1 00:00:00,040 --> 00:00:03,960 Let's be honest, 90% of all the AI in production is still just 2 00:00:04,120 --> 00:00:08,280 linear regression baby. The worth of knowing stuff in 3 00:00:08,280 --> 00:00:11,560 your head, memorizing it, that worth is only going down. 4 00:00:11,640 --> 00:00:15,640 Don't be a fool with a tool. That is one of the core pieces 5 00:00:15,640 --> 00:00:17,520 of advice I got started my career. 6 00:00:17,720 --> 00:00:19,800 If they understood the problem they wouldn't be hiring you, 7 00:00:19,800 --> 00:00:24,080 they'd have fixed it themselves. If I have skills and whatever to 8 00:00:24,080 --> 00:00:27,920 make a difference, but I choose not to, then what kind of man am 9 00:00:27,920 --> 00:00:30,520 I? Today I'm joined by Marian 10 00:00:30,520 --> 00:00:34,120 Marcus, AI lead over at Cap Gemini, and he spends his time 11 00:00:34,120 --> 00:00:36,280 in tech improving people's lives. 12 00:00:36,560 --> 00:00:39,920 We discuss how he does that, how he got there, and what skills 13 00:00:39,920 --> 00:00:41,880 you should focus on to get there as well. 14 00:00:42,280 --> 00:00:45,400 It's a conversation that's raw, authentic, with many hot takes 15 00:00:45,400 --> 00:00:47,720 and a little bit of controversy. So enjoy. 16 00:00:51,120 --> 00:00:53,400 Can you share a little bit more about some of the projects that 17 00:00:53,400 --> 00:00:54,680 you've been doing? Because I looked at your 18 00:00:54,680 --> 00:00:57,760 LinkedIn and one of your main passions is really leveraging 19 00:00:57,760 --> 00:01:00,520 data to better people's lives, and I'm curious what you've 20 00:01:00,520 --> 00:01:03,680 done. I started out in the scene as a 21 00:01:03,680 --> 00:01:08,320 statistician, which is not sexy statistics and and and data. 22 00:01:08,320 --> 00:01:11,120 Who who does that? And my dad for decades was for a 23 00:01:11,120 --> 00:01:12,920 decade was asking, hey, my dad, what? 24 00:01:12,920 --> 00:01:14,200 What, what the heck are you going to do? 25 00:01:14,200 --> 00:01:15,800 How are they going to pay you? How are you going to make a 26 00:01:15,800 --> 00:01:18,600 living? Then I did an internship. 27 00:01:18,720 --> 00:01:22,360 Everybody please do an internship at Pyro 5 for five. 28 00:01:22,360 --> 00:01:25,720 And I was modelling the spread of Ebola back then pandemic. 29 00:01:25,720 --> 00:01:28,720 And we were using, I was making the visuals, the maps. 30 00:01:28,720 --> 00:01:31,400 We were using that on the on the Dutch national news to raise 31 00:01:31,400 --> 00:01:33,800 funds. And we raised 10 dot 6 million 32 00:01:33,800 --> 00:01:36,840 euros to send emergency aid to West Africa back then. 33 00:01:36,840 --> 00:01:40,160 And I realized that the bullshit, pardon my French, but 34 00:01:40,160 --> 00:01:44,720 the the things I do with computers that was mocked at the 35 00:01:44,720 --> 00:01:47,720 time can actually help save people's lives. 36 00:01:48,800 --> 00:01:51,720 And that for me was the most empowering thing in my entire 37 00:01:51,720 --> 00:01:54,920 career, that hey, I can actually do something and I want to do 38 00:01:54,920 --> 00:01:59,640 this more. And then it happened that that's 39 00:01:59,640 --> 00:02:03,240 the statistics and data became data science, and then machine 40 00:02:03,240 --> 00:02:06,080 learning, and then suddenly everything is AI. 41 00:02:06,080 --> 00:02:08,680 And now, now I'm sexy. Yeah, you were doing AI before 42 00:02:08,680 --> 00:02:09,639 AI. Hell yeah. 43 00:02:10,120 --> 00:02:13,640 Back when it was just statistics and saying, let's be honest, 90% 44 00:02:13,640 --> 00:02:16,920 of all the AI in production is still just linear regression, 45 00:02:16,920 --> 00:02:21,720 baby. So I lucked out in that regard, 46 00:02:21,720 --> 00:02:25,520 that I learned something a few years before it became popular. 47 00:02:25,520 --> 00:02:27,320 And then they said, is that data science? 48 00:02:27,320 --> 00:02:29,520 And I said yes, 'cause I needed a damn job. 49 00:02:29,520 --> 00:02:32,080 Uh huh. So the trick is to know the 50 00:02:32,080 --> 00:02:36,240 trends coming up before the words for those trends even 51 00:02:36,240 --> 00:02:37,840 exist. Because by the time the word 52 00:02:37,840 --> 00:02:41,320 exists and it's all over over your social media feed, then 53 00:02:41,320 --> 00:02:44,600 you're already too late. So the trick is to convince 54 00:02:44,600 --> 00:02:48,200 people to invest in something before the buzzword arises. 55 00:02:48,840 --> 00:02:52,160 How consciously were you working on a skill that would later on 56 00:02:52,160 --> 00:02:55,880 be kind of marketable? Because it feels like there was 57 00:02:56,000 --> 00:02:58,880 partly luck, but also you put in the effort to put yourself in a 58 00:02:58,880 --> 00:03:00,480 position to be lucky in the first place. 59 00:03:00,760 --> 00:03:07,520 So especially in modern careers, we have this overfit on doing 60 00:03:07,520 --> 00:03:10,920 what everybody else is doing. So, oh, everybody learned 61 00:03:10,920 --> 00:03:12,640 Python. Everybody learned Python. 62 00:03:12,760 --> 00:03:14,920 Back in my day, it was RI still miss R. 63 00:03:15,240 --> 00:03:16,800 You need to know this package. You need to have that 64 00:03:16,800 --> 00:03:20,360 certification, but you will never out compete the the 65 00:03:20,360 --> 00:03:23,920 competition by doing the same thing as the competition. 66 00:03:24,600 --> 00:03:29,120 You will win by diversifying. I'm a social scientist, 67 00:03:29,120 --> 00:03:31,440 sociology, criminology, statistics. 68 00:03:31,920 --> 00:03:35,880 I will never be able to to out code a computer scientist. 69 00:03:36,160 --> 00:03:38,560 Heck no. But that's not what I focus on. 70 00:03:38,560 --> 00:03:43,760 But when I have projects where we were modelling HR data stuff 71 00:03:43,920 --> 00:03:47,640 with some computer scientists and somebody comes in and ask, 72 00:03:47,640 --> 00:03:50,760 hey, Are you sure we're we're not discriminating. 73 00:03:50,760 --> 00:03:52,480 Our algorithm isn't discriminating people. 74 00:03:52,480 --> 00:03:54,960 The computer scientists panicked and I said hell yeah. 75 00:03:54,960 --> 00:03:57,080 I studied complaining about discrimination. 76 00:03:57,800 --> 00:04:01,520 So the point here is to have your niche and your unique 77 00:04:01,520 --> 00:04:05,680 combination of skills that make you valuable in the labour 78 00:04:05,680 --> 00:04:07,280 market. And that in my case is a 79 00:04:07,280 --> 00:04:12,640 combination of, well, no, knowing a lot of knowing a lot 80 00:04:12,640 --> 00:04:18,000 of theory behind sociology, psychology, criminology, and 81 00:04:18,000 --> 00:04:21,920 applying that in a field where we're where we're modelling 82 00:04:21,920 --> 00:04:24,560 human behaviour at large, at large scale. 83 00:04:25,280 --> 00:04:31,400 I was doing stuff for the flow of refugees across borders years 84 00:04:31,400 --> 00:04:35,080 ago, and everybody was focusing on how can we get data on how 85 00:04:35,080 --> 00:04:38,840 many people cross. And I started talking about 86 00:04:39,080 --> 00:04:42,000 hamburger prices. They're like hamburger prices. 87 00:04:42,000 --> 00:04:44,680 And I'm like, yeah, 'cause I read this paper, Google Scholar 88 00:04:44,840 --> 00:04:48,720 back then from Norway, showing that hamburger prices, the Big 89 00:04:48,720 --> 00:04:52,680 Mac index is the one of the best indicators of people moving 90 00:04:52,680 --> 00:04:53,840 across borders. Why? 91 00:04:54,120 --> 00:04:57,640 Because the price of a Big Mac is the best proxy you have for 92 00:04:57,840 --> 00:05:01,040 economic prosperity in a country. 93 00:05:01,920 --> 00:05:04,120 Think about this. Most of the developing world, or 94 00:05:04,120 --> 00:05:07,360 at least parts, have governments that might not be fully 95 00:05:07,360 --> 00:05:09,080 objective in the numbers they publish. 96 00:05:09,320 --> 00:05:12,640 So you start using the price of the of the Big Mac instead 97 00:05:12,800 --> 00:05:16,600 'cause that's the the closest you can get to an actual real 98 00:05:16,600 --> 00:05:18,720 food price. And you see the price of the Big 99 00:05:18,720 --> 00:05:23,960 Mac go up from 2016 seventeen. That means prosperity is going 100 00:05:23,960 --> 00:05:26,600 down because food prices go up. That's your basic benchmark. 101 00:05:26,760 --> 00:05:30,120 Meaning the year after that, you will likely see more people from 102 00:05:30,120 --> 00:05:33,480 that country come to a western country because the economy 103 00:05:33,480 --> 00:05:36,880 there sucks. Outlier here is of course India, 104 00:05:36,880 --> 00:05:38,160 because they do not have a Big Mac. 105 00:05:38,160 --> 00:05:39,960 They have a big Maharaja with chicken. 106 00:05:40,520 --> 00:05:43,560 This, it almost sounds made-up, to be honest. 107 00:05:44,160 --> 00:05:46,200 It's incredible. Yeah, the Big Mac. 108 00:05:46,200 --> 00:05:48,120 Index is real. Yeah, I love that. 109 00:05:48,120 --> 00:05:51,440 I like that a lot earlier you were speaking about people 110 00:05:51,440 --> 00:05:53,880 finding their niche, not just following what other people do, 111 00:05:54,200 --> 00:05:56,560 but actually figuring out what their unique selling points is 112 00:05:56,560 --> 00:06:00,560 or what a combination of skills are that makes them unique in 113 00:06:00,840 --> 00:06:02,440 they can actually leverage in their career. 