1 00:00:03,274 --> 00:00:05,294 [music] 2 00:00:11,790 --> 00:00:13,810 [music] 3 00:00:15,120 --> 00:00:21,600 Well, uh, Vadim, thanks for joining me 4 00:00:18,480 --> 00:00:24,560 today. Uh, you are the founder and CEO 5 00:00:21,600 --> 00:00:26,640 of a company that's been around since, 6 00:00:24,560 --> 00:00:30,320 uh 2008 7 00:00:26,640 --> 00:00:32,399 um, around there. Defco. Uh, tell us 8 00:00:30,320 --> 00:00:34,239 about Defco. And, uh, I'm actually 9 00:00:32,399 --> 00:00:36,480 excited to have you on because you're, I 10 00:00:34,239 --> 00:00:38,800 guess I would say like a very 11 00:00:36,480 --> 00:00:41,120 experienced engineer, software engineer, 12 00:00:38,800 --> 00:00:42,399 and um [snorts] 13 00:00:41,120 --> 00:00:44,079 um, looking forward to this 14 00:00:42,399 --> 00:00:45,840 conversation. 15 00:00:44,079 --> 00:00:47,520 >> Yeah, I mean, thanks for inviting. Yeah, 16 00:00:45,840 --> 00:00:51,280 I know a few things about computers. 17 00:00:47,520 --> 00:00:54,879 Maybe that's can't confirm it. But um my 18 00:00:51,280 --> 00:00:56,640 company is uh uh diff we developing uh 19 00:00:54,879 --> 00:00:59,039 different software for different uh 20 00:00:56,640 --> 00:01:01,359 gross stage startups and enterprises and 21 00:00:59,039 --> 00:01:03,280 yeah we started 18 years ago. So in some 22 00:01:01,359 --> 00:01:07,680 countries we can all start drinking 23 00:01:03,280 --> 00:01:10,159 finally. uh but it been a journey uh and 24 00:01:07,680 --> 00:01:12,720 right now our main focus in gent 25 00:01:10,159 --> 00:01:14,560 development and building uh custom 26 00:01:12,720 --> 00:01:16,479 systems with enhanced development 27 00:01:14,560 --> 00:01:19,119 techniques and as well as uh building 28 00:01:16,479 --> 00:01:21,680 agents for many enterprises that we work 29 00:01:19,119 --> 00:01:23,759 with. So basically replenishing human 30 00:01:21,680 --> 00:01:26,880 workflows uh and automating human 31 00:01:23,759 --> 00:01:28,479 workflows with actually agents that 32 00:01:26,880 --> 00:01:31,200 works. 33 00:01:28,479 --> 00:01:32,880 Yeah. And you know, in our pre-all, I 34 00:01:31,200 --> 00:01:35,439 talked to you quite a bit and and it 35 00:01:32,880 --> 00:01:39,840 just seemed like you were good engineers 36 00:01:35,439 --> 00:01:42,079 before and now with AI tools, it just 37 00:01:39,840 --> 00:01:44,560 seems like, you know, everybody uses 38 00:01:42,079 --> 00:01:47,439 that term 10x engineer or 100x engineer, 39 00:01:44,560 --> 00:01:49,759 but I think in your case, in my opinion, 40 00:01:47,439 --> 00:01:52,880 it actually applies seeing some of the 41 00:01:49,759 --> 00:01:56,320 output that you guys have uh have done. 42 00:01:52,880 --> 00:01:59,119 What do you think about that? 43 00:01:56,320 --> 00:02:02,799 I think realistically when we talk about 44 00:01:59,119 --> 00:02:06,079 10x engineer 100x or like again you hear 45 00:02:02,799 --> 00:02:10,239 all over the the numbers realistically 46 00:02:06,079 --> 00:02:12,640 this is the question of um basically 47 00:02:10,239 --> 00:02:14,560 what type of project you have and what 48 00:02:12,640 --> 00:02:17,599 kind of and how you actually measuring 49 00:02:14,560 --> 00:02:20,720 this because first of all this is 50 00:02:17,599 --> 00:02:22,800 unrealistic metric the same as like you 51 00:02:20,720 --> 00:02:24,800 remember the on the old days was like oh 52 00:02:22,800 --> 00:02:26,319 this engineer delivered so many 53 00:02:24,800 --> 00:02:26,800 thousands lines of code. 54 00:02:26,319 --> 00:02:28,879 >> Yeah. 55 00:02:26,800 --> 00:02:29,680 >> And does it mean that this engineer is 56 00:02:28,879 --> 00:02:31,520 good 57 00:02:29,680 --> 00:02:34,959 >> and like how we can actually measure 58 00:02:31,520 --> 00:02:37,200 that we 10 times faster? 59 00:02:34,959 --> 00:02:41,200 >> It is an actually interesting type of 60 00:02:37,200 --> 00:02:44,000 illusion because in fact uh 61 00:02:41,200 --> 00:02:45,120 >> in many cases humans do not touch the 62 00:02:44,000 --> 00:02:48,560 tickets 63 00:02:45,120 --> 00:02:50,879 >> or like not touching the codes etc. So 64 00:02:48,560 --> 00:02:51,840 why and how we can actually measure 65 00:02:50,879 --> 00:02:54,959 this? Mhm. 66 00:02:51,840 --> 00:02:58,160 >> So it is an output per number of ticket. 67 00:02:54,959 --> 00:03:00,640 But this is also questionable. Can we 68 00:02:58,160 --> 00:03:04,319 measure this by speed of development of 69 00:03:00,640 --> 00:03:07,120 the project? Can yes but in reality 70 00:03:04,319 --> 00:03:10,080 architectural things in a project and 71 00:03:07,120 --> 00:03:13,519 product type of decisions is actually 72 00:03:10,080 --> 00:03:16,159 most uh limiting factor in many cases. 73 00:03:13,519 --> 00:03:18,800 So again, um I think when we're talking 74 00:03:16,159 --> 00:03:24,080 about those numbers, it's really great 75 00:03:18,800 --> 00:03:27,440 tagline for Techrunch article. Um but 76 00:03:24,080 --> 00:03:30,080 to actual folks on a field, it is more 77 00:03:27,440 --> 00:03:32,239 like a bluff and you like, yeah, great. 78 00:03:30,080 --> 00:03:33,360 Um we have the number, but what does 79 00:03:32,239 --> 00:03:35,440 actually mean? 80 00:03:33,360 --> 00:03:38,319 >> Well, what about let's let's go to 81 00:03:35,440 --> 00:03:40,239 something really pragmatic like costs 82 00:03:38,319 --> 00:03:44,720 for the same project. Would you say the 83 00:03:40,239 --> 00:03:46,080 cost is is less or more or the same with 84 00:03:44,720 --> 00:03:48,720 AI tools? 85 00:03:46,080 --> 00:03:50,560 >> I think realistically we cut down the 86 00:03:48,720 --> 00:03:53,200 cost 87 00:03:50,560 --> 00:03:53,760 uh for our clients by at least 4x. 88 00:03:53,200 --> 00:03:55,680 >> Wow. 89 00:03:53,760 --> 00:03:58,400 >> In most cases 90 00:03:55,680 --> 00:03:59,920 >> I would say in most cases it's 91 00:03:58,400 --> 00:04:02,720 >> still discussion. 92 00:03:59,920 --> 00:04:04,799 >> Um it's also really important to 93 00:04:02,720 --> 00:04:06,959 determine if it's a new project or if 94 00:04:04,799 --> 00:04:11,120 it's permale project. So it's existing 95 00:04:06,959 --> 00:04:13,200 project those completely different um an 96 00:04:11,120 --> 00:04:16,320 illusion that you have a like a 97 00:04:13,200 --> 00:04:19,440 20-year-old u legacy code base and you 98 00:04:16,320 --> 00:04:21,919 can really do wonderful things with AI 99 00:04:19,440 --> 00:04:24,320 about this. That's a really interesting 100 00:04:21,919 --> 00:04:26,880 uh myth. It's not in reality 101 00:04:24,320 --> 00:04:29,040 unfortunately. You still have a many 102 00:04:26,880 --> 00:04:31,680 many legacy issues as you would with 103 00:04:29,040 --> 00:04:34,400 humans as well and for many of those 104 00:04:31,680 --> 00:04:37,360 projects they not actually uh documented 105 00:04:34,400 --> 00:04:38,240 well. So you have a lot of uh surprises 106 00:04:37,360 --> 00:04:40,160 in along the way. 107 00:04:38,240 --> 00:04:41,520 >> So a lot of refactoring going on 108 00:04:40,160 --> 00:04:44,320 possibly. 