1 00:00:00,040 --> 00:00:02,440 Do you think that's the most effective way to create software 2 00:00:02,440 --> 00:00:05,040 within a team? I think that's the only way, 3 00:00:05,240 --> 00:00:07,560 being the architect, but also work with the teams and 4 00:00:07,840 --> 00:00:10,160 generally be part of the teams and code with them. 5 00:00:10,440 --> 00:00:13,960 They were just copy pasting all these folders and project 6 00:00:13,960 --> 00:00:16,360 structures without actually understanding why they were 7 00:00:16,360 --> 00:00:18,440 doing it. They end up being so dogmatic 8 00:00:18,440 --> 00:00:22,320 about that that they lose the capability of optimizing or 9 00:00:22,320 --> 00:00:25,080 simplifying that code base. I feel there will be a point of 10 00:00:25,080 --> 00:00:28,240 time in the future and I don't have a clue how fast that future 11 00:00:28,240 --> 00:00:30,400 is going to be that we don't care about the code anymore. 12 00:00:30,720 --> 00:00:34,080 If you use open source code, it may eventually become your own 13 00:00:34,080 --> 00:00:38,400 code that you have to maintain. It is almost like you give, but 14 00:00:38,400 --> 00:00:41,560 you get so much back from it. Knowledge, experience, new 15 00:00:41,560 --> 00:00:44,960 tools, new insights, feasibility of course, and that is very 16 00:00:44,960 --> 00:00:47,480 valuable. Today I'm joined by Dennis DOMA, 17 00:00:47,600 --> 00:00:52,080 Microsoft MVP, open source legend and Coding architect over 18 00:00:52,080 --> 00:00:54,960 at Aviva Solutions. Last time we spoke was four 19 00:00:54,960 --> 00:00:56,360 years ago. We talked about quality and 20 00:00:56,360 --> 00:00:59,440 software engineering and a lot has changed, which is exactly 21 00:00:59,440 --> 00:01:01,760 what we talk about. Strong conventions, battling 22 00:01:01,760 --> 00:01:05,200 complexity and the impact of AI on open source. 23 00:01:05,280 --> 00:01:10,720 So enjoy. You mentioned you would never do 24 00:01:10,720 --> 00:01:13,480 conferences full time because you love your job too much. 25 00:01:13,960 --> 00:01:17,560 Also because I Nah why would I do this full time? 26 00:01:17,560 --> 00:01:19,560 There's no point. I still want to be able to do 27 00:01:19,600 --> 00:01:22,880 things myself. I'm, I'd like to call myself a 28 00:01:23,440 --> 00:01:25,480 hands on architect or coding architect. 29 00:01:26,000 --> 00:01:31,080 And almost everything I talk about is based or fuelled by my 30 00:01:31,080 --> 00:01:33,560 experiences. And if you stop coding, if you 31 00:01:33,560 --> 00:01:35,800 actually stop building production systems and put them, 32 00:01:36,040 --> 00:01:38,480 you know, put them in on the Internet and everything, you 33 00:01:38,480 --> 00:01:40,200 lose that, that, that experience. 34 00:01:40,200 --> 00:01:43,880 I mean, let's imagine I would have stopped coding myself for 35 00:01:43,880 --> 00:01:46,280 the last three years, then I would have probably missed the 36 00:01:46,280 --> 00:01:50,400 whole AI scene. And I really, I'm 52 now. 37 00:01:50,400 --> 00:01:52,240 I've been doing this for almost 30 years. 38 00:01:52,720 --> 00:01:56,000 I can't even imagine stopping this for some reason. 39 00:01:56,000 --> 00:02:00,240 I've always been lucky where my job or my assignment is usually 40 00:02:00,240 --> 00:02:03,240 a combination of being the architect, but also work with 41 00:02:03,240 --> 00:02:06,560 the teams and generally be part of the teams and code with them. 42 00:02:06,760 --> 00:02:09,440 Putting stuff in production. You learn so much from that and 43 00:02:09,759 --> 00:02:14,040 being able to demonstrate that to developers like OK, if you 44 00:02:14,040 --> 00:02:17,000 actually take a slightly different approach, this will 45 00:02:17,000 --> 00:02:18,680 actually make your code more maintainable. 46 00:02:18,960 --> 00:02:22,240 That is so valuable, but also running into the limitations or 47 00:02:22,240 --> 00:02:25,960 learning new things about the framework or new techniques or a 48 00:02:25,960 --> 00:02:28,960 new insight or a new paradigm or something like that. 49 00:02:29,280 --> 00:02:30,840 I think that's extremely valuable. 50 00:02:30,840 --> 00:02:32,640 And I usually only talk about my experience. 51 00:02:32,640 --> 00:02:35,880 I don't talk about what's new in dot net 10 or something like 52 00:02:35,880 --> 00:02:36,800 that. There's plenty of people that 53 00:02:36,800 --> 00:02:39,400 can do that. I I like to talk about my 54 00:02:39,400 --> 00:02:43,520 experiences, the pains and the successes that I've experienced 55 00:02:43,520 --> 00:02:45,680 during my day job. Day job. 56 00:02:45,960 --> 00:02:49,320 This responsibility of also being responsible for 57 00:02:49,320 --> 00:02:51,400 architecture but still being hands on, do you think that's 58 00:02:51,400 --> 00:02:54,120 the most effective way to create software within a team? 59 00:02:54,440 --> 00:02:57,760 I think that's the only way. I can't imagine. 