. And so if you can be that person 1315 01:14:01,480 --> 01:14:04,880 that has some, you know, like domain knowledge, whatever it 1316 01:14:04,880 --> 01:14:08,720 is, pick what, pick anything, become an expert in some domain. 1317 01:14:09,280 --> 01:14:13,760 And if you know how to code you, you write your own ticket. 1318 01:14:14,200 --> 01:14:16,680 Yeah, I agree. And well like most like I don't 1319 01:14:16,680 --> 01:14:20,600 touch like I hate like most CMSS and things like that. 1320 01:14:20,600 --> 01:14:24,280 Like why is there so much integration to just track a 1321 01:14:24,280 --> 01:14:28,000 customer relationship? Like I I don't agree with it. 1322 01:14:28,040 --> 01:14:30,880 I don't want to solve the problem though either, but I 1323 01:14:30,880 --> 01:14:34,040 don't use the solutions because they're just bloated and 1324 01:14:34,040 --> 01:14:38,360 overcomplicated. I, I agree that's, so I think 1325 01:14:38,360 --> 01:14:40,720 that, you know, that's a, that's a, you know, a fun thing. 1326 01:14:40,720 --> 01:14:43,320 Like if you're, you know, somebody that like, take the CMS 1327 01:14:43,320 --> 01:14:46,160 thing, you knew CMSS, like you've been living that world 1328 01:14:46,160 --> 01:14:49,080 for forever. Like you could make a legitimate 1329 01:14:49,080 --> 01:14:52,960 business out of going to a bunch of small companies that couldn't 1330 01:14:52,960 --> 01:14:55,280 afford or didn't want to deal with the CMS and say, Hey, look, 1331 01:14:55,280 --> 01:14:57,800 I can build, I can build a bespoke one in a month for you. 1332 01:14:58,320 --> 01:15:00,400 That works how you need it to work. 1333 01:15:00,960 --> 01:15:02,680 And you know, what's going to happen is that they're going to 1334 01:15:02,680 --> 01:15:04,840 come back and they're going to ask for more features and you're 1335 01:15:04,840 --> 01:15:07,320 going to be able to build a, you know, a business around that. 1336 01:15:07,840 --> 01:15:10,800 And you know, those are that's just leveraging your domain 1337 01:15:10,800 --> 01:15:13,720 experience with your technical experience, putting it together 1338 01:15:14,040 --> 01:15:17,160 and becoming, you know, some type of an entrepreneur with 1339 01:15:17,160 --> 01:15:18,560 that. Yeah, I agree. 1340 01:15:18,960 --> 01:15:24,400 And let's stop the corporate form form 90s apps too, like. 1341 01:15:26,000 --> 01:15:28,120 Thank you. Get behind that. 1342 01:15:29,440 --> 01:15:33,480 All the old tech that's just over consumed in large 1343 01:15:33,480 --> 01:15:36,480 corporations because they just buy software and patch it 1344 01:15:36,480 --> 01:15:40,680 together. Yeah, Oh man, yeah. 1345 01:15:40,680 --> 01:15:44,160 I, I remember the, I, I remember sitting in a in a board meeting 1346 01:15:44,600 --> 01:15:49,280 my last company and going through the financials and the 1347 01:15:49,760 --> 01:15:54,720 line item for Salesforce came up and I was just like, really 1348 01:15:54,800 --> 01:15:58,040 like, that is a very, very, we weren't a huge company. 1349 01:15:58,040 --> 01:16:01,920 It was a very, very high 6 figure number for Salesforce, 1350 01:16:04,280 --> 01:16:07,280 but they had their hooks in so deep that there was just no way 1351 01:16:07,280 --> 01:16:10,960 to get out of it. Yeah, I was going to mention 1352 01:16:10,960 --> 01:16:13,160 Salesforce when I was talking about products that are just 1353 01:16:13,160 --> 01:16:18,320 like they do too much. And then and then like it's it's 1354 01:16:18,320 --> 01:16:21,560 like that and like products like SEM rush and things you just 1355 01:16:21,560 --> 01:16:27,040 like this upgrade plan of like, oh, you want our next tool here 1356 01:16:27,040 --> 01:16:29,560 Just, you know, it's just 300 more dollars a month. 