109 00:04:41,520 --> 00:04:46,320 >> Yeah. And this is really interesting 110 00:04:44,320 --> 00:04:49,520 conversation when what type of project 111 00:04:46,320 --> 00:04:52,639 you actually encounter and how to 112 00:04:49,520 --> 00:04:54,880 approach uh because if we have a new 113 00:04:52,639 --> 00:04:58,160 completely new project. Yeah. But like 114 00:04:54,880 --> 00:05:01,840 again in the same way I don't see people 115 00:04:58,160 --> 00:05:05,440 building um competitors of uh hotspot 116 00:05:01,840 --> 00:05:07,120 and salesforce in one day. That's not 117 00:05:05,440 --> 00:05:09,440 happening and I don't think it will 118 00:05:07,120 --> 00:05:11,039 happen next year or two. uh I don't 119 00:05:09,440 --> 00:05:14,160 think we're near uh those kind of 120 00:05:11,039 --> 00:05:18,880 things. So actual large piece of 121 00:05:14,160 --> 00:05:21,520 software it's not replaceable by uh AI 122 00:05:18,880 --> 00:05:25,680 at least at this moment. Can you develop 123 00:05:21,520 --> 00:05:27,759 really large system um with AI help and 124 00:05:25,680 --> 00:05:29,919 develop it faster than previously? 125 00:05:27,759 --> 00:05:32,880 Absolutely. Uh but again those things 126 00:05:29,919 --> 00:05:35,759 are completely different. 127 00:05:32,880 --> 00:05:39,440 So, I mean, one of the last questions I 128 00:05:35,759 --> 00:05:42,000 was going to ask you was like, what 129 00:05:39,440 --> 00:05:43,840 would you build today 130 00:05:42,000 --> 00:05:45,759 as a 131 00:05:43,840 --> 00:05:47,360 as a founder? Like, what would you 132 00:05:45,759 --> 00:05:48,639 build? Like, let's say you were starting 133 00:05:47,360 --> 00:05:50,400 your own project. What would you build 134 00:05:48,639 --> 00:05:51,919 today? Now, you know, I was going to 135 00:05:50,400 --> 00:05:54,320 leave that for last, but it sounds like 136 00:05:51,919 --> 00:05:56,479 it just fits right into this part of 137 00:05:54,320 --> 00:05:58,080 things. You know, if you're not going to 138 00:05:56,479 --> 00:06:02,280 build SAP or you're not going to build 139 00:05:58,080 --> 00:06:02,280 Salesforce, what are you going to build? 140 00:06:02,560 --> 00:06:08,560 I mean it's all comes down to the old 141 00:06:05,520 --> 00:06:10,240 matics of um system of decision, system 142 00:06:08,560 --> 00:06:11,199 of engagement and system of record. 143 00:06:10,240 --> 00:06:14,240 Yeah. 144 00:06:11,199 --> 00:06:17,039 >> Um and now we have also another thing 145 00:06:14,240 --> 00:06:20,000 that we call a system of intelligence 146 00:06:17,039 --> 00:06:23,120 which is basically uh system of agents 147 00:06:20,000 --> 00:06:26,560 or agentic systems which is a quite a 148 00:06:23,120 --> 00:06:28,960 new concept in a way and um it's really 149 00:06:26,560 --> 00:06:30,400 dependent on what uh could be done. 150 00:06:28,960 --> 00:06:34,160 reality it's a lot of different 151 00:06:30,400 --> 00:06:37,759 opportunities I would say um what to 152 00:06:34,160 --> 00:06:40,240 build I think realistically 153 00:06:37,759 --> 00:06:43,680 I would try to go against the most 154 00:06:40,240 --> 00:06:45,520 hardest um and most compliance-based 155 00:06:43,680 --> 00:06:49,039 industry 156 00:06:45,520 --> 00:06:51,440 because no one would look there and 157 00:06:49,039 --> 00:06:53,600 those things looks really tough in 158 00:06:51,440 --> 00:06:59,280 reality they in many many cases not 159 00:06:53,600 --> 00:07:02,720 really tough um and uh they all wants to 160 00:06:59,280 --> 00:07:08,000 automate um again banking, healthcare, 161 00:07:02,720 --> 00:07:12,479 insurance type um of uh businesses 162 00:07:08,000 --> 00:07:15,520 any like complex uh manufacturing u drug 163 00:07:12,479 --> 00:07:19,120 discoveries etc. So again when you go 164 00:07:15,520 --> 00:07:21,360 against those complex type of system you 165 00:07:19,120 --> 00:07:24,960 really finding an interesting 166 00:07:21,360 --> 00:07:28,800 opportunities I think and you really can 167 00:07:24,960 --> 00:07:32,639 defend your position much more than 168 00:07:28,800 --> 00:07:35,039 building let's say a simpler u tool 169 00:07:32,639 --> 00:07:36,319 market automation or something like this 170 00:07:35,039 --> 00:07:40,319 >> because [snorts] come on like how many 171 00:07:36,319 --> 00:07:44,400 marketmations tools are there and we all 172 00:07:40,319 --> 00:07:48,000 use 10 maybe maybe life. [clears throat] 173 00:07:44,400 --> 00:07:51,919 So again, I think this is uh my playbook 174 00:07:48,000 --> 00:07:54,479 because uh you do not look into the most 175 00:07:51,919 --> 00:07:56,080 brightest thing. You look into the most 176 00:07:54,479 --> 00:07:58,720 unsexy 177 00:07:56,080 --> 00:08:01,680 uh type of industry and you're like, 178 00:07:58,720 --> 00:08:04,319 "Okay, now I will get and own this." 179 00:08:01,680 --> 00:08:08,400 >> Yeah. It's funny that you say that 180 00:08:04,319 --> 00:08:12,080 because I I watched a a video with 181 00:08:08,400 --> 00:08:13,759 uh a fellow who is a billionaire, but 182 00:08:12,080 --> 00:08:16,720 you don't know he was because he just 183 00:08:13,759 --> 00:08:20,240 chose the most unsexy type of industries 184 00:08:16,720 --> 00:08:22,960 like uh piping and and logistics and and 185 00:08:20,240 --> 00:08:24,560 and different things like that. And it 186 00:08:22,960 --> 00:08:26,000 sounds like the software industry is 187 00:08:24,560 --> 00:08:29,120 going through the same sort of thing. 188 00:08:26,000 --> 00:08:32,800 It's like choose the most boring 189 00:08:29,120 --> 00:08:36,240 pro like part of software. the regulated 190 00:08:32,800 --> 00:08:39,760 areas, the compliance areas, the the 191 00:08:36,240 --> 00:08:43,440 piping, the the things that are 192 00:08:39,760 --> 00:08:46,160 I I guess used on on a mass scale and 193 00:08:43,440 --> 00:08:50,480 just build there. So that that's that's 194 00:08:46,160 --> 00:08:53,360 very interesting to see. Um, and there's 195 00:08:50,480 --> 00:08:56,399 a lot of people building out stuff in 196 00:08:53,360 --> 00:09:01,680 Silicon Valley that are, you know, the 197 00:08:56,399 --> 00:09:04,720 AI for HVAC companies, the a the AI for 198 00:09:01,680 --> 00:09:07,120 whatever. And I think you have a strong 199 00:09:04,720 --> 00:09:10,160 opinion on these things. Uh, just 200 00:09:07,120 --> 00:09:12,959 looking at your website. Um, and I know 201 00:09:10,160 --> 00:09:15,279 you're a you're a service provider, so 202 00:09:12,959 --> 00:09:19,360 you will build anything, right? But, uh, 203 00:09:15,279 --> 00:09:22,000 what do you enjoy building the most? 204 00:09:19,360 --> 00:09:23,961 Oh, okay. So, you really prepared the 205 00:09:22,000 --> 00:09:26,480 challenging questions. I like this. Um, 206 00:09:23,961 --> 00:09:30,320 [clears throat] thank you 207 00:09:26,480 --> 00:09:32,399 man. Like, um, 208 00:09:30,320 --> 00:09:36,800 again realistically 209 00:09:32,399 --> 00:09:39,600 um, I would kind of divide this into 210 00:09:36,800 --> 00:09:42,880 three type of categories um that would 211 00:09:39,600 --> 00:09:44,800 be uh, in a way interesting. I think 212 00:09:42,880 --> 00:09:49,839 still in enterprises it's a lot of use 213 00:09:44,800 --> 00:09:52,399 cases that um you can really solve and I 214 00:09:49,839 --> 00:09:56,880 really like to see the the case when you 215 00:09:52,399 --> 00:10:02,160 like take a large um system and you 216 00:09:56,880 --> 00:10:04,800 really um let's say we had 150 sales for 217 00:10:02,160 --> 00:10:08,320 working on the flow 218 00:10:04,800 --> 00:10:12,399 and the output was let's say 100 mil 219 00:10:08,320 --> 00:10:14,399 just as example uh and now we can take 220 00:10:12,399 --> 00:10:17,360 and free up 221 00:10:14,399 --> 00:10:19,519 basically more than half of the day of 222 00:10:17,360 --> 00:10:20,079 those people so they can sell more. 223 00:10:19,519 --> 00:10:24,959 >> Mhm. 224 00:10:20,079 --> 00:10:28,480 >> In reality, it's not 200 mil, it's u60, 225 00:10:24,959 --> 00:10:30,640 but it is good. And when you see those 226 00:10:28,480 --> 00:10:34,720 examples, you're like, okay, this is 227 00:10:30,640 --> 00:10:37,120 really the right flow of things. 