114 00:06:02,840 --> 00:06:04,960 I feel like I don't have a traditional computer science 115 00:06:04,960 --> 00:06:07,160 background. I kind of started in operations. 116 00:06:07,440 --> 00:06:10,480 I got rolled into software engineering, super excited about 117 00:06:10,480 --> 00:06:13,840 it, but I also had to look at my peers and be like, I don't have 118 00:06:13,840 --> 00:06:16,160 the same drive to go deep as they can. 119 00:06:16,320 --> 00:06:18,720 I would have to put in a lot more meters to get to that level 120 00:06:18,720 --> 00:06:20,960 and it it doesn't fulfill me to the same degree. 121 00:06:21,600 --> 00:06:23,920 I like communicating with people, like the people aspect a 122 00:06:23,920 --> 00:06:25,320 lot. I like certain outcomes. 123 00:06:25,320 --> 00:06:27,680 I'm very interested in the business side, in the product 124 00:06:27,680 --> 00:06:29,400 side. I'm still trying to figure out 125 00:06:29,400 --> 00:06:31,560 what my niche is going to be and I'm on that path. 126 00:06:32,080 --> 00:06:34,120 But what would your advice be? Because I feel like you 127 00:06:34,120 --> 00:06:37,040 definitely fit in a lot of things together and what you're 128 00:06:37,040 --> 00:06:38,640 doing nowadays sounds super excited. 129 00:06:38,640 --> 00:06:41,080 What would your advice be for people to find their niche? 130 00:06:42,280 --> 00:06:44,880 So finding your niche is not linear. 131 00:06:45,160 --> 00:06:49,480 I love assumptions of linearity in my models, but they're often 132 00:06:49,480 --> 00:06:52,360 false. The only way you can figure this 133 00:06:52,360 --> 00:06:54,880 out is by trying lots of different things. 134 00:06:54,880 --> 00:06:57,960 Imagine you're you just graduated. 135 00:06:58,560 --> 00:07:01,320 How are you going to figure out the one thing you want to do in 136 00:07:01,320 --> 00:07:05,200 life but by doing one thing for 2-3 years straight? 137 00:07:05,800 --> 00:07:09,360 Or will you go like, Oh, well, this was nice, this role, now I 138 00:07:09,360 --> 00:07:11,280 want to try something else. This is why I like the 139 00:07:11,280 --> 00:07:16,440 consulting field that you and I both work in, because you will 140 00:07:16,440 --> 00:07:20,040 get to do many more different things in a shorter time frame. 141 00:07:20,280 --> 00:07:24,200 Different, different projects, different roles, technical 142 00:07:24,400 --> 00:07:28,800 business, large companies, small company, very governmenty, very 143 00:07:28,800 --> 00:07:31,040 retail, very banking. And that's the only way you're 144 00:07:31,040 --> 00:07:35,480 going to figure out what you're good at, what you like to do, 2 145 00:07:35,480 --> 00:07:38,600 very different things, and where the coffee doesn't suck. 146 00:07:40,240 --> 00:07:44,880 I spent my first I, I've spent most of my career in consulting 147 00:07:45,200 --> 00:07:49,160 and I think I developed way faster because of that steeper 148 00:07:49,160 --> 00:07:53,920 learning curve, more hair loss. In my first year at Gap Gemini, 149 00:07:53,920 --> 00:07:56,840 for example, I, I was at 5 different clients, but because 150 00:07:56,840 --> 00:08:00,160 of that, I had much more of a learning curve was collecting 151 00:08:00,160 --> 00:08:03,680 data way faster than had I just worked for one of those clients 152 00:08:03,920 --> 00:08:07,160 directly for a year. And that way I could figure out 153 00:08:07,160 --> 00:08:11,200 what I'm actually good at and what I enjoy doing and where are 154 00:08:11,200 --> 00:08:15,200 the coffee? Oh man, that I'm public sector I 155 00:08:15,200 --> 00:08:17,920 love because I want to help people but the coffee is just 156 00:08:17,920 --> 00:08:19,560 terrible. I'm just I'm just saying, just 157 00:08:19,560 --> 00:08:21,720 saying. But you will only collect this 158 00:08:21,720 --> 00:08:26,160 data to get a better verdict by trying different things. 159 00:08:26,360 --> 00:08:31,480 So anyone still studying out there, do a minors in something 160 00:08:31,480 --> 00:08:34,280 completely different, do a master's in something different, 161 00:08:34,280 --> 00:08:37,159 do internships at completely different places. 162 00:08:37,360 --> 00:08:40,799 Not because you will pursue a career in every one of those, 163 00:08:40,799 --> 00:08:43,400 but to figure out also what you do not like. 164 00:08:43,799 --> 00:08:45,520 Because when else are you going to figure that out? 165 00:08:45,520 --> 00:08:49,560 When you're 4050, when you have a mortgage in kids, OK, most of 166 00:08:49,560 --> 00:08:51,160 you won't even get a mortgage at that age. 167 00:08:51,160 --> 00:08:54,160 But you, you get what you get. What I mean, you have to try 168 00:08:54,160 --> 00:08:55,720 different things and collect data. 169 00:08:55,720 --> 00:08:58,360 And the biggest problem I see both for individuals and for 170 00:08:58,360 --> 00:09:02,520 organizations today is that we're stuck in analysis 171 00:09:02,520 --> 00:09:04,640 paralysis, that we don't know what to do. 172 00:09:04,640 --> 00:09:08,240 So we just do nothing while the rest of the world keeps 173 00:09:08,240 --> 00:09:12,640 advancing and we're like, oh, but we could do A or B. 174 00:09:13,840 --> 00:09:17,840 To quote Slavo Shisek, sometimes doing nothing is the most 175 00:09:17,840 --> 00:09:25,760 violent thing to do. And we we ponder and we worry 176 00:09:25,760 --> 00:09:30,680 about the price of action, but we forget about the price of our 177 00:09:30,720 --> 00:09:33,600 inaction, both at the individual and at the collective level. 178 00:09:33,840 --> 00:09:36,160 Can you share a little bit more about that? 179 00:09:36,160 --> 00:09:38,880 When you see organizations or where you see teams or maybe 180 00:09:38,880 --> 00:09:42,640 even individuals that are stuck in analysis paralysis mode, too 181 00:09:42,640 --> 00:09:44,320 many options. I don't know what to do. 182 00:09:44,320 --> 00:09:48,000 I have trouble starting. How do you handle that or how do 183 00:09:48,000 --> 00:09:50,960 you help them? I'm going to use an old example 184 00:09:50,960 --> 00:09:54,680 here, Thesis. The the recommendation I give to 185 00:09:54,680 --> 00:09:58,320 all my interns and when they do a thesis, like a master's thesis 186 00:09:58,320 --> 00:10:00,520 or a bachelor or whatever, you get to spend half a year to a 187 00:10:00,520 --> 00:10:03,920 year of your life on it. You make sure the that you enjoy 188 00:10:03,920 --> 00:10:06,840 the topic because you're going to spend half a year to a year 189 00:10:07,000 --> 00:10:11,640 of your life on it. And if you enjoy that topic. 190 00:10:11,640 --> 00:10:15,240 I was doing analysis on crime back then for the Dutch police 191 00:10:15,400 --> 00:10:19,400 in my own hometown, Rotterdam, which was hella complicated. 192 00:10:19,400 --> 00:10:23,000 The data was horrifying, but I liked the topic. 193 00:10:23,000 --> 00:10:26,080 So I could keep pursuing that because I had that motivation. 194 00:10:26,560 --> 00:10:29,680 Doesn't mean you need to enjoy every aspect of it, of the 195 00:10:29,680 --> 00:10:33,720 methodology of the screaming at your computer, But at a basic 196 00:10:33,720 --> 00:10:36,400 level you need to have some interest in the topic, because 197 00:10:36,400 --> 00:10:38,040 that's the only way you can keep it up. 198 00:10:38,200 --> 00:10:42,440 Yeah, I like that a lot. Previously, very early when I 199 00:10:42,440 --> 00:10:45,720 joined consultancy, it was a question of OK, what do you want 200 00:10:45,720 --> 00:10:47,080 to work at? What are the skills you want to 201 00:10:47,080 --> 00:10:48,920 develop? But also, do you have any 202 00:10:48,920 --> 00:10:51,360 industries that you want to work in or not want to work in? 203 00:10:51,480 --> 00:10:55,000 And back then, early in career, me was like industries, like I 204 00:10:55,000 --> 00:10:56,800 had never given that a thought even. 205 00:10:57,280 --> 00:10:58,920 So I said, no, I said industry doesn't matter. 206 00:10:58,920 --> 00:11:00,200 I want to walk with very smart people. 207 00:11:00,200 --> 00:11:01,480 I want to grow as fast as possible. 208 00:11:01,480 --> 00:11:04,160 I want to have a lot of experiences and a lot of things 209 00:11:04,160 --> 00:11:05,840 so I can figure out what I like and what I don't like. 210 00:11:06,160 --> 00:11:09,000 And then I got into an industry and now it's on my resume. 211 00:11:09,000 --> 00:11:10,760 So you can probably figure out which one that is. 212 00:11:10,760 --> 00:11:12,240 And it's not something that I'm proud of. 213 00:11:12,480 --> 00:11:14,480 And I never even stood still at the time. 214 00:11:14,480 --> 00:11:17,640 And now I'm like, yeah, industry, it might matter. 215 00:11:17,640 --> 00:11:20,520 I'm now in healthcare and I see people come in, they say, well, 216 00:11:20,520 --> 00:11:23,600 I've taken a pay cut and things are definitely slower in this 217 00:11:23,600 --> 00:11:24,680 industry. It's in healthcare. 218 00:11:25,280 --> 00:11:27,520 But they like the goal, they like the purpose, they like the 219 00:11:27,520 --> 00:11:29,720 fact that whatever they're building, that it might help 220 00:11:29,800 --> 00:11:33,320 society, it might help people, it might help in the end, the 221 00:11:33,320 --> 00:11:35,960 Netherlands as a whole, which is very motivating. 