1 00:00:00,040 --> 00:00:03,480 Hi everyone, My name is Patrick Akio and joining me today is CEO 2 00:00:03,480 --> 00:00:06,560 and Co founder of Deploy Marta Stock. 3 00:00:06,720 --> 00:00:09,240 We discussed a few things, some of which you might think are 4 00:00:09,240 --> 00:00:12,280 controversial. First one, bottom up initiatives 5 00:00:12,280 --> 00:00:17,320 are meaningless without a strong vision, A dot on the horizon top 6 00:00:17,320 --> 00:00:21,560 down from leadership. Another one, the AI Act that the 7 00:00:21,560 --> 00:00:25,000 European Union has put into place is actually beneficial to 8 00:00:25,000 --> 00:00:28,440 all parties involved. And lastly, Marta shares how 9 00:00:28,440 --> 00:00:32,520 he's building this AI hub here in the Netherlands, in Amsterdam 10 00:00:32,840 --> 00:00:34,080 and what the benefits are going to be. 11 00:00:34,320 --> 00:00:41,720 So enjoy what made you start this AI Hub or AI community in 12 00:00:41,720 --> 00:00:44,720 the first place. I didn't really, I wouldn't say 13 00:00:44,720 --> 00:00:46,360 start. I think those things start to 14 00:00:46,360 --> 00:00:49,400 exist. So not giving myself all the 15 00:00:49,400 --> 00:00:53,080 credits or so, but if you talk especially about the physical 16 00:00:53,080 --> 00:00:58,120 location, I promised them as I started together with my car 17 00:00:58,120 --> 00:01:01,000 furnace to other companies and I promised to myself and also to 18 00:01:01,000 --> 00:01:03,280 my car furnace. I'm not going to spend a lot of 19 00:01:03,280 --> 00:01:05,400 time on on this. If there's no demand for 20 00:01:05,400 --> 00:01:09,240 anything, if it is a hub or something else and there's no 21 00:01:09,240 --> 00:01:11,360 proper demand. If nobody really wants this, I'm 22 00:01:11,680 --> 00:01:13,960 not going to spend any time on this because I have plenty of 23 00:01:13,960 --> 00:01:15,880 other things to do. And that's already hard enough. 24 00:01:16,400 --> 00:01:21,840 But then we found that more and more people recognise that in 25 00:01:21,840 --> 00:01:24,640 the nodes, we don't achieve a lot when it comes to AI at the 26 00:01:24,640 --> 00:01:26,880 moment or data or infrastructure in general. 27 00:01:28,200 --> 00:01:30,880 So for me, that was the reason to, to, to start talking with a 28 00:01:30,880 --> 00:01:34,360 lot of founders, like what is the, what are the challenges 29 00:01:34,360 --> 00:01:37,840 you're facing at the moment? And in January, we decided to, 30 00:01:37,840 --> 00:01:41,200 to host a dinner to get up with a few other founders who were 31 00:01:41,200 --> 00:01:42,520 happy to spend some time on this. 32 00:01:44,200 --> 00:01:48,080 We found Techie Peer to also be part of it and that was clear 33 00:01:48,080 --> 00:01:52,320 that everyone is in the end doing a bit their own thing in 34 00:01:52,480 --> 00:01:58,440 Eint, over in Rotterdam, in Amsterdam, in Delft, in here in, 35 00:01:58,440 --> 00:01:59,920 in so many different places and others. 36 00:01:59,920 --> 00:02:02,080 Everyone is kind of doing their own thing and that doesn't 37 00:02:02,080 --> 00:02:06,440 really work for AI because AI successful AI companies are 38 00:02:06,440 --> 00:02:08,600 scars I think. And you really need the top 39 00:02:08,840 --> 00:02:13,640 0.001% of people that are both really smart and passionate 40 00:02:13,640 --> 00:02:16,920 about something related to AI and entrepreneurial enough. 41 00:02:17,280 --> 00:02:20,160 And if that's Canada across a pretty small country, then you 42 00:02:20,160 --> 00:02:22,920 won't really find each other. And I strongly believe that if 43 00:02:22,920 --> 00:02:26,720 you, if you create one place where we are all together, where 44 00:02:26,720 --> 00:02:29,720 we can help each other, where we can, where we can help each 45 00:02:29,720 --> 00:02:31,960 other finding investment, where we can maybe invest in each 46 00:02:31,960 --> 00:02:35,360 other a bit, where successful start-ups will see spin outs out 47 00:02:35,360 --> 00:02:38,440 of the company. They create that flywheel and 48 00:02:38,440 --> 00:02:40,880 with that flywheel, we'll create bigger and bigger companies. 49 00:02:41,120 --> 00:02:44,600 Without that flywheel, everyone keeps on doing their own small 50 00:02:44,600 --> 00:02:47,240 things. And yeah, that's not going to 51 00:02:47,240 --> 00:02:49,080 work. So the, the, the, the polar 52 00:02:49,080 --> 00:02:53,240 model in that sense is and, and, and and and, and, and, and 53 00:02:53,600 --> 00:02:59,000 trying to do this in a fair and, and, and, and distributed way. 54 00:02:59,000 --> 00:03:01,760 I think that's not going to work for AI that we keep on thinking 55 00:03:01,760 --> 00:03:03,560 too small. We keep on acting too small. 56 00:03:03,560 --> 00:03:04,760 So for me, that was the main reason. 57 00:03:05,760 --> 00:03:08,000 And then a lot of people got quite unjust about it. 58 00:03:08,120 --> 00:03:10,960 And that's good to see. So that that that that motivated 59 00:03:10,960 --> 00:03:14,800 me to spend more time on this to, to, to together with those 60 00:03:14,800 --> 00:03:18,880 other founders to yeah, start searching for location, to 61 00:03:18,880 --> 00:03:21,800 negotiations, find out that the real estate world is a 62 00:03:21,800 --> 00:03:23,480 completely different world than the AI world. 63 00:03:24,880 --> 00:03:26,120 I'm not a big fan of it. Yeah. 64 00:03:26,600 --> 00:03:30,080 So I'm not sure if I want to do it again, but but that was the 65 00:03:30,080 --> 00:03:35,360 starting point for us. And what helped a lot for me is 66 00:03:35,360 --> 00:03:40,040 that with one of the companies we founded, Engines, we're based 67 00:03:40,080 --> 00:03:42,640 in both in Berlin and we are based in the AI campus. 68 00:03:42,640 --> 00:03:46,160 So we actually see how it works to be with a lot of AI companies 69 00:03:46,160 --> 00:03:51,120 in the same space to do meet ups together to basically help each 70 00:03:51,120 --> 00:03:52,920 other out, become each other's customers. 71 00:03:52,920 --> 00:03:54,320 We actually see it working there. 72 00:03:54,320 --> 00:03:57,840 So for us, that was like a great blueprint what we can achieve 73 00:03:57,840 --> 00:04:00,080 here and on us as well, I think. Yeah, I like that a lot. 74 00:04:00,280 --> 00:04:03,600 I mean, I like to start of how your thinking works in the 1st 75 00:04:03,600 --> 00:04:05,000 place. If there's no demand then why 76 00:04:05,000 --> 00:04:07,800 spend any time on it at all? Like what's the point? 77 00:04:07,800 --> 00:04:09,680 What's the purpose? I learned that the hard. 78 00:04:11,160 --> 00:04:15,120 Way and in general, I think one of the things that I love being 79 00:04:15,120 --> 00:04:18,519 in tech as a software engineer is people share a lot with 80 00:04:18,519 --> 00:04:21,959 regards to knowledge more so on the implementation side more 81 00:04:21,959 --> 00:04:24,880 than anything else. Look, I there was this problem. 82 00:04:24,880 --> 00:04:26,800 This is the context this is how we solve it. 83 00:04:26,960 --> 00:04:30,600 They do that for free articles, blogs, podcasts, videos, 84 00:04:30,600 --> 00:04:34,400 conferences, meetups, everything like that and this is kind of in 85 00:04:34,400 --> 00:04:37,760 a same vein, but then for AI startups with regards to 86 00:04:38,000 --> 00:04:40,360 challenges, funding, what problems are we solving? 87 00:04:40,360 --> 00:04:42,920 Learning from each other, being each other's customers sounds 88 00:04:42,920 --> 00:04:44,920 really cool man. I think we need it. 89 00:04:44,960 --> 00:04:47,840 I think we, we have seen it before that it actually works 90 00:04:48,520 --> 00:04:51,440 not even on purpose. Like we didn't create per SE a 91 00:04:51,440 --> 00:04:54,360 fintech ecosystem in them. So there was not, there was not 92 00:04:54,360 --> 00:04:56,560 one person who said let's create an ecosystem here. 93 00:04:56,840 --> 00:04:59,040 But it happens. It's whenever there are a few 94 00:04:59,040 --> 00:05:02,560 successful companies, you've got spinners, you've got, you got 95 00:05:02,560 --> 00:05:03,880 more money in the ecosystem as well. 96 00:05:03,880 --> 00:05:07,480 I think that's what we, that's what we lack in and out now. 97 00:05:07,480 --> 00:05:10,680 But if you talk about cloud infrastructure, data and now AI 98 00:05:10,680 --> 00:05:12,120 as well, we, we kind of missed that. 99 00:05:12,120 --> 00:05:15,840 I can't name, I can't remember a lot of companies at least that 100 00:05:15,840 --> 00:05:19,080 started here that have a lot of spin outs in that, in that 101 00:05:19,080 --> 00:05:22,480 vertical and we see now the the results of that I think. 102 00:05:22,680 --> 00:05:25,200 Got you. And you said NL is behind with 103 00:05:25,200 --> 00:05:27,680 regards to that flywheel and that ecosystem. 104 00:05:27,920 --> 00:05:30,640 Are there any countries that you say, OK, those are actually 105 00:05:30,640 --> 00:05:33,560 quite far ahead, those are companies or countries to look 106 00:05:33,560 --> 00:05:35,120 up to with the rest of their ecosystems? 107 00:05:35,160 --> 00:05:38,360 Yeah, I think if still work ecosystems is more cities or 108 00:05:40,000 --> 00:05:42,560 yeah, basically cities or the the, the, the large area around 109 00:05:42,560 --> 00:05:45,240 it. Of course, Paris and Berlin are 110 00:05:45,240 --> 00:05:47,960 the most obvious one I think in Europe, maybe next to London. 