228 00:10:34,720 --> 00:10:40,800 And you know that you can accumate more 229 00:10:37,120 --> 00:10:43,600 and you can get rid of uh even more 230 00:10:40,800 --> 00:10:45,120 things that people do not need there. 231 00:10:43,600 --> 00:10:47,360 And again I'm not talking about like 232 00:10:45,120 --> 00:10:51,120 replacing humans because I think in 233 00:10:47,360 --> 00:10:54,399 sales um you need human touch and I 234 00:10:51,120 --> 00:10:58,560 think it will be valued for a long time. 235 00:10:54,399 --> 00:11:00,880 I think the second thing is um really 236 00:10:58,560 --> 00:11:03,200 challenging R&D projects. 237 00:11:00,880 --> 00:11:06,640 >> Mhm. when you [clears throat] 238 00:11:03,200 --> 00:11:09,680 when you look into the specifics of the 239 00:11:06,640 --> 00:11:11,360 project and you like I have no freaking 240 00:11:09,680 --> 00:11:12,880 idea how to build this stuff. 241 00:11:11,360 --> 00:11:17,279 >> Mhm. 242 00:11:12,880 --> 00:11:19,760 >> And again, you know the direction 243 00:11:17,279 --> 00:11:22,959 but you have no idea how to build this 244 00:11:19,760 --> 00:11:27,120 properly. You you doubting 245 00:11:22,959 --> 00:11:30,320 um which approach you would choose. And 246 00:11:27,120 --> 00:11:33,680 this is more complex type of things uh 247 00:11:30,320 --> 00:11:37,120 where you do not trying to recreate the 248 00:11:33,680 --> 00:11:39,519 same um systems that exist. 249 00:11:37,120 --> 00:11:41,839 I think those are kind of like the most 250 00:11:39,519 --> 00:11:45,200 fun and challenging to work with. Um in 251 00:11:41,839 --> 00:11:48,000 a way um not necessarily the most 252 00:11:45,200 --> 00:11:49,760 profitable in the cases. However, I 253 00:11:48,000 --> 00:11:51,760 think it's a it's a lot of fun uh to 254 00:11:49,760 --> 00:11:55,680 work with those. 255 00:11:51,760 --> 00:12:00,880 Um beside this I think uh the last piece 256 00:11:55,680 --> 00:12:03,920 is uh is uh the projects that you kind 257 00:12:00,880 --> 00:12:06,399 of like you see the immediate result. Um 258 00:12:03,920 --> 00:12:08,560 so you build the system and you work 259 00:12:06,399 --> 00:12:10,240 with the founder. It's not about the 260 00:12:08,560 --> 00:12:13,120 product, it's about the founder in this 261 00:12:10,240 --> 00:12:14,959 case. Um because we work with a lot of 262 00:12:13,120 --> 00:12:17,200 growth stage startups and we work also 263 00:12:14,959 --> 00:12:19,600 with a lot of second time founders and 264 00:12:17,200 --> 00:12:22,639 I'm a big fan of working with those guys 265 00:12:19,600 --> 00:12:24,880 because they know what to do. Mhm. And 266 00:12:22,639 --> 00:12:26,480 they sometime you sometimes see on like 267 00:12:24,880 --> 00:12:30,480 some decision that they will make and 268 00:12:26,480 --> 00:12:34,880 you like why and it end up being a great 269 00:12:30,480 --> 00:12:36,399 deal. Um and again you learn from those 270 00:12:34,880 --> 00:12:38,800 experiences 271 00:12:36,399 --> 00:12:41,600 and I think this is the biggest kind of 272 00:12:38,800 --> 00:12:43,600 advantage of working in this industry 273 00:12:41,600 --> 00:12:45,440 that you can learn from so many folks in 274 00:12:43,600 --> 00:12:48,160 the same time 275 00:12:45,440 --> 00:12:50,480 and you have uh a lot of relationship 276 00:12:48,160 --> 00:12:52,959 that really kind of leading you to 277 00:12:50,480 --> 00:12:56,160 understanding that those type of 278 00:12:52,959 --> 00:13:00,320 projects really is something different. 279 00:12:56,160 --> 00:13:04,160 And when you see the napkin idea 280 00:13:00,320 --> 00:13:08,639 um coming back to an actual 100 million 281 00:13:04,160 --> 00:13:11,200 plus business in next couple of years. 282 00:13:08,639 --> 00:13:14,480 This is really amazing. 283 00:13:11,200 --> 00:13:16,880 It sounds like you're you're almost um 284 00:13:14,480 --> 00:13:22,160 taking an advisor consulting position as 285 00:13:16,880 --> 00:13:23,920 well. Uh are you also and I believe you 286 00:13:22,160 --> 00:13:25,600 said you were you're also an investor in 287 00:13:23,920 --> 00:13:27,920 some of these projects too if you like 288 00:13:25,600 --> 00:13:31,120 the idea. 289 00:13:27,920 --> 00:13:33,279 I um kind of like combining the advisor 290 00:13:31,120 --> 00:13:35,120 and investor role in some of those 291 00:13:33,279 --> 00:13:37,760 cases. 292 00:13:35,120 --> 00:13:40,160 In some cases we work as a service 293 00:13:37,760 --> 00:13:41,680 provider for them. In some cases we so 294 00:13:40,160 --> 00:13:42,959 completely separate. 295 00:13:41,680 --> 00:13:44,720 >> Okay. 296 00:13:42,959 --> 00:13:47,839 And I [snorts] also met mentor in 297 00:13:44,720 --> 00:13:53,360 Alchemist and few other places but yeah. 298 00:13:47,839 --> 00:13:55,279 >> Okay. Talk about your um experience as a 299 00:13:53,360 --> 00:14:00,920 investor. What do you like? What do you 300 00:13:55,279 --> 00:14:00,920 what do you what would you invest in 301 00:14:02,079 --> 00:14:07,120 >> on early stages? You mean? 302 00:14:03,839 --> 00:14:10,560 >> Yeah. Yeah. 303 00:14:07,120 --> 00:14:12,160 I think it's a kind of uh really 304 00:14:10,560 --> 00:14:15,839 straightforward playbook. So it's a 305 00:14:12,160 --> 00:14:16,959 always combination of timing uh market 306 00:14:15,839 --> 00:14:18,240 and founder. 307 00:14:16,959 --> 00:14:20,320 >> Yeah. 308 00:14:18,240 --> 00:14:23,920 >> Realistically for me the founder is more 309 00:14:20,320 --> 00:14:28,399 important than many other things. Um 310 00:14:23,920 --> 00:14:31,360 because I see that um 311 00:14:28,399 --> 00:14:33,040 sometimes the great founder will uh 312 00:14:31,360 --> 00:14:36,959 create the market or sometimes the great 313 00:14:33,040 --> 00:14:42,240 founder will adopt the product. And I 314 00:14:36,959 --> 00:14:45,199 think the the fact that again the 315 00:14:42,240 --> 00:14:48,399 founder is really of I mean founding 316 00:14:45,199 --> 00:14:52,240 team in reality. Um 317 00:14:48,399 --> 00:14:55,519 you want to see the folks that not just 318 00:14:52,240 --> 00:14:59,120 knows what they do because this is the 319 00:14:55,519 --> 00:15:01,760 boring part. Like if you have insurance 320 00:14:59,120 --> 00:15:04,240 folks trying to build insurance startup. 321 00:15:01,760 --> 00:15:06,240 Yeah. I mean they would build something 322 00:15:04,240 --> 00:15:08,000 but they typically would not build a 323 00:15:06,240 --> 00:15:10,000 disruptive thing. 