1357 01:16:29,560 --> 01:16:32,440 Like what? And then every seat you have to 1358 01:16:32,440 --> 01:16:35,440 add, like if you're a company that you need to have everyone 1359 01:16:35,440 --> 01:16:38,240 on your company using. Like Slack's a good example too. 1360 01:16:38,240 --> 01:16:42,760 Like I don't get it. I think Discord works a lot 1361 01:16:42,760 --> 01:16:45,800 better. And but so many companies like 1362 01:16:46,800 --> 01:16:50,520 I, I have Slack up right now because everyone uses it for, 1363 01:16:50,720 --> 01:16:52,320 you know, companies I have to interact with. 1364 01:16:52,960 --> 01:16:56,960 Yeah, yeah, I agree. They're not the worst offender 1365 01:16:56,960 --> 01:16:58,640 out there, but they're definitely on that path. 1366 01:16:59,120 --> 01:17:01,280 They can be depending on team size, right? 1367 01:17:01,760 --> 01:17:04,880 Yeah, Yeah, that's true. I just saw posted like like 1368 01:17:04,920 --> 01:17:08,160 multiple like I I want to say it was like a six figure maybe it 1369 01:17:08,160 --> 01:17:12,200 was a 5 figure bill for Slack. Which is wild to me because 1370 01:17:12,200 --> 01:17:14,200 there's free solutions out there. 1371 01:17:15,080 --> 01:17:16,800 There are, there are. Yeah, I could. 1372 01:17:17,280 --> 01:17:21,320 I could buy that. But maybe I'm just a cheap 1373 01:17:22,480 --> 01:17:25,920 software engineer that just wants the the open source 1374 01:17:25,920 --> 01:17:31,080 versions. Yeah, yeah, open source 1375 01:17:31,080 --> 01:17:34,480 discussion about how the economics of open source can or 1376 01:17:34,480 --> 01:17:37,760 should work is probably a whole other episode of of a podcast. 1377 01:17:37,760 --> 01:17:40,200 Cuz I, yeah, I have no opinion on it, honestly. 1378 01:17:40,200 --> 01:17:46,160 I, I understand both both sides of things and I, I like, I'm 1379 01:17:46,160 --> 01:17:49,200 very thankful for a lot of open source projects, obviously, 1380 01:17:49,600 --> 01:17:51,560 yeah. But I don't get the anger behind 1381 01:17:51,560 --> 01:17:54,920 it when they like try to close off features and like monetize 1382 01:17:54,920 --> 01:17:57,240 part of their product. Like they deserve that. 1383 01:17:57,920 --> 01:18:04,040 Yeah, yeah, we, there's a lot of freeloading on, on the backs of 1384 01:18:04,040 --> 01:18:06,560 things. And so I I don't know how to, I 1385 01:18:06,560 --> 01:18:10,520 don't know how to solve it. Yeah, there's a big dramatic 1386 01:18:10,520 --> 01:18:16,080 conversation on X as well around open source and like how they're 1387 01:18:16,080 --> 01:18:18,920 going to deal with LLM slop entering code base. 1388 01:18:19,520 --> 01:18:21,720 Yeah. And I, I don't, I don't know. 1389 01:18:21,720 --> 01:18:25,200 I think it opens the barrier of entry to less. 1390 01:18:27,160 --> 01:18:31,080 I don't know developer savvy people, but as long as they can 1391 01:18:31,080 --> 01:18:34,840 read code well, I think you can still keep your standards of in 1392 01:18:34,840 --> 01:18:37,400 the code base up, but I don't know. 1393 01:18:39,160 --> 01:18:41,400 Yeah, I mean, I don't know how to, I don't know how to solve 1394 01:18:41,400 --> 01:18:43,920 that piece. I think it's been interesting to 1395 01:18:43,920 --> 01:18:48,520 watch the, the, the furor around the whole tailwind thing, how 1396 01:18:48,560 --> 01:18:53,240 that's unfolded. And because that's one of the, I 1397 01:18:53,240 --> 01:18:57,560 mean, it's such a powerful tool set that's been built that so 1398 01:18:57,560 --> 01:19:00,400 many people use that I just hate to see him in a position like 1399 01:19:00,400 --> 01:19:03,160 that. Yeah, Yeah, it is. 1400 01:19:03,320 --> 01:19:09,760 It is unfortunate, but I feel like that happens in in most 1401 01:19:09,760 --> 01:19:12,880 popular open source projects. I don't know, it becomes a 1402 01:19:12,880 --> 01:19:15,160 problem. It is, yeah. 1403 01:19:15,160 --> 01:19:16,280 That's where I don't know how to solve it. 1404 01:19:16,280 --> 01:19:17,120 Like it. It doesn't. 1405 01:19:17,440 --> 01:19:21,440 There's not an obvious balancing act there to to get it back on 1406 01:19:21,440 --> 01:19:24,240 track. Hopefully up with something 1407 01:19:24,240 --> 01:19:29,160 though, because I know a lot of people, a lot of people depend 1408 01:19:29,160 --> 01:19:32,680 on things like this and a lot of people who have their careers 1409 01:19:32,680 --> 01:19:35,920 wrapped up and and income wrapped up and making it