222 00:11:36,120 --> 00:11:38,960 And I love the current experience that I have because I 223 00:11:38,960 --> 00:11:41,280 get to see that behind the scenes, I had a lot of 224 00:11:41,280 --> 00:11:43,720 assumptions in working in the public sector or working in a 225 00:11:43,720 --> 00:11:46,440 lot of industries that are good for people, good for your 226 00:11:46,440 --> 00:11:49,480 country or good for humanity. Might be very slow, might be 227 00:11:49,480 --> 00:11:52,720 very boring, might not have many opportunities for me to thrive 228 00:11:52,720 --> 00:11:54,800 in. And I don't know if that's true. 229 00:11:54,800 --> 00:11:56,720 I feel like it's still true, but it's definitely something that's 230 00:11:56,720 --> 00:11:59,240 been withholding me for from actually going there. 231 00:11:59,640 --> 00:12:02,840 I might go there maybe in a in a few more decades, But that's 232 00:12:02,840 --> 00:12:04,720 also a shame because I feel like that industry meets a lot of 233 00:12:04,720 --> 00:12:07,680 people and also people that are early in career, people of all 234 00:12:07,680 --> 00:12:09,520 kinds. So I'm curious from your 235 00:12:09,520 --> 00:12:12,760 perspective, what makes it that's super exciting within 236 00:12:12,760 --> 00:12:15,040 those industries or what has been the case for you? 237 00:12:15,840 --> 00:12:19,360 So career choices are not black, white, XY01. 238 00:12:19,520 --> 00:12:24,720 There's this time factor that we often forget because in the 239 00:12:24,720 --> 00:12:28,240 short term, and I've worked for for banks, for example, in, in 240 00:12:28,240 --> 00:12:31,960 the past, in the short term, yeah, I was not doing much for 241 00:12:31,960 --> 00:12:34,560 for society or for the world, but the stuff I was building 242 00:12:34,560 --> 00:12:37,160 there and learning over there, because banks are hey, way 243 00:12:37,160 --> 00:12:39,680 bigger budgets compared to some other sectors. 244 00:12:39,960 --> 00:12:43,600 The stuff I was learning there. Years later, I started applying 245 00:12:43,600 --> 00:12:47,160 in completely different fields like forecasting the yield for 246 00:12:47,160 --> 00:12:51,880 smallholder farmers in Kenya and India and and help it helping 247 00:12:52,160 --> 00:12:55,640 the worldwide organizations that help feed millions of people. 248 00:12:55,800 --> 00:12:59,120 And I would not be able to help them at that point in time, if 249 00:12:59,120 --> 00:13:03,280 not a few years before I was crunching similar numbers and 250 00:13:03,280 --> 00:13:06,160 models for for banking and for retail. 251 00:13:06,280 --> 00:13:11,000 So here's the short versus long term, even if you're not not 252 00:13:11,000 --> 00:13:13,240 directly helping people in the short term. 253 00:13:13,440 --> 00:13:16,200 Heck, I could have joined a soup kitchen instead of instead of 254 00:13:16,200 --> 00:13:20,120 studying, but I think I can contribute more to the world 255 00:13:20,360 --> 00:13:23,880 with the skill set that I've developed since rather than just 256 00:13:23,880 --> 00:13:27,040 just doing soup all day back then. 257 00:13:27,520 --> 00:13:30,320 And yeah, that that's my answer to your question. 258 00:13:30,320 --> 00:13:32,480 Time. The time dimension is often 259 00:13:32,480 --> 00:13:35,400 forgotten there. I can only do what I do now to 260 00:13:35,400 --> 00:13:39,680 help people because of the stuff I built in the past decade. 261 00:13:39,960 --> 00:13:44,280 Heck, remember, remember Andrew and G? 262 00:13:44,280 --> 00:13:50,000 We were talking about the before the 1st computer vision models. 263 00:13:50,000 --> 00:13:54,240 This is a dog, This is a cat. Those were developed in 2012, 264 00:13:54,240 --> 00:13:59,120 2013, the Atlas NVIDIA GPU, back when GPUs were still affordable. 265 00:14:00,560 --> 00:14:03,080 Those were the first models that could say, hey, this picture, 266 00:14:03,080 --> 00:14:05,960 this bundle of pixels is a car. This is a whatever. 267 00:14:06,160 --> 00:14:09,440 Nowadays, the same tech is in everybody's phone, in 268 00:14:09,440 --> 00:14:13,440 everybody's self driving car and in every autonomous killer AI 269 00:14:13,440 --> 00:14:18,800 drone that's out there out there flying to an alive people. 270 00:14:19,040 --> 00:14:24,760 Tech progresses so fast and we collectively as as a field can 271 00:14:24,760 --> 00:14:30,360 only help out with that because we did our homework and five to 272 00:14:30,360 --> 00:14:34,440 10 years ago, we're fitting our own neural Nets on, I don't 273 00:14:34,440 --> 00:14:38,480 know, local GPUs with, with your laptop that you had to leave, 274 00:14:38,600 --> 00:14:42,320 keep on running overnight because tech progresses so fast. 275 00:14:42,320 --> 00:14:45,400 And sometimes the best investment you can make is just 276 00:14:45,400 --> 00:14:49,360 to build something, home projects, hobby projects, just 277 00:14:49,360 --> 00:14:53,120 to keep up and have a little bit more hands on experience than 278 00:14:53,120 --> 00:14:56,720 the competitor who only read a paper about it or watched a 279 00:14:56,720 --> 00:14:59,480 YouTube video or asked Jack GPT about it about it. 280 00:14:59,760 --> 00:15:03,920 Because all diplomas and certifications, I'm sorry 281 00:15:03,920 --> 00:15:06,760 people, all diplomas and certifications are compensation 282 00:15:06,760 --> 00:15:11,040 for lack of experience. And this market mainly focuses 283 00:15:11,040 --> 00:15:16,440 on experience and you lack that. You get something that a proxy 284 00:15:16,440 --> 00:15:18,600 and you go get a proxy for experience. 285 00:15:18,600 --> 00:15:22,720 And that is why we focus so much on internship stuff you built in 286 00:15:22,720 --> 00:15:26,160 your spare time, stuff you have on git, anything you can, you 287 00:15:26,160 --> 00:15:28,320 can point at. And some of my biggest 288 00:15:28,320 --> 00:15:33,560 achievements in my, my entire career are actually things that 289 00:15:33,560 --> 00:15:35,560 I can just point out that you can Google. 290 00:15:35,560 --> 00:15:38,080 And I'm like, Hey, that's my got my name on it on that. 291 00:15:38,080 --> 00:15:40,240 And that's my next point of career advice for people. 292 00:15:40,520 --> 00:15:45,760 Focus on external stuff rather than internal stuff within your 293 00:15:45,920 --> 00:15:51,240 organization, because internal certifications or titles or blah 294 00:15:51,240 --> 00:15:54,760 blah of the of the quarter or God knows what you cannot take 295 00:15:54,760 --> 00:15:58,160 with you throughout your career. But if it's external, like 296 00:15:58,240 --> 00:16:00,960 standing on the stage, I'm a three time world world summer 297 00:16:00,960 --> 00:16:02,560 day. I speaker for example, I really 298 00:16:02,560 --> 00:16:07,480 focused on that because that is external and I can I will keep 299 00:16:07,480 --> 00:16:13,360 that with with me wherever I go. The last few guests have all 300 00:16:13,480 --> 00:16:15,000 kind of mentioned the same thing. 301 00:16:15,280 --> 00:16:18,760 We're moving from an experience based economy to a skills based 302 00:16:18,760 --> 00:16:20,480 economy. So the fact that you have a 303 00:16:20,480 --> 00:16:23,000 paper might not distinguish you anymore from someone that can 304 00:16:23,000 --> 00:16:26,400 show and point to some of the skills that they have and that 305 00:16:26,400 --> 00:16:29,080 they've built something with. It used to be more valuable, but 306 00:16:29,080 --> 00:16:31,880 right now we just have too many people with too many 307 00:16:31,880 --> 00:16:34,000 certification. So it evaluates. 308 00:16:34,000 --> 00:16:37,680 This is basic economics one O 1. Yeah, most of the people and 309 00:16:37,680 --> 00:16:39,120 most of the audience that's listening. 310 00:16:39,400 --> 00:16:41,520 I did a poll. What is your experience and how 311 00:16:41,520 --> 00:16:42,760 long have you been working in tech? 312 00:16:42,880 --> 00:16:46,040 And it actually varies, but the biggest slice of people have 313 00:16:46,040 --> 00:16:47,960 worked in tech for shorter than five years. 314 00:16:48,320 --> 00:16:50,520 I'm assuming mainly they are software engineers. 315 00:16:50,520 --> 00:16:53,680 And if we are talking about the skills that are for the future, 316 00:16:53,720 --> 00:16:55,800 you're saying technology accelerates quite quickly. 317 00:16:56,080 --> 00:16:58,520 What are some of the skills that you think people should focus on 318 00:16:58,760 --> 00:17:00,960 specifically for their career but also for like peace of 319 00:17:00,960 --> 00:17:02,920 fulfillment and looking towards the future? 320 00:17:03,000 --> 00:17:05,560 The ability to learn first and foremost. 321 00:17:05,560 --> 00:17:08,119 None of the tools I work with I I learned. 322 00:17:08,760 --> 00:17:11,640 I learned SPSS people back in the day. 323 00:17:11,880 --> 00:17:14,800 Some some people in the audience will be cringing right now. 324 00:17:15,359 --> 00:17:20,359 SPSS taught myself R to scrape Twitter back before it became 325 00:17:20,359 --> 00:17:27,319 shitter to to model the 2013 fourteen Syrian refugee crisis. 