111 00:05:48,360 --> 00:05:50,640 You do see other successful companies in other smaller 112 00:05:50,640 --> 00:05:52,480 cities sometimes. So it doesn't have to be a 113 00:05:52,480 --> 00:05:55,920 physical ecosystem I guess. But especially Paris is 114 00:05:55,920 --> 00:05:57,840 fascinating I think. And it's good to realise that it 115 00:05:57,840 --> 00:06:01,000 didn't start like five years ago, it started 25 years ago. 116 00:06:01,920 --> 00:06:06,280 Szechuan F was funded in the end by the founder of Scaleway and 117 00:06:06,280 --> 00:06:08,360 Scaleway exist for like 25 years, I believe. 118 00:06:10,480 --> 00:06:13,680 So you see that flywheel already started 2025 years ago with 119 00:06:13,680 --> 00:06:15,560 investments and successful companies in cloud 120 00:06:15,560 --> 00:06:17,720 infrastructure didn't start yesterday. 121 00:06:18,280 --> 00:06:20,720 And also don't believe like if we, if we're going to shout 122 00:06:20,720 --> 00:06:23,240 really loud that we should, should, should push now that 123 00:06:23,240 --> 00:06:24,760 doesn't work. There's a lot of groundwork. 124 00:06:24,760 --> 00:06:29,280 There's a lot of, it takes decades of them to really become 125 00:06:29,480 --> 00:06:32,040 very good in, in a certain industry as a country that that 126 00:06:32,040 --> 00:06:36,120 doesn't come overnight. And that's why I, I strongly 127 00:06:36,120 --> 00:06:38,520 believe in that the ecosystem, if we, if we get things 128 00:06:38,520 --> 00:06:41,440 together, if we start building more successful companies, that 129 00:06:41,440 --> 00:06:43,200 will lead to more successful companies that will lead to 130 00:06:43,200 --> 00:06:45,320 another wave. So you need a flywheel to create 131 00:06:45,320 --> 00:06:48,520 new waves and become better in a certain industry, in this case 132 00:06:48,520 --> 00:06:51,360 AII. Feel like you need a very strong 133 00:06:51,360 --> 00:06:55,720 vision on a long term horizon to kind of realise what you need to 134 00:06:55,720 --> 00:06:58,840 do now and plant seeds with people, talk to people and 135 00:06:58,840 --> 00:07:01,680 really start creating this. Is that usually how you think? 136 00:07:01,680 --> 00:07:04,200 Like more in the bigger picture, long term vision. 137 00:07:04,760 --> 00:07:07,720 I'm also very good to to worry about things yesterday. 138 00:07:08,160 --> 00:07:09,680 The day-to-day, yeah. Definitely. 139 00:07:10,640 --> 00:07:14,760 And I do think you do need to realize that it takes long. 140 00:07:14,880 --> 00:07:17,320 I don't think it needs to be one person with a vision or so. 141 00:07:17,320 --> 00:07:20,960 I think especially these kind of things can be organically 142 00:07:20,960 --> 00:07:25,640 multiple people working towards something fakely in like 5 or 10 143 00:07:25,640 --> 00:07:28,600 years, and then new people will come and they work for another 144 00:07:28,600 --> 00:07:30,320 10 years or something in the same direction. 145 00:07:30,640 --> 00:07:33,240 I think in that way you create a strong ecosystem. 146 00:07:33,400 --> 00:07:37,440 Yeah, yeah, you get that. What's your take on kind of how 147 00:07:37,440 --> 00:07:39,760 Netherlands and even EU is handling AI? 148 00:07:39,760 --> 00:07:41,960 Let's start with the A i.e. U Act. 149 00:07:42,280 --> 00:07:45,560 EUAI Act. It's a difficult 1, isn't it? 150 00:07:47,760 --> 00:07:50,000 Yeah, I'm probably one of the fewer engineers who is actually 151 00:07:50,000 --> 00:07:54,640 quite a favour of the AI Act. I've seen so many organisations 152 00:07:55,000 --> 00:08:00,280 struggle to get beyond POCMPP, innovation, PowerPoint, blah, 153 00:08:00,280 --> 00:08:03,520 blah. And the reason for that often is 154 00:08:03,520 --> 00:08:05,960 that someone in the organization at some point is going to say, 155 00:08:06,120 --> 00:08:08,360 yeah, that looks great, but you're not going to get get this 156 00:08:08,360 --> 00:08:10,520 direction. And it's often because something 157 00:08:10,520 --> 00:08:13,040 is lacking, which is actually written down the AI act, it's 158 00:08:13,040 --> 00:08:17,560 either oversight, monitoring, it's, it's, it's, it's either 159 00:08:17,560 --> 00:08:20,280 way too risky to get things beyond POC. 160 00:08:20,760 --> 00:08:23,760 And I think what we, what we define now in the AIA is quite a 161 00:08:23,760 --> 00:08:28,960 good set of rules and standards which are going to lead to, to, 162 00:08:29,080 --> 00:08:34,960 to proper safeguards and help us to get AI beyond POC, to get AI 163 00:08:34,960 --> 00:08:39,520 to production level software and systems. 164 00:08:39,919 --> 00:08:41,400 So that's why I'm actually a big fan of it. 165 00:08:41,400 --> 00:08:46,000 It creates standards which help us to get AI in especially more 166 00:08:46,000 --> 00:08:48,920 regulated and high risk environments, production and 167 00:08:49,240 --> 00:08:51,840 often the high risk service, also the high impact stuff. 168 00:08:51,840 --> 00:08:55,520 So that's the reason I'm, I'm quite a big fan of it that I've 169 00:08:55,520 --> 00:08:58,120 seen that having the proper governance and the proper 170 00:08:58,120 --> 00:09:01,520 controls in place helps to get things beyond POC into 171 00:09:01,520 --> 00:09:03,640 production. Yeah, help me deepen this out 172 00:09:03,640 --> 00:09:06,800 for the people that don't know, because the the EUAI Act 173 00:09:06,800 --> 00:09:10,960 categorises different types of AI in different fields of 174 00:09:10,960 --> 00:09:13,720 industry, right? If it's education is different, 175 00:09:13,720 --> 00:09:15,960 if it's risk and financial management, it's different. 176 00:09:15,960 --> 00:09:20,200 If it's consumer facing, spam or anything else, it's mainly the 177 00:09:20,200 --> 00:09:22,600 lower risk applications and then you have different 178 00:09:22,600 --> 00:09:25,600 classifications with regards to what you need to adhere to. 179 00:09:26,240 --> 00:09:28,080 True. Mary, first of all, disclaimer, 180 00:09:28,080 --> 00:09:30,120 I'm not a lawyer. Yeah, legal specialist. 181 00:09:30,360 --> 00:09:33,160 So we actually give training together with, with, with our 182 00:09:33,160 --> 00:09:37,920 partners and often they are more the legal specialist and more 183 00:09:37,920 --> 00:09:39,960 the, the, the technical person in the room. 184 00:09:40,360 --> 00:09:45,200 And we often joke that we always avoid doing anything legal 185 00:09:45,200 --> 00:09:46,800 because we hate it. And the other way around. 186 00:09:46,800 --> 00:09:48,680 They understand shit about engineering. 187 00:09:50,400 --> 00:09:55,960 But to summarise it in, in a way, in a way which is probably 188 00:09:55,960 --> 00:09:58,160 also more interesting for engineers, You do indeed have, 189 00:09:58,240 --> 00:10:00,560 have multiple categories. It's prohibited. 190 00:10:00,800 --> 00:10:02,320 Don't think about it. You're not going to do it 191 00:10:02,320 --> 00:10:04,600 anyway. I've hardly seen application 192 00:10:04,600 --> 00:10:05,960 which you fall in that category anyway. 193 00:10:06,560 --> 00:10:09,320 Then you have high risk and that actually entails quite a few 194 00:10:09,320 --> 00:10:11,200 things. And to summarise, it is 195 00:10:11,200 --> 00:10:15,040 everything that has an high impact on human rights, on 196 00:10:15,040 --> 00:10:17,640 basically the rights you have and, and I have, which are also 197 00:10:18,080 --> 00:10:20,200 a strong part of EU regulation in general. 198 00:10:21,000 --> 00:10:24,200 And there, there are quite some articles which dictate what you 199 00:10:24,200 --> 00:10:26,080 have to do. Basically the standards you have 200 00:10:26,080 --> 00:10:28,920 to take into account. So you have to explain for 201 00:10:28,920 --> 00:10:30,360 example, where your data is coming from. 202 00:10:30,360 --> 00:10:33,400 You have to, you have to explain and you have to document how you 203 00:10:33,400 --> 00:10:36,760 think about bias or ethical considerations. 204 00:10:36,760 --> 00:10:39,240 You have to have proper controls in place. 205 00:10:39,240 --> 00:10:42,440 That's called human oversight in this case, or effective human 206 00:10:42,440 --> 00:10:44,000 oversight. So there are a few things which 207 00:10:44,000 --> 00:10:46,560 you need to do if there's a high risk model or high risk system 208 00:10:46,560 --> 00:10:49,720 in general. And I believe that those things 209 00:10:49,720 --> 00:10:51,440 you should be willing to do anyway. 210 00:10:52,240 --> 00:10:54,480 But it helps now to have a standard of like, OK, these 211 00:10:54,480 --> 00:10:56,800 eight points with like sub bullets are the things you have 212 00:10:56,800 --> 00:10:58,480 to do. And there's often software for 213 00:10:58,480 --> 00:10:59,720 that. There's a reason we started 214 00:10:59,720 --> 00:11:03,160 deploy is to basically make sure that you have all the tools in 215 00:11:03,160 --> 00:11:06,840 place to adhere to the things you need to do for anything that 216 00:11:06,840 --> 00:11:09,800 has some risk with it. And you should do that anyway 217 00:11:09,800 --> 00:11:12,800 because if you don't do it and there wouldn't be an AI act, you 218 00:11:12,800 --> 00:11:15,280 would still probably be in the news if you if you basically 219 00:11:15,280 --> 00:11:17,120 fuck up with with a certain system. 220 00:11:17,120 --> 00:11:21,240 So I think it really helps to create a standard and I don't 221 00:11:21,240 --> 00:11:23,040 really think it's going to block innovation. 222 00:11:23,720 --> 00:11:26,280 Is that the main counter argument why people don't like 223 00:11:26,280 --> 00:11:28,160 it? That it's kind of putting the 224 00:11:28,160 --> 00:11:30,320 brakes on innovation. Because I agree with what you 225 00:11:30,320 --> 00:11:32,720 say, a framework can be a good thing, right? 