60 00:02:58,200 --> 00:03:02,000 I mean, I do work with quite a lot of like, let's say 61 00:03:02,040 --> 00:03:06,200 enterprise architects in the companies that I work with, but 62 00:03:06,200 --> 00:03:08,440 they deal with the high level stuff they talk about. 63 00:03:08,440 --> 00:03:11,320 They, they deal with policies and maybe information 64 00:03:11,320 --> 00:03:14,840 architecture, but they don't actually have the experience 65 00:03:14,840 --> 00:03:17,800 anymore to judge whether a certain solution will work 66 00:03:17,800 --> 00:03:20,520 correctly or what's the consequence of using a 67 00:03:20,680 --> 00:03:24,120 particular direction on the long term maintainability and 68 00:03:24,120 --> 00:03:29,520 sustainability of a solution. So for me being an architect, I 69 00:03:29,520 --> 00:03:31,360 mean, if if you're a software architect, I mean, there are 70 00:03:31,360 --> 00:03:33,520 different types of architecture, there's a software architect. 71 00:03:34,200 --> 00:03:37,400 The only way to be effective is to actually know what you're 72 00:03:37,400 --> 00:03:42,360 doing, experiment, see how a particular solution that you 73 00:03:42,360 --> 00:03:46,000 imagined is being used in practice. 74 00:03:47,360 --> 00:03:49,400 Yeah. I mean, just fixing bugs in an 75 00:03:49,400 --> 00:03:52,240 existing code base is already valuable, very valuable because 76 00:03:52,240 --> 00:03:53,440 you learn a lot of stuff for that. 77 00:03:54,280 --> 00:03:57,200 And I've, I've had, I had quite, I don't know, we're going down 78 00:03:57,200 --> 00:04:00,840 all kinds of directions, but I had quite a lot of experiences 79 00:04:00,840 --> 00:04:04,200 where I joined a client because I'm a consultant. 80 00:04:04,200 --> 00:04:06,440 I work for a small consultancy firm, Afiva. 81 00:04:07,160 --> 00:04:11,920 And you go into a team and you look at the code base and then 82 00:04:11,960 --> 00:04:13,840 you see they're struggling with the code base. 83 00:04:13,840 --> 00:04:17,800 They struggling to keep it maintainable, too much coupling, 84 00:04:17,800 --> 00:04:21,200 not enough cohesion. Well, you know, all these terms 85 00:04:21,200 --> 00:04:24,760 or principles and then you, you look at that code base from a 86 00:04:25,000 --> 00:04:28,760 higher perspective and suddenly you realise like them, that 87 00:04:28,760 --> 00:04:31,200 looks like an, an union architecture, clean 88 00:04:31,200 --> 00:04:33,560 architecture. And then you discover that 89 00:04:33,720 --> 00:04:37,320 apparently somebody in the past had a kind of vision, but the 90 00:04:37,320 --> 00:04:39,440 current development team completely lost track of what 91 00:04:39,440 --> 00:04:42,320 the architecture was about. Like what was the underlying 92 00:04:42,320 --> 00:04:44,440 architectural style? They didn't even know it 93 00:04:44,440 --> 00:04:46,800 anymore. They were just copy pasting all 94 00:04:46,800 --> 00:04:50,120 these folders and the project structures without actually 95 00:04:50,120 --> 00:04:51,440 understanding why they were doing it. 96 00:04:51,880 --> 00:04:54,680 And they, they, they, they end up being so dogmatic about that, 97 00:04:55,320 --> 00:05:00,000 that they lose, yeah, the capability of, of optimizing or 98 00:05:00,000 --> 00:05:02,000 simplifying that code base because they're just copy 99 00:05:02,000 --> 00:05:05,720 pasting a particular pattern. And I think if you're not 100 00:05:05,800 --> 00:05:08,320 actually in that code base, you will not be able to detect, 101 00:05:08,320 --> 00:05:11,440 detect that. And if you're an architect and 102 00:05:11,440 --> 00:05:13,680 you're, you're a Greenfield architect where you, where you 103 00:05:13,680 --> 00:05:18,480 get the time to build up something from scratch, you make 104 00:05:18,480 --> 00:05:22,600 assumptions when you architect, you probably, you know, you know 105 00:05:22,600 --> 00:05:25,240 about the functional non functionals, you talk to the 106 00:05:25,240 --> 00:05:28,200 stakeholders, you try to come up with an architecture that makes 107 00:05:28,200 --> 00:05:31,200 sense. But actually making that a 108 00:05:31,200 --> 00:05:34,040 reality, implementing that. I, I don't believe you can write 109 00:05:34,040 --> 00:05:37,480 it down in such a way that the development team just says, Oh 110 00:05:37,480 --> 00:05:40,560 yeah, we know how to do that because they will make mistakes. 