326 00:17:27,319 --> 00:17:29,680 I hated Putin before I knew where Ukraine was on the map. 327 00:17:31,840 --> 00:17:36,480 But from our I went into Python And now Python is the end all be 328 00:17:36,480 --> 00:17:38,320 all. But I'm from an era where people 329 00:17:38,320 --> 00:17:40,840 were still asking like, Are you sure? 330 00:17:40,840 --> 00:17:44,000 Should you be learning SAS? Because SAS is better in the 331 00:17:44,000 --> 00:17:47,240 market right now. So God knows what we'll be using 332 00:17:47,240 --> 00:17:52,000 in 10 in 10 years, unless all the all the LLMS start 333 00:17:52,000 --> 00:17:55,840 overfitting on on just languages that are already in use. 334 00:17:55,960 --> 00:17:58,560 We might actually see a stagnation in the development in 335 00:17:58,560 --> 00:18:02,760 the development of coding languages now, since Claude and 336 00:18:02,760 --> 00:18:06,240 Gemini will not be able to adopt to new languages as much as 337 00:18:06,240 --> 00:18:09,920 produce more Java horrors. Meaning we will be stuck with 338 00:18:09,920 --> 00:18:13,480 Java until the death of the universe, I'm pretty sure. 339 00:18:14,480 --> 00:18:17,800 But what I'm trying to show here is that those tools, those 340 00:18:17,800 --> 00:18:21,520 techniques, cloud platforms, they change so fast that you 341 00:18:21,520 --> 00:18:23,560 can't focus on learning a technique. 342 00:18:23,880 --> 00:18:26,920 You've got to learn a method to learn new stuff. 343 00:18:26,920 --> 00:18:30,640 That is the most single most important skill in the entire 344 00:18:30,640 --> 00:18:35,920 tech field to be able to learn not even master, but just work 345 00:18:35,920 --> 00:18:39,680 with new tech Anytime I have a new project, oh God, it's a new 346 00:18:39,680 --> 00:18:41,440 cloud platform I haven't worked with. 347 00:18:41,760 --> 00:18:45,400 OK, well AWS, Google, Azure, I'm pretty sure it has most of the 348 00:18:45,400 --> 00:18:48,440 same functionalities. They they just call every damn 349 00:18:48,440 --> 00:18:51,440 button differently and that's how they vendor lock. 350 00:18:51,440 --> 00:18:54,360 So, so you can't go go to the other cloud platform because the 351 00:18:54,360 --> 00:18:58,720 names is different, which is less of a hurdle than you would 352 00:18:58,720 --> 00:19:01,120 think. And then there's of course 353 00:19:01,120 --> 00:19:04,080 nuance, like the billing models are different, the scaling is 354 00:19:04,080 --> 00:19:07,240 different. The, the, the type of servers 355 00:19:07,240 --> 00:19:11,120 you can run it in is different. You cannot local host, I mean 356 00:19:11,200 --> 00:19:17,640 sovereign this this version. Remember that people sovereignty 357 00:19:17,640 --> 00:19:19,800 is the new local host. You're just running your own 358 00:19:19,800 --> 00:19:22,720 damn model on your own damn computer. 359 00:19:22,720 --> 00:19:25,720 But you don't call it local host you you call it sovereign. 360 00:19:25,720 --> 00:19:27,680 The client will love it, thank me later. 361 00:19:28,840 --> 00:19:32,200 So it's and that's the other thing. 362 00:19:32,400 --> 00:19:37,280 And I mentioned that before understanding the words because 363 00:19:37,280 --> 00:19:40,160 buzzwords will come, but they usually mean something that 364 00:19:40,160 --> 00:19:44,400 wasn't new at all. Again, sovereignty is just local 365 00:19:44,400 --> 00:19:47,840 host but new. Everybody screams as as was 366 00:19:47,840 --> 00:19:51,200 screaming AI, but we mainly meant just deep learning on 367 00:19:51,200 --> 00:19:53,680 steroids with a lot a lot of GPU and compute. 368 00:19:53,960 --> 00:19:57,480 Big data was just data that you store on multiple different 369 00:19:57,480 --> 00:19:59,760 servers because you don't have a single server that can contain 370 00:19:59,760 --> 00:20:02,240 it all. But nowadays my laptop has more 371 00:20:02,240 --> 00:20:05,920 storage than a whole complex of servers a decade ago. 372 00:20:05,920 --> 00:20:07,400 Meaning. Big data is dead. 373 00:20:07,400 --> 00:20:11,160 Long live Long live big data. The core problem here is the 374 00:20:11,160 --> 00:20:14,600 core problem of humanity, which is that sometimes we say the 375 00:20:14,600 --> 00:20:18,840 same words but we mean different things, and sometimes we say 376 00:20:18,840 --> 00:20:21,240 different words but we mean the same thing. 377 00:20:21,240 --> 00:20:24,160 And all of miscommunication stems from this. 378 00:20:24,480 --> 00:20:29,160 And in our field, it is important to figure out not just 379 00:20:29,560 --> 00:20:32,640 what words people are using, which is what sales heavily 380 00:20:32,640 --> 00:20:36,680 focuses on, but what their actual problem is and what they 381 00:20:36,680 --> 00:20:39,280 actually mean, which can be, which can be completely 382 00:20:39,280 --> 00:20:41,200 different from the words they're actually using. 383 00:20:41,440 --> 00:20:45,320 Because I've had clients that that ask for AI and it can mean 384 00:20:45,360 --> 00:20:49,320 a random tree, a random forest, It can mean you have some kind 385 00:20:49,320 --> 00:20:52,960 of image recognition model. It can mean some type of speech 386 00:20:52,960 --> 00:20:56,480 synthesis. It can mean producing synthetic 387 00:20:56,480 --> 00:21:00,640 data of bananas using 3D models. It can mean so many different 388 00:21:00,640 --> 00:21:03,080 things, but they all use the same word and it's our job to 389 00:21:03,080 --> 00:21:06,160 figure out what the heck they want and then to build that. 390 00:21:06,440 --> 00:21:10,480 Not just to receive orders and build something, because that 391 00:21:10,480 --> 00:21:15,440 job will be outsourced and automated to hell him back by 392 00:21:15,440 --> 00:21:22,080 whatever cloud copilot thing they they buy and or consume. 393 00:21:22,320 --> 00:21:26,400 But the core skill to figure out what the fluff they actually 394 00:21:26,400 --> 00:21:29,160 want and need and what is the problem behind the problem. 395 00:21:29,960 --> 00:21:32,120 Because if they understood the problem, they wouldn't be hiring 396 00:21:32,120 --> 00:21:33,560 you, they'd have fixed it themselves. 397 00:21:33,760 --> 00:21:36,200 But they do not understand the problem, do not understand the 398 00:21:36,200 --> 00:21:38,320 solution. It's our job to figure out that 399 00:21:38,320 --> 00:21:41,200 solution, then do that. And to do that we need to be 400 00:21:41,200 --> 00:21:43,720 able to learn new tech and learn new stuff. 401 00:21:43,960 --> 00:21:48,160 And that's why learning is key. I want to zoom in on that 402 00:21:48,160 --> 00:21:51,320 because I, I'm a big believer that compounding knowledge is 403 00:21:51,320 --> 00:21:53,120 one of the most valuable things you can have. 404 00:21:53,440 --> 00:21:56,520 And you can only compound knowledge if you bit by bit make 405 00:21:56,560 --> 00:21:59,400 ownership or take ownership of certain sources of information. 406 00:21:59,400 --> 00:22:02,680 Really familiarize yourself with it to a certain foundational 407 00:22:02,680 --> 00:22:04,360 level. And I'm assuming that a lot of 408 00:22:04,360 --> 00:22:05,680 listeners are going to hear you speak. 409 00:22:05,680 --> 00:22:09,560 They're going to say he seems very well educated, 410 00:22:09,560 --> 00:22:12,240 communicated, all of that and much, much more. 411 00:22:12,240 --> 00:22:15,360 So they are seeing you from a level of where you are now, but 412 00:22:15,360 --> 00:22:17,360 they haven't seen this journey of how you got here. 413 00:22:18,040 --> 00:22:20,520 How do you actually get here, or how do you take ownership and 414 00:22:20,520 --> 00:22:22,640 get good at learning in the first place? 415 00:22:23,560 --> 00:22:26,200 How do you get good people? That's that's the question. 416 00:22:26,680 --> 00:22:29,960 Well, all the gamers out there, how do you get good? 417 00:22:29,960 --> 00:22:32,320 You practice. Look, I'm not talking about 418 00:22:32,320 --> 00:22:35,040 video games right now because I get slapped and clapped by 13 419 00:22:35,040 --> 00:22:37,560 year olds on online games nowadays because they got way 420 00:22:37,560 --> 00:22:42,040 better reaction speeds than me at the at this at this age. 421 00:22:42,040 --> 00:22:47,040 But again, practice and you shouldn't obsess over the short 422 00:22:47,040 --> 00:22:51,520 term and that you can't succeed all at once because learning all 423 00:22:51,520 --> 00:22:53,880 of this learning is not a Sprint. 424 00:22:53,880 --> 00:22:56,560 It's a marathon. Unfortunately, we mainly teach 425 00:22:56,560 --> 00:23:00,080 people, kids to Sprint in education because it's to your 426 00:23:00,080 --> 00:23:04,040 next exam, to your next test and and then you go drink beer 427 00:23:04,040 --> 00:23:08,080 again. But it's about acquiring 428 00:23:08,080 --> 00:23:11,960 knowledge over time. I I always felt stupid back when 429 00:23:11,960 --> 00:23:15,440 I was a student because again, my teachers and my fellow 430 00:23:15,440 --> 00:23:18,920 students always knew more than me about this one topic. 431 00:23:19,280 --> 00:23:24,040 And then in my in my actual job, I have to explain how burnouts 432 00:23:24,040 --> 00:23:27,680 work and what's, what's HR data forecasts that because there's a 433 00:23:27,680 --> 00:23:29,840 lot of stuff you can do to actually see that coming. 