324 00:15:08,000 --> 00:15:12,720 >> When you have real outsiders come into 325 00:15:10,000 --> 00:15:16,000 this market and they really create 326 00:15:12,720 --> 00:15:19,440 something new and they don't know that 327 00:15:16,000 --> 00:15:21,760 no one done this before like creating 328 00:15:19,440 --> 00:15:25,600 the accounting systems in a way that no 329 00:15:21,760 --> 00:15:28,320 one done before. you like 330 00:15:25,600 --> 00:15:30,320 um why you have not done this before 331 00:15:28,320 --> 00:15:33,360 like oh no one actually suggested this 332 00:15:30,320 --> 00:15:37,040 but this is I think a good thing why you 333 00:15:33,360 --> 00:15:40,560 skipped this oh uh no one actually 334 00:15:37,040 --> 00:15:42,399 suggested this so again uh I think 335 00:15:40,560 --> 00:15:46,320 challenging the start school is 336 00:15:42,399 --> 00:15:48,639 important there and uh founders that 337 00:15:46,320 --> 00:15:50,399 really kind of bold and really 338 00:15:48,639 --> 00:15:53,920 persistent 339 00:15:50,399 --> 00:15:56,240 um and really seen the future. I think 340 00:15:53,920 --> 00:16:00,000 this is really really important and not 341 00:15:56,240 --> 00:16:04,560 being afraid again on losing because 342 00:16:00,000 --> 00:16:06,800 it's always the conversation I you 343 00:16:04,560 --> 00:16:12,240 uh 344 00:16:06,800 --> 00:16:14,480 have fear of losing or you have a um uh 345 00:16:12,240 --> 00:16:18,399 basically passion to win. 346 00:16:14,480 --> 00:16:23,120 >> And in both cases it's okay. And it's 347 00:16:18,399 --> 00:16:27,920 just two type of of founders and you in 348 00:16:23,120 --> 00:16:29,839 my cases I think you can find a right 349 00:16:27,920 --> 00:16:32,880 playbook 350 00:16:29,839 --> 00:16:34,240 but you need to see those eyes really 351 00:16:32,880 --> 00:16:37,360 really 352 00:16:34,240 --> 00:16:39,360 >> you need to see the fire there because 353 00:16:37,360 --> 00:16:42,399 if you don't see the fire 354 00:16:39,360 --> 00:16:44,639 >> it's it's all long done. That that that 355 00:16:42,399 --> 00:16:47,360 that makes a lot of sense. You need to 356 00:16:44,639 --> 00:16:49,279 look into a founders's eyes and see like 357 00:16:47,360 --> 00:16:52,160 that man is really or that woman is 358 00:16:49,279 --> 00:16:54,079 really crazy and they're going to take a 359 00:16:52,160 --> 00:16:56,160 big risk or they're going to do what it 360 00:16:54,079 --> 00:16:57,360 takes to to make this company succeed. 361 00:16:56,160 --> 00:16:59,759 So it's like a passion. 362 00:16:57,360 --> 00:17:00,959 >> Yeah. Because the founders journey is 363 00:16:59,759 --> 00:17:03,199 really lonely. 364 00:17:00,959 --> 00:17:05,439 >> The entrepreneurship in general really 365 00:17:03,199 --> 00:17:08,079 really lonely and really really painful. 366 00:17:05,439 --> 00:17:11,520 >> Mhm. People think about like smoothies 367 00:17:08,079 --> 00:17:15,679 and all the funny things and uh looking 368 00:17:11,520 --> 00:17:18,079 into the wonderful things of uh uh TV 369 00:17:15,679 --> 00:17:19,120 shows etc. In reality it's freaking 370 00:17:18,079 --> 00:17:19,919 broil guys. 371 00:17:19,120 --> 00:17:22,319 >> Mhm. 372 00:17:19,919 --> 00:17:24,240 >> And they like oh yeah we will raise so 373 00:17:22,319 --> 00:17:27,039 much money. We'll do such wonderful 374 00:17:24,240 --> 00:17:29,840 things. Come on. Like first of all only 375 00:17:27,039 --> 00:17:31,919 one out of 10 will raise anything. um 376 00:17:29,840 --> 00:17:34,400 second one of the trainer will raise 377 00:17:31,919 --> 00:17:37,200 something else uh on the second round 378 00:17:34,400 --> 00:17:40,160 and uh we can continue 379 00:17:37,200 --> 00:17:43,679 >> and uh if you relying on the on 380 00:17:40,160 --> 00:17:47,280 investments you already lost. So it's 381 00:17:43,679 --> 00:17:50,282 it's many other things that exist in 382 00:17:47,280 --> 00:17:51,440 this playbook that important to change. 383 00:17:50,282 --> 00:17:54,480 [snorts] 384 00:17:51,440 --> 00:17:56,799 Yep. And um 385 00:17:54,480 --> 00:18:00,960 I think you said it right there. It's a 386 00:17:56,799 --> 00:18:03,679 it's a it's a journey and it's something 387 00:18:00,960 --> 00:18:06,720 like a lot of people don't take into 388 00:18:03,679 --> 00:18:11,840 account. It's funny. Uh, did you see the 389 00:18:06,720 --> 00:18:14,880 post by, uh, Gary Tan out of, uh, Y 390 00:18:11,840 --> 00:18:17,600 Cominator talking about, 391 00:18:14,880 --> 00:18:20,240 you know, uh, 392 00:18:17,600 --> 00:18:22,559 founders who are in their 40s and 50s 393 00:18:20,240 --> 00:18:25,280 being a little bit more resilient than 394 00:18:22,559 --> 00:18:27,679 founders uh, well well, he didn't make a 395 00:18:25,280 --> 00:18:30,640 comparison. He just made the statement. 396 00:18:27,679 --> 00:18:33,760 And you know, I just thought that was 397 00:18:30,640 --> 00:18:36,480 cool because there's people like myself 398 00:18:33,760 --> 00:18:42,400 included um that have a lot of life 399 00:18:36,480 --> 00:18:45,760 experience and uh that sort of goes into 400 00:18:42,400 --> 00:18:48,880 your skill set and you know why 401 00:18:45,760 --> 00:18:51,679 Cominator is known for just having 18y 402 00:18:48,880 --> 00:18:55,600 olds, right? and and 23 year olds. But 403 00:18:51,679 --> 00:18:59,039 to have the the founder and the the 404 00:18:55,600 --> 00:19:00,799 manager managing partner of Y Cominator 405 00:18:59,039 --> 00:19:02,559 say something like that was was kind of 406 00:19:00,799 --> 00:19:05,360 like it made me smile a little bit. What 407 00:19:02,559 --> 00:19:07,919 what are your thoughts about that? 408 00:19:05,360 --> 00:19:10,640 I don't think it's specifically about 409 00:19:07,919 --> 00:19:13,280 age or like specific type of 410 00:19:10,640 --> 00:19:14,480 experiences. I think it's all about the 411 00:19:13,280 --> 00:19:16,160 energy that you have. 412 00:19:14,480 --> 00:19:20,080 >> Yeah. 413 00:19:16,160 --> 00:19:23,600 Because again I see absolutely dumb 20 414 00:19:20,080 --> 00:19:27,120 year olds and absolutely dumb 40 year 415 00:19:23,600 --> 00:19:30,160 olds and 50 etc. And I have absolutely 416 00:19:27,120 --> 00:19:35,600 brilliant folks in the same categories 417 00:19:30,160 --> 00:19:38,559 and I see that they like really like 418 00:19:35,600 --> 00:19:40,160 I see so many founders being so careful. 419 00:19:38,559 --> 00:19:43,679 >> Mhm. 420 00:19:40,160 --> 00:19:47,280 >> You not winning by being careful. 421 00:19:43,679 --> 00:19:50,320 >> Yeah. You win it by betting and betting 422 00:19:47,280 --> 00:19:52,559 again and again. The business is all 423 00:19:50,320 --> 00:19:55,360 about experiments. 424 00:19:52,559 --> 00:19:58,320 It's like, hey, I have a theory that 425 00:19:55,360 --> 00:20:01,440 this stuff will work. I have no idea it 426 00:19:58,320 --> 00:20:03,600 will. I have only limited control, but I 427 00:20:01,440 --> 00:20:06,080 will do whatever it takes to really 428 00:20:03,600 --> 00:20:08,400 prove the theory, 429 00:20:06,080 --> 00:20:11,440 and it will not work by itself, and I 430 00:20:08,400 --> 00:20:14,720 will push it as hard as I freaking can. 431 00:20:11,440 --> 00:20:17,280 And the only way I succeeded if I really 432 00:20:14,720 --> 00:20:20,000 analyze if this is actually a good 433 00:20:17,280 --> 00:20:23,120 outcome or not. And I will be wrong in 434 00:20:20,000 --> 00:20:26,880 many many cases. And this would be a fun 435 00:20:23,120 --> 00:20:30,320 journey because I will fail so many many 436 00:20:26,880 --> 00:20:32,000 times. And my successes actually comes 437 00:20:30,320 --> 00:20:34,400 from those failures that I will learn 438 00:20:32,000 --> 00:20:37,600 and I will adapt and will send up again 439 00:20:34,400 --> 00:20:39,520 and we'll do whatever it takes. So I 440 00:20:37,600 --> 00:20:43,760 think this is the kind of like the 441 00:20:39,520 --> 00:20:47,840 journey and I think I have a lot of uh 442 00:20:43,760 --> 00:20:51,679 folks in the 40s and 50s and even 60s 443 00:20:47,840 --> 00:20:54,400 that I think is a great founders. Um 444 00:20:51,679 --> 00:20:57,280 and in the same time I have the same 445 00:20:54,400 --> 00:21:01,039 category that really kind of careful and 446 00:20:57,280 --> 00:21:02,960 like oh I don't know if I should oh let 447 00:21:01,039 --> 00:21:04,320 me take a vacation and think about this 448 00:21:02,960 --> 00:21:07,200 stuff. 