111 00:05:40,560 --> 00:05:42,560 You make mistakes, you make assumptions. 112 00:05:42,840 --> 00:05:46,680 You may actually, you may have thought like, OK, I need a 113 00:05:46,680 --> 00:05:49,360 particular abstraction or I'm going to use, I don't know, 114 00:05:49,360 --> 00:05:54,000 event sourcing or something I've done in the past or CQRS based 115 00:05:54,000 --> 00:05:56,960 on those non functionals. And maybe also based on this, 116 00:05:57,280 --> 00:05:59,480 this gut feeling that you built after 30 years. 117 00:06:00,040 --> 00:06:03,120 And then you discover that either it's too complicated for 118 00:06:03,120 --> 00:06:07,240 the team or it's not. It's not performant enough. 119 00:06:07,240 --> 00:06:11,920 You have to make changes or the team is not capable of making 120 00:06:11,920 --> 00:06:14,720 decisions on their own that they're not experienced enough 121 00:06:14,720 --> 00:06:17,480 to understand when that pattern applies or when not. 122 00:06:17,840 --> 00:06:19,960 That is something you need to experience on the job. 123 00:06:19,960 --> 00:06:22,560 And then reiterate, you know, it's, it's really like agile 124 00:06:22,560 --> 00:06:24,040 development. You look at it, you implement 125 00:06:24,040 --> 00:06:28,360 it, you you reflect on it, you make changes and that's part of 126 00:06:28,360 --> 00:06:30,680 the job. Yeah, I truly believe that 127 00:06:30,880 --> 00:06:34,440 software engineering or creating software how it is or how it 128 00:06:34,440 --> 00:06:38,120 used to be as well as how it will be in the future is all 129 00:06:38,120 --> 00:06:41,360 about paddling or maintaining kind of this balance between 130 00:06:41,360 --> 00:06:43,840 what you have and the complexity that you build up. 131 00:06:44,200 --> 00:06:46,480 So if you lose a fundamental understanding of certain 132 00:06:46,480 --> 00:06:49,840 decisions that have been made in the past and you're dogmatic 133 00:06:49,840 --> 00:06:52,320 about certain decisions and you don't really understand, but you 134 00:06:52,320 --> 00:06:55,520 keep doing it conventionally, you might build up a lot of 135 00:06:55,760 --> 00:06:57,880 complexity. Or you should be actually 136 00:06:57,880 --> 00:07:00,560 keeping that in mind and kind of battling it as you go. 137 00:07:00,920 --> 00:07:04,440 Because time is the factor here and complexity is just to it can 138 00:07:04,440 --> 00:07:05,960 be make or break for your software at some. 139 00:07:06,040 --> 00:07:08,800 Point and and if you, if you detect that, if you acknowledge 140 00:07:08,800 --> 00:07:12,840 the fact that you have unnecessary complexity, some, 141 00:07:12,840 --> 00:07:15,400 some people actually call it accidental complexity because 142 00:07:15,400 --> 00:07:18,120 everybody's trying to do their best, but they sometimes don't 143 00:07:18,120 --> 00:07:21,080 realize they make, they, they introduce something that makes 144 00:07:21,080 --> 00:07:23,960 the code base more complicated. And if you don't need that, you 145 00:07:23,960 --> 00:07:27,080 need to continuously invest in that, try to remove it. 146 00:07:27,880 --> 00:07:31,640 But then usually you end up in in this discussion like, OK, you 147 00:07:31,640 --> 00:07:33,840 know, we have built this system, different modules. 148 00:07:34,080 --> 00:07:37,520 This module has been has become too complicated, but it does 149 00:07:37,520 --> 00:07:40,760 follow the same pattern. Should we then remove the 150 00:07:40,760 --> 00:07:42,360 pattern and make this module simpler? 151 00:07:42,520 --> 00:07:44,000 Yeah. But then this code base becomes 152 00:07:44,000 --> 00:07:46,000 inconsistent. You know, that that is an all 153 00:07:46,000 --> 00:07:48,440 sort of thing. And then some developers will 154 00:07:48,440 --> 00:07:51,960 say or some experienced tech leads will say, yeah, but I 155 00:07:51,960 --> 00:07:54,640 prefer the consistency. So let's keep that complexity 156 00:07:54,960 --> 00:07:58,800 because for aspiring developers, it might be easy to understand. 157 00:07:58,800 --> 00:08:01,560 I'm more the person that will say, no, no, we treat this 158 00:08:01,560 --> 00:08:05,880 module as a boundary. And the consistency that we want 159 00:08:05,880 --> 00:08:10,320 to keep applies only to this boundary, this module. 160 00:08:10,320 --> 00:08:14,000 So we keep, we make this module simpler because I know that if 161 00:08:14,000 --> 00:08:16,640 you make it simpler, it will also be much easier to maintain. 162 00:08:17,200 --> 00:08:19,320 You know, a lot of developers are struggling like, OK, what 163 00:08:19,320 --> 00:08:20,840 does, why do I need this abstraction? 