434 00:23:31,120 --> 00:23:33,000 Yeah, I read a bunch of papers about that. 435 00:23:33,000 --> 00:23:37,480 I, I wrote some, I built some models and I was just explaining 436 00:23:37,480 --> 00:23:40,800 that to an audience. And I only realized how much I 437 00:23:40,800 --> 00:23:44,040 knew about this once. I had to explain it to an 438 00:23:44,040 --> 00:23:48,000 audience who did not know, which was very special to me because 439 00:23:48,000 --> 00:23:50,840 that was the point in time that I realized that, hey, I actually 440 00:23:50,840 --> 00:23:54,880 know some stuff. And I warmly wish for the for 441 00:23:54,880 --> 00:23:58,160 the audience and the crowd and all the listeners to experience 442 00:23:58,160 --> 00:24:01,200 that, that that feeling of what's the word? 443 00:24:01,200 --> 00:24:05,960 Not justification, but just realizing your value, that you 444 00:24:05,960 --> 00:24:09,160 have value, that your skills and your knowledge have value, that 445 00:24:09,160 --> 00:24:12,920 you can be an expert on something beyond, beyond 446 00:24:12,920 --> 00:24:18,840 slapping people in Dota. And I love that. 447 00:24:18,840 --> 00:24:20,120 I don't want, I want more of that. 448 00:24:20,160 --> 00:24:22,640 Yeah. And that keeps me going to do 449 00:24:22,640 --> 00:24:26,840 that again over time, OK, then IoT becomes a thing. 450 00:24:26,840 --> 00:24:29,640 I'm going to get a Raspberry Pi. I'm going to get all the all the 451 00:24:29,640 --> 00:24:31,160 sensors. I'm going to stick that on 452 00:24:31,160 --> 00:24:33,360 everything in my house, including my cat. 453 00:24:33,520 --> 00:24:36,400 And before I know it, I know something about it. 454 00:24:36,400 --> 00:24:39,760 I I can talk about it, OK. And we're going to, we're going 455 00:24:39,760 --> 00:24:42,640 to, we're doing, we're doing time series models on things for 456 00:24:42,640 --> 00:24:47,040 banks, on things for retail. And, and then we need to do that 457 00:24:47,320 --> 00:24:49,640 when lives are at stake. And I'm like, hey, I've done 458 00:24:49,640 --> 00:24:53,080 this before. I can do this and I can help 459 00:24:53,480 --> 00:24:54,520 people. Yeah. 460 00:24:55,040 --> 00:24:59,760 And that is just wild because that's what I want to do, do 461 00:24:59,760 --> 00:25:02,480 data to help people. Because apparently doing data 462 00:25:02,480 --> 00:25:04,520 stuff and talking about it is what I'm good at. 463 00:25:05,600 --> 00:25:09,800 And I just want people to find the things they're good at and 464 00:25:09,800 --> 00:25:15,000 to then apply that to help people because we have these, we 465 00:25:15,000 --> 00:25:19,160 have the IT field here in the Netherlands and we talk about 466 00:25:19,160 --> 00:25:22,160 innovation. And usually that's counting the 467 00:25:22,160 --> 00:25:26,520 amount of AIPOCS you've done. We call that AI maturity. 468 00:25:27,160 --> 00:25:29,200 That was last year. This year we count the amount of 469 00:25:29,200 --> 00:25:32,640 Co piloted licenses you have in use and we call that AI 470 00:25:32,640 --> 00:25:34,760 maturity. But most of the actual 471 00:25:34,760 --> 00:25:38,680 innovation does not come from banking or retail or or 472 00:25:38,680 --> 00:25:40,400 governments. Most of the actual innovation 473 00:25:40,400 --> 00:25:43,920 comes from the three fields we do not want to talk about, which 474 00:25:43,920 --> 00:25:47,600 is crime, warfare and the adult entertainment industry. 475 00:25:48,800 --> 00:25:53,240 But like seriously, most of cybersecurity bit Bitcoin, 476 00:25:53,240 --> 00:26:00,360 streaming technology, 3D, HDVR and deepfakes all stem from 477 00:26:00,360 --> 00:26:02,640 California's most famous industry. 478 00:26:03,360 --> 00:26:07,480 And so if I want to keep up with new stuff, I'm more focusing on 479 00:26:07,480 --> 00:26:11,960 the fields where rules and laws matter less because that's where 480 00:26:11,960 --> 00:26:14,120 this stuff is invented and scaled. 481 00:26:14,520 --> 00:26:17,720 And then that is transferred to the to the fields where we 482 00:26:17,720 --> 00:26:22,280 actually work in. And then lastly, there's the pro 483 00:26:22,280 --> 00:26:24,960 bono, there's the NGOs and there's the large governments 484 00:26:25,480 --> 00:26:29,280 that make the most of a difference to people in need, 485 00:26:29,280 --> 00:26:32,000 people who are hungry, people who are in depth, people who we 486 00:26:32,000 --> 00:26:34,880 want to help. But to help them with those 487 00:26:34,880 --> 00:26:41,040 skills, we first need to acquire those skills in other sectors. 488 00:26:41,040 --> 00:26:43,320 And that is usually, again, the banking, the retail, the 489 00:26:43,320 --> 00:26:47,720 whatever. And we glean those, those skills 490 00:26:47,720 --> 00:26:50,160 and those techniques from the sectors that are most ahead. 491 00:26:50,640 --> 00:26:52,560 But unfortunately, then I'm going to talk about the 492 00:26:52,560 --> 00:26:55,880 Rotterdam drug traffic industry. Cause hey, we have the biggest 493 00:26:55,880 --> 00:26:58,520 port in Europe. We have thus the most cocaine 494 00:26:58,520 --> 00:27:02,640 traffic in Europe. Hey, represent Rotterdam and and 495 00:27:02,640 --> 00:27:07,440 or Modern Warfare because yeah, everything we see on drones 496 00:27:07,440 --> 00:27:13,480 nowadays, everything we thing we see on 3D printing and ARAI 497 00:27:13,520 --> 00:27:16,320 autonomous warfare, that's all happening in Ukraine. 498 00:27:16,800 --> 00:27:19,240 That's not happening in the Netherlands. 499 00:27:19,240 --> 00:27:22,840 That's not happening in America as much as they they try to 500 00:27:22,840 --> 00:27:25,760 claim otherwise. Just this morning I saw the 501 00:27:25,760 --> 00:27:28,400 American military brag about their new drones they're 502 00:27:28,400 --> 00:27:33,680 acquiring for like 3330 KA piece even though Ukraine can build 503 00:27:33,680 --> 00:27:35,360 the same drones for less than 1000. 504 00:27:35,560 --> 00:27:42,040 But in warfare scale matters and down the line we will see pizza 505 00:27:42,040 --> 00:27:44,680 delivery drones or medevac drones. 506 00:27:44,680 --> 00:27:48,640 So ground based drones with wheels that can evacuate a 507 00:27:48,640 --> 00:27:50,760 person. That's healthcare too. 508 00:27:51,400 --> 00:27:54,680 Those are all being developed and tested live at the front 509 00:27:54,680 --> 00:27:58,960 line to the extent that any any serious drone company that does 510 00:27:58,960 --> 00:28:03,040 not supply their stuff to Ukraine is not a serious drone 511 00:28:03,040 --> 00:28:06,040 company because they haven't tested and proven their stuff. 512 00:28:07,440 --> 00:28:09,920 To digress back to point to the point I was trying to make, 513 00:28:09,920 --> 00:28:12,440 there is fields where rules do not matter and that's where all 514 00:28:12,440 --> 00:28:15,560 the new stuff comes from. And This is why I professionally 515 00:28:15,560 --> 00:28:18,440 watch Borne either. That's we why, why I we as 516 00:28:18,640 --> 00:28:22,280 professionals should at least look at these fields because 517 00:28:22,280 --> 00:28:24,040 that's where the new stuff is coming from. 518 00:28:24,040 --> 00:28:27,960 And that will trickle down into our banking, retail, whatever 519 00:28:27,960 --> 00:28:30,320 sectors. And down the line we use those 520 00:28:30,320 --> 00:28:34,000 skills that we acquire here to do good. 521 00:28:34,000 --> 00:28:38,040 And that is the eternal cycle. And we all have to find some 522 00:28:38,040 --> 00:28:44,560 kind of balance for that as as techies, as I tears as humans, 523 00:28:44,920 --> 00:28:49,920 so that we can learn and develop ourselves and also be able to 524 00:28:49,920 --> 00:28:52,920 give back. Because it pains me to some 525 00:28:52,920 --> 00:28:56,640 extent that the newest generative AI stuff that we 526 00:28:56,640 --> 00:29:01,960 develop as humans is mainly used to generate porn and not to do 527 00:29:01,960 --> 00:29:04,240 good. Yeah, but there's a lot of that 528 00:29:04,240 --> 00:29:05,360 porn. There's more porn being 529 00:29:05,360 --> 00:29:08,920 generated nowadays than is actually being produced by 530 00:29:08,920 --> 00:29:10,280 humans. Think about that. 531 00:29:10,280 --> 00:29:12,560 And it's and it's really good. Oh, OK. 532 00:29:13,840 --> 00:29:16,280 Think about this. I, I've known that certain 533 00:29:16,280 --> 00:29:20,280 industries are way ahead in technology and I don't know how 534 00:29:20,280 --> 00:29:21,880 I've known that. Maybe it's just anchored and I 535 00:29:21,880 --> 00:29:24,080 listen to a lot of podcasts, but it was always in there. 536 00:29:24,640 --> 00:29:26,960 And then you can kind of see what will trickle down into 537 00:29:26,960 --> 00:29:29,280 retail and banking and all that sort of stuff. 538 00:29:29,760 --> 00:29:32,520 But there was something that you mentioned where you are now 539 00:29:32,840 --> 00:29:36,160 equipped and also well equipped throughout history with regards 540 00:29:36,160 --> 00:29:38,640 to the skills that you've obtained to now help people's 541 00:29:38,640 --> 00:29:41,360 lives. Doesn't it frighten you 542 00:29:41,400 --> 00:29:44,040 sometimes the position you are or what you're working on, that 543 00:29:44,040 --> 00:29:46,320 when you make a mistake or where there's a few mistakes that are 544 00:29:46,320 --> 00:29:49,160 compounding that, it can also negatively impact people's 545 00:29:49,160 --> 00:29:50,800 lives. Oh hell yeah. 546 00:29:50,800 --> 00:29:53,000 But that means you can make a difference. 