226 00:11:32,720 --> 00:11:34,640 And now this is a legal framework, so it's kind of 227 00:11:34,640 --> 00:11:38,040 mandatory and you have to comply to it, but it's still a 228 00:11:38,040 --> 00:11:40,480 framework. And if those are a means to an 229 00:11:40,480 --> 00:11:43,000 end to create a good system, then it could be a good 230 00:11:43,000 --> 00:11:45,080 framework as well. But that's subjective. 231 00:11:45,840 --> 00:11:48,880 Yeah, the the only other thing that's going to happen if you 232 00:11:48,880 --> 00:11:52,360 don't have an AI act is every single member state is going to 233 00:11:52,880 --> 00:11:55,400 bring up an AI act kind of version. 234 00:11:55,800 --> 00:11:58,000 You see the same now in the US If you want to bring something 235 00:11:58,000 --> 00:12:00,960 to the US market, you have to adhere probably to the New York 236 00:12:00,960 --> 00:12:03,840 State regulation on AI, which is more about bias and ethical 237 00:12:03,840 --> 00:12:06,640 considerations of fairness. You have to do, you have to act. 238 00:12:06,640 --> 00:12:08,280 You have to comply to the one in Illinois. 239 00:12:08,280 --> 00:12:10,600 You have to comply to the one in California, which actually looks 240 00:12:10,600 --> 00:12:13,320 quite similar to the AI. So it's just a patchwork of 241 00:12:13,320 --> 00:12:16,720 different regulations, often also conflicting on some points. 242 00:12:17,160 --> 00:12:18,960 It's way harder. And the same is going to happen 243 00:12:18,960 --> 00:12:20,680 in Europe. Like France probably going to 244 00:12:20,680 --> 00:12:24,040 protect tech companies so they will be quite, quite, quite 245 00:12:24,680 --> 00:12:27,080 relaxed about a lot of things. But Italy doesn't. 246 00:12:27,080 --> 00:12:28,640 So Italy's going to be super strict on things. 247 00:12:28,640 --> 00:12:30,000 It's going to be a mess in the end. 248 00:12:30,000 --> 00:12:32,280 And then we're all going to complain like, yeah, it's such a 249 00:12:32,280 --> 00:12:34,480 mess in Europe. We need to have harmonise things 250 00:12:34,480 --> 00:12:36,520 and it's all over again, the same thing. 251 00:12:36,520 --> 00:12:39,040 So I'm really happy we actually have some clarity. 252 00:12:39,600 --> 00:12:41,880 We have standards. It's going to help actually also 253 00:12:41,880 --> 00:12:43,640 because you will do those things anyway. 254 00:12:43,640 --> 00:12:46,880 You know, the opposite is exactly what I explained. 255 00:12:46,880 --> 00:12:48,800 It's it's, it's basically what's happening in the US. 256 00:12:49,120 --> 00:12:51,240 Yeah. And I wonder why people are not 257 00:12:51,240 --> 00:12:53,760 a fan of it, right, If it's otherwise, kind of, we don't 258 00:12:53,760 --> 00:12:55,720 have many other options. And each country is going to do 259 00:12:55,720 --> 00:12:58,280 their own thing. There will be something because 260 00:12:58,280 --> 00:13:00,960 I feel like people need to be protected towards the 261 00:13:00,960 --> 00:13:03,920 applications that they use, they consume, especially when they're 262 00:13:03,920 --> 00:13:06,800 more on the human rights side. With regards to the impact and 263 00:13:07,400 --> 00:13:10,080 yeah, mainly impact, I guess. So there will be something. 264 00:13:10,080 --> 00:13:12,560 So this might be the least of the evil I guess. 265 00:13:12,720 --> 00:13:14,600 Yeah. And I think, I think which, so I 266 00:13:14,600 --> 00:13:17,480 do agree that we see large differences between the US and 267 00:13:17,480 --> 00:13:20,320 Europe, but I think it doesn't come from regulation per SE. 268 00:13:20,600 --> 00:13:23,800 I think it has to do with the way we we deal with regulation. 269 00:13:24,480 --> 00:13:25,840 So we can use it in the next use. 270 00:13:26,120 --> 00:13:28,800 We can always use regulation as an excuse, like I'm not sure if 271 00:13:28,800 --> 00:13:31,680 that's allowed by GDPRARAX, whatever you can think of. 272 00:13:32,720 --> 00:13:37,280 Or you can try to comply as good as possible in a practical way 273 00:13:37,280 --> 00:13:38,800 and just go ahead with innovation. 274 00:13:39,320 --> 00:13:41,920 And I have the feeling the latter happens in other parts of 275 00:13:41,920 --> 00:13:43,760 the world and it first happens here. 276 00:13:43,760 --> 00:13:46,560 So if you can just use legislation as an as an excuse, 277 00:13:46,560 --> 00:13:47,960 yeah, we don't have to do anything. 278 00:13:48,400 --> 00:13:50,520 So it's easier. We have to work less, we have to 279 00:13:50,520 --> 00:13:53,880 spend less hours on the office. But that's more cultural thing 280 00:13:53,880 --> 00:13:55,400 than anything related to regulation. 281 00:13:55,400 --> 00:13:58,040 So if we wouldn't have the AI, we would find another reason or 282 00:13:58,040 --> 00:14:00,040 those people would find another reason not to innovate. 283 00:14:00,960 --> 00:14:03,760 So we have to fix the cultural path, which is the the hardest 284 00:14:03,760 --> 00:14:05,400 part by the way. But I think that's the thing we 285 00:14:05,400 --> 00:14:07,000 need to fix if we want to innovate quicker. 286 00:14:07,000 --> 00:14:09,080 But regulation itself, I think it's not the bottleneck. 287 00:14:09,960 --> 00:14:12,560 You also mentioned that this might be a means to an end to 288 00:14:12,560 --> 00:14:15,560 help go from proof of concept to something that's actually 289 00:14:15,880 --> 00:14:17,880 production ready that people can use. 290 00:14:18,440 --> 00:14:21,600 I've been in my previous assignment at a bank where I've 291 00:14:21,600 --> 00:14:24,600 seen many proof of concepts, but then they didn't comply to let's 292 00:14:24,600 --> 00:14:26,520 say the standards of the environment with regards to 293 00:14:26,520 --> 00:14:28,360 technology and what you can or cannot use. 294 00:14:28,640 --> 00:14:31,440 It was just proof of value. All right, we see this value 295 00:14:31,640 --> 00:14:32,760 now. Can we do it with the tool set 296 00:14:32,760 --> 00:14:35,320 that we have? And then if it's a no, then it's 297 00:14:35,320 --> 00:14:38,520 kind of yeah, there's stakeholder expectations there. 298 00:14:38,520 --> 00:14:40,400 There's an investment. There's no return on investment 299 00:14:40,400 --> 00:14:43,080 because the technology is not compliant. 300 00:14:43,080 --> 00:14:46,120 So it's very difficult to then go from proof of concept to 301 00:14:46,120 --> 00:14:49,480 something that is running an application in in production. 302 00:14:49,880 --> 00:14:52,360 I don't know if many other or at least from the organizations 303 00:14:52,360 --> 00:14:55,320 that that you've been in and out of are struggling with something 304 00:14:55,320 --> 00:14:58,360 similar when it comes to actually we see proof of value 305 00:14:58,360 --> 00:15:01,080 and then we're trying to go to production, but technology yes 306 00:15:01,080 --> 00:15:03,720 or no or we're behind or there's a lot of risk and low risk 307 00:15:03,720 --> 00:15:06,520 appetite. Yeah. 308 00:15:08,440 --> 00:15:12,080 I, I see those things a lot of course that that that I think it 309 00:15:12,080 --> 00:15:14,720 has to do with bottom up is this top down approach. 310 00:15:14,760 --> 00:15:19,480 A lot of organizations are hiring a few smart people also 311 00:15:19,480 --> 00:15:22,520 very young, and they kind of have to figure out what they 312 00:15:22,520 --> 00:15:25,280 have to do with AI or data products. 313 00:15:25,640 --> 00:15:27,840 Just build something, be innovative, experiment, 314 00:15:27,840 --> 00:15:30,280 whatever. And I never saw that really 315 00:15:30,280 --> 00:15:33,280 working. So if there's no pretty clear 316 00:15:33,280 --> 00:15:36,280 picture where you want to go to a top down in an organization, 317 00:15:36,720 --> 00:15:39,680 it doesn't work. It it just ends up in MPPS and 318 00:15:39,680 --> 00:15:42,760 PO CS and proof of failure or whatever other names you want to 319 00:15:42,760 --> 00:15:44,520 give it. But it's not going to prediction 320 00:15:44,520 --> 00:15:48,960 because it's at the end of the day, it probably doesn't really 321 00:15:48,960 --> 00:15:51,800 fit in whatever, whatever strategic direction you're 322 00:15:52,160 --> 00:15:54,600 deciding to go to. So it's really also the 323 00:15:54,600 --> 00:15:59,560 responsibility of board level or whatever you, whatever is 324 00:15:59,560 --> 00:16:03,240 leading organization, it's their responsibility to really think 325 00:16:03,240 --> 00:16:06,160 about, OK, this is where we go to in like one or two or three 326 00:16:06,160 --> 00:16:08,480 or five years. These are the products we need 327 00:16:08,480 --> 00:16:10,480 to get there. This is the reason we are doing 328 00:16:10,480 --> 00:16:12,760 it. And that can drill down to the 329 00:16:12,760 --> 00:16:14,960 rest of the organization. They can figure out how to build 330 00:16:14,960 --> 00:16:18,040 it, but in the end I think you need to really have a good 331 00:16:18,040 --> 00:16:20,120 design where you want to be as an organization in like 3 or 332 00:16:20,120 --> 00:16:22,360 five years. Same as building houses or 333 00:16:22,360 --> 00:16:24,320 building suburbs. You also don't start bottom up, 334 00:16:24,320 --> 00:16:26,320 you start top down. You need to have a proper plan 335 00:16:26,320 --> 00:16:28,800 of a suburb, then a proper plan of how you build a house, and 336 00:16:28,800 --> 00:16:31,320 then you can actually start building bottom up. 337 00:16:31,320 --> 00:16:35,160 But otherwise it doesn't work. Yeah, maybe in Belgium, but in 338 00:16:35,160 --> 00:16:39,240 most countries it doesn't work. Especially in some industries, 339 00:16:39,320 --> 00:16:42,680 leadership teams have always struggled with the role of tech 340 00:16:42,680 --> 00:16:46,160 product, digital products. And now you have AI there as 341 00:16:46,160 --> 00:16:47,320 well. And you mentioned that this 342 00:16:47,320 --> 00:16:50,400 vision, this kind of dot on the horizon, where do we want to be 343 00:16:50,400 --> 00:16:52,280 as an organization? How do we leverage technology 344 00:16:52,280 --> 00:16:55,640 that's now out there? New is now also their 345 00:16:55,640 --> 00:16:58,400 responsibility. It's quite challenging to I, 346 00:16:58,640 --> 00:17:02,560 from my perspective, kind of see that take shape in some of the 347 00:17:02,560 --> 00:17:04,520 organizations or industries that I've been in. 348 00:17:04,960 --> 00:17:07,319 I do see people experiment bottom up. 349 00:17:07,319 --> 00:17:10,240 But then in the end, like you mentioned, the effect of that or 350 00:17:10,240 --> 00:17:12,599 the outcome of that might be that it doesn't reach production 351 00:17:12,599 --> 00:17:14,480 at all. But I do like the 352 00:17:14,480 --> 00:17:17,200 experimentation mindset, just within boundaries I guess. 