164 00:08:20,840 --> 00:08:22,920 Why do I, how do I unit test this? 165 00:08:22,920 --> 00:08:24,040 What's the scope of the unit test? 166 00:08:24,040 --> 00:08:25,720 And we talked about this, I think 4 years ago. 167 00:08:26,200 --> 00:08:29,280 This is still the thing, you know, how do you apply DRY? 168 00:08:29,800 --> 00:08:32,520 Do you actually duplicate code? I'm a person that will duplicate 169 00:08:32,520 --> 00:08:35,320 a lot of code because it reduces coupling. 170 00:08:35,720 --> 00:08:37,760 But I do that consciously. It's not like, OK, I duplicate 171 00:08:37,760 --> 00:08:39,760 everything. Well, I never duplicate. 172 00:08:39,760 --> 00:08:43,840 And that is the challenge that I see a lot of developers, 173 00:08:44,800 --> 00:08:47,800 especially if they reach the, I don't know if I'm generalizing, 174 00:08:47,800 --> 00:08:52,040 but between 10:15 years experience, they start to learn 175 00:08:52,040 --> 00:08:55,560 the patterns, they start to appreciate them, but then they 176 00:08:55,560 --> 00:08:58,960 start to over overzealously apply them and stick to that. 177 00:08:58,960 --> 00:09:02,360 And and if you then say, Abbott, we don't need this abstraction. 178 00:09:02,560 --> 00:09:06,040 Yeah, but this is this is open for extension, close for 179 00:09:06,040 --> 00:09:08,080 modification. You know, this is a single 180 00:09:08,080 --> 00:09:09,720 responsibility principle. Yeah, but you don't need this 181 00:09:09,720 --> 00:09:11,360 abstraction. Yeah, but it's ready for the 182 00:09:11,360 --> 00:09:13,400 future. You don't need this abstraction. 183 00:09:14,400 --> 00:09:17,520 I've seen this so many times. And it's impossible to 184 00:09:17,520 --> 00:09:20,520 experience them because those developers will then, you know, 185 00:09:20,520 --> 00:09:23,000 wave the SOLID principles in your face like, you know. 186 00:09:23,320 --> 00:09:25,080 Yeah. But dude, you don't need this 187 00:09:25,080 --> 00:09:27,720 complexity right now. You know, do you actually need 188 00:09:27,720 --> 00:09:29,960 to support different databases? No. 189 00:09:30,360 --> 00:09:32,160 But we may have to do that in the future. 190 00:09:32,560 --> 00:09:34,040 Yeah. But do you actually believe that 191 00:09:34,040 --> 00:09:37,200 you can create an abstraction that has where the semantics 192 00:09:37,200 --> 00:09:39,040 don't even change between different databases? 193 00:09:39,480 --> 00:09:41,240 I don't know. But I've defined, you know, I've 194 00:09:41,240 --> 00:09:43,920 defined this abstraction. I've decoupled my code from the 195 00:09:43,920 --> 00:09:45,040 database. Yeah. 196 00:09:45,040 --> 00:09:48,800 But I. Mean it would be easy if there's 197 00:09:48,800 --> 00:09:50,720 like a playbook or a rule book, right? 198 00:09:50,720 --> 00:09:53,440 We followed this. If we just follow this and we 199 00:09:53,440 --> 00:09:56,080 keep following this, then everything is like really good. 200 00:09:56,120 --> 00:09:59,560 It's. Called Sequel and every database 201 00:09:59,800 --> 00:10:01,520 implements the exact same no of course. 202 00:10:01,600 --> 00:10:04,560 Yeah, no, like it's, it's, I would love that. 203 00:10:04,560 --> 00:10:06,520 I mean, no, I don't think I would love that because I think 204 00:10:06,520 --> 00:10:07,920 that would make things boring as well. 205 00:10:08,440 --> 00:10:10,600 We can use. AI then to implement it. 206 00:10:10,600 --> 00:10:12,080 In fact, we can already do that, but. 207 00:10:12,080 --> 00:10:14,520 Yeah. I mean, this is for me the, the 208 00:10:14,520 --> 00:10:16,960 interesting part in those conversations about kind of 209 00:10:17,440 --> 00:10:20,520 conventions versus being pragmatic versus, OK, what do we 210 00:10:20,520 --> 00:10:23,680 do for the future or what do we do for now and how much is it 211 00:10:23,680 --> 00:10:26,360 going to speed us up versus how much complexity is going to add 212 00:10:26,360 --> 00:10:29,120 that we have to navigate. I've always really enjoyed those 213 00:10:29,120 --> 00:10:31,600 conversations. And now with AI assisted 214 00:10:31,600 --> 00:10:34,120 development tooling, I wonder if we're going to have 215 00:10:34,600 --> 00:10:36,440 conversations in a different matter. 216 00:10:36,640 --> 00:10:40,920 Because how we do things, I feel like is accelerating to a point 217 00:10:40,920 --> 00:10:43,720 where it might be really good and we might not really have to 218 00:10:43,720 --> 00:10:47,680 care about how things are done if it still is maintainable in 219 00:10:47,680 --> 00:10:52,920 the end. I think I, yeah, realistically, 220 00:10:52,920 --> 00:10:55,720 I feel there will be a point of time in the future and I don't 221 00:10:55,720 --> 00:10:58,200 have a clue how how fast that future is going to be. 222 00:10:59,040 --> 00:11:00,840 Is that we don't care about the code anymore. 