547 00:29:53,000 --> 00:29:57,480 If you are afraid of responsibility or the 548 00:29:57,480 --> 00:30:00,240 consequences of responsibility, then you should go sit in the 549 00:30:00,240 --> 00:30:04,120 corner and do nothing. A year ago, well, well over a 550 00:30:04,120 --> 00:30:06,040 year ago, 'cause this is about taking heat. 551 00:30:06,440 --> 00:30:09,600 A friend of mine on Discord tried to contact me saying, 552 00:30:09,600 --> 00:30:12,920 like, hey, Marine, your account's getting getting 553 00:30:13,160 --> 00:30:14,600 banned. Something happens, you need to 554 00:30:14,600 --> 00:30:17,360 report to the Discord help desk. And I'm like, I don't want to 555 00:30:17,360 --> 00:30:18,760 lose my memes. Yeah. 556 00:30:18,960 --> 00:30:22,000 So I contact Natalie Rosser to the Discord help desk. 557 00:30:22,240 --> 00:30:25,320 Please don't ban me. It's like, OK, can you identify 558 00:30:25,320 --> 00:30:28,400 yourself? I'm like, identify myself. 559 00:30:28,400 --> 00:30:30,240 Yeah, you place a birth, blah, blah, blah. 560 00:30:30,240 --> 00:30:33,000 And like, Discord should already know this stuff, right? 561 00:30:33,000 --> 00:30:36,080 I just pull it up, right? I started to notice that Natalie 562 00:30:36,080 --> 00:30:39,160 Rosser is not actually Google, Google a bull. 563 00:30:39,880 --> 00:30:43,160 And that her answers were always after exactly 20 seconds, 564 00:30:43,160 --> 00:30:45,760 regardless of the answer being 2 words or 20. 565 00:30:45,800 --> 00:30:47,880 I saw a pattern. I was talking to a bot. 566 00:30:47,880 --> 00:30:50,480 Yeah, this was a reason. Sorry Sir, No no, this was 567 00:30:51,000 --> 00:30:54,600 almost a year ago before we started calling every goddamn 568 00:30:54,600 --> 00:30:57,680 bot an agent. I was just talking to a bot 569 00:30:57,920 --> 00:31:03,480 impersonating a human but also impersonating my friend because 570 00:31:03,480 --> 00:31:09,400 my friend got hacked and it will and LLM was told to talk like 571 00:31:09,400 --> 00:31:11,800 him. Terrible spelling Unicorn 572 00:31:11,800 --> 00:31:13,680 smileys. No, no caps. 573 00:31:16,560 --> 00:31:19,600 And had I identified myself to Natalie Rosser, they would have 574 00:31:19,600 --> 00:31:23,960 taken my account, talked to Discord saying like I am Mariah 575 00:31:23,960 --> 00:31:26,520 Marcus, this is my personally identifiable information. 576 00:31:26,520 --> 00:31:30,240 They would have tried taken my account and tried the same trick 577 00:31:30,520 --> 00:31:35,040 with everybody in my list. Now I have no clue who was 578 00:31:35,040 --> 00:31:37,680 trying to do this. Again, this was before we 579 00:31:37,680 --> 00:31:41,200 started calling everything an agent, but I was already talking 580 00:31:41,200 --> 00:31:44,200 about agents. Basically because you do not 581 00:31:44,200 --> 00:31:47,720 need a Nigerian Prince on the other line or somebody 582 00:31:48,320 --> 00:31:51,360 pretending to be a Nigerian Prince, you can just use an LLM. 583 00:31:51,640 --> 00:31:56,800 Foreign agents of the future will be code and prompts and we 584 00:31:56,800 --> 00:32:00,400 are not even ready to start considering the implications of 585 00:32:00,400 --> 00:32:02,480 this. Actually, we saw this last week 586 00:32:02,480 --> 00:32:06,040 with X Turn. Turns out that the biggest MAGA 587 00:32:06,040 --> 00:32:08,480 influencers in America, as well as some of the biggest 588 00:32:08,480 --> 00:32:12,520 nationalists in Poland and the Netherlands, were actually 589 00:32:12,520 --> 00:32:16,040 accounts based in Bangladesh and Nigeria and stuff. 590 00:32:16,480 --> 00:32:20,400 Again, this is basic data analytics, but it's either 591 00:32:20,400 --> 00:32:23,720 humans who are just profiting off of outrage or not even 592 00:32:23,720 --> 00:32:27,000 humans, just bots and agents and whatever. 593 00:32:27,200 --> 00:32:31,040 And again, we, we pretend that agents are anything new, but 594 00:32:31,040 --> 00:32:34,400 they're all over our social media for years already because 595 00:32:34,400 --> 00:32:37,800 the criminal sector and the hybrid war sector have adopted 596 00:32:37,800 --> 00:32:44,040 these technologies long before our our banking and our retail 597 00:32:44,040 --> 00:32:46,240 and our shops will adapt to this. 598 00:32:46,240 --> 00:32:50,840 And This is why I obsess over disinformation about bots trying 599 00:32:50,840 --> 00:32:52,680 to hack me. Because this comes back to your 600 00:32:52,680 --> 00:32:56,080 question about heat. Yes, I should be scared about 601 00:32:56,080 --> 00:32:59,920 taking heat, but that is the price I pay for making a 602 00:32:59,920 --> 00:33:02,000 difference. Then hell yeah, I had to 603 00:33:02,000 --> 00:33:05,480 calculate. I was involved in calculating 604 00:33:05,480 --> 00:33:10,280 the, well, how to feed people who suffer from malnutrition. 605 00:33:10,800 --> 00:33:13,880 That means you're calculating how many people you can save. 606 00:33:15,880 --> 00:33:18,040 It also means calculating how many you cannot. 607 00:33:18,720 --> 00:33:21,280 But that is where I can make the biggest difference with my 608 00:33:21,280 --> 00:33:24,920 skills in ways that I never could if I were applying that 609 00:33:24,920 --> 00:33:28,080 for just marketing purposes to sell candy. 610 00:33:29,040 --> 00:33:32,200 And that's where I want to be. That's where I can make a 611 00:33:32,200 --> 00:33:36,720 difference, and if I have skills and whatever to make a 612 00:33:36,720 --> 00:33:41,040 difference, but I choose not to, then what kind of man am I? 613 00:33:42,040 --> 00:33:45,560 For people listening, let's say it's very inspirational what 614 00:33:45,560 --> 00:33:48,160 you're saying and I want to progress my career in a way 615 00:33:48,160 --> 00:33:50,880 where I can also do that. How do people get there? 616 00:33:50,960 --> 00:33:53,760 Are there any actionable steps? You already mentioned building 617 00:33:53,760 --> 00:33:57,280 up the skills first is actually your career is non linear. 618 00:33:57,400 --> 00:33:59,640 So whether you do it now or whether you do it later, it can 619 00:33:59,640 --> 00:34:02,720 still come up on your path if that's really what you want. 620 00:34:03,120 --> 00:34:05,560 But how can people actively work towards getting themselves into 621 00:34:05,560 --> 00:34:08,440 a position to actually with their career and what they do on 622 00:34:08,440 --> 00:34:11,159 the day today, help whoever they want to. 623 00:34:11,159 --> 00:34:13,040 Careers are not linear. Yeah. 624 00:34:13,040 --> 00:34:15,360 They're also not even branching paths. 625 00:34:15,440 --> 00:34:19,080 Careers are roguelikes. OK, like the video game where 626 00:34:19,080 --> 00:34:22,159 you can branching paths, but you go back to the start at some 627 00:34:22,159 --> 00:34:26,800 point because then suddenly data analytics is dead, but data 628 00:34:26,800 --> 00:34:29,719 science becomes a thing. Now everybody is an AI engineer. 629 00:34:29,719 --> 00:34:32,480 And I don't know what Pokémon names we'll be using next 630 00:34:32,800 --> 00:34:36,080 because the kids I educate these days never played Pokémon, so my 631 00:34:36,080 --> 00:34:40,120 Pokémon jokes fall flat. But the important point here is 632 00:34:40,440 --> 00:34:44,120 that we keep reinventing the titles and the career paths 633 00:34:44,360 --> 00:34:46,920 faster than we can walk those paths. 634 00:34:46,920 --> 00:34:54,800 So it's more important to focus on on skills without expiration 635 00:34:54,800 --> 00:34:59,200 dates because you can master some type type of coding or some 636 00:34:59,200 --> 00:35:02,800 kind of language like English or French or Japanese. 637 00:35:03,080 --> 00:35:07,400 But it's more important that you learn how to learn new languages 638 00:35:07,400 --> 00:35:09,880 because the more languages you speak, the the more easily 639 00:35:09,880 --> 00:35:12,480 you'll pick up the next. The same goes for coding. 640 00:35:12,480 --> 00:35:17,960 I've seen people go all in on on their learning COBOL way back in 641 00:35:17,960 --> 00:35:19,840 the day. But if that's the only thing 642 00:35:19,840 --> 00:35:21,920 they they know, then they're stuck. 643 00:35:22,200 --> 00:35:25,400 If you're go all in on learning one cloud platform, I know 644 00:35:25,400 --> 00:35:30,600 everything about Azure, then the company you works on my the 645 00:35:30,600 --> 00:35:33,040 company you work for migrates from Azure. 646 00:35:33,160 --> 00:35:35,280 You're stuck going all in on SAP. 647 00:35:35,280 --> 00:35:40,560 Never go all in on one thing that has an expiration date. 648 00:35:40,920 --> 00:35:44,280 Focus on skills that do not have expiration. 