449 00:21:04,320 --> 00:21:12,000 No, no. You you you know that this will 450 00:21:07,200 --> 00:21:17,919 not work. Like you need to battle this 451 00:21:12,000 --> 00:21:22,240 today. Mhm. Like it's 452 00:21:17,919 --> 00:21:26,000 need to be uh done immediately. Uh not 453 00:21:22,240 --> 00:21:31,679 like because like the point is 454 00:21:26,000 --> 00:21:33,919 if you keep um uh combining those issues 455 00:21:31,679 --> 00:21:36,000 um you will have a lot of issues 456 00:21:33,919 --> 00:21:40,320 >> if you try to battle them and you 457 00:21:36,000 --> 00:21:43,760 pertize them and you play right um sure 458 00:21:40,320 --> 00:21:47,039 and it's not about the rules either like 459 00:21:43,760 --> 00:21:49,600 again your own rules 460 00:21:47,039 --> 00:21:54,400 again you need to remember that history 461 00:21:49,600 --> 00:21:56,640 written by winners There's no losers. 462 00:21:54,400 --> 00:21:59,760 I love your perspective. Like I I I 463 00:21:56,640 --> 00:22:02,159 thank you for correcting me on that 464 00:21:59,760 --> 00:22:04,960 generalization I made, which was clearly 465 00:22:02,159 --> 00:22:08,640 a generalization. And it's it's hard to 466 00:22:04,960 --> 00:22:10,799 to uh say anything across the board like 467 00:22:08,640 --> 00:22:12,880 that. But uh you're right. You're right. 468 00:22:10,799 --> 00:22:15,440 It is about attitude really. It's about 469 00:22:12,880 --> 00:22:18,799 attitude, sense of urgency, 470 00:22:15,440 --> 00:22:23,039 um, instinct, many different things that 471 00:22:18,799 --> 00:22:30,280 that happen. Um, anyways, let's move on. 472 00:22:23,039 --> 00:22:30,280 U, you've been previbe code, initial 473 00:22:30,480 --> 00:22:33,840 foundations of vibe coding, and now 474 00:22:32,640 --> 00:22:36,880 we're getting to the point where we're 475 00:22:33,840 --> 00:22:37,919 having cloud agents just not just doing 476 00:22:36,880 --> 00:22:40,720 work for you, and you're just coming 477 00:22:37,919 --> 00:22:42,640 back and hitting the PR button or uh, 478 00:22:40,720 --> 00:22:45,280 merge button. [snorts] 479 00:22:42,640 --> 00:22:47,840 um where 480 00:22:45,280 --> 00:22:50,720 what's wrong how has software changed 481 00:22:47,840 --> 00:22:52,400 and and how is what is the career 482 00:22:50,720 --> 00:22:55,039 progression for a software engineer 483 00:22:52,400 --> 00:22:56,480 these days like h how what is what's the 484 00:22:55,039 --> 00:22:59,480 best way to get started on all these 485 00:22:56,480 --> 00:22:59,480 things 486 00:23:00,000 --> 00:23:03,200 I want to be honest I have no idea 487 00:23:02,065 --> 00:23:07,200 [laughter] 488 00:23:03,200 --> 00:23:09,840 it's changing right yeah I 489 00:23:07,200 --> 00:23:12,400 been asked this same question around 490 00:23:09,840 --> 00:23:12,400 nine months ago 491 00:23:12,480 --> 00:23:22,400 um oh 7 months ago. So basically the 492 00:23:16,400 --> 00:23:25,679 OPUS was just released and um 493 00:23:22,400 --> 00:23:30,559 we seen completely different flows of 494 00:23:25,679 --> 00:23:35,440 agentic workflows again um it's all just 495 00:23:30,559 --> 00:23:38,400 like started basically and um one person 496 00:23:35,440 --> 00:23:40,640 that I think is 18 or 17 asked me about 497 00:23:38,400 --> 00:23:42,799 like hey I study computer science what 498 00:23:40,640 --> 00:23:46,000 do you think what is the next steps uh 499 00:23:42,799 --> 00:23:47,679 will look like and I'm like man I have 500 00:23:46,000 --> 00:23:49,679 no idea 501 00:23:47,679 --> 00:23:52,799 >> because again I've been running this 502 00:23:49,679 --> 00:23:55,280 business for 16 years or for 18 years 503 00:23:52,799 --> 00:23:59,840 and we're the reality is I have no idea 504 00:23:55,280 --> 00:24:01,600 where it will lead us in next um year. I 505 00:23:59,840 --> 00:24:03,039 don't know where we will be in next 506 00:24:01,600 --> 00:24:07,760 months. [laughter] 507 00:24:03,039 --> 00:24:10,080 Um again the um horizontal 508 00:24:07,760 --> 00:24:12,159 um 509 00:24:10,080 --> 00:24:16,740 planning in many companies now like 510 00:24:12,159 --> 00:24:18,400 weeks. Yeah. And it it is long because 511 00:24:16,740 --> 00:24:21,520 [laughter] you literally don't know what 512 00:24:18,400 --> 00:24:24,320 will happen tomorrow. So I think like 513 00:24:21,520 --> 00:24:28,080 the reality of um many of those use 514 00:24:24,320 --> 00:24:31,679 cases that I think the fundamentals is 515 00:24:28,080 --> 00:24:36,559 important and right now how we see this 516 00:24:31,679 --> 00:24:38,559 is um we don't need coders anymore. 517 00:24:36,559 --> 00:24:40,559 Like we don't need people who write the 518 00:24:38,559 --> 00:24:41,520 code because we're not writing code 519 00:24:40,559 --> 00:24:43,039 anymore 520 00:24:41,520 --> 00:24:46,720 >> specifically. 521 00:24:43,039 --> 00:24:49,520 we need to understand uh the system. So 522 00:24:46,720 --> 00:24:52,480 we need people who can design the 523 00:24:49,520 --> 00:24:55,279 system. So we need architects, we need 524 00:24:52,480 --> 00:24:57,919 product people [snorts] and uh we need 525 00:24:55,279 --> 00:25:01,120 to create a right 526 00:24:57,919 --> 00:25:03,039 harnessing uh for them to build those 527 00:25:01,120 --> 00:25:07,600 systems 528 00:25:03,039 --> 00:25:11,200 not other way around. So 529 00:25:07,600 --> 00:25:13,120 in many cases it's all about the context 530 00:25:11,200 --> 00:25:16,960 and all about putting the right 531 00:25:13,120 --> 00:25:20,240 structure of AI to build us. So in the 532 00:25:16,960 --> 00:25:22,559 software industry I think architects 533 00:25:20,240 --> 00:25:26,400 still will be important. 534 00:25:22,559 --> 00:25:28,640 I think really understanding and playing 535 00:25:26,400 --> 00:25:32,720 really really fast and learning really 536 00:25:28,640 --> 00:25:35,760 fast is also really important because 537 00:25:32,720 --> 00:25:38,880 like it's not longer enough and not 538 00:25:35,760 --> 00:25:44,159 longer actually making sense to just go 539 00:25:38,880 --> 00:25:48,320 ahead and check your uh whatever 540 00:25:44,159 --> 00:25:50,720 uh books uh and like learn something for 541 00:25:48,320 --> 00:25:52,320 a year. No, those skills doesn't really 542 00:25:50,720 --> 00:25:54,640 apply. 543 00:25:52,320 --> 00:25:56,159 You need to learn really really fast and 544 00:25:54,640 --> 00:25:58,000 you need to adapt. 545 00:25:56,159 --> 00:26:00,000 >> The fun fact like again I don't care 546 00:25:58,000 --> 00:26:02,880 which language you're learning. It 547 00:26:00,000 --> 00:26:08,559 doesn't really apply anymore 548 00:26:02,880 --> 00:26:12,159 because like again we took many projects 549 00:26:08,559 --> 00:26:17,279 in the last 6 months for tech stack that 550 00:26:12,159 --> 00:26:20,320 we had um limited capabilities for. 551 00:26:17,279 --> 00:26:24,463 The fun fact, we actually never wrote 552 00:26:20,320 --> 00:26:25,520 any code. Oh wow. It was all by a 553 00:26:24,463 --> 00:26:28,240 [clears throat] 554 00:26:25,520 --> 00:26:30,559 we found many issues in a legacy system. 555 00:26:28,240 --> 00:26:32,480 We fixed those. We applied those things. 556 00:26:30,559 --> 00:26:35,120 We're not writing this manually because 557 00:26:32,480 --> 00:26:37,679 it doesn't really make sense because it 558 00:26:35,120 --> 00:26:40,480 is expensive. Um first of all, it's 559 00:26:37,679 --> 00:26:44,400 slow. And second, we will make mistakes. 560 00:26:40,480 --> 00:26:46,720 Yeah. Will AI make mistakes? Sure. But 561 00:26:44,400 --> 00:26:49,760 if we compare a number of mistakes of AI 562 00:26:46,720 --> 00:26:53,200 to human, we actually will see that we 563 00:26:49,760 --> 00:26:55,440 actually do pretty good on AI. 564 00:26:53,200 --> 00:26:58,400 >> So I think like 565 00:26:55,440 --> 00:27:00,000 like yeah you have so many issues with 566 00:26:58,400 --> 00:27:02,320 yeah but like you developing 10 times 567 00:27:00,000 --> 00:27:05,039 faster or 100 times faster. 