1 00:00:00,040 --> 00:00:02,080 If they understood the problem, they wouldn't be hiring you, 2 00:00:02,080 --> 00:00:04,920 they'd have fixed it themselves. Simple is complicated enough, 3 00:00:05,080 --> 00:00:07,600 especially at scale. There's a lot of people trying 4 00:00:07,600 --> 00:00:10,160 to adopt new technology just for the sake of it, and that is 5 00:00:10,160 --> 00:00:12,240 something that a good engineer doesn't do. 6 00:00:12,240 --> 00:00:14,480 The good architects are usually the ones where magically 7 00:00:14,480 --> 00:00:17,120 everything goes well and nobody knows exactly why. 8 00:00:17,440 --> 00:00:20,560 What is your advice for people that really want to become great 9 00:00:20,560 --> 00:00:22,160 software engineers? What do they need to do? 10 00:00:25,800 --> 00:00:29,000 What is your advice for people that really want to become great 11 00:00:29,000 --> 00:00:30,600 software engineers? What do they need to do? 12 00:00:31,960 --> 00:00:33,920 Increase your breadth. There's a lot of focus on the 13 00:00:33,920 --> 00:00:38,040 depth early on in your career, which is important, and there's 14 00:00:38,040 --> 00:00:43,800 an advice that perpetuates which says focus, true focus. 15 00:00:44,200 --> 00:00:46,640 But at the same time don't dismiss everything else that's 16 00:00:46,640 --> 00:00:48,360 happening. Increase your breadth. 17 00:00:49,000 --> 00:00:53,400 Personally, I optimised one thing that really got me far, 18 00:00:53,680 --> 00:00:57,320 which is having the ability to learn effectively in very short 19 00:00:57,320 --> 00:01:00,840 periods of time. And that's why whenever I tackle 20 00:01:00,840 --> 00:01:04,280 a problem, I don't have a problem digging really deep into 21 00:01:04,280 --> 00:01:09,160 anything UI encounter from quantum mechanics all the way to 22 00:01:09,880 --> 00:01:13,200 any technical problem that I face on a day-to-day basis. 23 00:01:14,840 --> 00:01:17,000 I'm a, you know, I'm a university dropout. 24 00:01:17,240 --> 00:01:20,080 I never finished my computer science education, but that did 25 00:01:20,080 --> 00:01:24,400 not deter me from picking up that education myself and really 26 00:01:24,400 --> 00:01:27,400 digging deep into all sorts of theoretical topics, but also 27 00:01:27,400 --> 00:01:30,960 practical areas and topics. And I'm a perpetual learner. 28 00:01:31,920 --> 00:01:36,240 So in my opinion, the next phase where we're heading, it's going 29 00:01:36,240 --> 00:01:39,560 to require people to be able to learn really fast, get out of 30 00:01:39,560 --> 00:01:43,800 their comfort zone really fast, dig deep really fast, and attain 31 00:01:43,800 --> 00:01:46,720 high levels of functional proficiency in any new area. 32 00:01:47,360 --> 00:01:51,320 And then attaining mastery is not always required in 33 00:01:51,320 --> 00:01:53,360 everything. So just pick a few topics where 34 00:01:53,360 --> 00:01:55,480 you attain mastery in but also increase your breadth. 35 00:01:56,120 --> 00:01:59,320 And if we are talking about the skills that are for the future, 36 00:01:59,360 --> 00:02:01,400 you're saying technology accelerates quite quickly. 37 00:02:01,680 --> 00:02:04,120 What are some of the skills that you think people should focus on 38 00:02:04,360 --> 00:02:06,560 specifically for their career but also for like peace of 39 00:02:06,560 --> 00:02:08,560 fulfilment and looking towards the future? 40 00:02:08,600 --> 00:02:11,160 The ability to learn first and foremost. 41 00:02:11,160 --> 00:02:16,200 None of the tools I work with I I learned I learned SPSS people 42 00:02:16,240 --> 00:02:19,440 back in the day, some some people in the audience will be 43 00:02:19,440 --> 00:02:24,320 cringing right now. SPSS taught myself R to scrape 44 00:02:24,320 --> 00:02:30,240 Twitter back before it became shitter to to model the 2013 45 00:02:30,240 --> 00:02:34,200 fourteen Syrian refugee crisis. I hated Putin before I knew 46 00:02:34,200 --> 00:02:39,680 where Ukraine was on the map, but from our I went into Python 47 00:02:39,680 --> 00:02:42,200 And now Python is the end all be all. 48 00:02:42,200 --> 00:02:45,920 But I'm from an era where people were still asking like Are you 49 00:02:45,920 --> 00:02:47,920 sure? Should you be learning SAS? 50 00:02:48,040 --> 00:02:51,040 Because SAS is better in the market right now. 51 00:02:51,400 --> 00:02:54,160 So God knows what we'll be using in 10 in 10 years. 52 00:02:54,160 --> 00:02:59,400 Unless all the all the LLM start over fitting on on just 53 00:02:59,640 --> 00:03:01,440 languages that are already in use. 54 00:03:01,560 --> 00:03:04,080 We might actually see a stagnation in the development, 55 00:03:04,080 --> 00:03:08,360 in the development of coding languages now, since Claude and 56 00:03:08,360 --> 00:03:11,880 Gemini will not be able to adopt to new languages as much as 57 00:03:11,880 --> 00:03:15,560 produce more Java horrors, meaning we will be stuck with 58 00:03:15,560 --> 00:03:19,080 Java until the he death of the universe, I'm pretty sure. 59 00:03:20,000 --> 00:03:23,400 But what I'm trying to show here is that those tools, those 60 00:03:23,400 --> 00:03:27,120 techniques, cloud platforms, they change so fast that you 61 00:03:27,120 --> 00:03:29,200 can't focus on learning a technique. 62 00:03:29,480 --> 00:03:32,520 You got to learn a method to learn new stuff. 63 00:03:32,520 --> 00:03:36,280 That is the most single and most important skill in the entire 64 00:03:36,280 --> 00:03:41,560 tech field to be able to learn, not even master, but just work 65 00:03:41,560 --> 00:03:44,480 with new tech. Anytime I have a new project, oh 66 00:03:44,480 --> 00:03:47,040 God, it's a new cloud platform I haven't worked with. 67 00:03:47,360 --> 00:03:51,040 OK, well AWS, Google, Azure, I'm pretty sure it has most of the 68 00:03:51,040 --> 00:03:54,040 same functionalities. They they just call every damn 69 00:03:54,040 --> 00:03:57,080 button differently and that's how they vendor lock. 70 00:03:57,080 --> 00:03:59,960 So, so you can't go go to the other cloud platform because the 71 00:03:59,960 --> 00:04:04,360 names is different, which is less of a hurdle than you would 72 00:04:04,360 --> 00:04:06,720 think. And then there's of course 73 00:04:06,720 --> 00:04:09,680 nuance, like the billing models are different, the scaling is 74 00:04:09,680 --> 00:04:12,880 different. The, the, the type of servers 75 00:04:12,880 --> 00:04:16,760 you can run it in is different. You cannot local host, I mean 76 00:04:16,800 --> 00:04:23,240 sovereign this this version. Remember that people sovereignty 77 00:04:23,240 --> 00:04:25,400 is the new local host. You're just running your own 78 00:04:25,400 --> 00:04:28,320 damn model on your own damn computer. 