work. 1410 01:19:35,920 --> 01:19:37,800 So we got to find some type of solution for it. 1411 01:19:37,960 --> 01:19:41,440 Yeah, one of the like one of the things like it used to be like 1412 01:19:41,440 --> 01:19:44,680 large corporations put a significant portion of their 1413 01:19:44,680 --> 01:19:47,720 resources into letting developers contribute to open 1414 01:19:47,720 --> 01:19:50,120 source. And I think the big fear is 1415 01:19:50,120 --> 01:19:52,480 they're seeing that dry up because now these big 1416 01:19:52,480 --> 01:19:58,040 corporations are rolling out their own in house solutions and 1417 01:19:58,800 --> 01:20:02,280 and then there goes, you know, sponsorships and open source 1418 01:20:02,280 --> 01:20:05,520 projects. Yeah, yeah. 1419 01:20:05,520 --> 01:20:07,720 And I wonder a little bit if they like, if some of it isn't 1420 01:20:07,720 --> 01:20:11,720 healthy as well, because there's so many like little, a little 1421 01:20:11,720 --> 01:20:13,800 open source projects where it's just like, do you really need a 1422 01:20:13,800 --> 01:20:16,320 shared project to do that? And it's like, you can, you can 1423 01:20:16,320 --> 01:20:19,360 have, you know, you can actually create bespoke code to, to solve 1424 01:20:19,360 --> 01:20:20,920 some of these problems in a few minutes. 1425 01:20:21,440 --> 01:20:25,240 And so, you know, I, I think that that is another maybe 1426 01:20:25,240 --> 01:20:27,840 filtering effect where, hey, is this really something that 1427 01:20:27,840 --> 01:20:30,000 should exist? Like are you really moving the 1428 01:20:30,000 --> 01:20:32,640 needle and creating something interesting or is it just some 1429 01:20:32,640 --> 01:20:39,680 like random little function? Yeah, I think like I'm a big 1430 01:20:39,680 --> 01:20:43,320 believer in don't pull random dependencies into projects. 1431 01:20:43,320 --> 01:20:49,480 Like I pull them in around security and like currency and 1432 01:20:49,480 --> 01:20:53,240 financial stuff mostly. But for a lot of it you can self 1433 01:20:53,240 --> 01:20:55,560 roll your own solutions pretty easily. 1434 01:20:56,000 --> 01:20:58,560 Yeah, I don't know if you've like, I went back and looked at 1435 01:20:58,560 --> 01:21:04,040 some of the mixed depths, some Phoenix projects over the over 1436 01:21:04,040 --> 01:21:08,680 the years. And it's crazy how much smaller 1437 01:21:08,680 --> 01:21:12,680 the dependency list is just in this last six to nine months. 1438 01:21:13,880 --> 01:21:16,800 Still really complicated, still doing a lot of things, but it's 1439 01:21:16,800 --> 01:21:20,760 like a 10th the size. And I think a lot of it is that 1440 01:21:20,760 --> 01:21:24,200 exact thing where, you know, we're just building it and you 1441 01:21:24,200 --> 01:21:27,560 can actually point something at A at a project and go like, I 1442 01:21:27,560 --> 01:21:29,400 just need this piece. Can you can you just give me 1443 01:21:29,400 --> 01:21:31,680 that piece? And I do that a lot. 1444 01:21:32,800 --> 01:21:35,160 Yeah, yeah, same. I do too. 1445 01:21:36,240 --> 01:21:39,680 Let's see Pablo, how you created your startup USP. 1446 01:21:40,240 --> 01:21:42,000 What are your thoughts you went through? 1447 01:21:43,400 --> 01:21:48,160 USP. I'm not sure what USP is, maybe 1448 01:21:48,160 --> 01:21:50,480 we missed something. Is that a typo Pablo? 1449 01:21:53,600 --> 01:21:56,960 Well, anyways, we'll let him re answer that and we'll just 1450 01:21:56,960 --> 01:21:59,800 continue. Yeah, it'd be interesting to see 1451 01:21:59,800 --> 01:22:04,960 what the question was. Oh, what another thing, 'cause 1452 01:22:04,960 --> 01:22:07,360 you raised, so you raised seed money, right? 