223 00:11:01,720 --> 00:11:04,880 I mean, I don't know if you ever watch Star Trek TV series or 224 00:11:04,880 --> 00:11:06,760 anything like that, but they don't code. 225 00:11:07,040 --> 00:11:11,640 They either tell the computer to do something or they, you know, 226 00:11:11,800 --> 00:11:14,120 press some buttons and suddenly everything works differently. 227 00:11:14,120 --> 00:11:15,800 Of course, that was based on the past. 228 00:11:15,800 --> 00:11:18,520 They probably the original authors wouldn't even be able to 229 00:11:18,520 --> 00:11:22,360 imagine what we can do today. But I, I'm approaching this 230 00:11:22,360 --> 00:11:27,560 point pretty quickly. I have, last September, I think 231 00:11:27,960 --> 00:11:31,000 I've build a new open source project, some kind of HTP 232 00:11:31,000 --> 00:11:36,160 mocking libraryfor.net. And I started with like at a 233 00:11:36,160 --> 00:11:37,760 client project. We had to do a lot of HTTP 234 00:11:37,760 --> 00:11:40,840 marketing because we were doing unit testing and at the right 235 00:11:40,840 --> 00:11:42,680 scale, I would say. And I was using some other 236 00:11:42,680 --> 00:11:44,280 libraries and I didn't like them. 237 00:11:44,280 --> 00:11:47,280 I started building abstractions on top of them to make them 238 00:11:47,280 --> 00:11:49,400 easier to work with. I was missing features. 239 00:11:49,400 --> 00:11:51,920 So at some point I thought like, OK, you know what? 240 00:11:52,080 --> 00:11:54,200 I've been an open source developer for 15 years. 241 00:11:55,160 --> 00:11:57,800 I can probably build something similar, maybe even better. 242 00:11:58,320 --> 00:12:00,880 Arrogant. I don't know naive myself. 243 00:12:00,880 --> 00:12:03,440 So I started collecting all these notes in GitHub in an 244 00:12:03,440 --> 00:12:05,280 issue. I created the repo, it was 245 00:12:05,280 --> 00:12:08,400 empty, almost empty. I used some kind of template 246 00:12:08,400 --> 00:12:12,160 that I built myself and at some point the issue in GitHub 247 00:12:12,160 --> 00:12:15,360 contained basically all the features that are needed, plus a 248 00:12:15,360 --> 00:12:17,880 couple of examples like I I actually thought about what the 249 00:12:17,880 --> 00:12:22,240 API should look like and just out of fun the curiosity, I 250 00:12:22,240 --> 00:12:25,840 decided to assign that GitHub issue to GitHub copilot. 251 00:12:25,840 --> 00:12:30,320 You can do that inside GitHub. Yeah, to my surprise, it 252 00:12:30,320 --> 00:12:35,800 produced a beautiful open source project implementing basically 253 00:12:35,800 --> 00:12:39,480 all the features that I needed using my own conventions because 254 00:12:39,480 --> 00:12:43,040 I already had set up editor configs and Roslin analysis so 255 00:12:43,320 --> 00:12:46,360 to define what my coding conventions are and it was 256 00:12:46,360 --> 00:12:50,040 exactly what I needed. After that I took the code. 257 00:12:50,040 --> 00:12:53,400 Oh, no, actually when the when git up created the pool request, 258 00:12:53,880 --> 00:12:56,240 I started, you know, asking additional questions. 259 00:12:56,240 --> 00:12:59,280 So like I was looking at some of the unit tests it created based 260 00:12:59,280 --> 00:13:02,800 on my examples and I thought like, actually, you know, this, 261 00:13:02,800 --> 00:13:05,360 there's some additional steps. I want you to remove this. 262 00:13:05,840 --> 00:13:09,720 And oh, by the way, can you also create some assertion methods as 263 00:13:09,720 --> 00:13:13,000 part of the library? So it's easier to say, OK, did 264 00:13:13,000 --> 00:13:16,040 the mock actually send or was there actually a request to a 265 00:13:16,040 --> 00:13:18,800 particular endpoint with a particular route and particular 266 00:13:18,800 --> 00:13:20,840 body? And I thought about it and then 267 00:13:20,840 --> 00:13:24,320 I thought like, OK, can you, maybe, you know, I have an open 268 00:13:24,320 --> 00:13:25,800 source project called Fluid assertions. 269 00:13:25,800 --> 00:13:28,800 There's now 2 versions. Version 7 is still open source 270 00:13:28,800 --> 00:13:31,240 fully version A's is like a dual license. 271 00:13:31,240 --> 00:13:35,200 And I wanted to support both. So I, I left a, a pull request 272 00:13:35,200 --> 00:13:38,480 comment, like, can you actually create 2 sets of assertions, one 273 00:13:38,480 --> 00:13:40,960 for seven and one of eight? And it did that. 274 00:13:41,440 --> 00:13:44,240 And then I said, like, OK, can you actually split that up into 275 00:13:44,240 --> 00:13:45,760 projects? So I can create 2 new cat 276 00:13:45,760 --> 00:13:47,600 packages from it. It did it as well. 277 00:13:48,360 --> 00:13:51,400 And I said, can you also update to read me with some examples? 