649 00:35:44,280 --> 00:35:47,680 That we see this so much nowadays with the with the Gen. 650 00:35:47,680 --> 00:35:53,800 AI hype about learning this prompt or that even though those 651 00:35:53,800 --> 00:35:58,760 prompts, even though those prompt tips will be useless two 652 00:35:58,760 --> 00:36:01,600 months down the line when the model gets updated and it all 653 00:36:01,600 --> 00:36:04,520 falls flat and you have to and they no longer do the weird 654 00:36:04,520 --> 00:36:08,640 dashes, you know? So there the skill is not to 655 00:36:08,640 --> 00:36:12,840 know on top of your head all these prompting skills, but to 656 00:36:13,120 --> 00:36:17,640 learn to be able to reverse engineer how the prompting 657 00:36:17,640 --> 00:36:20,480 works. That is a far more valuable 658 00:36:20,480 --> 00:36:22,480 skill. In the same time, in the same 659 00:36:22,480 --> 00:36:28,680 way that writing good speeches, anyone can generate a good 660 00:36:28,680 --> 00:36:33,000 speech, but how to deliver that, how to talk and how to connect 661 00:36:33,000 --> 00:36:36,320 with your audience and all that, that is still a far more 662 00:36:36,320 --> 00:36:39,240 immaterial skill. And a lot of things we are 663 00:36:39,240 --> 00:36:42,680 automating right now, in the same way that steam, the steam 664 00:36:42,680 --> 00:36:46,520 engine and radio and electricity and the Internet helped us 665 00:36:46,520 --> 00:36:50,200 automate parts of tasks, but that only made the other parts 666 00:36:50,200 --> 00:36:53,200 of tasks more important. We have silly things like 667 00:36:53,200 --> 00:36:56,800 marketing nowadays that we never would have invented if not for 668 00:36:57,000 --> 00:36:59,680 for the past few industrial revolutions. 669 00:36:59,840 --> 00:37:01,400 Hell, we used to call it propaganda. 670 00:37:01,400 --> 00:37:04,520 Now we only call it propaganda when we don't like it, even 671 00:37:04,520 --> 00:37:07,520 though both books were written by the same Frankfurt schuler 672 00:37:07,720 --> 00:37:10,480 back in the day. I also did a minor in propaganda 673 00:37:10,480 --> 00:37:14,320 studies, which is why I obsess over this information nowadays. 674 00:37:14,320 --> 00:37:17,240 And we see that these models are actually now starting to 675 00:37:17,240 --> 00:37:21,320 regurgitate disinformation and propaganda from totalitarian 676 00:37:21,320 --> 00:37:24,160 states. Meaning it still becomes more 677 00:37:24,160 --> 00:37:29,040 important for us humans to have the skill to filter information 678 00:37:29,040 --> 00:37:33,280 and to give verdicts and to give judgements about pieces of of 679 00:37:33,280 --> 00:37:36,680 information rather than to just copy paste whatever the heck 680 00:37:36,680 --> 00:37:40,480 ChatGPT said. That goes both for your coding 681 00:37:40,840 --> 00:37:42,840 and for your business analysis skills. 682 00:37:44,040 --> 00:37:49,840 Drop Mike. I was on Twitter and I same as 683 00:37:49,840 --> 00:37:50,720 you. I'm so sorry for you. 684 00:37:51,080 --> 00:37:54,200 Same as you, I got reached out to several companies actually, 685 00:37:54,200 --> 00:37:57,360 and this was in the span of a few weeks all with regards to we 686 00:37:57,360 --> 00:37:59,960 love what you're doing on the Internet, podcasting, social 687 00:37:59,960 --> 00:38:02,560 presence, anything like that. We want to sponsor you. 688 00:38:04,080 --> 00:38:06,720 I thought, well, I've made it very cool companies, companies 689 00:38:06,720 --> 00:38:08,480 that I look up to, I used to gain. 690 00:38:08,480 --> 00:38:10,640 Razor was one of them. GoPro was one of them, Udemy, 691 00:38:10,640 --> 00:38:13,560 which I thought great fit. My sweet summer. 692 00:38:13,560 --> 00:38:14,840 Child. All of them. 693 00:38:14,920 --> 00:38:17,000 All of them scams. You thought you were talking to 694 00:38:17,000 --> 00:38:17,960 humans. Exactly. 695 00:38:18,000 --> 00:38:19,680 All of them scams. So that that was one. 696 00:38:19,680 --> 00:38:22,160 But this was also somewhere last year. 697 00:38:22,800 --> 00:38:26,320 And the obvious one was they sent me a signing link to one of 698 00:38:26,320 --> 00:38:28,480 their NDA's or a contract or whatever. 699 00:38:28,720 --> 00:38:30,960 And it was just a forefront, which is very well done to be 700 00:38:30,960 --> 00:38:33,480 honest compared to the actual the real thing. 701 00:38:33,800 --> 00:38:36,640 Now I made a post of that. I reported everyone, etcetera, 702 00:38:36,640 --> 00:38:38,960 etcetera. And now in the most recent one, 703 00:38:39,560 --> 00:38:41,080 it was a YouTube that reached that. 704 00:38:41,080 --> 00:38:42,920 We're going to do this YouTube collab, We're going to do a 705 00:38:42,920 --> 00:38:44,400 podcast. It's going to be for a good 706 00:38:44,400 --> 00:38:46,920 cause. All money will be donated and 707 00:38:46,920 --> 00:38:49,480 we'll pay you X. And I'm like, the money will be 708 00:38:49,480 --> 00:38:51,480 donated. And then how do I get, how does 709 00:38:51,480 --> 00:38:53,520 this work? And then I'm in a group chat 710 00:38:53,520 --> 00:38:56,440 with everyone from the US, none of them in Europe, and they all 711 00:38:56,440 --> 00:38:59,120 say when they're going to be online, but everyone uses CET 712 00:38:59,160 --> 00:39:01,480 and I'm like, that doesn't make any sense. 713 00:39:01,480 --> 00:39:04,320 So why would you translate that specifically to me, who I'm the 714 00:39:04,320 --> 00:39:06,760 only one there? And then the same thing signing 715 00:39:06,760 --> 00:39:10,880 link or and like I report everyone, these people have real 716 00:39:10,880 --> 00:39:13,120 Twitter, well, real Twitter accounts, but they also have 717 00:39:13,360 --> 00:39:15,560 links to their YouTube channel. And that same YouTube channel 718 00:39:15,560 --> 00:39:18,320 who has millions of subscribers also links back to the same 719 00:39:18,320 --> 00:39:20,800 Twitter account. So I haven't figured out exactly 720 00:39:20,800 --> 00:39:24,320 what's at play. But in essence, this frightens 721 00:39:24,320 --> 00:39:25,760 me. Like, I'm excited for a lot of 722 00:39:25,760 --> 00:39:28,480 things in the future, but the fact that there is such 723 00:39:28,480 --> 00:39:31,960 manipulation of information and misinformation, that is the part 724 00:39:31,960 --> 00:39:34,920 that frightens me. This is the information age, not 725 00:39:34,920 --> 00:39:39,480 the knowledge age, and the skill to distinguish truth true from 726 00:39:39,480 --> 00:39:42,640 false will only become more important in the future, 727 00:39:42,800 --> 00:39:47,480 especially as no single point of information can be trusted 728 00:39:47,480 --> 00:39:49,480 anymore. We can create fakes that look 729 00:39:49,480 --> 00:39:54,160 more real than real than real things by now, but our ability 730 00:39:54,720 --> 00:39:59,280 must be to triangulate between different points of data and to 731 00:39:59,280 --> 00:40:01,600 then go like. Yeah, well, actually this 732 00:40:01,600 --> 00:40:04,600 doesn't make any sense given from what I previously know, 733 00:40:04,880 --> 00:40:07,840 that will only become more important, even though back in 734 00:40:07,840 --> 00:40:12,840 the day it was viewed as not important because anybody can 735 00:40:12,840 --> 00:40:16,360 take your photo and make you make a video of you doing very, 736 00:40:16,360 --> 00:40:20,960 very weird things. I read that last month over 15% 737 00:40:20,960 --> 00:40:23,720 of all American high schoolers have seen deepfakes of 738 00:40:23,720 --> 00:40:27,480 classmates doing very inappropriate things, and I was 739 00:40:27,480 --> 00:40:29,520 surprised because I thought that number would be higher. 740 00:40:30,320 --> 00:40:32,160 Really. Yeah, man. 741 00:40:32,600 --> 00:40:36,200 See, that's the part where and mainly bring it back to anything 742 00:40:36,200 --> 00:40:39,560 more from a career perspective, people are saying very early on, 743 00:40:39,560 --> 00:40:42,880 topics are getting automated now by agents, which means that the 744 00:40:42,880 --> 00:40:45,360 people that are educating themselves and that usually 745 00:40:45,360 --> 00:40:48,160 that's a step for them to build up the skills to do very, very 746 00:40:48,160 --> 00:40:52,400 cool things that help society. That step is now changing and 747 00:40:52,400 --> 00:40:54,880 people are asking, well, what do I do then to make myself not 748 00:40:54,880 --> 00:40:56,640 obsolete? And we're talking about anything 749 00:40:56,640 --> 00:40:58,800 with regards to communication and soft skills and learning. 750 00:40:59,040 --> 00:41:01,560 I agree with all of that. And then there's still a tooling 751 00:41:01,560 --> 00:41:02,960 aspect. Like I don't know what I should 752 00:41:02,960 --> 00:41:07,360 advise people with regards to, OK, how do you go from junior to 753 00:41:07,360 --> 00:41:08,720 actually in the end becoming senior? 754 00:41:08,720 --> 00:41:11,160 Because it's such great area that I don't know yet. 755 00:41:11,200 --> 00:41:13,200 I feel like we're really figuring things out. 756 00:41:13,480 --> 00:41:17,600 And the only advice that I have is be as open minded as you can. 757 00:41:17,600 --> 00:41:20,280 Try and familiarize yourself with a lot of things. 758 00:41:20,280 --> 00:41:22,480 Try and build as much foundational knowledge as you 759 00:41:22,480 --> 00:41:25,560 can because tooling does change. I'm wondering what your 760 00:41:25,560 --> 00:41:29,000 perspective is on that. Don't be a fool with a tool. 761 00:41:30,240 --> 00:41:33,640 That is one of the core pieces of advice I got. 762 00:41:33,640 --> 00:41:36,560 I started my career so over 90 years ago. 763 00:41:36,560 --> 00:41:39,440 God, don't be a fool with a tool. 764 00:41:40,760 --> 00:41:44,920 Back then we refer to only knowing our only knowing Python 765 00:41:44,920 --> 00:41:48,120 or only back then. Do you remember RPA people? 