1453 01:22:08,880 --> 01:22:10,800 So I was super fortunate on this. 1454 01:22:11,880 --> 01:22:15,400 I did some work with Sidewalk Infrastructure Partners, which 1455 01:22:15,400 --> 01:22:18,920 is a Google spin out many years ago and built some relationships 1456 01:22:18,920 --> 01:22:21,920 there. And one of the one of the 1457 01:22:21,920 --> 01:22:27,080 principals joined Montauk Capital and they had an idea for 1458 01:22:27,080 --> 01:22:30,840 this hyper scale like transmission interconnection 1459 01:22:30,840 --> 01:22:34,040 problem. And so they reached out to me a 1460 01:22:34,040 --> 01:22:36,480 few months ago and said, hey, you know, we've got this idea 1461 01:22:36,480 --> 01:22:38,800 for a company, we would like to bring in a couple founders to 1462 01:22:38,800 --> 01:22:40,640 start it and it comes with a seed check. 1463 01:22:41,400 --> 01:22:45,200 So I had like the most enviable position that I could ever ask 1464 01:22:45,200 --> 01:22:49,160 for of being able to come in as you know, you know, full on 1465 01:22:49,160 --> 01:22:54,480 founder type of role and not have to go raise seed money. 1466 01:22:54,480 --> 01:22:57,960 The check was right there and that's awesome gift. 1467 01:22:58,800 --> 01:23:01,840 Yeah, that's amazing because like, I feel like especially 1468 01:23:01,840 --> 01:23:04,760 these days, like the climate of raising money is a lot more 1469 01:23:04,760 --> 01:23:10,040 difficult. It, it is and come back and, and 1470 01:23:10,040 --> 01:23:14,440 we'll talk in a few months when like after we've done our Series 1471 01:23:14,680 --> 01:23:16,640 A. But I think it's all about like 1472 01:23:17,080 --> 01:23:19,240 what's happening? Like are you in an area that 1473 01:23:19,240 --> 01:23:22,400 actually has tailwind? Because right now there's a ton 1474 01:23:22,400 --> 01:23:26,320 of companies that are wanting to talk to us about investing in, 1475 01:23:26,480 --> 01:23:29,600 you know, Series A and wanting to have nothing to do with 1476 01:23:29,600 --> 01:23:33,080 actually cutting a check. But it's completely different 1477 01:23:33,080 --> 01:23:35,360 than a lot of the people that I know that are out there trying 1478 01:23:35,360 --> 01:23:37,760 to raise money where they're having a hard time even getting 1479 01:23:37,760 --> 01:23:40,120 meetings. I mean, I guess that means 1480 01:23:40,120 --> 01:23:43,440 you're solving A viable solution That's a that's a good 1481 01:23:43,440 --> 01:23:45,840 temperature. I hope so. 1482 01:23:46,080 --> 01:23:48,240 I really do. But I mean, it does seem like a 1483 01:23:48,240 --> 01:23:50,160 real problem. Everybody we talked to that is 1484 01:23:50,160 --> 01:23:52,840 much more knowledgeable about it than I am that has been living 1485 01:23:52,840 --> 01:23:54,480 this. We talked to them and they're 1486 01:23:54,480 --> 01:23:57,640 like, if what you can, if what you're showing me is actually 1487 01:23:57,640 --> 01:24:00,240 real and I can say that it is because I built it, they're like 1488 01:24:00,240 --> 01:24:04,200 this is going to go somewhere. And so that's exactly like I 1489 01:24:04,280 --> 01:24:08,600 think when you're in a environment where capital is a 1490 01:24:08,600 --> 01:24:12,680 little bit constrained or hard to get at, it's really have to 1491 01:24:12,680 --> 01:24:17,040 be focused on being in an area that that people want to invest 1492 01:24:17,040 --> 01:24:20,680 in just throw. AI in your company description. 1493 01:24:21,720 --> 01:24:23,880 No, it's so funny that you say that. 1494 01:24:24,080 --> 01:24:27,400 I would. I think it was actually one of 1495 01:24:27,400 --> 01:24:30,400 Lenny's podcasts just recently. It was talking about that where 1496 01:24:31,040 --> 01:24:36,000 there is an investor on it that said, you know, it's actually 1497 01:24:36,000 --> 01:24:40,480 becoming a little bit of a red flag because they're like a lot 1498 01:24:40,480 --> 01:24:42,240 of investors. You know, there's a whole cod 1499 01:24:42,240 --> 01:24:45,480 ray of investors now that are kind of viewing AI companies as 1500 01:24:45,480 --> 01:24:46,720 like they're all going to go to 0. 1501 01:24:48,200 --> 01:24:50,680 All these companies that not the core ones that are like building 1502 01:24:50,680 --> 01:24:53,480 the infrastructure, but you're like I'm you know, like AI for 1503 01:24:53,480 --> 01:24:56,400 this and it's like no, you're. Not all the bubble companies 1504 01:24:56,400 --> 01:25:03,120 that just like it's the same crap that happened with NFTS and 1505 01:25:03,760 --> 01:25:07,800 and then crypto before that and. It's exactly. 