79 00:04:28,320 --> 00:04:31,320 But you don't call it local host you you call it sovereign. 80 00:04:31,320 --> 00:04:33,280 The client will love it, thank me later. 81 00:04:34,440 --> 00:04:37,800 So it's the and that's the other thing. 82 00:04:38,000 --> 00:04:42,880 And I mentioned that before understanding the words because 83 00:04:42,880 --> 00:04:45,760 buzzwords will come, but they usually mean something that 84 00:04:45,760 --> 00:04:50,000 wasn't new at all. Again, sovereignty is just local 85 00:04:50,000 --> 00:04:53,440 host but new. Everybody screams as as was 86 00:04:53,440 --> 00:04:56,800 screaming AI, but we mainly meant just deep learning on 87 00:04:56,800 --> 00:04:59,280 steroids with a lot a lot of GPU and compute. 88 00:04:59,520 --> 00:05:03,080 Big data was just data that you store on multiple different 89 00:05:03,080 --> 00:05:05,360 servers because you don't have a single server that can contain 90 00:05:05,360 --> 00:05:07,840 it all. But nowadays my laptop has more 91 00:05:07,840 --> 00:05:11,520 storage than a whole complex of servers a decade ago. 92 00:05:11,800 --> 00:05:14,760 Big data is dead. Long live big data. 93 00:05:15,040 --> 00:05:18,800 The core problem here is the core problem of humanity, which 94 00:05:18,800 --> 00:05:22,200 is that sometimes we say the same words but we mean different 95 00:05:22,200 --> 00:05:26,160 things, and sometimes we say different words but we mean the 96 00:05:26,160 --> 00:05:28,560 same thing. And all of miscommunication 97 00:05:28,720 --> 00:05:32,400 stems from this. And in our field, it is 98 00:05:32,400 --> 00:05:37,080 important to figure out not just what words people are using, 99 00:05:37,080 --> 00:05:40,400 which is what sales heavily focuses on, but what their 100 00:05:40,400 --> 00:05:43,760 actual problem is and what they actually mean, which can be, 101 00:05:43,760 --> 00:05:46,000 which can be completely different from the words they're 102 00:05:46,000 --> 00:05:48,040 actually using. Because I've had clients that 103 00:05:48,040 --> 00:05:53,320 that ask for AI and it can mean a random tree, a random forest, 104 00:05:53,320 --> 00:05:56,800 It can mean you have some kind of image recognition model. 105 00:05:56,800 --> 00:05:59,240 It can mean some type of speech synthesis. 106 00:05:59,240 --> 00:06:05,000 It can mean producing synthetic data of bananas using 3D models. 107 00:06:05,000 --> 00:06:07,160 It can mean so many different things, but they all use the 108 00:06:07,160 --> 00:06:10,120 same word and it's our job to figure out what the heck they 109 00:06:10,120 --> 00:06:14,000 want and then to build that. Not just to receive orders and 110 00:06:14,000 --> 00:06:18,640 build something, because that job will be outsourced and 111 00:06:18,800 --> 00:06:22,880 automated to hell him back by whatever cloud copilot thing 112 00:06:22,880 --> 00:06:30,120 they they buy and or consume. But the core skill to figure out 113 00:06:30,120 --> 00:06:33,320 what the fluff they actually want and need and what is the 114 00:06:33,320 --> 00:06:36,600 problem behind the problem. Because if they understood the 115 00:06:36,600 --> 00:06:38,560 problem, they wouldn't be hiring you, they'd have fixed it 116 00:06:38,560 --> 00:06:40,520 themselves. But they do not understand the 117 00:06:40,520 --> 00:06:42,240 problem, do not understand the solution. 118 00:06:42,240 --> 00:06:45,080 It's our job to figure out that solution, then do that. 119 00:06:45,360 --> 00:06:48,680 And to do that we need to be able to learn new tech and learn 120 00:06:48,680 --> 00:06:52,200 new stuff. And that's why learning is key. 121 00:06:53,000 --> 00:06:56,240 What have you seen in how architects execute that makes 122 00:06:56,240 --> 00:06:58,280 them really good architects for the people that actually go on 123 00:06:58,280 --> 00:07:00,320 that career path? Yeah, I always say, I mean, 124 00:07:00,320 --> 00:07:03,880 first comment would be it's not a sort of binary choice or I 125 00:07:03,880 --> 00:07:07,360 don't think that being an architect is determined by 126 00:07:07,360 --> 00:07:10,600 what's on your business card. So I think you can be a 127 00:07:10,600 --> 00:07:16,120 technical product person with a very architecture mindset, right 128 00:07:16,120 --> 00:07:18,560 for me. Well, let me start the other way 129 00:07:18,560 --> 00:07:20,240 around. The the bad architects are 130 00:07:20,240 --> 00:07:22,760 easier to spot. The good architects are usually 131 00:07:22,760 --> 00:07:25,320 the ones where magically everything goes well and nobody 132 00:07:25,320 --> 00:07:28,800 knows exactly why the bad architects are easier to spot. 133 00:07:28,800 --> 00:07:32,240 And it's like people who spew out a lot of buzzwords, you 134 00:07:32,240 --> 00:07:34,840 know, it's like, oh, everything must be cloud native or loosely 135 00:07:34,840 --> 00:07:37,240 coupled. Now I don't need an architect to 136 00:07:37,240 --> 00:07:39,040 tell me that. I can just put a poster on the 137 00:07:39,040 --> 00:07:40,680 wall. And then why? 138 00:07:40,680 --> 00:07:43,680 Why do I need this? This person or people who 139 00:07:43,680 --> 00:07:46,080 believe they should have all the decision power. 140 00:07:46,400 --> 00:07:48,600 It's like, oh developers, I'm here to guide you. 141 00:07:48,600 --> 00:07:50,800 You should make 3 components, right? 142 00:07:50,800 --> 00:07:54,320 No more, no less. And those I find the not so 143 00:07:54,560 --> 00:07:59,000 stellar architects, the mantra that we have in the workshop and 144 00:07:59,000 --> 00:08:03,520 that I, that I very much follow is that architects shouldn't try 145 00:08:03,520 --> 00:08:06,600 to be the smartest people, but they should make everybody else 146 00:08:06,600 --> 00:08:08,800 smarter. All right, that's an amplifier. 147 00:08:09,120 --> 00:08:11,320 You don't want to be a kind of Oracle. 148 00:08:11,320 --> 00:08:13,520 When people come with the questions and look for magic 149 00:08:13,520 --> 00:08:16,320 answers, Oh, what should we do? People come to me, right? 150 00:08:16,320 --> 00:08:19,720 And I'm like, well, it's kind of charming, but how, how do I know 151 00:08:19,720 --> 00:08:21,640 it's your project? It's your application, right? 152 00:08:21,640 --> 00:08:24,120 How can I make all the decisions for you? 153 00:08:24,440 --> 00:08:27,640 But what I can do is help you make a, a better decision. 