568 00:27:02,320 --> 00:27:07,600 >> Of course you will have issues. Yeah it 569 00:27:05,039 --> 00:27:09,919 is expected you should have issues. 570 00:27:07,600 --> 00:27:12,480 Basically you will have security 571 00:27:09,919 --> 00:27:14,159 problems. You have bugs. you will have 572 00:27:12,480 --> 00:27:18,320 some things that doesn't really make 573 00:27:14,159 --> 00:27:20,960 sense. Sure, fix the prompting, fix the 574 00:27:18,320 --> 00:27:23,440 harnessing and memory and yeah, go ahead 575 00:27:20,960 --> 00:27:26,400 and uh fix those things uh in the 576 00:27:23,440 --> 00:27:29,600 system, but do not touch the code. If 577 00:27:26,400 --> 00:27:31,919 you touch the code, you kind of like 578 00:27:29,600 --> 00:27:34,880 broke the idea of agentic development 579 00:27:31,919 --> 00:27:38,960 basically. 580 00:27:34,880 --> 00:27:41,279 Oh my goodness. So it went from 581 00:27:38,960 --> 00:27:44,480 I I guess what you're saying is like it 582 00:27:41,279 --> 00:27:47,679 maybe a year and a half ago it went from 583 00:27:44,480 --> 00:27:50,880 yeah I'm using an AI coding assistant 584 00:27:47,679 --> 00:27:52,720 it's helping me write 20% of my code 30% 585 00:27:50,880 --> 00:27:56,640 it's checking things it's validating 586 00:27:52,720 --> 00:27:58,720 things to maybe 587 00:27:56,640 --> 00:28:03,440 8 months ago when Opus came out to oh 588 00:27:58,720 --> 00:28:07,840 it's writing code it's writing 60 70% 589 00:28:03,440 --> 00:28:10,399 and now you're seeing it's 100%. 590 00:28:07,840 --> 00:28:12,399 Um, or close to 100%. I mean, it can it 591 00:28:10,399 --> 00:28:16,240 can be argued it's it's it's always 592 00:28:12,399 --> 00:28:19,520 going to be in between 85 and 100%. But 593 00:28:16,240 --> 00:28:23,039 you're saying as an experienced vendor, 594 00:28:19,520 --> 00:28:25,039 an experienced shop, you're nearing the 595 00:28:23,039 --> 00:28:27,440 100 mark or you're closer to the 100 596 00:28:25,039 --> 00:28:30,480 mark depending on the project just to be 597 00:28:27,440 --> 00:28:32,080 clear. Wow. It is some projects that you 598 00:28:30,480 --> 00:28:35,760 cannot do this. It's some limitations 599 00:28:32,080 --> 00:28:38,720 that applies etc. But in many many cases 600 00:28:35,760 --> 00:28:44,240 especially for new projects we we have 601 00:28:38,720 --> 00:28:49,760 many uh smaller projects uh below let's 602 00:28:44,240 --> 00:28:53,520 say 100k uh uh budgeting wise that is 603 00:28:49,760 --> 00:28:56,559 basically one month one month's built 604 00:28:53,520 --> 00:28:59,200 uh all agentic development. 605 00:28:56,559 --> 00:29:00,559 Yeah, we spend a lot of time on creating 606 00:28:59,200 --> 00:29:03,919 the right documentation and the right 607 00:29:00,559 --> 00:29:07,919 harnessing and right uh logic to 608 00:29:03,919 --> 00:29:10,640 actually create the flows there um and 609 00:29:07,919 --> 00:29:12,640 ensure that AI knows everything and yeah 610 00:29:10,640 --> 00:29:14,000 we mention EI to develop this basically 611 00:29:12,640 --> 00:29:18,000 in a matter of days. 612 00:29:14,000 --> 00:29:19,840 >> Wow. So the most time spent actually on 613 00:29:18,000 --> 00:29:23,520 discussing functionality, making product 614 00:29:19,840 --> 00:29:27,360 decisions, um all of those things, 615 00:29:23,520 --> 00:29:29,830 writing with a genic systems, it's it's 616 00:29:27,360 --> 00:29:30,960 really quick. Mhm. 617 00:29:29,830 --> 00:29:36,799 >> [gasps] 618 00:29:30,960 --> 00:29:41,440 >> So I remember working for a company 619 00:29:36,799 --> 00:29:43,919 uh maybe about 7 8 years ago [snorts] 620 00:29:41,440 --> 00:29:47,919 and there they used to charge I don't 621 00:29:43,919 --> 00:29:50,480 know 20,000 30,000 for a website you 622 00:29:47,919 --> 00:29:53,120 know maybe it's got a store on it maybe 623 00:29:50,480 --> 00:29:55,760 it gets up to 40,000 624 00:29:53,120 --> 00:29:58,559 um 625 00:29:55,760 --> 00:30:01,440 it seems to me like that if somebody got 626 00:29:58,559 --> 00:30:04,320 that quote today they would just go no 627 00:30:01,440 --> 00:30:08,640 that's you want a simple website with a 628 00:30:04,320 --> 00:30:10,640 an e-commerce store and a CMS and and uh 629 00:30:08,640 --> 00:30:13,919 for for 40,000 are you kidding me you 630 00:30:10,640 --> 00:30:16,399 you can do that now for probably in days 631 00:30:13,919 --> 00:30:19,200 and and like 4x less the cost like you 632 00:30:16,399 --> 00:30:22,799 said right so so where is the what is 633 00:30:19,200 --> 00:30:24,720 the unique selling proposition like you 634 00:30:22,799 --> 00:30:27,360 you it's different for you like you you 635 00:30:24,720 --> 00:30:29,840 have a you have a business that has 18 636 00:30:27,360 --> 00:30:32,399 years worth of clients your your your 637 00:30:29,840 --> 00:30:35,840 reputation is amazing in the industry 638 00:30:32,399 --> 00:30:38,720 but where does the business of software 639 00:30:35,840 --> 00:30:40,880 development go from here and and you 640 00:30:38,720 --> 00:30:42,799 know maybe you don't know like like you 641 00:30:40,880 --> 00:30:44,960 said before but what is the edge that 642 00:30:42,799 --> 00:30:46,960 people can have 643 00:30:44,960 --> 00:30:49,919 >> I don't think anyone knows 644 00:30:46,960 --> 00:30:52,000 >> this is the fun fact [snorts] 645 00:30:49,919 --> 00:30:56,240 >> I mean 646 00:30:52,000 --> 00:30:59,039 listen um if we're talking about uh the 647 00:30:56,240 --> 00:31:00,559 websites we still have websites that 648 00:30:59,039 --> 00:31:05,679 do for 100k. 649 00:31:00,559 --> 00:31:10,399 >> Okay. Um it's not a simple promo website 650 00:31:05,679 --> 00:31:13,279 obviously and it is many cases when AI 651 00:31:10,399 --> 00:31:16,320 is not doing the best. Um it still 652 00:31:13,279 --> 00:31:20,080 exists again it's not the silver bullet 653 00:31:16,320 --> 00:31:24,000 for everything. Um pixel perfect is not 654 00:31:20,080 --> 00:31:26,880 the best thing for AI. Um and uh again 655 00:31:24,000 --> 00:31:30,240 we build system to leverage as much as 656 00:31:26,880 --> 00:31:33,360 possible. However, we limited into 657 00:31:30,240 --> 00:31:36,559 ability of doing this just because again 658 00:31:33,360 --> 00:31:41,200 AI is not perfect there and if we have a 659 00:31:36,559 --> 00:31:46,159 projects where uh client will tell us 660 00:31:41,200 --> 00:31:48,799 hey this uh shade is 30% it's supposed 661 00:31:46,159 --> 00:31:51,600 to be 45 662 00:31:48,799 --> 00:31:54,640 you kind of like you cannot use the for 663 00:31:51,600 --> 00:31:56,559 this will we will think it's okay 664 00:31:54,640 --> 00:31:57,519 >> because it's really it's okay no one 665 00:31:56,559 --> 00:31:58,960 cares 666 00:31:57,519 --> 00:32:01,679 But someone [gasps] 667 00:31:58,960 --> 00:32:03,840 >> someone in some enterprises will. So 668 00:32:01,679 --> 00:32:07,120 realistically 669 00:32:03,840 --> 00:32:09,840 um we have limitations there obviously 670 00:32:07,120 --> 00:32:11,600 but uh for many projects those things 671 00:32:09,840 --> 00:32:14,559 not really applies and not really 672 00:32:11,600 --> 00:32:18,080 important because the time of the 673 00:32:14,559 --> 00:32:20,399 feature that will leave is really low. 674 00:32:18,080 --> 00:32:22,240 Uh and yeah you can adapt and you can 675 00:32:20,399 --> 00:32:24,399 fix those things later etc. So it's a 676 00:32:22,240 --> 