353 00:17:17,400 --> 00:17:22,240 Absolutely. So my advice to to more junior 354 00:17:22,240 --> 00:17:24,520 people in the organization, if you feel there's no direction, 355 00:17:25,599 --> 00:17:29,760 stop developing and start, stop developing and start asking 356 00:17:29,760 --> 00:17:31,240 those questions because it's never going to work. 357 00:17:31,240 --> 00:17:33,080 It's going to be very, very energy draining. 358 00:17:33,080 --> 00:17:35,520 I think if you work on something for six months and it's going to 359 00:17:35,520 --> 00:17:38,280 be thrown away and the other way around. 360 00:17:38,280 --> 00:17:41,160 If you're really on top of the organization, if you're on the 361 00:17:41,160 --> 00:17:43,600 board, if you're a director or whatever. 362 00:17:44,080 --> 00:17:46,840 And then you have the feeling that for some reason you don't 363 00:17:46,840 --> 00:17:49,080 really understand what you're working on or you don't really 364 00:17:49,080 --> 00:17:52,360 understand anything about digital products, data, AI. 365 00:17:52,880 --> 00:17:56,240 Either find a person and give them a board seat or Start 366 00:17:56,240 --> 00:17:59,400 learning how it works and start gaining knowledge about it. 367 00:18:00,360 --> 00:18:04,000 But doing neither and hope that bottom up stuff is going to 368 00:18:04,000 --> 00:18:07,440 happen. I've, I haven't seen it working 369 00:18:07,440 --> 00:18:10,800 for the past 1520 years. Is not going to work in the 370 00:18:10,800 --> 00:18:12,680 upcoming years either if you don't have a vision. 371 00:18:12,920 --> 00:18:16,280 Yeah, and I see that also if you talk about other geographies, I 372 00:18:16,280 --> 00:18:19,560 see younger people actually leading companies than in most 373 00:18:19,560 --> 00:18:21,360 parts of Europe. We tend to have pretty old 374 00:18:21,360 --> 00:18:24,240 people in parties, especially outside of the Netherlands I 375 00:18:24,240 --> 00:18:27,640 think not doing that bad, but I've seen statistics of that and 376 00:18:27,920 --> 00:18:29,400 the average age is just pretty old. 377 00:18:29,400 --> 00:18:31,320 Also here of people running companies. 378 00:18:32,320 --> 00:18:33,800 And that's a bad thing, you think? 379 00:18:34,000 --> 00:18:36,200 Yeah. Yeah, I think you need a few 380 00:18:36,600 --> 00:18:40,200 younger people sometimes as well on the top of an organization. 381 00:18:40,440 --> 00:18:43,600 Yeah, I haven't figured out how to get there yet, except like 382 00:18:43,600 --> 00:18:45,280 this. Oh, you start something from 383 00:18:45,280 --> 00:18:47,640 scratch with the years, and then you're also old by the time you 384 00:18:47,640 --> 00:18:49,840 have something of scale and you get there. 385 00:18:50,640 --> 00:18:53,560 There are really, really big companies in the US where the 386 00:18:53,560 --> 00:18:55,480 CFO is like 35 years old. Yeah. 387 00:18:55,600 --> 00:18:56,600 I think that's going to be really. 388 00:18:56,600 --> 00:18:58,360 Helpful. Yeah, very good there. 389 00:18:59,600 --> 00:19:02,600 What do you think of when people say culture eats strategy for 390 00:19:02,600 --> 00:19:04,120 breakfast? Because I've been in a lot of 391 00:19:04,120 --> 00:19:07,120 environments where people really kind of hamper on their culture 392 00:19:08,120 --> 00:19:10,520 to a point where indeed they might not even have a strategy. 393 00:19:10,520 --> 00:19:12,360 And then sometimes it just feels like anarchy. 394 00:19:13,360 --> 00:19:15,200 I don't know. I think it goes hand in hand. 395 00:19:15,760 --> 00:19:17,640 If you write something down when it's not a lift, it doesn't 396 00:19:17,640 --> 00:19:19,120 work. Yeah, If you never write 397 00:19:19,120 --> 00:19:21,280 something down and it's all captured, your culture, it also 398 00:19:21,280 --> 00:19:23,640 doesn't doesn't scale. At least there's a lot of 399 00:19:23,640 --> 00:19:27,800 unclarity. We at deploy, we have, we have a 400 00:19:27,800 --> 00:19:30,600 hybrid model, so people come to the office and that helps in 401 00:19:30,600 --> 00:19:34,200 culture and also work at home. And so we need to write things 402 00:19:34,200 --> 00:19:36,040 down. And I have a lot of respect for 403 00:19:36,040 --> 00:19:39,320 companies who are remote only because then you have to write 404 00:19:39,320 --> 00:19:41,000 things down. Culture's pretty hard. 405 00:19:41,000 --> 00:19:43,400 In the end, you can, you can work around if it's really hard. 406 00:19:44,320 --> 00:19:45,440 So you have to write things down. 407 00:19:45,440 --> 00:19:47,760 And that strategy I think is super, super important. 408 00:19:47,760 --> 00:19:51,160 Vision, mission, strategy, goals, targets, etcetera. 409 00:19:52,400 --> 00:19:54,200 And if you're always in the office, I think the culture is 410 00:19:54,200 --> 00:19:56,880 really important. I see that here when I walk 411 00:19:56,880 --> 00:20:01,200 around and obviously culture is super important and, and both 412 00:20:01,200 --> 00:20:03,480 ways can work. But I think if you if you have 413 00:20:03,480 --> 00:20:05,600 some kind of hybrid model, it also goes hand in hand to have 414 00:20:05,600 --> 00:20:09,360 both strategy and culture pretty much aligned in that sense. 415 00:20:09,440 --> 00:20:12,800 Yeah, Yeah, you mentioned that bottom up initiatives and bottom 416 00:20:12,800 --> 00:20:16,080 up experimentation, it, it just doesn't go anywhere without a 417 00:20:16,080 --> 00:20:18,400 strong vision, right. And then, at least from my 418 00:20:18,400 --> 00:20:21,440 perspective, you can experiment within those boundaries and see 419 00:20:21,760 --> 00:20:24,080 if the outcomes are a means to an end. 420 00:20:24,080 --> 00:20:26,840 In the end, if they help the business, the company, the 421 00:20:26,840 --> 00:20:29,600 organization and the group of people, because organisations 422 00:20:29,600 --> 00:20:32,840 are still just groups of people, reach whatever their vision is, 423 00:20:32,840 --> 00:20:34,920 or at least take a right step towards that. 424 00:20:36,280 --> 00:20:38,120 That's super easy I think I just. 425 00:20:38,160 --> 00:20:39,600 I just don't understand why it's. 426 00:20:39,600 --> 00:20:43,120 It sounds complex to me. You define where you want to be 427 00:20:43,120 --> 00:20:45,040 5 years from now. That should be doable for most 428 00:20:45,040 --> 00:20:47,440 board members. I guess you drill down what 429 00:20:47,440 --> 00:20:50,960 initiatives need to be taken and you see where AI or data or 430 00:20:50,960 --> 00:20:53,800 digital products fit in. But how do you take people along 431 00:20:53,800 --> 00:20:55,760 with you, right? If there's a vision where they 432 00:20:55,760 --> 00:20:57,200 don't see themselves, you lose them. 433 00:20:57,640 --> 00:20:59,160 That's true. So that's easier. 434 00:20:59,160 --> 00:21:00,480 So it's smaller companies, of course. 435 00:21:01,200 --> 00:21:03,560 I think, I think that the more you go to done, the more, the 436 00:21:03,560 --> 00:21:05,920 more you can involve people. I think in any way you can 437 00:21:05,920 --> 00:21:09,360 involve them. But in the end, I think clarity 438 00:21:09,360 --> 00:21:13,840 needs to come from a few people. In the end, if you don't provide 439 00:21:13,840 --> 00:21:16,480 a clarity, everyone is swimming a bit in the end. 440 00:21:16,680 --> 00:21:19,440 Yeah, yeah, yeah. Which is super frustrating. 441 00:21:19,720 --> 00:21:22,560 The thing that I've seen the most, and this is indeed coming 442 00:21:22,560 --> 00:21:25,120 from board level and organisations, is we see 443 00:21:25,120 --> 00:21:27,720 companies doing stuff with AI. We need to do something like 444 00:21:27,720 --> 00:21:31,080 this decent of urgency that just trickles down and there's like, 445 00:21:31,080 --> 00:21:35,080 we need to act something with AI on the product sense, embed Gen. 446 00:21:35,080 --> 00:21:37,840 AI into whatever we have, but they don't even use Gen. 447 00:21:37,840 --> 00:21:38,560 AI. It's just AI. 448 00:21:38,560 --> 00:21:41,360 Everything, everything gets under this umbrella of AI. 449 00:21:41,360 --> 00:21:43,640 Doesn't matter if it's predictive AI or generative AI, 450 00:21:43,800 --> 00:21:45,960 That's one. And then there's the second one, 451 00:21:45,960 --> 00:21:48,080 which is more for the people that already have a lot of 452 00:21:48,440 --> 00:21:49,960 software engineering capabilities. 