766 00:41:48,120 --> 00:41:51,640 Do you remember back when RPA was a was a thing? 767 00:41:52,000 --> 00:41:55,600 I was using a Python package back back in the day to to 768 00:41:55,600 --> 00:41:58,240 scrape Twitter. I forgot forgot the name 769 00:41:58,280 --> 00:42:01,000 Selenium. Yeah, yeah, I was learning 770 00:42:01,000 --> 00:42:04,800 Selenium because that was one of the first Python based browser. 771 00:42:04,800 --> 00:42:07,160 This automate the browser and click down. 772 00:42:07,920 --> 00:42:13,480 Back then those were bots, but I knew people who doubled down on 773 00:42:13,480 --> 00:42:15,920 only learning Selenium and all the insurance and outs. 774 00:42:16,240 --> 00:42:18,200 But Selenium was open source and it was great. 775 00:42:18,200 --> 00:42:20,960 You could do everything with it. But then the blue prisms came 776 00:42:20,960 --> 00:42:26,520 in, which was basically the the company or the company 777 00:42:26,520 --> 00:42:29,480 organization equivalent of it. But it was reliable because you 778 00:42:29,480 --> 00:42:31,280 could sue somebody if it didn't work. 779 00:42:31,680 --> 00:42:34,280 And suddenly the whole Selenium market disappeared because 780 00:42:34,280 --> 00:42:38,200 people went for for larger infrastructure tools instead. 781 00:42:38,200 --> 00:42:41,600 And the guy who knew everything about Selenium was stuck and the 782 00:42:41,600 --> 00:42:44,360 people who learned Blue Prism were laughing at him. 783 00:42:44,520 --> 00:42:47,720 Nowadays we laugh at the people who only know Blue Prism because 784 00:42:47,720 --> 00:42:50,880 the cycle continues. It's not about learning anyone 785 00:42:50,880 --> 00:42:54,720 tool, it's about being able to learn in general. 786 00:42:54,840 --> 00:42:58,000 And that involves learning the insurance and outs of tools. 787 00:42:58,000 --> 00:43:05,640 I think it we are switching from an explicit knowledge economy to 788 00:43:05,680 --> 00:43:09,360 an implicit knowledge economy and our education systems 789 00:43:09,360 --> 00:43:13,480 haven't adjusted to that, that the worth of knowing stuff in 790 00:43:13,480 --> 00:43:16,920 your head, memorizing it, that worth is only going down. 791 00:43:16,920 --> 00:43:22,000 But the ability to look up stuff that was introduced like what, 792 00:43:22,000 --> 00:43:25,400 25 years ago when Google hit the market, before we had that, we 793 00:43:25,400 --> 00:43:32,080 had start start that has been introduced during our lifetimes. 794 00:43:32,080 --> 00:43:35,720 But the value of that, of finding stuff in large amounts 795 00:43:35,720 --> 00:43:40,480 of information, that is only going up, especially in the age 796 00:43:40,480 --> 00:43:43,880 of disinformation, because we are drowning in information. 797 00:43:44,200 --> 00:43:49,440 But truth and wisdom has only increased invaluable and has 798 00:43:49,800 --> 00:43:53,760 become relatively more scarce because now there's more bots 799 00:43:53,760 --> 00:43:58,200 and more fake stuff flooding the Internet than actual real stuff. 800 00:43:59,120 --> 00:44:03,280 And the only way we can still distinguish truth from fake, 801 00:44:03,440 --> 00:44:06,760 because I see so many deep fakes now of Zelensky, of Putin, of 802 00:44:06,760 --> 00:44:10,600 F-30 fives that crashed in Iran. And we could only tell they were 803 00:44:10,600 --> 00:44:13,720 fake because there was still a a human in the cockpit, because 804 00:44:13,720 --> 00:44:16,880 they forgot to prompt it remove the human from the cockpit. 805 00:44:17,200 --> 00:44:20,640 We can only tell truth from fake because of human error. 806 00:44:20,640 --> 00:44:23,880 And that too will end. So what we need to learn, again, 807 00:44:23,880 --> 00:44:27,360 because learning is important, is to triangulate information to 808 00:44:27,360 --> 00:44:30,600 figure out what's real. And This is why I I'm frustrated 809 00:44:30,600 --> 00:44:34,120 with modern media and journalism that thinks that just 810 00:44:34,120 --> 00:44:38,240 regurgitating and repeating what no liar say is somehow 811 00:44:38,240 --> 00:44:40,840 journalism. I forgot the quote, but somebody 812 00:44:41,200 --> 00:44:44,440 once wrote this. If one source tells you it's 813 00:44:44,440 --> 00:44:47,040 raining and the other says there's sunshine, your job as a 814 00:44:47,040 --> 00:44:50,880 journalist is not to report both, but to open a damn window. 815 00:44:52,840 --> 00:44:55,600 And that's what I'm missing. Yeah, I love that. 816 00:44:55,760 --> 00:44:59,320 One of the final thoughts that I had was when you say don't be a 817 00:44:59,320 --> 00:45:01,000 fool with a tool, I really resonate with that. 818 00:45:01,400 --> 00:45:03,760 But then I look out in the market, I get bombarded, 819 00:45:03,880 --> 00:45:07,160 especially nowadays with tools, with so many tools, which means. 820 00:45:07,160 --> 00:45:10,040 Both human tools and software. Tools like like anything. 821 00:45:10,080 --> 00:45:13,040 And that means that there's definitely a market for tools. 822 00:45:13,320 --> 00:45:16,120 Are you saying don't use tools, don't master tools? 823 00:45:16,120 --> 00:45:18,120 Because I also love mastering my tools. 824 00:45:18,600 --> 00:45:21,560 I take pride in being efficient with regards to whatever I do on 825 00:45:21,560 --> 00:45:24,880 the keyboard, but in the end it's still a tool. 826 00:45:25,200 --> 00:45:29,480 Creating tools is cheaper than ever and let's be honest, 90% of 827 00:45:29,480 --> 00:45:31,800 all Gen. AI tools and and Gen. 828 00:45:31,800 --> 00:45:37,040 AI startups are just some front and API call to to ChatGPT, you 829 00:45:37,040 --> 00:45:40,760 know so tools became cheaper and easier to make. 830 00:45:40,920 --> 00:45:43,760 So as a consequence, we have more tools flooding the market 831 00:45:43,760 --> 00:45:47,920 than ever and 99 out of the out of 100, they're all the same or 832 00:45:47,920 --> 00:45:51,080 they have similar functionality mean it's a harder than ever to 833 00:45:51,080 --> 00:45:54,880 stand out because again, they're just a bunch of generated 834 00:45:55,160 --> 00:45:58,320 things. And why the heck should I go for 835 00:45:58,320 --> 00:46:02,440 a generated for a generated third party tool when I can just 836 00:46:02,440 --> 00:46:08,080 vibe code my own tool and get to the 80% by myself where usually 837 00:46:08,080 --> 00:46:10,240 I would need months of developing time in an entire 838 00:46:10,240 --> 00:46:12,560 team? Sure, I will waste two or three 839 00:46:12,560 --> 00:46:16,440 times as much money on the last 20% to get it to get it home 840 00:46:16,440 --> 00:46:20,280 free and bug fix. And why can I upload 10 or 20 841 00:46:20,280 --> 00:46:24,400 pictures but not 15? But overall I'm still 842 00:46:24,400 --> 00:46:28,240 accelerating, meaning the value of tools then decreases even 843 00:46:28,240 --> 00:46:32,560 further because I creating my own is technically now also way 844 00:46:32,560 --> 00:46:35,280 cheaper. It's just most organizations 845 00:46:36,000 --> 00:46:43,120 only have people in in their HR in their HR database that no 846 00:46:43,120 --> 00:46:47,520 specific tools stand out. Rise above the tool. 847 00:46:47,640 --> 00:46:52,840 The client doesn't care that I am the toilet duck experts. 848 00:46:53,040 --> 00:46:56,400 They say ain't that I know everything about Azure, or that 849 00:46:56,400 --> 00:47:00,480 I know everything about Edo S or about Google or IBM, but that I 850 00:47:00,480 --> 00:47:04,160 can distinguish between them and tell them which which one sucks 851 00:47:04,160 --> 00:47:06,920 at what task and which one they should go for. 852 00:47:06,920 --> 00:47:10,800 And that's not some kind of Pokémon Dragon Ball Z, my dad 853 00:47:10,800 --> 00:47:13,480 can beat up your dad competition. 854 00:47:13,720 --> 00:47:16,960 But they all have their niches and they all have their own 855 00:47:17,400 --> 00:47:19,760 delicacies and this one is better than that. 856 00:47:19,760 --> 00:47:23,040 That one has better image recognition service and that one 857 00:47:23,040 --> 00:47:26,400 you can saw is that one you can local host. 858 00:47:26,400 --> 00:47:31,840 I mean has sovereign options and AWS has not until they do of 859 00:47:31,840 --> 00:47:35,240 course, in two months time, but that's how we keep up. 860 00:47:35,520 --> 00:47:39,120 Nobody will appreciate you for just knowing the tool, but for 861 00:47:39,120 --> 00:47:43,240 your ability to distinguish between them, especially over 862 00:47:43,240 --> 00:47:48,200 time as tools and methods and techniques and insights change. 863 00:47:48,200 --> 00:47:51,360 And that's why you should listen to podcasts like these. 864 00:47:52,760 --> 00:47:54,640 But I thank you so much for going on and sharing. 865 00:47:54,800 --> 00:47:56,520 This has been a real blast. Thank you. 866 00:47:56,840 --> 00:47:57,880 Awesome. We're going to round it off 867 00:47:57,880 --> 00:47:59,200 here. If you're still with us, let us 868 00:47:59,200 --> 00:48:01,240 know in the comments section what you thought of this episode 869 00:48:01,520 --> 00:48:02,480 and we'll see you in the next one.