1506 01:25:08,040 --> 01:25:11,560 Coins and all this garbage. Well, and just a personal 1507 01:25:11,560 --> 01:25:14,000 example, like when I joined this, this start up, it was 1508 01:25:14,000 --> 01:25:19,080 called Voltage AI and we changed the name so that we didn't have 1509 01:25:19,080 --> 01:25:22,080 that moniker on there because it was just, we didn't want to be a 1510 01:25:22,400 --> 01:25:24,960 like just pegged as like that type of company. 1511 01:25:24,960 --> 01:25:28,320 We felt like that was a piece like we are absolutely going to 1512 01:25:28,320 --> 01:25:31,360 use AI in a lot of areas to move, to move things forward, 1513 01:25:31,560 --> 01:25:33,240 but we don't want to be an AI company. 1514 01:25:34,560 --> 01:25:39,040 Yeah, I think that was, that's a smart move because everyone is 1515 01:25:40,880 --> 01:25:44,280 what's the word. I think everyone's getting a 1516 01:25:44,280 --> 01:25:50,160 little tired of hearing about AI like the hype, the hype train is 1517 01:25:50,160 --> 01:25:52,200 over. Like, I think we're at a point 1518 01:25:52,200 --> 01:25:57,880 now where we're leveling, we're levering the strengths of AI, 1519 01:25:57,880 --> 01:26:00,360 but we're sick of hearing about it as like the hype. 1520 01:26:01,240 --> 01:26:03,120 That's exactly right. Like just use it for what it 1521 01:26:03,120 --> 01:26:05,320 actually moves the needle for and don't make your company 1522 01:26:05,320 --> 01:26:07,560 about that. I mean, there are a few if 1523 01:26:07,560 --> 01:26:10,320 you're, you know, anthropic or something like that, you're 1524 01:26:10,320 --> 01:26:14,600 about AI, but that's a capital intensive business to be in the 1525 01:26:14,600 --> 01:26:16,320 AI infrastructure that that's a whole different. 1526 01:26:16,800 --> 01:26:18,320 Like I don't that's not my space. 1527 01:26:18,320 --> 01:26:21,320 I will never be that. Well, and like, I don't want to 1528 01:26:21,320 --> 01:26:24,480 knock companies like obviously, like there's a lot of awesome 1529 01:26:24,480 --> 01:26:28,840 tools out there that are leveraging AI to basically 1530 01:26:28,840 --> 01:26:33,280 remove friction, right? The ability to make your team 1531 01:26:33,280 --> 01:26:35,360 smaller. And I think there's, there's a 1532 01:26:35,360 --> 01:26:39,680 lot of awesome projects out there, whether it's like the 1533 01:26:39,680 --> 01:26:42,160 marketing space or whatever it may be. 1534 01:26:42,480 --> 01:26:45,960 Like I don't want to sell them short, but like, as long as 1535 01:26:45,960 --> 01:26:49,120 you're like creating powerful tools that are actually 1536 01:26:49,120 --> 01:26:51,680 beneficial, I think you're in a good spot. 1537 01:26:51,680 --> 01:26:54,680 But if you're just throwing AI in your product to get eyeballs 1538 01:26:54,680 --> 01:26:57,400 and attention, yeah, you're going to 0. 1539 01:26:58,080 --> 01:27:00,480 Yep, 100% I couldn't agree more. That's. 1540 01:27:01,880 --> 01:27:06,040 Oh, USP unique selling point. See, I'm not in on the lingo. 1541 01:27:06,040 --> 01:27:10,080 I had no idea. Those TLA's are hard to keep up 1542 01:27:10,080 --> 01:27:12,920 with. So what's our what was the 1543 01:27:12,920 --> 01:27:16,200 question around unique so? How you how you created your 1544 01:27:16,200 --> 01:27:19,360 startup unique selling point? What are like thoughts you went 1545 01:27:19,360 --> 01:27:23,120 through? Well, that, that's a, you know, 1546 01:27:23,120 --> 01:27:27,600 I think going back to what, and when I, when I came in, like 1547 01:27:27,640 --> 01:27:30,640 there was an idea for a company like, and it was not my original 1548 01:27:30,640 --> 01:27:33,480 idea. They, they, the investment firm 1549 01:27:33,480 --> 01:27:35,960 had this idea, but it was a little bit different than what 1550 01:27:35,960 --> 01:27:37,840 it is. And we've gone through a few 1551 01:27:37,840 --> 01:27:44,360 different iterations of this and a lot of it came down to, you 1552 01:27:44,360 --> 01:27:47,360 know, just like this is nothing special. 