453 00:21:50,360 --> 00:21:53,640 We need to be more efficient with regards to how we use AI so 454 00:21:53,640 --> 00:21:56,800 we can improve our output regardless of whatever outcomes 455 00:21:56,800 --> 00:21:58,680 we're trying to achieve. We need to optimize the 456 00:21:58,680 --> 00:22:00,120 efficiencies and stuff like that. 457 00:22:00,440 --> 00:22:02,680 Those are the two. And then especially if there's 458 00:22:02,680 --> 00:22:06,320 no good framework, if there's no good story, even the people that 459 00:22:06,320 --> 00:22:09,760 communicate that are lacking in communication skills or even 460 00:22:09,760 --> 00:22:12,640 sometimes just charisma, they lose the buy in. 461 00:22:13,000 --> 00:22:15,720 There's been a bottom up kind of challenge and push back on, OK, 462 00:22:15,720 --> 00:22:17,120 we don't even have tools available. 463 00:22:17,120 --> 00:22:20,040 How do you want us to work with AI or how are we going to be 464 00:22:20,040 --> 00:22:22,360 enabled in this, this mismatch and expectations. 465 00:22:22,760 --> 00:22:25,200 And then it just sandwiches and middle management. 466 00:22:25,200 --> 00:22:29,240 It's trying to middle manage. I agree, but maybe to turn it 467 00:22:29,240 --> 00:22:32,080 around and make it positive. I've seen really, really good 468 00:22:32,080 --> 00:22:34,880 examples as well. That's, that's the, that's the 469 00:22:34,880 --> 00:22:37,400 fortunate part of working on something like D players that 470 00:22:37,400 --> 00:22:40,840 you're early adopters are the ones that actually have a strong 471 00:22:40,840 --> 00:22:43,400 vision, that actually know where they want to be in two years 472 00:22:43,400 --> 00:22:45,280 from now, that actually are growing really quickly. 473 00:22:45,280 --> 00:22:47,200 That understand that efficiency is super important. 474 00:22:47,200 --> 00:22:51,760 The ones that understand that having a few killer AI features, 475 00:22:51,800 --> 00:22:54,320 AI products in the offering is going to be super important. 476 00:22:54,320 --> 00:22:58,040 So it's fascinating to see them operating actually, they are 477 00:22:58,040 --> 00:23:01,520 actually really growing quickly because they they do define 478 00:23:01,520 --> 00:23:05,040 where what they need in like one or two years from now. 479 00:23:06,360 --> 00:23:08,600 So I would just turn around and look at the ones that grow 480 00:23:08,600 --> 00:23:11,440 quickly what they are doing. And often they're quite open to 481 00:23:11,440 --> 00:23:13,160 talk about it. They often share quite a lot of 482 00:23:13,160 --> 00:23:15,480 things. But those are actually really 483 00:23:15,480 --> 00:23:18,200 interesting to watch, I think. And you don't have to do that 484 00:23:18,200 --> 00:23:21,120 only in the Netherlands, but if you look in, for example, the UK 485 00:23:21,120 --> 00:23:25,200 and how Fintech and also share tech now is taking over quite a 486 00:23:25,200 --> 00:23:27,880 big market share from the incumbents in the market. 487 00:23:28,640 --> 00:23:30,880 You can learn a lot from them. And it's super interesting, I 488 00:23:30,880 --> 00:23:32,920 think to see how they are operating at the moment. 489 00:23:33,080 --> 00:23:35,640 Can you elaborate on that? You said already it starts with 490 00:23:35,640 --> 00:23:38,040 a strong vision of where do you want to be in one or two years, 491 00:23:38,040 --> 00:23:41,240 but what are they doing that is exceptional that allows them to 492 00:23:41,240 --> 00:23:44,120 excel in that way? It's not even exceptional. 493 00:23:44,120 --> 00:23:46,360 I think it's it's, it's so if you think about it, it's not 494 00:23:46,360 --> 00:23:52,400 even, I think it's not that they actually spend time, I think on, 495 00:23:52,400 --> 00:23:55,960 on, on defining where they want to be in one or two or three 496 00:23:55,960 --> 00:23:59,720 years. And they do quite some sessions, 497 00:23:59,720 --> 00:24:02,800 I think together with their colleagues on discussing and 498 00:24:02,800 --> 00:24:04,480 arguing whether it's going to be. 499 00:24:06,280 --> 00:24:09,560 So if you talk about banking, for example, and your banks, you 500 00:24:09,560 --> 00:24:11,800 see that they basically take the whole process. 501 00:24:11,800 --> 00:24:13,640 Every company is a is a little factory in the end. 502 00:24:13,920 --> 00:24:16,600 They take the whole process and define and every step of the 503 00:24:16,600 --> 00:24:19,240 process they define here and here and here and here. 504 00:24:19,240 --> 00:24:22,720 We feel that with relatively small investment, we can 505 00:24:22,720 --> 00:24:26,000 actually make things much more efficient or much more effective 506 00:24:26,000 --> 00:24:29,640 for basically make a business case in the end that that skills 507 00:24:29,640 --> 00:24:33,760 and that's what they are doing. They basically make a blueprint 508 00:24:33,760 --> 00:24:37,280 of how a bank works and then find all the places where they 509 00:24:37,280 --> 00:24:39,880 are more efficient of more effective or can create new 510 00:24:39,880 --> 00:24:43,240 products and new offerings which are you know, which which they 511 00:24:43,240 --> 00:24:44,640 expect is going to be demand for. 512 00:24:45,120 --> 00:24:47,320 So that's what they're basically doing and that's often was 513 00:24:47,320 --> 00:24:49,760 lacking in all organisations. So if you're going to do the 514 00:24:49,760 --> 00:24:53,720 same thing, I'm 100% sure that's also we have the right people in 515 00:24:53,720 --> 00:24:55,640 the organization, but often you need less people than what you 516 00:24:55,640 --> 00:24:57,640 think you can actually achieve the same thing. 517 00:24:57,640 --> 00:25:00,680 And that's why I why I mentioned fintech and, and in shirt tech 518 00:25:00,680 --> 00:25:02,560 in the UK, they're much further than here. 519 00:25:03,480 --> 00:25:05,480 They also have already much bigger market share. 520 00:25:05,480 --> 00:25:08,240 And you can, you can learn quite a lot by looking at the 521 00:25:08,800 --> 00:25:12,240 presentations, meet up slides, those kind of things. 522 00:25:13,600 --> 00:25:15,000 We're going to learn a lot from that, I think. 523 00:25:15,640 --> 00:25:20,120 How much can you still gather from this kind of competitive 524 00:25:20,120 --> 00:25:21,880 analysis? Because you mentioned, OK, look 525 00:25:21,880 --> 00:25:24,960 at your own organizations, look at your whether it's process 526 00:25:24,960 --> 00:25:28,440 journeys or customer journeys or workflows that you have and see 527 00:25:28,440 --> 00:25:31,760 where kind of the most time is spent where you can make a solid 528 00:25:31,760 --> 00:25:34,800 business case where you can rely on technologies like these to be 529 00:25:34,800 --> 00:25:36,160 more effective in what you do, right? 530 00:25:36,160 --> 00:25:40,880 That's very much, I feel like looking at your organization in 531 00:25:40,880 --> 00:25:42,880 a nutshell and trying to optimize in that way. 532 00:25:43,240 --> 00:25:45,440 What I've seen other organizations do is they see 533 00:25:45,440 --> 00:25:49,920 what someone has created in certain output. 534 00:25:49,920 --> 00:25:52,880 It might be chatbot is always the kind of go to example. 535 00:25:53,200 --> 00:25:55,560 Someone sees a chatbot, they can be like, OK, well we can 536 00:25:55,560 --> 00:25:57,440 actually also do that and it's going to plug and play. 537 00:25:57,880 --> 00:26:00,320 But it might not be the best for their organization or for their 538 00:26:00,320 --> 00:26:03,760 consumer base or for the branding in the 1st place, but 539 00:26:03,760 --> 00:26:06,160 that's what they do. I feel like what you mentioned 540 00:26:06,160 --> 00:26:08,080 in looking at your own organization, that's kind of the 541 00:26:08,080 --> 00:26:09,840 better way to do it. But how much can you then still 542 00:26:09,840 --> 00:26:16,400 learn from your competitors? I think how they how much time 543 00:26:16,400 --> 00:26:18,400 they spent on making the blueprint, there's something you 544 00:26:18,400 --> 00:26:20,280 can learn from them. The exact blueprints or the 545 00:26:20,280 --> 00:26:22,640 exact use cases. That's something I think you can 546 00:26:22,640 --> 00:26:26,680 better define yourself because every bank is slightly different 547 00:26:26,680 --> 00:26:28,320 with different kind of customers, different kind of 548 00:26:28,320 --> 00:26:29,560 offering, different market positioning. 549 00:26:30,600 --> 00:26:32,520 So you have to make your own blueprint in the end. 550 00:26:34,360 --> 00:26:36,320 But the fact that you have to make a blueprint that you have 551 00:26:36,320 --> 00:26:37,920 to think about it, that's something you can learn from 552 00:26:37,920 --> 00:26:39,920 them. I think if they have time to 553 00:26:39,920 --> 00:26:44,200 spend spend hours and days on on on defining where they want to 554 00:26:44,200 --> 00:26:47,600 be in a few years from now, I think also other organizations 555 00:26:47,600 --> 00:26:51,240 can do that, especially prioritising innovation and the 556 00:26:51,840 --> 00:26:54,680 stuff of the future of everything that happened 557 00:26:54,680 --> 00:26:57,640 yesterday. It feels like there's no easy 558 00:26:57,640 --> 00:27:00,400 answer and you just have to go through the motions and step by 559 00:27:00,400 --> 00:27:03,640 step, kind of figure out where the value is for your own 560 00:27:03,640 --> 00:27:05,440 context and for your own organization. 561 00:27:05,880 --> 00:27:07,960 People are looking for a quick win though, especially. 