154 00:08:28,040 --> 00:08:31,320 And I think to me, that is one of the characteristics of a, of 155 00:08:31,320 --> 00:08:34,320 a great architect. You can absorb the context, 156 00:08:34,320 --> 00:08:38,159 things that people explain to you, but you can uncover blind 157 00:08:38,159 --> 00:08:41,760 spots or maybe help people see different points of view, 158 00:08:41,760 --> 00:08:45,960 different angles, distilled trade-offs that they might be 159 00:08:45,960 --> 00:08:50,080 implicitly making, but they're not aware of those kind of 160 00:08:50,080 --> 00:08:51,880 examples. I think that makes a great 161 00:08:51,880 --> 00:08:55,440 architect and it makes a nice architect if you wish, because 162 00:08:55,680 --> 00:08:58,720 you're not trying to be the the ivory tower big boss, but he 163 00:08:58,720 --> 00:09:01,240 actually an amplifier to the team. 164 00:09:01,880 --> 00:09:04,520 Any other foundations or fundamentals you want to 165 00:09:04,520 --> 00:09:08,360 highlight that distinguishes great engineers from not so 166 00:09:08,360 --> 00:09:13,520 great ones? If I look at what is happening 167 00:09:13,520 --> 00:09:18,680 in, in, in picnic context again, what what really strikes me is 168 00:09:18,680 --> 00:09:25,000 that our our engineering culture is relatively uniform in the 169 00:09:25,000 --> 00:09:27,200 sense that we started only 10 years ago. 170 00:09:27,480 --> 00:09:32,080 We have a very strong developer platform that was developed from 171 00:09:32,080 --> 00:09:36,560 the get go. And there, what I see in some 172 00:09:36,560 --> 00:09:41,520 other companies and other contexts is that there's a lot 173 00:09:41,520 --> 00:09:46,320 of people trying to adopt new technology just for the sake of 174 00:09:46,320 --> 00:09:49,280 it. And that is something that I 175 00:09:49,400 --> 00:09:51,600 feel that a good engineer doesn't do right. 176 00:09:51,600 --> 00:09:55,120 So if you adopt new technology, if you find out about a new 177 00:09:55,120 --> 00:10:00,040 technology, you need to understand whether it's actually 178 00:10:00,040 --> 00:10:03,960 an improvement of what you have. Make very clear also why that is 179 00:10:03,960 --> 00:10:08,080 the case and not just jump on any, any bandwagon. 180 00:10:08,080 --> 00:10:11,600 And that's something that I feel that very good engineers are 181 00:10:11,600 --> 00:10:14,480 good at, right? So they there's this sites, I'm 182 00:10:14,480 --> 00:10:16,880 not sure if you know it, the boring technology dot COP. 183 00:10:17,440 --> 00:10:19,080 I don't know it, but that sounds really funny. 184 00:10:19,080 --> 00:10:23,240 Yeah, yeah. Which sort of advocates for use 185 00:10:23,240 --> 00:10:27,840 what works, use what is proven use and then use it in a very 186 00:10:28,200 --> 00:10:30,840 effective way. As you mentioned that that is I 187 00:10:30,840 --> 00:10:34,600 think what what good engineers do and discerning between OK, 188 00:10:34,600 --> 00:10:37,720 but what new technology is that interesting to jump onto 189 00:10:37,720 --> 00:10:39,360 etcetera. And that is super hard, of 190 00:10:39,360 --> 00:10:43,440 course, and also maybe a bit more of an art than than a 191 00:10:43,440 --> 00:10:46,640 science sometimes. But it's precisely that's kind 192 00:10:46,640 --> 00:10:52,320 of gut feeling, I would say that's that that knowledge, that 193 00:10:52,320 --> 00:10:55,720 wisdom that you have as a, as a, as a Craftsman, as an engineer, 194 00:10:55,720 --> 00:10:59,400 that sets you apart from people who, yeah, who don't have that 195 00:10:59,400 --> 00:11:02,320 taste or who don't have that knowledge to, to make a good 196 00:11:02,320 --> 00:11:05,000 decision. And then sometimes you end up in 197 00:11:05,000 --> 00:11:08,680 situations where there's companies that are much smaller 198 00:11:08,680 --> 00:11:12,800 than than the company I work at that have like dozens of 199 00:11:12,800 --> 00:11:18,440 different technology stacks because, well, people decided 200 00:11:18,440 --> 00:11:21,440 that in some kind of context, it was better to use Rust or to use 201 00:11:21,440 --> 00:11:23,960 Go or to. And then you end up with this 202 00:11:24,640 --> 00:11:27,720 fantastic collection of technologies that nobody really 203 00:11:27,720 --> 00:11:29,800 knows how to manage anymore, really know how to, how to 204 00:11:29,800 --> 00:11:32,520 operate anymore, really knows how to, how to bring this 205 00:11:32,520 --> 00:11:36,200 together. And yeah, that's, I think one of 206 00:11:36,200 --> 00:11:38,360 the hallmarks to come back to your question of a good engineer 207 00:11:38,360 --> 00:11:42,560 is to actually also limit this and, and, and, and stick to, 208 00:11:43,440 --> 00:11:44,960 yeah, proven things and what works. 209 00:11:45,160 --> 00:11:47,560 Yeah. Is this then your advice as 210 00:11:47,560 --> 00:11:49,960 well? So to stick with what you have, 211 00:11:49,960 --> 00:11:53,120 start with the initial design that is simple and then stick 212 00:11:53,120 --> 00:11:56,520 with it until you actually reach certain limits and you can kind 213 00:11:56,520 --> 00:11:59,080 of based on your usage, forecast what the next. 214 00:11:59,080 --> 00:12:02,320 Iteration would precisely if I was a start up founder or acto 215 00:12:02,320 --> 00:12:05,520 of a start up right? I would never go and start 216 00:12:05,520 --> 00:12:08,720 building for 100 X scale. Even if I'm like a firm believer 217 00:12:08,720 --> 00:12:12,480 in my idea and know that it's going to reach the masses and 218 00:12:12,480 --> 00:12:15,000 it's going to be global. I would never design for that. 219 00:12:15,320 --> 00:12:19,760 I would design for getting 100 users or maybe even 1000, right? 220 00:12:20,160 --> 00:12:22,320 And I would just run everything on a single VM, to be honest. 221 00:12:22,320 --> 00:12:25,120 Maybe 2, make it a bit more available. 222 00:12:25,680 --> 00:12:29,280 I would not go and build massive database clusters. 223 00:12:29,600 --> 00:12:32,640 Just start with one, maybe 2 nodes tops. 224 00:12:32,840 --> 00:12:35,960 Replication, that's it. And this will like carry you 225 00:12:35,960 --> 00:12:41,360 really, really far. To be honest, once I attain or 226 00:12:41,360 --> 00:12:44,880 I'm get, I'm like close to 80% of these limits of what my 227 00:12:44,880 --> 00:12:47,320 resources can do. And by the way, vertical scaling 228 00:12:47,320 --> 00:12:50,280 can take you a long way. Like you have no idea. 229 00:12:50,280 --> 00:12:55,760 We have my SQL nodes that have hundreds of CP us on them. 230 00:12:57,240 --> 00:13:00,880 So you don't really need to go and start sharding and doing all 231 00:13:00,880 --> 00:13:04,120 sorts of fancy stuff when vertical scaling can still carry 232 00:13:04,120 --> 00:13:08,160 you. And right now with cloud native 233 00:13:08,160 --> 00:13:09,800 solutions, you can go pretty far. 234 00:13:09,800 --> 00:13:13,800 Like you can get a one VM with 120 CP cores, right? 