278 00:13:51,400 --> 00:13:53,000 And it was incredible what it did. 279 00:13:53,600 --> 00:13:56,480 And that's the point when I thought, OK, now I have the, 280 00:13:57,640 --> 00:13:59,480 the, the bulk of the implementation. 281 00:13:59,840 --> 00:14:01,440 Now I'm going to continue myself. 282 00:14:02,080 --> 00:14:04,200 And from that point on, I started to treat that code as 283 00:14:04,200 --> 00:14:09,520 mine and started to evolve. But still, I sometimes use 284 00:14:09,520 --> 00:14:12,640 Copilot in GitHub to add new features like I create an issue, 285 00:14:12,880 --> 00:14:16,280 assign it to it, but sometimes I do it from my IDE. 286 00:14:16,320 --> 00:14:21,240 So I I now have a mix of both. What's interesting is that I now 287 00:14:21,240 --> 00:14:26,360 have codes in my library which doesn't entirely meet my 288 00:14:26,360 --> 00:14:29,880 standards. Copilot tends to create pretty 289 00:14:29,880 --> 00:14:33,160 long methods, especially if there's a lot of sequential 290 00:14:33,160 --> 00:14:37,600 related to logic, and I've kind of accepted that because I 291 00:14:37,600 --> 00:14:41,080 maintain it myself. I've accepted the fact that this 292 00:14:41,080 --> 00:14:44,360 code will be there for a while until the point that I'm going 293 00:14:44,360 --> 00:14:47,600 to touch that and I'm going to refactor it to make to make sure 294 00:14:47,600 --> 00:14:51,000 it follows my standards. So I kind of embrace this. 295 00:14:51,000 --> 00:14:53,360 So I have copilot that generates implementation. 296 00:14:53,760 --> 00:14:57,120 I sometimes use copilot in GitHub to add new features. 297 00:14:58,280 --> 00:15:02,200 Then I go back to writers my my preferred editor and there I use 298 00:15:02,200 --> 00:15:06,840 Juni and Claude like a mix, mix and match to also make changes 299 00:15:06,840 --> 00:15:08,880 for me and a lot of things I do myself. 300 00:15:10,040 --> 00:15:12,120 I think this is quite interesting because I feel like 301 00:15:12,120 --> 00:15:15,160 this also highlights then still the pragmatic side of things, 302 00:15:15,160 --> 00:15:16,600 right? You could have said, OK, this 303 00:15:16,600 --> 00:15:18,800 doesn't follow my standards, therefore I'm going to change 304 00:15:18,800 --> 00:15:20,520 it. But the pragmatic thing is, if 305 00:15:20,520 --> 00:15:24,240 it works, if all the tests a green basically, then in the end 306 00:15:24,240 --> 00:15:26,880 it does what it needs to do and I don't really have to touch it, 307 00:15:26,880 --> 00:15:28,200 right? This is more of a subjective 308 00:15:28,200 --> 00:15:31,160 thing, or it's becoming more of a more of a subjective. 309 00:15:31,160 --> 00:15:33,240 It is, it is. And of course, the code is not 310 00:15:33,240 --> 00:15:34,560 bad. I mean, it does follow. 311 00:15:34,640 --> 00:15:37,520 I've set up all kinds of conventions so it works and I'm 312 00:15:37,520 --> 00:15:40,080 still experimenting, for example, with instructions files 313 00:15:40,240 --> 00:15:42,080 and now we have skills. I haven't tried that. 314 00:15:42,080 --> 00:15:43,480 It's to my list to try that as well. 315 00:15:43,960 --> 00:15:47,080 Yeah, I I've read about it. I still haven't figured out how 316 00:15:47,080 --> 00:15:51,440 would I use this with for example a jet brace Juni also 317 00:15:51,560 --> 00:15:53,040 because I switch agents all the time. 318 00:15:53,040 --> 00:15:55,200 Within writer you can you can choose what you want and then 319 00:15:55,200 --> 00:15:58,200 you have 15 different LMS to choose from. 320 00:15:58,200 --> 00:15:59,400 So it's. It's a lot. 321 00:15:59,720 --> 00:16:02,760 It's everyday is like a surprise, like what will happen? 322 00:16:03,920 --> 00:16:06,320 What was I saying? But I do review the tests. 323 00:16:06,680 --> 00:16:09,760 So even if copilot or an AI generate some of the code, 324 00:16:10,000 --> 00:16:12,640 including the test, I do review the test to make sure like OK, 325 00:16:12,960 --> 00:16:15,480 do I understand the test? Do they make, do they make 326 00:16:15,480 --> 00:16:17,600 sense? Do they actually test what I 327 00:16:17,600 --> 00:16:19,360 expect? And I use that also to look at 328 00:16:19,360 --> 00:16:21,800 the behaviour. So is the implementation 329 00:16:21,800 --> 00:16:25,720 actually doing what I expected? Does the API look the way I want 330 00:16:25,720 --> 00:16:28,400 it to be? And then I continue to go back 331 00:16:28,400 --> 00:16:29,760 and forth. But whenever I these days, 332 00:16:29,760 --> 00:16:33,240 whenever