562 00:27:07,960 --> 00:27:10,880 That doesn't work. That's OK, That doesn't work. 563 00:27:11,280 --> 00:27:16,960 So maybe that's that's the thing I'm trying to tell you is just 564 00:27:16,960 --> 00:27:19,320 saying that we have to push something very quickly. 565 00:27:19,680 --> 00:27:23,680 It's often a very stupid strategy that I hardly ever sort 566 00:27:23,680 --> 00:27:24,480 of working. No. 567 00:27:25,080 --> 00:27:27,720 What are some of the implementations or integrations 568 00:27:27,720 --> 00:27:30,640 that you've seen that are actually greatly valuable? 569 00:27:30,640 --> 00:27:33,400 Is there something you can share in that aspect, some of the AI 570 00:27:33,400 --> 00:27:36,440 features that you have seen go live and actually have impact? 571 00:27:37,040 --> 00:27:41,920 Yeah, I think it's good to to to distinguish different kind of AI 572 00:27:41,920 --> 00:27:43,440 applications. And so you have a few AI 573 00:27:43,440 --> 00:27:47,560 applications which are basically in the current process, often 574 00:27:47,560 --> 00:27:51,040 predictive AI, not even generative AI, but every process 575 00:27:51,040 --> 00:27:57,240 that is super repetitive, fraud detection or claim handling a 576 00:27:57,240 --> 00:28:01,600 certain insurance company or you know, whatever. 577 00:28:01,600 --> 00:28:04,760 So many of those repetitive tasks, you see a repetitive task 578 00:28:04,760 --> 00:28:07,080 which is done by people at the moment, which is slightly more 579 00:28:07,080 --> 00:28:10,320 complex than normal automation. It's often a very good candidate 580 00:28:10,320 --> 00:28:15,400 for machine learning and AI. If you talk about creating stuff 581 00:28:15,400 --> 00:28:19,960 which is, which is based on facts, that's also of course 582 00:28:19,960 --> 00:28:23,520 very easy to, to, to, to use. Actually at Deplo we are, we are 583 00:28:23,520 --> 00:28:28,320 now working with an AUS company which is doing quite a lot of 584 00:28:28,320 --> 00:28:32,080 our development work, which is I think super fascinating to see 585 00:28:32,080 --> 00:28:35,200 how the learning curve's going. So also with development work, 586 00:28:35,200 --> 00:28:37,720 you can even go beyond Copilot, you can actually work with 587 00:28:37,720 --> 00:28:41,360 companies that take over quite large chunks of your whole 588 00:28:41,360 --> 00:28:43,920 development work because you have the context, you have the 589 00:28:43,920 --> 00:28:46,760 code base, you have especially all the documents you can give 590 00:28:46,760 --> 00:28:48,600 your own developers. You can give that also through 591 00:28:48,680 --> 00:28:52,120 AI and with all the cartwheels they already have built and have 592 00:28:52,120 --> 00:28:54,480 in place. You can, you can actually 593 00:28:54,480 --> 00:28:56,800 automate quite a lot of development work, not as an 594 00:28:56,800 --> 00:28:58,960 assistant, but actually taking over your development work. 595 00:28:59,640 --> 00:29:02,520 So we see that working on our own organization as well. 596 00:29:04,440 --> 00:29:07,560 But all the other more repetitive stuff, I would also 597 00:29:07,560 --> 00:29:11,600 have a look at it. But all of them in the end 598 00:29:12,400 --> 00:29:15,400 require all the guardrails and the governments we talked about. 599 00:29:15,840 --> 00:29:17,920 If you want to go into the direction, invest in all the 600 00:29:17,920 --> 00:29:20,800 groundwork and all the the the infrastructure around it to do 601 00:29:20,800 --> 00:29:25,760 it in a good way. I think that's it's never a 602 00:29:25,760 --> 00:29:27,640 waste of money. It's the same as investing in 603 00:29:27,640 --> 00:29:29,720 cybersecurity 15 or 20 years ago. 604 00:29:30,000 --> 00:29:33,600 Whenever you want to go fully digital, you do need those, 605 00:29:33,680 --> 00:29:36,200 those cart fields basically to to go beyond. 606 00:29:37,320 --> 00:29:39,800 Yeah, just some experimentation work. 607 00:29:40,040 --> 00:29:43,320 Yeah, you mentioned there's an ex, Did I understand that 608 00:29:43,320 --> 00:29:44,720 correctly? There's a company helping you 609 00:29:44,720 --> 00:29:48,000 with the development work that you have on deploy itself as an 610 00:29:48,000 --> 00:29:50,480 application. So of course everyone is using 611 00:29:50,480 --> 00:29:53,520 Co pilots of rabbits. That's not really taking off all 612 00:29:53,520 --> 00:29:55,840 your development. That's just someone piloting. 613 00:29:55,840 --> 00:29:57,440 Yeah. You're, you're still operating 614 00:29:57,440 --> 00:29:59,600 in the AI system we actually create. 615 00:30:01,360 --> 00:30:04,360 We we are basically doing the design work, defining what we 616 00:30:04,360 --> 00:30:07,640 need and then send it off to another company and they do all 617 00:30:07,640 --> 00:30:11,000 the work until merch request and then we accept it or a review of 618 00:30:11,000 --> 00:30:12,080 the list. Gotcha. 619 00:30:12,080 --> 00:30:14,440 But isn't that your core? Isn't it scary to? 620 00:30:15,040 --> 00:30:17,080 Kind of scary. Give that, yeah, like or have 621 00:30:17,080 --> 00:30:19,480 that in collaboration with someone else, but it's working. 622 00:30:19,520 --> 00:30:22,280 It's also interesting because we have to think even better about 623 00:30:22,280 --> 00:30:24,120 the design. Yeah, we have to think even 624 00:30:24,120 --> 00:30:27,200 better about the requirements. We have to review it even better 625 00:30:27,200 --> 00:30:29,640 than before, but it actually creates better code. 626 00:30:30,320 --> 00:30:34,480 Does it a lot of companies invest in how much their 627 00:30:34,480 --> 00:30:36,760 development team understands the business, understands the 628 00:30:36,760 --> 00:30:38,280 product, right? Because then you can think 629 00:30:38,280 --> 00:30:41,920 along, understanding why we do things makes you challenge what 630 00:30:41,920 --> 00:30:44,040 we need to do in the 1st place. And then if the requirements 631 00:30:44,040 --> 00:30:47,080 don't make sense from your lens, also looking at diversity, then 632 00:30:47,080 --> 00:30:49,360 you can challenge that. But I feel like with this 633 00:30:49,360 --> 00:30:51,080 construction, you might miss that. 634 00:30:51,600 --> 00:30:54,640 Yet you're still saying kind of quality wise we have seen 635 00:30:54,640 --> 00:30:56,720 increased? For some. 636 00:30:56,840 --> 00:30:58,920 So I do get your point. There's also the worry we have. 637 00:30:58,920 --> 00:31:04,800 So if it is, if it requires a lot of context and expert 638 00:31:04,800 --> 00:31:07,720 knowledge, it's harder, then it often doesn't really work yet. 639 00:31:07,760 --> 00:31:11,600 Yeah. Although we can, the more we do 640 00:31:11,600 --> 00:31:13,680 with them, the more context they have, so the better they will 641 00:31:13,680 --> 00:31:19,520 understand it as well. But if it is, if it is a feature 642 00:31:19,760 --> 00:31:23,920 which is pretty well scout upfront and has less of those 643 00:31:24,960 --> 00:31:28,120 business context on it is actually quite doable. 644 00:31:29,720 --> 00:31:33,600 And because we're not doing the development work itself for that 645 00:31:33,600 --> 00:31:36,720 feature, we actually are more critical in the review, we're 646 00:31:36,720 --> 00:31:38,120 actually more critical in the design. 647 00:31:38,440 --> 00:31:39,880 So it actually creates better features. 648 00:31:39,960 --> 00:31:41,840 Interesting. Is this like an agency or a 649 00:31:41,840 --> 00:31:43,600 consultancy that does this construction? 650 00:31:43,800 --> 00:31:45,840 I don't know how to classify it. We'll discuss it actually. 651 00:31:45,840 --> 00:31:47,040 Is it a supplier? Is it? 652 00:31:47,240 --> 00:31:49,480 Outsourcing is it. An AI tool is that freelance is 653 00:31:49,480 --> 00:31:53,400 a combination of those things. Yeah, it's pretty cool zooming 654 00:31:53,400 --> 00:31:55,880 in into AI features in production. 655 00:31:55,880 --> 00:31:58,040 You mentioned a few things with regards to effectiveness and 656 00:31:58,040 --> 00:31:59,960 efficiency. What are some of the things 657 00:31:59,960 --> 00:32:03,240 that, if you're talking about a production system, needs to be 658 00:32:03,240 --> 00:32:04,920 there with regards to observability? 659 00:32:05,280 --> 00:32:09,160 I talked to Saud up earlier and he has a very distinct view on 660 00:32:10,560 --> 00:32:13,240 AI and operations, let's say. But I'm wondering what your take 661 00:32:13,240 --> 00:32:15,160 is. What are the metrics that I need 662 00:32:15,160 --> 00:32:18,840 to think of or what do I need to have in observability to be in 663 00:32:18,840 --> 00:32:21,040 control? That's a good question. 664 00:32:21,040 --> 00:32:23,720 It depends a bit again on the on the type of application of the 665 00:32:23,720 --> 00:32:26,640 type of technology you're using. If the generative AI or 666 00:32:27,520 --> 00:32:29,960 predative AI. I actually listened to the 667 00:32:29,960 --> 00:32:31,640 podcast you you met with sub OPS. 668 00:32:32,280 --> 00:32:33,600 I'm trying to recall everything now. 669 00:32:35,000 --> 00:32:38,800 I think what is so you can measure all the technical 670 00:32:38,800 --> 00:32:44,480 aspects, both the the, the, the, the, the more the metrics around 671 00:32:44,480 --> 00:32:47,560 traffic, but also the ones which try to capture performance. 672 00:32:48,720 --> 00:32:51,760 There's a bit like you, your your critical question like 2 673 00:32:51,760 --> 00:32:56,480 minutes ago. A lot of things are more require 674 00:32:56,480 --> 00:32:58,280 more context and simple statistics. 