235 00:13:13,800 --> 00:13:17,280 You can get terabytes of memory, Yeah. 236 00:13:17,960 --> 00:13:20,880 And that can do a lot for you. You don't need to start doing 237 00:13:20,880 --> 00:13:23,360 horizontal scaling before you get you really start hitting 238 00:13:23,360 --> 00:13:25,160 these limits. That's what I've seen where it's 239 00:13:25,160 --> 00:13:28,200 challenging for people, where they say OK, I'm a little bit of 240 00:13:28,200 --> 00:13:30,920 a perfectionist. I value quality in what I do on 241 00:13:30,920 --> 00:13:33,720 a day-to-day. My job is basically part of who 242 00:13:33,720 --> 00:13:35,880 I am. And if I don't deliver the 243 00:13:35,880 --> 00:13:38,280 highest quality that I can in this aspect, like I'm doing 244 00:13:38,280 --> 00:13:41,600 myself with this justice. And it's very like, if I lay it 245 00:13:41,600 --> 00:13:44,200 out like that, it might sound selfish, but those people have 246 00:13:44,200 --> 00:13:46,040 no clue that what they're saying is selfish. 247 00:13:46,040 --> 00:13:48,200 They genuinely think it's best for the organization. 248 00:13:48,520 --> 00:13:50,640 There's just some misinformation and misunderstanding. 249 00:13:50,640 --> 00:13:53,760 Because there's a dogmatic aspect of software engineering 250 00:13:53,760 --> 00:13:57,280 and it was perpetuated in the previous era into two thousands. 251 00:13:57,360 --> 00:14:01,680 2000 I when I ventured in working professionals of Engine 252 00:14:01,720 --> 00:14:04,360 2007 8-9, there was a lot of dogma. 253 00:14:04,360 --> 00:14:06,840 Like there's a lot of this is how you do things and this is 254 00:14:06,840 --> 00:14:08,960 how you should do things. And if you do things any other 255 00:14:08,960 --> 00:14:13,680 way, you're wrong. And part of that is elegance of 256 00:14:13,680 --> 00:14:15,760 code, code cleanliness, code quality. 257 00:14:15,760 --> 00:14:18,000 You're all raised on this, these these concepts, right? 258 00:14:18,000 --> 00:14:20,920 And they're important. And everybody always complained 259 00:14:20,920 --> 00:14:23,240 about bad code and horrible code because it was difficult to 260 00:14:23,240 --> 00:14:24,880 maintain. Nobody enjoys that. 261 00:14:24,920 --> 00:14:26,760 You're going to make it very difficult for the future 262 00:14:26,760 --> 00:14:28,120 engineers to solve these problems. 263 00:14:29,080 --> 00:14:33,160 But sometimes, you know, you don't have future engineers to 264 00:14:33,160 --> 00:14:34,920 solve these problems. Sometimes nobody's going to 265 00:14:34,920 --> 00:14:36,960 touch this code for decades to come. 266 00:14:37,320 --> 00:14:39,920 It's OK if it's just, it just works. 267 00:14:40,520 --> 00:14:42,800 And then we ended up on the complete other end of the 268 00:14:42,800 --> 00:14:46,480 spectrum where we tried to clean up everything that we started 269 00:14:46,480 --> 00:14:49,920 adopting all sorts of design overcomplicated abstractions 270 00:14:49,920 --> 00:14:53,560 that are are just horrible to read, horrible to reason about. 271 00:14:53,960 --> 00:14:57,640 It takes you decades to become proficient in just reading that 272 00:14:57,640 --> 00:15:00,800 type of code. And then when you write it, it 273 00:15:00,800 --> 00:15:03,400 gets even more complicated and it's very difficult to maintain. 274 00:15:04,400 --> 00:15:06,760 And big tech and I always tell my colleagues and the juniors 275 00:15:06,760 --> 00:15:11,240 that report to me, simple is complicated enough, especially 276 00:15:11,240 --> 00:15:14,360 at scale. So just write it in the dumbest, 277 00:15:14,560 --> 00:15:17,240 most simple way possible. This is why I also love Go, for 278 00:15:17,240 --> 00:15:19,160 example as a language. Right. 279 00:15:20,280 --> 00:15:21,720 It's amazing. It's dumb. 280 00:15:22,240 --> 00:15:25,680 You don't have to think too much, it's just all in your 281 00:15:25,680 --> 00:15:27,760 face. Whenever you want to understand 282 00:15:27,760 --> 00:15:29,520 how something is built, just trace it around. 283 00:15:29,520 --> 00:15:32,080 It's there. It's there, fantastic. 284 00:15:32,160 --> 00:15:35,720 I love that. And when the situations where 285 00:15:35,720 --> 00:15:38,200 we're building at scale, I don't want to reason about 286 00:15:38,200 --> 00:15:42,080 abstractions, I don't want to think about edge cases, I don't 287 00:15:42,080 --> 00:15:45,880 want to think about complicated memory allocations and 288 00:15:45,880 --> 00:15:49,280 complicated memory structures. You know, I just want things to 289 00:15:49,280 --> 00:15:52,960 be simple. Memory leaks happen often and 290 00:15:53,000 --> 00:15:55,240 they are very catastrophic at scale. 291 00:15:55,840 --> 00:15:57,040 I don't want to think about these things. 292 00:15:57,040 --> 00:15:59,160 I don't want to think about how my garbage collector is going 293 00:15:59,160 --> 00:16:02,560 to, you know, shut down the whole world and start causing 294 00:16:02,920 --> 00:16:06,800 multi second latency on my API requests, which at scale again, 295 00:16:06,800 --> 00:16:08,200 become a gigantic problem for me. 296 00:16:08,440 --> 00:16:13,200 I just want things to be simple, easy to reason about and that's 297 00:16:13,200 --> 00:16:15,320 it. What is then your perspective on 298 00:16:15,320 --> 00:16:17,160 building systems that can evolve? 299 00:16:17,160 --> 00:16:19,720 Because some people you can say, OK, this is the system designed 300 00:16:19,720 --> 00:16:24,000 that we have and we're going to continue on this until we reach 301 00:16:24,000 --> 00:16:26,360 a certain level of scale, and then we're going to revamp and 302 00:16:26,360 --> 00:16:28,440 say this would be the next iteration. 303 00:16:28,760 --> 00:16:31,800 But some people also say, OK, we have a system now and this is 304 00:16:31,800 --> 00:16:34,400 evolvable to whatever skill we need. 305 00:16:34,400 --> 00:16:36,800 And then they put in the extra effort to make sure that that is 306 00:16:36,800 --> 00:16:41,160 there from the beginning. In our industry, very few 307 00:16:41,160 --> 00:16:45,840 technologies last for like a decade plus and the rest is 308 00:16:46,200 --> 00:16:51,840 changing often and that's why I personally don't invest or I 309 00:16:51,840 --> 00:16:54,480 wouldn't design a solution that will last like 1020, thirty 310 00:16:54,480 --> 00:16:57,280 years, unlike what we did in the 80s or 90s. 311 00:16:57,280 --> 00:16:59,360 That's why that's how they that was the design mentality. 