00:32:26,320 iterative process 677 00:32:24,399 --> 00:32:29,279 where the software development will go 678 00:32:26,320 --> 00:32:32,159 it is the question that I have no idea 679 00:32:29,279 --> 00:32:35,679 uh because I think the how I see the 680 00:32:32,159 --> 00:32:37,519 industry right now um everyone using it 681 00:32:35,679 --> 00:32:40,080 the 682 00:32:37,519 --> 00:32:42,080 number of people that using this in 683 00:32:40,080 --> 00:32:44,880 advanced level 684 00:32:42,080 --> 00:32:47,519 so let's say we have uh the first level 685 00:32:44,880 --> 00:32:50,640 of uh uh quitting assistance kind of 686 00:32:47,519 --> 00:32:54,640 like a copilot etc. Um, [clears throat] 687 00:32:50,640 --> 00:32:57,440 this is probably 80% of the market 688 00:32:54,640 --> 00:32:59,039 or even more. And you still have, by the 689 00:32:57,440 --> 00:33:01,039 way, a lot of folks that believe that 690 00:32:59,039 --> 00:33:04,000 they can do things better than AI. And 691 00:33:01,039 --> 00:33:06,799 thank you. I really appreciate that 692 00:33:04,000 --> 00:33:10,399 those folks exist. They really helping 693 00:33:06,799 --> 00:33:11,877 us to grow our business. Um, I I'm 694 00:33:10,399 --> 00:33:12,960 really happy for this. 695 00:33:11,877 --> 00:33:16,320 [cough and clears throat] 696 00:33:12,960 --> 00:33:18,480 And last year we had those folks that we 697 00:33:16,320 --> 00:33:21,360 had inside of the company that no longer 698 00:33:18,480 --> 00:33:24,960 was a company because we don't need the 699 00:33:21,360 --> 00:33:28,240 folks to really write better than AI. We 700 00:33:24,960 --> 00:33:31,279 need folks to create systems for AI to 701 00:33:28,240 --> 00:33:33,200 create systems that better. So it's 702 00:33:31,279 --> 00:33:35,120 simple 703 00:33:33,200 --> 00:33:39,679 where it will go. It's a good question 704 00:33:35,120 --> 00:33:43,919 that I think I have really a lot of open 705 00:33:39,679 --> 00:33:46,480 um uh open-ended uh discussions because 706 00:33:43,919 --> 00:33:48,799 realistically 707 00:33:46,480 --> 00:33:50,480 we don't really know if we will have a 708 00:33:48,799 --> 00:33:52,559 system that will allow us to create 709 00:33:50,480 --> 00:33:54,080 everything from scratch in the matter of 710 00:33:52,559 --> 00:33:57,600 seconds. 711 00:33:54,080 --> 00:33:59,679 I think a few kind of like ideas where I 712 00:33:57,600 --> 00:34:01,200 think we kind of like are going. I think 713 00:33:59,679 --> 00:34:06,240 in 714 00:34:01,200 --> 00:34:08,480 um real life um what I see is the model 715 00:34:06,240 --> 00:34:12,399 advancements that happening at least 716 00:34:08,480 --> 00:34:15,119 like right now doesn't really affect the 717 00:34:12,399 --> 00:34:17,280 quality of output so dramatically 718 00:34:15,119 --> 00:34:20,839 because again the difference between 719 00:34:17,280 --> 00:34:20,839 feeble meas 720 00:34:21,359 --> 00:34:27,760 etc 721 00:34:23,200 --> 00:34:29,839 is basically really minimal if you have 722 00:34:27,760 --> 00:34:33,040 the right heart if you have the right 723 00:34:29,839 --> 00:34:34,879 system and you're not stupid enough to 724 00:34:33,040 --> 00:34:38,399 not create the documentation in a stupid 725 00:34:34,879 --> 00:34:41,359 way. Um, in this case, you will have 726 00:34:38,399 --> 00:34:43,040 pretty much a good output and you can 727 00:34:41,359 --> 00:34:46,560 use 728 00:34:43,040 --> 00:34:50,000 open source models as well and have cost 729 00:34:46,560 --> 00:34:51,839 like that 10 times less. [clears throat] 730 00:34:50,000 --> 00:34:55,520 I think this is where we kind of like 731 00:34:51,839 --> 00:34:57,680 leading this into. So the better models 732 00:34:55,520 --> 00:35:00,400 doesn't equal the better performance 733 00:34:57,680 --> 00:35:01,680 because the problem is actually people 734 00:35:00,400 --> 00:35:02,400 that prompting this. 735 00:35:01,680 --> 00:35:04,160 >> Mhm. 736 00:35:02,400 --> 00:35:07,040 >> And I'm not saying that the prompting is 737 00:35:04,160 --> 00:35:09,520 a problem. I'm saying that many people 738 00:35:07,040 --> 00:35:12,800 just do not give AI enough information 739 00:35:09,520 --> 00:35:15,920 what they need to do. They give AI two 740 00:35:12,800 --> 00:35:17,440 lines. Hey, build me 741 00:35:15,920 --> 00:35:20,800 wherever 742 00:35:17,440 --> 00:35:23,599 um this system um and I build me a 743 00:35:20,800 --> 00:35:27,200 website for my company. I'm doing 744 00:35:23,599 --> 00:35:30,079 wherever balloon uh sales for holiday 745 00:35:27,200 --> 00:35:33,839 parties. Great. 746 00:35:30,079 --> 00:35:36,320 It's not enough to describe the task. 747 00:35:33,839 --> 00:35:38,960 You probably should really think through 748 00:35:36,320 --> 00:35:41,200 the content first before you ask about 749 00:35:38,960 --> 00:35:45,760 uh the website as example. in this case 750 00:35:41,200 --> 00:35:48,240 is a simple uh um example and you need 751 00:35:45,760 --> 00:35:53,599 to go through the understanding what the 752 00:35:48,240 --> 00:35:56,960 actual flows um uh of users should be. 753 00:35:53,599 --> 00:35:59,359 So I think this is not going away and AI 754 00:35:56,960 --> 00:36:02,160 with better models will not solve this 755 00:35:59,359 --> 00:36:03,920 and in large systems it's still the 756 00:36:02,160 --> 00:36:06,560 question of really deep understanding of 757 00:36:03,920 --> 00:36:10,480 the users and really deep understanding 758 00:36:06,560 --> 00:36:11,520 of um what people will buy and what they 759 00:36:10,480 --> 00:36:12,079 actually need. 760 00:36:11,520 --> 00:36:17,040 >> Mhm. 761 00:36:12,079 --> 00:36:20,240 >> And in many cases um it is the question 762 00:36:17,040 --> 00:36:22,079 of harnessing and engineering uh those 763 00:36:20,240 --> 00:36:25,280 systems. 764 00:36:22,079 --> 00:36:30,640 Will we have uh more faster development? 765 00:36:25,280 --> 00:36:33,599 Absolutely. Yes. Um I think um the cost 766 00:36:30,640 --> 00:36:36,560 um like we kind of like have a kind of 767 00:36:33,599 --> 00:36:38,400 like pricing crisis uh that coming 768 00:36:36,560 --> 00:36:41,200 probably to the market 769 00:36:38,400 --> 00:36:44,800 >> as well because um the cost is kind of 770 00:36:41,200 --> 00:36:47,839 like going down in a way. Um 771 00:36:44,800 --> 00:36:52,480 again right now I think many uh folks 772 00:36:47,839 --> 00:36:55,040 set in more advanced levels um um uh of 773 00:36:52,480 --> 00:36:59,920 uh agentic development systems have a 774 00:36:55,040 --> 00:37:04,320 age of uh creating u faster and better 775 00:36:59,920 --> 00:37:07,440 uh and I think the smaller dev shops and 776 00:37:04,320 --> 00:37:09,920 uh folks that don't really s deep into I 777 00:37:07,440 --> 00:37:12,880 think those guys 778 00:37:09,920 --> 00:37:15,920 will get to this point. Uh but at this 779 00:37:12,880 --> 00:37:19,359 point it uh we'll have more and more um 780 00:37:15,920 --> 00:37:22,240 on um uh uh the folks that already on 781 00:37:19,359 --> 00:37:25,839 the uh bigger level but we definitely 782 00:37:22,240 --> 00:37:28,240 will have on like midmarket and lower 783 00:37:25,839 --> 00:37:31,359 tier market we'll have a lot of folks 784 00:37:28,240 --> 00:37:34,240 that kind of hey we'll do everything for 785 00:37:31,359 --> 00:37:38,480 5,000 or whatever. Yeah, we'll have 786 00:37:34,240 --> 00:37:41,200 those cases and again you kind of like 787 00:37:38,480 --> 00:37:44,960 looking into this and with the same kind 788 00:37:41,200 --> 00:37:47,280 of mindset as like can you let's say we 789 00:37:44,960 --> 00:37:50,160 like 5 years ago. Yeah. Can we go to 790 00:37:47,280 --> 00:37:52,400 cheaper developer in different countries 791 00:37:50,160 --> 00:37:53,599 and build systems 792 00:37:52,400 --> 00:37:56,400 >> for cheaper? 