675 00:32:58,640 --> 00:33:01,520 So we are deeper strongly believe that one of the biggest 676 00:33:01,520 --> 00:33:03,440 things you need to measure, one of the most important things you 677 00:33:03,440 --> 00:33:06,440 need to measure is some kind of feedback loop, either directly 678 00:33:06,440 --> 00:33:08,480 or indirectly. So indirectly could be that you 679 00:33:08,480 --> 00:33:11,800 choose the second option or start arguing with an, with an 680 00:33:11,800 --> 00:33:15,160 NLM that's, there's also a kind of feedback loop, but having 681 00:33:15,160 --> 00:33:17,600 some kind of feedback loop and monitoring the feedback loop and 682 00:33:17,600 --> 00:33:19,240 act on it. Whenever you see that things get 683 00:33:19,240 --> 00:33:23,000 worse or when in a certain subcategory you see more 684 00:33:23,000 --> 00:33:27,680 feedback could, could mean that the, that the system is biased 685 00:33:27,680 --> 00:33:30,280 on some subgroup. I think that's the most 686 00:33:30,280 --> 00:33:32,040 important thing you need to measure next to all the 687 00:33:32,040 --> 00:33:34,640 statistical stuff and all the the technical stuff. 688 00:33:35,160 --> 00:33:37,120 The real thing you need to measure is in the end, human 689 00:33:37,120 --> 00:33:39,520 feedback, because that's, that's basically interaction between 690 00:33:39,520 --> 00:33:41,600 humans and machines or human and AI systems. 691 00:33:42,720 --> 00:33:45,080 And that's often pretty easy in practice to, to measure. 692 00:33:45,080 --> 00:33:48,840 It's if it is a, a predictive AI model, it could be feeding back 693 00:33:48,840 --> 00:33:52,480 actuals or feeding back expert opinions or overalls. 694 00:33:52,480 --> 00:33:55,360 If it is an, if there's a generative AI system, it's 695 00:33:55,360 --> 00:33:59,080 basically all the arguing that this happening between an LM and 696 00:33:59,080 --> 00:34:03,120 a human being or over ruling decisions or trying to correct 697 00:34:03,120 --> 00:34:04,720 things. Those, those things are all 698 00:34:04,720 --> 00:34:07,400 forms of feedback. And if you get feedback, find an 699 00:34:07,400 --> 00:34:10,239 API and start measuring that. Then you can act whenever you 700 00:34:10,239 --> 00:34:12,800 see that things don't really work in certain applications or 701 00:34:12,800 --> 00:34:16,159 some subgroups, or it's getting worse over time, or it's getting 702 00:34:16,159 --> 00:34:19,120 better. The positive side, I think 703 00:34:19,120 --> 00:34:21,480 that's the most important one because a lot of things are not 704 00:34:21,480 --> 00:34:24,560 captured in statistics. Yeah, I really like that. 705 00:34:24,800 --> 00:34:27,600 Like the only Gen. AI feature that I worked with in 706 00:34:27,600 --> 00:34:29,600 the end, it was going to make people more effective. 707 00:34:29,920 --> 00:34:32,840 So work that they were doing, spending hours on it, we had 708 00:34:32,840 --> 00:34:35,520 something that would pre fill a form instead of them doing the 709 00:34:35,520 --> 00:34:38,520 work, they would kind of have their own position of reviewing. 710 00:34:38,960 --> 00:34:42,800 And I saw a lot of people talk about, OK, if the answer is 711 00:34:42,800 --> 00:34:45,080 correct, then in the end it's correct, right? 712 00:34:45,080 --> 00:34:48,600 And we can even have a model to say if what a person changed, if 713 00:34:48,600 --> 00:34:51,679 the intent was exactly the same. And for me is the fact that 714 00:34:51,679 --> 00:34:54,360 someone already changed it, if it's not the way they like it or 715 00:34:54,360 --> 00:34:57,120 the way they want to kind of have that work be put forward 716 00:34:57,120 --> 00:34:59,520 because that goes in a whole chain and other people are going 717 00:34:59,520 --> 00:35:01,240 to see it. It's not good enough, right? 718 00:35:01,240 --> 00:35:03,800 It's kind of this user acceptance that also needs to 719 00:35:04,040 --> 00:35:06,360 take part of that, which I think is interesting. 720 00:35:06,480 --> 00:35:10,520 And I think that we still call it explainability, but it's more 721 00:35:10,520 --> 00:35:11,960 than just Shepherd Lyme or whatever. 722 00:35:12,280 --> 00:35:16,360 But but but explain ability or support for a certain decision 723 00:35:16,360 --> 00:35:18,320 or certain output is super important as well. 724 00:35:19,480 --> 00:35:22,080 So if you talk about it and I'm giving the support for certain 725 00:35:22,080 --> 00:35:25,640 outcomes, it's super important because the outcome can be, can 726 00:35:25,640 --> 00:35:27,800 be good, can be a good recommendation for something, 727 00:35:28,200 --> 00:35:30,200 but the reason behind it could be completely wrong. 728 00:35:30,880 --> 00:35:33,600 So I think that support is also important to make sure you 729 00:35:33,600 --> 00:35:37,040 capture the feedback about the reasoning as well, which is 730 00:35:37,040 --> 00:35:39,960 harder for a Netherland because asking it for an explanation is 731 00:35:39,960 --> 00:35:42,320 not an explanation of the model itself or the underlying 732 00:35:42,360 --> 00:35:45,720 technology. No, it's could be just as well 733 00:35:45,720 --> 00:35:48,320 hallucination. Yeah, yeah, yeah. 734 00:35:49,000 --> 00:35:53,240 Reasoning with regards to why an answer is the answer it is I 735 00:35:53,240 --> 00:35:56,040 think is incredibly challenging. But if you've if you have a good 736 00:35:56,040 --> 00:35:59,160 way and it also makes sense from a user perspective continuously, 737 00:35:59,520 --> 00:36:00,880 then I feel like you have a great outcome. 738 00:36:01,080 --> 00:36:04,480 I think that's an it's going to be an an ongoing discussion for 739 00:36:04,480 --> 00:36:07,240 the upcoming years. But it's basically the 740 00:36:07,680 --> 00:36:12,560 discussion between closed source systems, AI systems and open 741 00:36:12,560 --> 00:36:14,600 source models. I'm not sure which one is going 742 00:36:14,600 --> 00:36:15,880 to be leading in a few years from now. 743 00:36:15,880 --> 00:36:18,320 So you see open source getting closer to close source. 744 00:36:18,680 --> 00:36:22,600 We also see that the business value of it is is pretty small 745 00:36:22,600 --> 00:36:25,680 at the moment to be more open about which code you use, which 746 00:36:25,680 --> 00:36:30,880 data you use, etcetera. But if we go more towards or if 747 00:36:30,880 --> 00:36:34,080 we demand more transparency than it could be that we in the end 748 00:36:34,080 --> 00:36:37,000 move more towards using open source models for repetitive 749 00:36:37,000 --> 00:36:41,080 tasks or for certain use cases. And with open source models, 750 00:36:41,080 --> 00:36:46,920 LLMS, you can to a certain extent explain why, why a 751 00:36:46,960 --> 00:36:49,960 certain outcome was the way it was, like which tokens were 752 00:36:49,960 --> 00:36:55,000 important, for example, or saliency of, of of of a certain 753 00:36:55,000 --> 00:36:58,400 outcome. And I think that can help to 754 00:36:58,400 --> 00:37:01,840 really understand why the lamps are doing some really strange 755 00:37:01,840 --> 00:37:05,360 stuff, sometimes better than just asking for an explanation. 756 00:37:06,160 --> 00:37:08,800 So if transparency is getting more important, if control and 757 00:37:08,800 --> 00:37:11,120 source code is getting more important, if IP is getting more 758 00:37:11,120 --> 00:37:15,400 important, then we might move back towards more open source 759 00:37:16,000 --> 00:37:19,320 systems. If we for some reason decide 760 00:37:19,320 --> 00:37:23,680 that it's not important, then we probably stay working with 761 00:37:23,680 --> 00:37:25,400 closed source systems and it means that we have less 762 00:37:25,400 --> 00:37:27,840 transparency. We have to rely on the 763 00:37:27,840 --> 00:37:30,720 explanation we get from those providers. 764 00:37:32,040 --> 00:37:36,240 Right now like those are I think big ifs and I could see them, I 765 00:37:36,240 --> 00:37:38,760 have assumption on the way they are going, but are there any 766 00:37:38,760 --> 00:37:42,520 incentives to use open source versus closed source currently? 767 00:37:42,960 --> 00:37:45,560 Yeah, we see this quite a bit in, in in health tech. 768 00:37:47,480 --> 00:37:50,680 You don't want data to be shared to those really large tech 769 00:37:50,680 --> 00:37:52,920 companies. You do want to have more 770 00:37:52,920 --> 00:37:55,320 controls. You do want to really go deep 771 00:37:55,320 --> 00:37:59,680 into the the LLM system itself. You probably want to use smaller 772 00:37:59,720 --> 00:38:02,320 language models, still pretty big of course, with smaller 773 00:38:02,320 --> 00:38:04,440 language models because you have simply more control on what's 774 00:38:04,440 --> 00:38:06,560 happening. So there we actually see quite 775 00:38:06,560 --> 00:38:11,680 some companies working with their own open source LLMS which 776 00:38:11,680 --> 00:38:14,920 they fine tune and run use case because they simply have more 777 00:38:14,920 --> 00:38:16,480 control on what's happening. Interesting. 778 00:38:16,480 --> 00:38:19,040 And I think the same will happen for the high risk use cases. 779 00:38:19,280 --> 00:38:21,240 Yeah. Is it because I haven't 780 00:38:21,240 --> 00:38:24,000 experimented as much with open source variants? 781 00:38:24,000 --> 00:38:25,680 I don't know if they need to be self hosted. 