312 00:16:59,360 --> 00:17:01,800 But things were way less complicated than they are today 313 00:17:03,080 --> 00:17:07,520 and the scale was different. So I would always come with the 314 00:17:07,520 --> 00:17:09,920 mindset that I'm just going to design for the next order of 315 00:17:09,920 --> 00:17:11,920 magnitude and then I'm going to evolve it from there. 316 00:17:11,920 --> 00:17:14,839 And that's going to require another round of investment. 317 00:17:14,960 --> 00:17:17,960 There's going to be continual investments in this and that. 318 00:17:17,960 --> 00:17:20,280 Business people are not going to love this because that makes 319 00:17:21,000 --> 00:17:23,880 accounting much harder. That makes strategic finance 320 00:17:23,880 --> 00:17:27,280 much harder because they cannot predict or anticipate the costs 321 00:17:27,280 --> 00:17:29,840 that might come in the future. They cannot extrapolate from 322 00:17:29,840 --> 00:17:31,880 today. You know they don't like this 323 00:17:31,880 --> 00:17:34,000 fuzziness. Of software, it's all risk. 324 00:17:34,120 --> 00:17:36,160 Yes. So and, and also like they 325 00:17:36,160 --> 00:17:38,520 cannot calculate like their bottom line or projections for 326 00:17:38,520 --> 00:17:40,840 the future, which are things that are very realistic. 327 00:17:40,840 --> 00:17:44,320 Like public, publicly traded companies want to be able to do 328 00:17:44,320 --> 00:17:46,840 these projections and they want to be able to say that, hey, 329 00:17:46,840 --> 00:17:50,560 we're not going to be spending more here in on evolving our 330 00:17:50,560 --> 00:17:52,960 software, but we're going to make much more revenue and 331 00:17:52,960 --> 00:17:54,360 increase our bottom line and margin. 332 00:17:54,680 --> 00:17:57,680 That's what businesses look for. So it's understandable. 333 00:17:57,680 --> 00:18:00,400 But from a software perspective, there's needs to be regular 334 00:18:00,400 --> 00:18:01,880 investments. And I think the only 335 00:18:01,880 --> 00:18:04,080 approachable way, which is something we do at GitHub is to 336 00:18:04,080 --> 00:18:09,640 continuously revisit these investments on a quarterly by 337 00:18:09,640 --> 00:18:13,040 quarterly, half a year or a year basis, right? 338 00:18:13,600 --> 00:18:17,120 We can make projections for the next 3-4, five years, but the 339 00:18:17,120 --> 00:18:19,680 confidence in these projections is less and less and less the 340 00:18:19,680 --> 00:18:22,520 more time grows and I think a lot more companies need to start 341 00:18:22,520 --> 00:18:24,800 doing the same. We're always working in teams. 342 00:18:25,000 --> 00:18:28,040 Also from a business standpoint, how would I, for example, say 343 00:18:28,280 --> 00:18:30,560 we're going to design for this until we reach scale and then 344 00:18:30,560 --> 00:18:32,680 we're going to need to revamp? And this is what that's going to 345 00:18:32,880 --> 00:18:35,400 entail for people that don't understand how to build 346 00:18:35,400 --> 00:18:37,080 software. Understand the business 347 00:18:37,080 --> 00:18:38,800 constraints. So you're never going to be able 348 00:18:38,800 --> 00:18:42,440 to convince the business folks about of something technically. 349 00:18:43,000 --> 00:18:44,640 They're just. It's too abstract. 350 00:18:45,440 --> 00:18:48,880 Even if it feels too intuitive for us, for them it's. 351 00:18:49,480 --> 00:18:50,400 Magic. Yeah. 352 00:18:51,000 --> 00:18:54,160 And even if they try to understand how things work, 353 00:18:54,280 --> 00:18:56,160 they're never going to comprehend the magnitude of 354 00:18:56,160 --> 00:18:58,800 problems and they're never going to have have this intuition 355 00:18:58,800 --> 00:19:03,400 about how software is and how not hard, you know how how 356 00:19:03,400 --> 00:19:07,920 evolvable it is. So it's our job to get closer to 357 00:19:07,920 --> 00:19:09,920 their side. Some people say, no, we want the 358 00:19:09,920 --> 00:19:13,480 business to understand they're not gonna it's trickier. 359 00:19:13,480 --> 00:19:17,000 It's easier for you to acquire or assimilate some of the not 360 00:19:17,000 --> 00:19:20,560 everything, some of the business constraints and discuss them 361 00:19:20,560 --> 00:19:22,640 from that perspective. And that's why a lot of 362 00:19:22,640 --> 00:19:26,160 engineers who grow are very good at understanding the business 363 00:19:26,160 --> 00:19:29,280 constraints and then discussing the technical solutions within 364 00:19:29,280 --> 00:19:33,120 those constraints. So for example, I used to work 365 00:19:33,120 --> 00:19:35,080 in the container terminal industry for a while. 366 00:19:35,080 --> 00:19:37,720 We built a huge container terminal. 367 00:19:38,000 --> 00:19:40,600 We furnished everything from the fibre optics layer all the way 368 00:19:40,600 --> 00:19:42,960 to the three tier data centre with all of the enterprise 369 00:19:42,960 --> 00:19:45,960 applications running, running on it and the operating system for 370 00:19:45,960 --> 00:19:47,680 the container terminal. And it's one of the most massive 371 00:19:47,680 --> 00:19:52,000 operations you can ever imagine, one of the most hazardous as 372 00:19:52,000 --> 00:19:55,560 well. So the folks there, they're as 373 00:19:55,560 --> 00:19:59,400 far away from tech as possible. You're not never going to find 374 00:19:59,400 --> 00:20:01,360 the terminal operator who's going to understand what it 375 00:20:01,360 --> 00:20:04,360 means that your database is running hot and you cannot 376 00:20:04,360 --> 00:20:06,640 support their keys. All they care about is, I have a 377 00:20:06,640 --> 00:20:10,800 ship here who's docked. There's like 10,000 containers 378 00:20:10,800 --> 00:20:12,480 on it. I need to empty them into my 379 00:20:12,480 --> 00:20:14,560 yard. I have a very fixed window of 380 00:20:14,560 --> 00:20:17,080 time and if I don't do it in that window of time, this is 381 00:20:17,080 --> 00:20:21,400 going to cost me money. We were operating at 12% 382 00:20:21,400 --> 00:20:23,320 capacity. That terminal was generating 383 00:20:23,320 --> 00:20:31,600 about $50,000 per hour, so 12%. So you can imagine if we get 384 00:20:31,600 --> 00:20:35,040 delayed, every minute of delay is actually money that's 385 00:20:35,080 --> 00:20:37,880 disappearing. So this is how you need to 386 00:20:37,880 --> 00:20:41,280 discuss problems with and and investments and solutions. 