I don't know how we end on this discussion, but whenever 333 00:16:33,240 --> 00:16:37,160 I feel like OK this is repetitive and I think Journey 334 00:16:37,160 --> 00:16:39,720 can do better or cloud, then I'll let it do it. 335 00:16:39,800 --> 00:16:41,960 OK, that's the. That's the rule. 336 00:16:41,960 --> 00:16:44,760 That's, that's kind of how I do it now these days, even for the 337 00:16:44,760 --> 00:16:49,160 projects like this morning, I even asked the, the, the, the AI 338 00:16:49,160 --> 00:16:52,400 to generate some additional tests to make sure coverage was 339 00:16:52,400 --> 00:16:54,720 complete. But then I went back and then 340 00:16:54,720 --> 00:16:58,320 reviewed every test to make sure they did what they did or they 341 00:16:58,320 --> 00:17:02,240 did what they're supposed to do. And also to, to actually to use 342 00:17:02,240 --> 00:17:05,760 the same patterns that I liked. And now what I do is then the 343 00:17:05,760 --> 00:17:08,680 first thing I ask them again to make some adjustments, but also 344 00:17:08,680 --> 00:17:11,200 at the same time update the instructions file to make sure 345 00:17:11,200 --> 00:17:13,720 the next time it will follow this more closely. 346 00:17:14,240 --> 00:17:16,400 That's not a guarantee, but that's the whole way I work now, 347 00:17:16,520 --> 00:17:18,800 yeah. Yeah, I, I have been in teams 348 00:17:18,800 --> 00:17:23,960 where during refinement the team preference is to completely fill 349 00:17:23,960 --> 00:17:27,680 in a solution direction in the task, in the ticket itself with 350 00:17:27,680 --> 00:17:30,120 regards to what are we going to change, kind of where is it? 351 00:17:30,120 --> 00:17:32,880 These are the conventions, these are the patterns making a lot of 352 00:17:32,880 --> 00:17:36,080 things explicit and without AI assist development tooling. 353 00:17:36,400 --> 00:17:38,520 I did not like this way of working, but the only thing you 354 00:17:38,520 --> 00:17:40,560 would have to do is pick up the ticket, read what has to be 355 00:17:40,560 --> 00:17:43,920 done, and then execute. And I feel like nowadays the 356 00:17:43,920 --> 00:17:47,680 execution side of things, that's accelerating, so you don't have 357 00:17:47,680 --> 00:17:50,480 to necessarily do all the execution yourself anymore. 358 00:17:50,760 --> 00:17:53,040 Which brings an interesting question for me, because build 359 00:17:53,040 --> 00:17:55,840 is accelerating. I see tools pop up out of the 360 00:17:55,840 --> 00:17:58,840 ground like mushrooms, basically out of nothing, out of thin air. 361 00:17:59,560 --> 00:18:03,080 And my assumption is that research and planning, figuring 362 00:18:03,080 --> 00:18:05,320 out what problems you need to solve is going to become more 363 00:18:05,320 --> 00:18:07,880 and more important. Build is going to accelerate so 364 00:18:07,880 --> 00:18:10,880 we can build really fast. And then distribution and 365 00:18:10,880 --> 00:18:14,120 adoption is going to be more and more important than it now is. 366 00:18:14,360 --> 00:18:16,800 Because if you build a tool like that, like it's just a finger 367 00:18:16,800 --> 00:18:19,640 snap, how are you going to get your users Or what problems does 368 00:18:19,640 --> 00:18:22,520 it actually solve if they can also build a tool for themselves 369 00:18:22,520 --> 00:18:25,000 like you did with a mocking library, and then solve their 370 00:18:25,000 --> 00:18:30,440 own problems by themselves? Yeah, that is something I'm 371 00:18:30,440 --> 00:18:33,880 wondering, like, is it still worthwhile building open source 372 00:18:33,880 --> 00:18:37,200 projects or something like that? But in the end, yeah, not 373 00:18:37,200 --> 00:18:41,600 everybody has that capability. I think we sometimes forget that 374 00:18:41,880 --> 00:18:45,400 we live in a bubble. And especially if I look at 375 00:18:45,400 --> 00:18:48,800 myself, I, I, I'm very lucky that I get to speak at, at 376 00:18:48,800 --> 00:18:51,120 different conferences. And I'm part of an open source 377 00:18:51,120 --> 00:18:54,800 community that, you know, a lot of people that are in the same 378 00:18:55,160 --> 00:18:59,280 space, which are always working, are always trying to be at the 379 00:18:59,280 --> 00:19:03,240 brink of technology, doing what they can to learn as much as 380 00:19:03,240 --> 00:19:05,760 possible. And also use conferences to be 381 00:19:05,760 --> 00:19:08,520 able to speak to other people. I mean, a lot of discussions are 382 00:19:08,520 --> 00:19:11,040 happening, you know, during the, the speakers dinner and the 383 00:19:11,040 --> 00:19:13,480 evenings where you sit with people that are very 384 00:19:13,480 --> 00:19:16,320 knowledgeable in their fields and you learn a lot of that. 385 00:19:16,640 --> 00:19:19,040 And sometimes you forget there's a lot of people, a lot of 386 00:19:19,040 --> 00:19:22,280 developers in our profession that don't have that luxury or 387 00:19:22,280 --> 00:19:23,840 don't even know how to get that knowledge. 