1553 01:27:47,360 --> 01:27:50,240 Like, I feel like, almost like I'm just repeating a, a, a 1554 01:27:50,440 --> 01:27:53,720 ridiculous mantra. Like we went out and we talked 1555 01:27:53,720 --> 01:27:57,960 to developers, we talked to investors in development in, in 1556 01:27:58,040 --> 01:28:01,280 development projects for like data centers, but not, we talked 1557 01:28:01,360 --> 01:28:03,920 to the people that are managing the interconnection queues on 1558 01:28:03,920 --> 01:28:07,520 the utilities. We, you know, hired like all 1559 01:28:07,520 --> 01:28:10,080 these companies like alpha sites and stuff that give you like 1560 01:28:10,120 --> 01:28:13,240 access to like hourly discussions with industry 1561 01:28:13,240 --> 01:28:15,800 experts. We did a bunch of those. 1562 01:28:16,440 --> 01:28:20,480 And it just becomes clear over time that you, you hear the 1563 01:28:20,480 --> 01:28:24,920 same, you hear the same sets of problems, you hear the same 1564 01:28:24,920 --> 01:28:28,560 worries, you hear the same, you know, frustrations about 1565 01:28:28,560 --> 01:28:33,440 existing tools. And we just honed closer and 1566 01:28:33,440 --> 01:28:36,240 closer into the essence of what we were hearing there. 1567 01:28:36,680 --> 01:28:39,080 And we were lucky that we had, we were around an area where 1568 01:28:39,080 --> 01:28:41,200 it's like, no, there's a massive amount of money being spent on 1569 01:28:41,200 --> 01:28:42,760 this. People are pissed off. 1570 01:28:43,040 --> 01:28:45,360 People are, you know, they, they don't like the way it's working. 1571 01:28:45,360 --> 01:28:51,320 So clearly we need a solution and it was just opening your 1572 01:28:51,320 --> 01:28:54,160 ears and listening to what, what's really going on out there 1573 01:28:54,160 --> 01:28:57,240 and not trying to, to be too beholden to, because like our 1574 01:28:57,240 --> 01:29:02,040 initial was to actually have a really strong presence in being 1575 01:29:02,040 --> 01:29:07,400 able to lease the mobile gear to solve the, the violations. 1576 01:29:08,240 --> 01:29:10,240 And the problem with that, it's like that's not, we're going to 1577 01:29:10,240 --> 01:29:12,960 do it eventually, but it's so capital intensive and there's 1578 01:29:12,960 --> 01:29:17,680 such a lead time to it that it's not the proper use of VC money 1579 01:29:17,680 --> 01:29:21,000 to get to get it off the ground. And so, and that was one of the 1580 01:29:21,000 --> 01:29:24,880 things that you know that we, you know, edited out of our of 1581 01:29:24,880 --> 01:29:26,760 our first round of the product and. 1582 01:29:27,480 --> 01:29:31,680 It's that MVP really got a skim. Skim the feature list. 1583 01:29:32,240 --> 01:29:35,880 You do, you do. Although I'll, I'll be honest, 1584 01:29:35,880 --> 01:29:41,040 like we're getting to the point where we're actually able to, to 1585 01:29:41,040 --> 01:29:44,120 release features pretty quickly and like it's easy to just. 1586 01:29:44,880 --> 01:29:47,000 Add another one and add another one. 1587 01:29:48,400 --> 01:29:50,520 Have you noticed that you've been able to get away from the 1588 01:29:50,520 --> 01:29:54,680 concept of a backlog And like, we don't have a backlog anymore 1589 01:29:54,680 --> 01:29:56,880 because if there's something important to build, we just 1590 01:29:56,880 --> 01:30:00,680 build it and, you know, a day or two usually to build it and, you 1591 01:30:00,680 --> 01:30:03,280 know, maybe a couple days to refine it and it's like. 1592 01:30:03,720 --> 01:30:05,920 What does your backlog look like for bugs though? 1593 01:30:12,600 --> 01:30:15,520 It's deeper than I would like based on where we're at because 1594 01:30:15,520 --> 01:30:16,920 we've had to move pretty quickly. 