387 00:20:41,640 --> 00:20:44,240 You need to understand what the impact is, what the impact is on 388 00:20:44,240 --> 00:20:45,960 the operators. And sometimes it's life 389 00:20:45,960 --> 00:20:50,200 threatening impact, right, If your software fails when ASTS, 390 00:20:50,760 --> 00:20:54,960 when the crane is offloading a container, this can be a very 391 00:20:55,200 --> 00:20:58,600 life threatening situation. You cannot really have your 392 00:20:58,600 --> 00:21:00,040 software malfunction at these points. 393 00:21:00,040 --> 00:21:04,160 So snappiness of the software, sub millisecond latencies, all 394 00:21:04,160 --> 00:21:07,000 of these things matter, right? So you need to approach the 395 00:21:07,000 --> 00:21:09,840 discussions with the stakeholders by understanding 396 00:21:09,840 --> 00:21:11,400 them first. So what I've done for example, 397 00:21:11,400 --> 00:21:16,200 in that industry is I've literally sat on a truck and 398 00:21:16,200 --> 00:21:18,760 went there and I sat with every single individual who works in 399 00:21:18,760 --> 00:21:20,640 that terminal to really understand what they go through 400 00:21:20,640 --> 00:21:23,920 on a daily basis. So a truck driver who's driving 401 00:21:23,920 --> 00:21:27,280 a truck for a six 7-8 hours and it was in Saudi Arabia in 402 00:21:27,280 --> 00:21:30,720 scorching heat is not going to understand why the button that 403 00:21:30,720 --> 00:21:34,960 they clicked is just is so slow. Doesn't matter. 404 00:21:35,200 --> 00:21:37,840 Doesn't he doesn't care, right? He just wants it to just work. 405 00:21:38,840 --> 00:21:41,120 So you got to understand all of these perspectives and then when 406 00:21:41,120 --> 00:21:43,280 you do, you can start approaching discussions from 407 00:21:43,280 --> 00:21:45,920 that vantage point. And this applies to banks, this 408 00:21:45,920 --> 00:21:48,520 applies to hospitals, to any other industry. 409 00:21:48,520 --> 00:21:50,920 So understand the industry, understand the business 410 00:21:50,920 --> 00:21:53,720 constraints, and then any technical solution you want to 411 00:21:53,720 --> 00:21:56,240 discuss an investment, you can approach it from that vantage 412 00:21:56,240 --> 00:21:59,240 point and you're going to have much more success in these 413 00:21:59,240 --> 00:22:01,520 conversations. When you were speaking about 414 00:22:01,520 --> 00:22:05,640 hard skills, how much of enterprise architecture or being 415 00:22:05,640 --> 00:22:08,320 an architect in essence has changed nowadays? 416 00:22:08,720 --> 00:22:10,880 Because I feel like I get bombarded with tools. 417 00:22:11,200 --> 00:22:14,200 There's a lot more possible, specifically with large language 418 00:22:14,200 --> 00:22:17,520 models, but in essence, like the Phantom sketch Artist, 419 00:22:18,080 --> 00:22:20,320 everything underneath kind of still holds true. 420 00:22:20,840 --> 00:22:24,080 From your perspective, does architecture evolve quickly? 421 00:22:24,520 --> 00:22:26,440 Because I feel like tooling does at a more rapid. 422 00:22:26,440 --> 00:22:30,040 Pace, definitely the environment in which you operate has 423 00:22:30,040 --> 00:22:31,760 changed. So there's a couple of 424 00:22:32,240 --> 00:22:36,160 assumptions that architects used to be able to make to make their 425 00:22:36,160 --> 00:22:39,280 lives easier that no longer hold right. 426 00:22:39,280 --> 00:22:42,120 And the one is like snapshotting, like the world is 427 00:22:42,120 --> 00:22:44,680 moving too fast to be able to pretend. 428 00:22:44,680 --> 00:22:47,040 It's like, OK, let's just assume nothing changes. 429 00:22:47,440 --> 00:22:50,480 Well, that turns out to be a very poor assumption in today's 430 00:22:50,480 --> 00:22:53,080 environment. So that, for example, means 431 00:22:53,080 --> 00:22:58,840 exercises like trying to draw a giant ID landscape, you know, or 432 00:22:59,040 --> 00:23:02,280 cataloguing all the applications you have and putting that into 433 00:23:02,280 --> 00:23:03,960 some tools. Taking a few months for that, 434 00:23:04,000 --> 00:23:05,360 yeah. And then a few months, by the 435 00:23:05,360 --> 00:23:07,640 time you're done, this has already changed. 436 00:23:07,640 --> 00:23:11,560 So for me, the assumption of being able to snapshot and 437 00:23:11,560 --> 00:23:14,600 being, I call this the cartographer, being the person 438 00:23:14,600 --> 00:23:18,800 who makes all the great landscapes, that is very useful, 439 00:23:18,800 --> 00:23:21,560 but it's no longer practical because stuff changes too 440 00:23:21,560 --> 00:23:23,800 quickly. So one of the metaphors I use 441 00:23:23,800 --> 00:23:27,440 which captures the change in role for Enterprise Architects 442 00:23:27,440 --> 00:23:30,840 is rather than trying to be the cartographer who walks around 443 00:23:30,840 --> 00:23:34,560 with a giant map of all the pieces, you need to be much more 444 00:23:34,560 --> 00:23:38,520 like a scout. You have a objective in mind. 445 00:23:38,760 --> 00:23:41,440 You want to cross the river or beat the enemy or like move 446 00:23:41,440 --> 00:23:42,920 somewhere or like whatever, right? 447 00:23:42,920 --> 00:23:45,880 You have some objective, you send some scouts, right? 448 00:23:45,880 --> 00:23:47,440 They're going to go look at things. 449 00:23:47,440 --> 00:23:50,400 They figure out what works, what won't work, right? 450 00:23:50,400 --> 00:23:53,880 And they come back to you with a very simple map, but a very 451 00:23:53,880 --> 00:23:56,760 timely and very expressive map, right? 452 00:23:56,760 --> 00:23:58,280 So you want to cross the river. It's like, oh, here's the 453 00:23:58,280 --> 00:24:01,000 mountain, here's the enemy, the bridge is gone, but here's not 454 00:24:01,000 --> 00:24:02,480 so deep. We can cross there, we can go 455 00:24:02,480 --> 00:24:04,760 around whatever it is, right? But they have a very 456 00:24:04,760 --> 00:24:08,160 situational, they have a sort of purpose driven map. 457 00:24:08,680 --> 00:24:13,520 The map is not trying to be like the perfect map of all the exact 458 00:24:13,520 --> 00:24:15,800 landscape. They only depict what's it's 459 00:24:15,800 --> 00:24:20,520 relevant to your move and I think that is becoming a more 460 00:24:20,760 --> 00:24:24,280 viable metaphor. What enterprise architects 461 00:24:24,280 --> 00:24:27,800 should do. It's no longer plausible to like 462 00:24:27,960 --> 00:24:32,080 maintain this giant map where everything is on and everything 463 00:24:32,080 --> 00:24:35,000 is up to date. Rather you need to have a 464 00:24:35,000 --> 00:24:36,760 purpose. You need to have a direction of 465 00:24:36,760 --> 00:24:38,280 mind. Like LLMS Gen. 466 00:24:38,280 --> 00:24:40,400 EI is a great example. I don't get de enterprise 467 00:24:40,400 --> 00:24:42,800 architects, what's our strategy for this? 468 00:24:43,000 --> 00:24:45,160 Where should we find the best first use cases? 