793 00:37:53,599 --> 00:37:59,520 >> Absolutely. Uh will this be a good 794 00:37:56,400 --> 00:38:03,119 quality product? No. 795 00:37:59,520 --> 00:38:05,359 Like I mean some exceptions, sure, but 796 00:38:03,119 --> 00:38:08,240 generally speaking, if you're hiring 797 00:38:05,359 --> 00:38:10,720 someone for 15 bucks an hour, 798 00:38:08,240 --> 00:38:12,400 do not expect them to build you a next 799 00:38:10,720 --> 00:38:13,359 unicorn. 800 00:38:12,400 --> 00:38:16,240 >> Yeah. 801 00:38:13,359 --> 00:38:18,560 >> Like don't don't be like ridiculous 802 00:38:16,240 --> 00:38:20,800 because people thinking that developers 803 00:38:18,560 --> 00:38:22,880 for 200 bucks an hour and 20 bucks an 804 00:38:20,800 --> 00:38:24,880 hour are the same thing. No, they're 805 00:38:22,880 --> 00:38:25,520 like completely different. 806 00:38:24,880 --> 00:38:28,240 >> Yeah. 807 00:38:25,520 --> 00:38:32,079 >> Can both of them be stupid? Yes. But 808 00:38:28,240 --> 00:38:34,400 generally this is this is a difference. 809 00:38:32,079 --> 00:38:36,960 Yeah, that that that that makes absolute 810 00:38:34,400 --> 00:38:39,680 sense and I see how your your your 811 00:38:36,960 --> 00:38:43,200 advisory consulting investing sort of 812 00:38:39,680 --> 00:38:45,359 like fit hand inand with that because uh 813 00:38:43,200 --> 00:38:49,599 you definitely 814 00:38:45,359 --> 00:38:51,599 uh they say taste and judgment is a big 815 00:38:49,599 --> 00:38:54,400 thing these days and I think with your 816 00:38:51,599 --> 00:38:58,720 experience it comes a lot of that. So, 817 00:38:54,400 --> 00:39:03,280 um I just wanted to to to end out with 818 00:38:58,720 --> 00:39:06,079 um I know that 819 00:39:03,280 --> 00:39:07,359 you I've said this a couple times 820 00:39:06,079 --> 00:39:09,760 already that you don't know where the 821 00:39:07,359 --> 00:39:12,560 industry is going, but if there was one 822 00:39:09,760 --> 00:39:16,000 bold prediction that you can make for 823 00:39:12,560 --> 00:39:20,000 the next 12 to 18 months that you can 824 00:39:16,000 --> 00:39:25,520 say, you know, I I I I I'll put my name 825 00:39:20,000 --> 00:39:28,960 on this prediction. What would it be? 826 00:39:25,520 --> 00:39:33,280 >> Oh boy. Uh, and after people will quote 827 00:39:28,960 --> 00:39:34,560 me and I would be like, um, I I'll I'll 828 00:39:33,280 --> 00:39:39,040 clip this one. 829 00:39:34,560 --> 00:39:40,560 >> Oh no. Um, okay. I I think I I will give 830 00:39:39,040 --> 00:39:42,079 you two just in case. 831 00:39:40,560 --> 00:39:44,480 >> Okay. [laughter] 832 00:39:42,079 --> 00:39:48,960 Uh, to to choose which one will work. 833 00:39:44,480 --> 00:39:52,720 Um, probably no. Uh, none of them. I 834 00:39:48,960 --> 00:39:55,280 think uh the first logic that we would 835 00:39:52,720 --> 00:39:58,079 see is 836 00:39:55,280 --> 00:39:59,119 that we'll have more and more um open 837 00:39:58,079 --> 00:39:59,520 source usage. 838 00:39:59,119 --> 00:40:03,200 >> Okay. 839 00:39:59,520 --> 00:40:07,040 >> Uh so basically we'll have um 840 00:40:03,200 --> 00:40:11,119 multiidentic systems on um many open 841 00:40:07,040 --> 00:40:13,440 source uh models that um you're not 842 00:40:11,119 --> 00:40:16,320 giving away your data. Again, if if 843 00:40:13,440 --> 00:40:19,520 you're a small person like company 844 00:40:16,320 --> 00:40:21,760 building some not like life science type 845 00:40:19,520 --> 00:40:23,760 of stuff and you building another S for 846 00:40:21,760 --> 00:40:25,680 X Y and Z things, it doesn't really 847 00:40:23,760 --> 00:40:28,800 matter. But when we're talking about 848 00:40:25,680 --> 00:40:31,280 like really companies with a lot of uh 849 00:40:28,800 --> 00:40:33,119 private data, hey, open source is a way 850 00:40:31,280 --> 00:40:36,800 to go and with the right harnessing, 851 00:40:33,119 --> 00:40:40,240 it's actually not much worse than um the 852 00:40:36,800 --> 00:40:42,880 best models on the market. Um so it is 853 00:40:40,240 --> 00:40:45,440 and much cheaper like it it is really 854 00:40:42,880 --> 00:40:47,760 important in terms of the inference. I 855 00:40:45,440 --> 00:40:51,680 think this is the one prediction that I 856 00:40:47,760 --> 00:40:55,440 see. The second is I think we'll have 857 00:40:51,680 --> 00:40:57,359 more and more systems that um have a 858 00:40:55,440 --> 00:41:01,520 different 859 00:40:57,359 --> 00:41:05,760 um user flow experiences where we 860 00:41:01,520 --> 00:41:07,599 actually have less uh people um 861 00:41:05,760 --> 00:41:10,160 interacting with the system and doing 862 00:41:07,599 --> 00:41:13,200 whatever system need like inputting 863 00:41:10,160 --> 00:41:15,440 information inside of the CRM etc. So 864 00:41:13,200 --> 00:41:18,880 you basically will have a system that 865 00:41:15,440 --> 00:41:20,400 will give you uh tasks that already 866 00:41:18,880 --> 00:41:22,960 completed 867 00:41:20,400 --> 00:41:24,560 or suggest tasks to do. So it will be 868 00:41:22,960 --> 00:41:27,520 something like okay so you had a 869 00:41:24,560 --> 00:41:30,240 conversation with John uh so I logged 870 00:41:27,520 --> 00:41:32,319 this conversation based on this uh 871 00:41:30,240 --> 00:41:34,480 here's the proposal that I generated 872 00:41:32,319 --> 00:41:36,720 please review and approve and here's the 873 00:41:34,480 --> 00:41:40,079 contract please review and approve as 874 00:41:36,720 --> 00:41:42,720 well as uh here's a few tasks to tell 875 00:41:40,079 --> 00:41:46,319 the product team to potentially add and 876 00:41:42,720 --> 00:41:48,720 I already send those so this is the way 877 00:41:46,319 --> 00:41:51,280 how we would see the interactions with u 878 00:41:48,720 --> 00:41:53,200 mini interfaces in the future I'm not 879 00:41:51,280 --> 00:41:56,000 saying that all will be text. I think it 880 00:41:53,200 --> 00:41:58,319 will be still interfaces and UI etc. I 881 00:41:56,000 --> 00:42:00,800 think we need the uh visual 882 00:41:58,319 --> 00:42:04,480 representation of those things. However, 883 00:42:00,800 --> 00:42:06,880 I don't think we need um actually 884 00:42:04,480 --> 00:42:08,640 manually click through those uh system 885 00:42:06,880 --> 00:42:12,640 and interfaces. 886 00:42:08,640 --> 00:42:15,760 And I would pay for this CRM uh more 887 00:42:12,640 --> 00:42:18,000 than I would pay for the regular one 888 00:42:15,760 --> 00:42:20,720 because it's actually reducing my uh 889 00:42:18,000 --> 00:42:22,400 flows. And you can do this inside of 890 00:42:20,720 --> 00:42:24,079 Hotspot. You can do this inside of 891 00:42:22,400 --> 00:42:26,480 Salesforce right now if you set up this 892 00:42:24,079 --> 00:42:28,160 right, but it's not a default. 893 00:42:26,480 --> 00:42:31,760 >> Interesting. 894 00:42:28,160 --> 00:42:34,560 Yeah, decision data is important. Yep. 895 00:42:31,760 --> 00:42:38,160 Um, thank you VM for your your time 896 00:42:34,560 --> 00:42:40,079 today. Uh, it was a very enlightening 897 00:42:38,160 --> 00:42:41,839 conversation for me and a very fun one 898 00:42:40,079 --> 00:42:44,079 because it's about software development 899 00:42:41,839 --> 00:42:46,560 which uh 900 00:42:44,079 --> 00:42:49,200 it's been a long-term passion of mine. 901 00:42:46,560 --> 00:42:53,200 So, uh, have yourself a great day, sir. 902 00:42:49,200 --> 00:42:53,200 And, uh, maybe stay on for a second.