1595 01:30:17,680 --> 01:30:22,160 But I don't think it's necessarily bugs like it's more 1596 01:30:22,160 --> 01:30:25,120 like known problems like hey, we just didn't take that feature to 1597 01:30:25,120 --> 01:30:28,120 it's fine to well. I don't mean that in a negative 1598 01:30:28,120 --> 01:30:29,880 manner. Like I think it's impossible to 1599 01:30:29,880 --> 01:30:32,360 write code and have it say it's a bug free. 1600 01:30:32,400 --> 01:30:34,960 I don't think that's a realistic statement, you know? 1601 01:30:35,280 --> 01:30:38,680 Never going to happen no, but I, I think I do think that that, 1602 01:30:38,680 --> 01:30:41,440 you know, probably you got to keep a closer eye on that 1603 01:30:41,440 --> 01:30:44,600 because you can like drive yourself right up to the edge of 1604 01:30:44,600 --> 01:30:49,600 the Cliff pretty quickly without knowing that's in with this with 1605 01:30:49,600 --> 01:30:53,680 this approach, but you can also swerve away from that Cliff, 1606 01:30:53,760 --> 01:30:57,080 that Cliff a lot quicker with this approach too, so. 1607 01:30:59,440 --> 01:31:02,680 Also to to like reiterate on your answer a little bit like 1608 01:31:02,680 --> 01:31:05,600 the unique selling point, like how do you get an idea for a 1609 01:31:05,600 --> 01:31:07,840 product? I think you just, it's like the 1610 01:31:07,840 --> 01:31:12,480 most important thing is you have to be solving a problem and you 1611 01:31:12,480 --> 01:31:14,600 have to make sure you're not inventing the problem. 1612 01:31:14,600 --> 01:31:17,960 Because I've seen so many startups try to invent a problem 1613 01:31:18,280 --> 01:31:19,920 that isn't actually a pain point. 1614 01:31:19,920 --> 01:31:23,720 And then sometimes they're lucky enough to raise money, but like, 1615 01:31:23,720 --> 01:31:28,160 it's still, you know, whether it's five years down the road, 1616 01:31:28,240 --> 01:31:30,400 they, they, they're going to close their doors. 1617 01:31:31,520 --> 01:31:33,800 I agree. I'll and it's, it's an overused 1618 01:31:33,800 --> 01:31:35,800 term, but I think it's, it's so, so clear. 1619 01:31:35,800 --> 01:31:40,000 It's like, don't ever build a company that makes vitamins. 1620 01:31:40,000 --> 01:31:41,880 You always want to make painkillers. 1621 01:31:44,280 --> 01:31:46,840 I mean, and it's just like so many companies out there, 1622 01:31:46,840 --> 01:31:49,440 they're, they're trying to make vitamins and it's like, yes, 1623 01:31:50,320 --> 01:31:52,200 people say they want to feel better. 1624 01:31:52,280 --> 01:31:56,240 People say they want to be healthier, but in reality, 1625 01:31:56,400 --> 01:32:00,440 everybody that's that's writing checks is trying to avoid pain. 1626 01:32:00,680 --> 01:32:03,520 They're trying to avoid embarrassment, they're trying to 1627 01:32:03,520 --> 01:32:07,800 avoid getting fired. And so if you play off of those, 1628 01:32:08,120 --> 01:32:11,320 I think you're, you're, you stand a much better chance. 1629 01:32:12,400 --> 01:32:16,640 And then to maybe just one more thing on that, if you get those 1630 01:32:16,800 --> 01:32:20,200 painkiller ideas. The other thing is learn how to 1631 01:32:20,200 --> 01:32:23,080 sell a little bit and realize that you don't. 1632 01:32:23,080 --> 01:32:28,200 Nobody buys on facts. They might choose to not buy 1633 01:32:28,280 --> 01:32:31,480 your stuff because of a fact, but they will never choose to 1634 01:32:31,480 --> 01:32:34,960 buy your stuff because you listed a laundry list of facts. 1635 01:32:35,200 --> 01:32:36,920 They don't. Want that emotion, you know? 1636 01:32:37,240 --> 01:32:39,600 Exactly. And I think one of the best 1637 01:32:39,600 --> 01:32:43,360 things that I did in my recent career was I took a couple years 1638 01:32:43,360 --> 01:32:47,760 off and I went to film school. Just random weirdness thing. 1639 01:32:47,760 --> 01:32:50,360 My wife and I both took some time off and went and did that 1640 01:32:50,360 --> 01:32:51,880 together. That's. 1641 01:32:52,160 --> 01:32:54,360 Fun. Learn how to tell a story. 1642 01:32:55,240 --> 01:32:58,600 And stories are the only thing that move people, really move 1643 01:32:58,600 --> 01:33:01,920 people. And that is something that it, 1644 01:33:02,080 --> 01:33:04,840 even if you're not a traditional salesperson, if that's not your 1645 01:33:04,840 --> 01:33:09,400 role, you're gonna have a responsibility to sell an idea 1646 01:33:09,840 --> 01:33:14,520 or a new technology or whatever it is in your company. 