469 00:24:45,160 --> 00:24:47,600 How do we integrate this with the rest of our system? 470 00:24:47,960 --> 00:24:51,760 We know LMS have super high rate of evolution, right? 471 00:24:51,760 --> 00:24:55,200 So how do we keep that change out of other systems, right? 472 00:24:55,680 --> 00:24:58,680 What are we doing with with Agents and Agentec, right? 473 00:24:58,840 --> 00:25:01,800 Is that just another form of workflow integration or is that 474 00:25:01,800 --> 00:25:04,160 something novel, right? Give me a point of view. 475 00:25:04,400 --> 00:25:08,640 But it's very purpose driven. And as you will quickly see, you 476 00:25:08,640 --> 00:25:11,320 can only do that if you understand the subject matter, 477 00:25:11,720 --> 00:25:14,000 right? You can sort of arm wave your 478 00:25:14,000 --> 00:25:17,360 way through that, right? You need to have, you know, the 479 00:25:17,360 --> 00:25:20,720 hard tech skills as well. But for me, it's very purpose 480 00:25:20,720 --> 00:25:22,360 driven. What should we do? 481 00:25:22,360 --> 00:25:24,280 Do we have a point of view, right? 482 00:25:24,480 --> 00:25:28,320 Do we have a question in mind? Too many architects try to find 483 00:25:28,360 --> 00:25:30,560 answers when they don't have a question. 484 00:25:30,560 --> 00:25:33,560 And so it's like, oh, here's the thing that answers all possible 485 00:25:33,560 --> 00:25:35,960 questions. I'm like, that would be nice. 486 00:25:37,040 --> 00:25:39,400 Wouldn't that be, wouldn't that be easy? 487 00:25:39,600 --> 00:25:43,680 That doesn't work anymore. You got to start with a question 488 00:25:43,680 --> 00:25:46,200 in your mind. And then like the scout, I was 489 00:25:46,200 --> 00:25:47,560 like, whatever, can we cross the river? 490 00:25:47,560 --> 00:25:49,920 I was OK, come back with the answer to this. 491 00:25:49,920 --> 00:25:52,680 It's got to be situational and timely. 492 00:25:52,960 --> 00:25:56,440 Otherwise it becomes hard work. And we might like our hard work, 493 00:25:56,600 --> 00:25:57,920 but I always say we're not a museum. 494 00:25:58,160 --> 00:26:00,680 We're the IT you. Can't be a historian, yeah. 495 00:26:00,800 --> 00:26:02,080 Exactly. We can't be a historian. 496 00:26:02,080 --> 00:26:03,680 We're not collecting modern art, right? 497 00:26:04,000 --> 00:26:06,200 Our diagrams serve a purpose, right? 498 00:26:06,200 --> 00:26:09,160 They help us make better decisions, get a better 499 00:26:09,160 --> 00:26:11,480 decision, Transparency and discipline. 500 00:26:11,840 --> 00:26:17,120 So a the decision is is clearer and everybody's on the same page 501 00:26:17,120 --> 00:26:19,920 to what the decision was actually what drove the 502 00:26:19,920 --> 00:26:21,920 decision. What trade-offs are we willing 503 00:26:21,920 --> 00:26:24,760 to make? That's what adding the value and 504 00:26:24,760 --> 00:26:27,360 that the picture just has to be a means to that end. 505 00:26:27,600 --> 00:26:30,520 If that end isn't there, the pictures also worthless. 506 00:26:30,840 --> 00:26:34,800 Then it just becomes modern arts and boxes and triangles and kind 507 00:26:34,800 --> 00:26:36,880 of kind of things. From a career perspective, 508 00:26:36,880 --> 00:26:40,600 people are saying very early on, topics are getting automated now 509 00:26:40,600 --> 00:26:43,640 by agents, which means that the people that are educating 510 00:26:43,640 --> 00:26:46,120 themselves and that usually that's a step for them to build 511 00:26:46,120 --> 00:26:49,680 up the skills to do very, very cool things that help society. 512 00:26:50,000 --> 00:26:53,120 That step is now changing and people are asking, well, what do 513 00:26:53,120 --> 00:26:54,800 I do then to make myself not obsolete? 514 00:26:54,800 --> 00:26:57,320 And we're talking about anything with regards to communication 515 00:26:57,320 --> 00:26:59,480 and soft skills and learning. I agree with all of that. 516 00:26:59,880 --> 00:27:01,440 And then there's still a tooling aspect. 517 00:27:01,440 --> 00:27:04,240 Like I don't know what I should advise people with regards to, 518 00:27:05,560 --> 00:27:07,920 OK, how do you go from junior to actually in the end becoming 519 00:27:07,920 --> 00:27:09,720 senior? Because it's such Gray area that 520 00:27:09,720 --> 00:27:11,640 I don't know yet. I feel like we're really 521 00:27:11,640 --> 00:27:14,320 figuring things out. And the only advice that I have 522 00:27:14,320 --> 00:27:19,120 is be as open minded as you can. Try and familiarize yourself 523 00:27:19,120 --> 00:27:20,720 with a lot of things. Try and build as much 524 00:27:20,720 --> 00:27:23,800 foundational knowledge as you can because tooling does change. 525 00:27:23,800 --> 00:27:25,920 I'm wondering what your perspective is on that. 526 00:27:26,240 --> 00:27:32,200 Don't be a fool with a tool. That is one of the core pieces 527 00:27:32,200 --> 00:27:36,040 of advice I got started my career over 9 years ago. 528 00:27:36,040 --> 00:27:38,920 God, don't be a fool with a tool. 529 00:27:40,120 --> 00:27:44,400 Back then we refer to only knowing our only knowing Python 530 00:27:44,400 --> 00:27:47,640 or only back then. Do you remember RPA people? 531 00:27:47,640 --> 00:27:51,160 Do you remember back when RPA was a was a thing? 