782 00:38:25,680 --> 00:38:30,280 For me closed source model is just an API Callaway and that's 783 00:38:30,280 --> 00:38:31,760 it. And if it's compliant within the 784 00:38:31,760 --> 00:38:35,360 organization then we use it. That's like and we have 785 00:38:35,360 --> 00:38:37,760 experimented with other open source models, but I haven't had 786 00:38:37,760 --> 00:38:40,840 any hands on experience. Do people usually self hosted or 787 00:38:40,840 --> 00:38:43,320 how much expertise is needed to get that up and running do you 788 00:38:43,320 --> 00:38:45,760 know? Yeah, I've seen it with some 789 00:38:45,760 --> 00:38:49,640 companies that differs a bit I think, but you do, it is of 790 00:38:49,640 --> 00:38:52,240 course way more investment in the end to, to make that work. 791 00:38:52,240 --> 00:38:55,680 So the ones I, I mentioned are often focusing on one single use 792 00:38:55,680 --> 00:38:58,480 case, which which is basically the company in the end as well. 793 00:38:58,480 --> 00:39:01,360 So they provide a certain service in a certain step of 794 00:39:01,480 --> 00:39:03,160 certain process. So if you talk about healthcare, 795 00:39:03,160 --> 00:39:06,960 they're doing like one care pathway, 2 care pathways and 796 00:39:06,960 --> 00:39:09,520 they optimise a model for that. And that on itself is already a 797 00:39:09,960 --> 00:39:11,720 big enough business case to run a company on. 798 00:39:12,160 --> 00:39:14,600 So if that's your whole company, then it makes sense to to 799 00:39:14,600 --> 00:39:18,320 provide all the card rails. If you want to scale that to 800 00:39:18,320 --> 00:39:21,360 like all the care pathways or use it in healthcare in general, 801 00:39:21,360 --> 00:39:23,320 then I think it's going to be super, super complex and it's 802 00:39:23,320 --> 00:39:25,120 requiring a lot of money, a lot of investment. 803 00:39:25,560 --> 00:39:29,040 But the other way around, just relying on a big U.S. company to 804 00:39:29,040 --> 00:39:31,240 do this for you is also another way forward, I guess. 805 00:39:33,120 --> 00:39:36,600 So yes, it does require quite some investment in in multiple 806 00:39:36,600 --> 00:39:41,120 different ways, but it's also the only way forward, I guess in 807 00:39:41,120 --> 00:39:45,560 those in those users. So especially in healthcare, I 808 00:39:45,560 --> 00:39:49,680 think you do want to make sure that you understand why system 809 00:39:49,680 --> 00:39:52,600 makes certain decisions. And hence you do need to have 810 00:39:52,600 --> 00:39:56,000 access to source code. So that's what what we at least 811 00:39:56,000 --> 00:39:57,200 see. But they're not honesty. 812 00:39:57,360 --> 00:40:00,000 If I touch card within deploy, I think our developers are going 813 00:40:00,000 --> 00:40:05,280 to kill me so. For me, like government 814 00:40:05,680 --> 00:40:08,800 healthcare, if you're saying like these are the typical use 815 00:40:08,800 --> 00:40:10,480 cases, it's quite specialistic work. 816 00:40:11,000 --> 00:40:13,560 But then the people that want to work in those environments, they 817 00:40:13,560 --> 00:40:16,880 might not want to work in healthcare governments because 818 00:40:16,880 --> 00:40:19,400 it's also like slow organizations, very slow moving 819 00:40:19,640 --> 00:40:22,040 people that have been there for 20 years and a part of the 820 00:40:22,040 --> 00:40:24,000 walls, let's say. And they don't have the same 821 00:40:24,000 --> 00:40:27,040 experimental or innovation mindset that other startups 822 00:40:27,040 --> 00:40:28,320 might have. So it's kind of this chicken and 823 00:40:28,320 --> 00:40:30,120 egg problem. I don't know how to get out of 824 00:40:30,120 --> 00:40:32,080 that yet. I feel like we do need 825 00:40:32,080 --> 00:40:35,400 innovation in specifically those fields, but the way we've been 826 00:40:35,400 --> 00:40:37,280 doing it so far, it's just not effective. 827 00:40:38,200 --> 00:40:39,720 Yeah. And like these kind of 828 00:40:39,720 --> 00:40:43,120 innovations often start with things being under pressure. 829 00:40:43,120 --> 00:40:46,360 So in healthcare we will run into a point where it's simply 830 00:40:46,360 --> 00:40:51,360 not affordable anymore. And I think that will, that will 831 00:40:51,360 --> 00:40:53,120 make it easier on the end to innovate. 832 00:40:53,120 --> 00:40:56,720 That will, that will help letting organization make 833 00:40:56,720 --> 00:40:58,560 decision that has to do with risk. 834 00:40:58,560 --> 00:41:01,560 So if we don't want to take risk, then we then it's hard to 835 00:41:01,560 --> 00:41:03,160 innovate. At some point we need to take 836 00:41:03,160 --> 00:41:05,800 the risk because the other, if we don't make that decision, 837 00:41:05,800 --> 00:41:07,960 then the decision basically we can't provide healthcare for 838 00:41:07,960 --> 00:41:09,920 anyone if not for everyone anymore. 839 00:41:11,840 --> 00:41:14,840 So I think because both government and Healthcare is 840 00:41:14,840 --> 00:41:17,800 probably under pressure of the upcoming decades in Europe, we 841 00:41:17,800 --> 00:41:20,360 will make decisions which stand towards innovation and will 842 00:41:20,360 --> 00:41:22,480 become easier to innovate and will become easier to both 843 00:41:22,480 --> 00:41:25,480 companies and that in that industry. 844 00:41:26,440 --> 00:41:29,320 But I completely realise that at the moment it's super, super, 845 00:41:29,320 --> 00:41:32,600 super hard to build a health tech company or Med tech company 846 00:41:32,600 --> 00:41:37,120 or company which is providing software to the government, not 847 00:41:37,120 --> 00:41:39,960 consulting with actual software, you know, super hard. 848 00:41:40,600 --> 00:41:43,480 Which is strange, right? Because in your scenario, 849 00:41:43,480 --> 00:41:46,200 there's like, OK, there's this building pressure, which means 850 00:41:46,400 --> 00:41:48,120 we don't have an option anymore and we have to. 851 00:41:48,120 --> 00:41:51,600 It's like back against the wall. Whereas now I feel like we still 852 00:41:51,600 --> 00:41:54,600 have time, yet it's not working as well as it should. 853 00:41:54,800 --> 00:41:56,440 But right now would be the best time. 854 00:41:57,160 --> 00:42:01,000 The best time for a lot of things, we wait till it's almost 855 00:42:01,000 --> 00:42:02,360 too late and then we start innovating. 856 00:42:03,080 --> 00:42:05,840 But it, yeah, it has to do with our personal horizon, I think. 857 00:42:05,840 --> 00:42:07,480 Do you know what you're going to do in five years from now? 858 00:42:07,480 --> 00:42:10,640 No, I hate that question. Whenever you've been talking 859 00:42:10,640 --> 00:42:13,160 about these long timeline horizons and I'm like, Oh my 860 00:42:13,160 --> 00:42:17,240 God, like I have no clue. Most people don't, I think, so 861 00:42:17,240 --> 00:42:18,880 they're not optimising for the long term. 862 00:42:19,400 --> 00:42:22,960 Do you like, do you have that strong vision for you as a 863 00:42:22,960 --> 00:42:26,960 person or even you as a company? I am my my girlfriend at home. 864 00:42:26,960 --> 00:42:30,800 She often finds it very annoying that I'm thinking I had for 865 00:42:30,840 --> 00:42:32,280 quite some years. Quite some. 866 00:42:32,280 --> 00:42:33,960 Because it makes it less romantic I guess. 867 00:42:36,600 --> 00:42:38,640 Yeah. So both personally and 868 00:42:38,640 --> 00:42:39,800 professional, it's something you do. 869 00:42:40,640 --> 00:42:43,240 Yeah, I also like it. Mary's a bit of dreaming 870 00:42:43,240 --> 00:42:47,320 sometimes about about some stuff as well, but I think it can also 871 00:42:47,320 --> 00:42:49,560 be really important to at least have some fake idea where you 872 00:42:49,560 --> 00:42:53,280 are in a few years from now. And it would indeed be helpful 873 00:42:53,280 --> 00:42:59,240 to do that in, you know, in with the government, in healthcare, 874 00:42:59,560 --> 00:43:02,680 with hospitals, in the financial system because they're also 875 00:43:02,680 --> 00:43:06,480 going to be under pressure. We talked a bit about fintechs 876 00:43:06,480 --> 00:43:09,200 initiatives which are definitely going to take market share. 877 00:43:09,200 --> 00:43:14,880 So if I think it's going to be crucial for them as well to put 878 00:43:14,880 --> 00:43:18,080 innovation high on the agenda and not on like on Friday 879 00:43:18,080 --> 00:43:21,640 afternoon, also for them, it's going to be critical. 880 00:43:21,640 --> 00:43:25,800 But probably you'll see that when it's almost too late, 881 00:43:26,200 --> 00:43:27,480 people really start to innovate. Yeah. 882 00:43:27,680 --> 00:43:29,960 People will move. And that's happened over the 883 00:43:29,960 --> 00:43:34,200 past decades or centuries as well, I think, over and over 884 00:43:34,200 --> 00:43:35,400 again. I can see that. 885 00:43:35,560 --> 00:43:37,280 Yeah. I've really enjoyed this 886 00:43:37,280 --> 00:43:39,480 conversation, Martha. It's fascinating the way you 887 00:43:39,480 --> 00:43:42,560 think, the way you have a strong vision and kind of are driving 888 00:43:42,560 --> 00:43:45,360 forward this community and in and now and I'm some 889 00:43:45,360 --> 00:43:47,520 specifically, I'm going to look forward to it. 890 00:43:47,800 --> 00:43:50,200 We'll, we'll keep in touch. Hopefully it's going to be 891 00:43:50,200 --> 00:43:51,480 successful. Hopefully it's going to be 892 00:43:51,480 --> 00:43:53,360 successful. We'll see you as we go. 893 00:43:53,360 --> 00:43:54,920 Thank you so much for coming on and sharing. 894 00:43:55,280 --> 00:43:56,440 Thanks a lot this. Was a blast. 895 00:43:56,560 --> 00:43:58,480 Well, round off here. Thank you so much for listening. 896 00:43:58,480 --> 00:44:00,480 If you're still here, let us know in the comments section 897 00:44:00,480 --> 00:44:02,880 what you thought of this episode and we'll see you in the next 898 00:44:02,880 --> 00:44:03,040 one.