1647 01:33:14,560 --> 01:33:18,520 And if you don't know how to hit those story beats and to pull on 1648 01:33:18,520 --> 01:33:21,480 the heartstrings of that person sitting across the table, then 1649 01:33:21,480 --> 01:33:24,000 you're gonna lose an opportunity to do something that may be 1650 01:33:24,000 --> 01:33:26,760 really powerful just because tell the story. 1651 01:33:26,920 --> 01:33:30,680 Not because your idea was bad, but you couldn't tell a story 1652 01:33:30,680 --> 01:33:34,080 that moved people. And if you can learn one trait, 1653 01:33:34,320 --> 01:33:38,760 I think that will solve so many problems for you in the world. 1654 01:33:38,760 --> 01:33:43,120 It is learn to be a storyteller and that'll that'll unlock so 1655 01:33:43,120 --> 01:33:46,280 many doors. I think that, yeah, that's great 1656 01:33:46,280 --> 01:33:49,120 advice. I don't know where my skill set 1657 01:33:49,120 --> 01:33:52,440 lies there, but it's always something I feel like I have to 1658 01:33:52,440 --> 01:33:52,960 work on. I. 1659 01:33:54,400 --> 01:33:57,000 I think we all do like it. If we're all great at it, then 1660 01:33:57,000 --> 01:33:59,800 everybody would be authors and movie makers and whatnot. 1661 01:33:59,800 --> 01:34:03,280 And it's hard. Like it's just really hard. 1662 01:34:03,560 --> 01:34:05,400 But you can also always get a little better. 1663 01:34:05,600 --> 01:34:08,080 You don't one. Thing I'm really good at is 1664 01:34:08,880 --> 01:34:13,240 asking stupid questions and I've gotten really good at that, but 1665 01:34:13,560 --> 01:34:16,960 people seem to enjoy them. Being comfortable with that is 1666 01:34:16,960 --> 01:34:17,840 the only way you're going to learn. 1667 01:34:18,280 --> 01:34:24,120 Yeah, I I agree. Well, like The thing is, I know 1668 01:34:24,120 --> 01:34:28,160 a lot of really smart people and I always find it fascinating the 1669 01:34:28,600 --> 01:34:33,120 vast difference in our backgrounds, like where where 1670 01:34:33,120 --> 01:34:36,440 our expertise lies and like, and then we all can join and we're 1671 01:34:36,440 --> 01:34:39,560 kind of doing the same job. You know, like if I look at all 1672 01:34:39,560 --> 01:34:42,800 the back end engineers that I'm working with at Collects right 1673 01:34:42,800 --> 01:34:45,840 now, we have such different backgrounds, but we're all 1674 01:34:45,840 --> 01:34:47,360 working on the same project together. 1675 01:34:47,760 --> 01:34:51,320 And it's cool to kind of see skill sets converge like that. 1676 01:34:51,840 --> 01:34:54,120 Yeah, that is really cool. That's really cool. 1677 01:34:55,000 --> 01:34:58,520 Yeah. But all right, I think this was 1678 01:34:58,520 --> 01:34:59,960 awesome. And then we can kind of wrap 1679 01:34:59,960 --> 01:35:02,120 this up. We've been talking for an hour 1680 01:35:02,120 --> 01:35:05,240 and a half and if you ever want to come back on, please reach 1681 01:35:05,240 --> 01:35:06,360 out. This is a lot of fun. 1682 01:35:06,720 --> 01:35:09,040 I really appreciate it was fantastic getting to know you a 1683 01:35:09,040 --> 01:35:11,920 little bit and getting to talk. Anything I can ever do for you, 1684 01:35:12,200 --> 01:35:14,120 don't hesitate to reach out all. Right. 1685 01:35:14,120 --> 01:35:16,240 Yeah. Well, and if when I start my 1686 01:35:16,240 --> 01:35:20,400 second podcast, we'll have to fly you out to Utah because it's 1687 01:35:20,400 --> 01:35:22,680 going to be an impersonal 1. I love it. 1688 01:35:23,080 --> 01:35:26,640 That'd be fantastic. Awesome, thank you so much. 1689 01:35:26,640 --> 01:35:29,280 And for everyone listening, thanks for tuning in and I love 1690 01:35:29,280 --> 01:35:31,800 when the chat is kind of bumping off it. 1691 01:35:32,480 --> 01:35:34,560 I like it. It's more fun including people 1692 01:35:34,560 --> 01:35:36,400 that are are watching the stream. 1693 01:35:36,400 --> 01:35:39,880 So I enjoy it. But all right, guys, thank you 1694 01:35:39,880 --> 01:35:41,800 so much and everyone have a good weekend. 1695 01:35:41,800 --> 01:35:46,240 Take care testing.