532 00:27:51,480 --> 00:27:55,120 I was using a Python package back back in the day to to 533 00:27:55,120 --> 00:27:57,720 scrape Twitter. I forgot forgot the name 534 00:27:57,800 --> 00:27:59,360 Selenium. Yeah, yeah. 535 00:27:59,760 --> 00:28:02,880 I was learning Selenium because that was one of the first Python 536 00:28:02,880 --> 00:28:06,000 based browser. This automate the browser and 537 00:28:06,000 --> 00:28:10,840 click down. Back then those were bots, but I 538 00:28:10,840 --> 00:28:14,560 knew people who doubled down on only learning Selenium and all 539 00:28:14,560 --> 00:28:17,240 the insurance and outs. But Selenium was open source and 540 00:28:17,240 --> 00:28:18,800 it was great. You could do everything with it. 541 00:28:18,800 --> 00:28:22,720 But then the blue prisms came in, which was basically the the 542 00:28:22,720 --> 00:28:27,680 company or the company organization equivalent of it. 543 00:28:27,680 --> 00:28:30,400 But it was reliable because you could sue somebody if it didn't 544 00:28:30,400 --> 00:28:32,560 work. And suddenly the whole Selenium 545 00:28:32,560 --> 00:28:36,040 market disappeared because people went for for larger 546 00:28:36,040 --> 00:28:38,920 infrastructure tools instead. And the guy who knew everything 547 00:28:38,920 --> 00:28:42,880 about Selenium was stuck and the people who learned Blue Prism 548 00:28:42,880 --> 00:28:45,520 were laughing at him. Nowadays we laugh at the people 549 00:28:45,520 --> 00:28:48,400 who only know Blue Prism because the cycle continues. 550 00:28:48,400 --> 00:28:52,640 It's not about learning anyone tool, it's about being able to 551 00:28:52,640 --> 00:28:56,160 learn in general. And that involves learning the 552 00:28:56,160 --> 00:29:00,880 insurance and outs of tools. I think it we are switching from 553 00:29:00,880 --> 00:29:07,240 an explicit knowledge economy to an implicit knowledge economy 554 00:29:07,240 --> 00:29:10,600 and our education systems haven't adjusted to that, that 555 00:29:10,760 --> 00:29:15,040 the worth of knowing stuff in your head, memorizing it, that 556 00:29:15,040 --> 00:29:19,280 worth is only going down. But the ability to look up stuff 557 00:29:20,160 --> 00:29:23,320 that was introduced like what, 25 years ago when Google hit the 558 00:29:23,320 --> 00:29:29,760 market, before we had that, we had start start that has been 559 00:29:29,760 --> 00:29:33,360 introduced during our lifetimes. But the value of that, of 560 00:29:33,360 --> 00:29:37,560 finding stuff in large amounts of information, that is only 561 00:29:37,560 --> 00:29:41,560 going up, especially in the age of disinformation, because we 562 00:29:41,560 --> 00:29:47,160 are drowning in information. But truth and wisdom has only 563 00:29:47,160 --> 00:29:51,640 increased invaluable and has become relatively more scarce 564 00:29:51,640 --> 00:29:55,320 because now there's more bots and more fake stuff flooding the 565 00:29:55,320 --> 00:30:00,160 Internet than actual real stuff. And the only way we can still 566 00:30:00,880 --> 00:30:04,440 distinguish truth from fake, because I see so many deep fakes 567 00:30:04,440 --> 00:30:08,840 now of Zelensky, of Putin, of F-30 fives that crashed in Iran. 568 00:30:08,840 --> 00:30:11,840 And we could only tell they were fake because there was still a a 569 00:30:11,840 --> 00:30:15,080 human in the cockpit, because they forgot to prompt it remove 570 00:30:15,080 --> 00:30:18,640 the human from the cockpit. We can only tell truth from fake 571 00:30:18,800 --> 00:30:21,360 because of human error. And that too will end. 572 00:30:22,000 --> 00:30:24,400 So what we need to learn, again, because learning is important, 573 00:30:24,400 --> 00:30:27,920 is to triangulate information to figure out what's real. 574 00:30:28,240 --> 00:30:32,600 And This is why I I'm frustrated with modern media and journalism 575 00:30:32,600 --> 00:30:35,600 that thinks that just regurgitating and repeating what 576 00:30:35,600 --> 00:30:38,440 no liar say is somehow journalism. 577 00:30:38,760 --> 00:30:41,600 I forgot the quote, but somebody once wrote this. 578 00:30:41,840 --> 00:30:45,080 If one source tells you it's raining and the other says 579 00:30:45,080 --> 00:30:48,120 there's sunshine, your job as a journalist is not to report 580 00:30:48,120 --> 00:30:53,480 both, but to open a damn window. And that's what I'm missing. 581 00:30:53,560 --> 00:30:57,160 What is your perspective on AI tooling in this building frame? 582 00:30:57,160 --> 00:30:59,800 Because if you're looking towards end goals, if my end 583 00:30:59,800 --> 00:31:02,400 goal is to build a product or an e-commerce website, I can, I can 584 00:31:02,400 --> 00:31:05,640 generate a lot of that stuff. But would I then learn how to 585 00:31:05,640 --> 00:31:08,080 actually do that? No, I would achieve my goal in a 586 00:31:08,080 --> 00:31:10,800 different manner. Would people from your 587 00:31:10,800 --> 00:31:12,560 perspective that Start learning now? 588 00:31:12,560 --> 00:31:15,640 Should they leverage AI tooling or how can they use it to learn 589 00:31:15,640 --> 00:31:18,160 effectively? That's a great question and 590 00:31:18,160 --> 00:31:22,440 we're still finding that out. I think so. 591 00:31:22,640 --> 00:31:28,840 So my take is yes, you can use it to to set up a lot of things 592 00:31:28,840 --> 00:31:32,960 for you. I would more use it as a tool 593 00:31:32,960 --> 00:31:35,680 for validation rather than something for, for you to 594 00:31:35,680 --> 00:31:40,480 generate things right. So if you are not able to 595 00:31:40,560 --> 00:31:43,640 actually explain what comes out of this model and, and if you 596 00:31:43,640 --> 00:31:47,160 are not able to understand why the code is there, then you will 597 00:31:47,160 --> 00:31:50,480 have a very rough problem also maintaining that and extending 598 00:31:50,480 --> 00:31:52,320 that. And of course then you can say, 599 00:31:52,320 --> 00:31:55,560 OK, well, then I have some agents, AI agents do that for 600 00:31:55,560 --> 00:31:58,120 me. But ultimately there, there's, 601 00:31:58,240 --> 00:32:00,440 there will be a gap of understanding where the model 602 00:32:00,440 --> 00:32:03,320 trips up and, and you are not able to get it, catch it. 603 00:32:04,520 --> 00:32:09,600 So use, use AR models to validate your understanding, try 604 00:32:09,600 --> 00:32:13,360 to generate maybe a few things and see, do I actually 605 00:32:13,440 --> 00:32:16,240 understand what's happening here and not ask the model right that 606 00:32:16,240 --> 00:32:20,640 that's great. But don't focus on on getting a 607 00:32:20,640 --> 00:32:23,280 polished product out of it and then getting sort of the, the, 608 00:32:24,800 --> 00:32:28,880 yeah, the products there. In the end, it's, it's really a 609 00:32:28,880 --> 00:32:32,160 tool that you can use, especially if you're like taxi 610 00:32:32,160 --> 00:32:34,320 driver who's who's going to learn coding. 611 00:32:34,680 --> 00:32:37,280 It's a tool that you can use to, to build your knowledge. 612 00:32:37,800 --> 00:32:40,520 You can also use very much to cheat, of course, and to just 613 00:32:40,520 --> 00:32:43,160 sidestep all of this hard work of, of trying to understand it. 614 00:32:44,280 --> 00:32:46,280 But I don't think that will in the long run help you.