388 00:19:23,840 --> 00:19:26,520 They, they just need to get it from doing the work. 389 00:19:26,520 --> 00:19:28,800 I mean, I've seen a lot of developers that are completely 390 00:19:28,800 --> 00:19:31,840 stuck because they work at the company where there's nobody 391 00:19:31,840 --> 00:19:33,840 that actually brings in that knowledge and they don't have 392 00:19:33,840 --> 00:19:37,960 that, that, that passion or, or drive themselves to fetch 393 00:19:37,960 --> 00:19:41,240 information. So what I'm trying to say is 394 00:19:41,240 --> 00:19:45,320 that we're probably doing it not just for, for us like the people 395 00:19:45,320 --> 00:19:47,720 in our little bubble, we're doing for the rest of the, the 396 00:19:47,720 --> 00:19:52,080 people that are doing it. And for example, like Mockley, I 397 00:19:52,080 --> 00:19:55,360 basically told like in my client projects, the teams like it, 398 00:19:55,360 --> 00:19:57,160 they enjoy how to use it or they use it. 399 00:19:57,160 --> 00:20:00,600 And now we just use the AI to say, OK, rewrite some of these 400 00:20:00,600 --> 00:20:03,680 tests, replace all the old mockling stuff and replace with 401 00:20:03,680 --> 00:20:07,760 Mockli, or create some tests for this Azure function that 402 00:20:07,760 --> 00:20:11,520 interacts with, I don't know, some external surface and use 403 00:20:11,520 --> 00:20:13,840 that library to implement it. And it will implement that 404 00:20:14,320 --> 00:20:16,920 because it's still the whole goal that I always have is to 405 00:20:16,920 --> 00:20:18,800 make my code as readable as possible. 406 00:20:19,600 --> 00:20:22,240 And you probably know from the past, I really enjoy test driven 407 00:20:22,240 --> 00:20:24,560 development, the unit testing. And one of the things that I'm 408 00:20:24,560 --> 00:20:27,720 trying to do is make sure that my tests are as self-explanatory 409 00:20:27,720 --> 00:20:30,200 as possible. They don't contain noise. 410 00:20:30,200 --> 00:20:33,880 I have all kinds of experiences and principles that I follow to 411 00:20:33,880 --> 00:20:36,280 make my make sure my test actually can be used as 412 00:20:36,280 --> 00:20:39,120 documentation. So using libraries that will 413 00:20:39,120 --> 00:20:43,000 make these things more explicit without really convoluted calls 414 00:20:43,000 --> 00:20:46,240 with mocking where you build up Jason strings and stuff, which 415 00:20:46,600 --> 00:20:49,440 is not always useful. That's basically drive. 416 00:20:49,440 --> 00:20:52,160 So I'm basically doing it for myself now that I think about 417 00:20:52,160 --> 00:20:53,760 it. I'm building these things for 418 00:20:53,760 --> 00:20:55,480 myself. And then I think like, OK, let's 419 00:20:55,480 --> 00:20:58,480 pull it out of there, put it out there and see if it's actually 420 00:20:58,480 --> 00:21:00,160 pick picks up if people enjoy it. 421 00:21:00,920 --> 00:21:04,520 I mean, I, I resonate with that. I used to take a lot of pride in 422 00:21:04,520 --> 00:21:07,320 kind of how my tests are organised and how my tests are 423 00:21:07,320 --> 00:21:10,000 structured, not just for me, but for my future self, for my 424 00:21:10,000 --> 00:21:13,000 future colleagues. I try and add as much context 425 00:21:13,000 --> 00:21:16,600 that I can in pull requests, in commit messages, individual and 426 00:21:16,600 --> 00:21:18,880 atomic commits that makes sense and are coherent. 427 00:21:19,280 --> 00:21:23,440 I spent extra effort on that. And I feel like nowadays when 428 00:21:23,440 --> 00:21:26,200 you're looking at your your tests and the quality that you 429 00:21:26,200 --> 00:21:28,480 put in your code base, that is kind of your safety net. 430 00:21:28,880 --> 00:21:32,560 And if you black box the software that you create and the 431 00:21:32,560 --> 00:21:35,440 safety net is your tests, then definitely you should spend 432 00:21:35,440 --> 00:21:37,800 extra attention on that and make sure those are robust because 433 00:21:37,800 --> 00:21:40,600 then you can trust whatever is behind it, whether it's created 434 00:21:40,600 --> 00:21:43,720 by human, by you, not by you or by an agent. 435 00:21:43,720 --> 00:21:46,400 In the end, it doesn't matter. So I feel like this is going to 436 00:21:46,400 --> 00:21:49,960 be quite crucial to make sure that your tests make sense, you 437 00:21:49,960 --> 00:21:51,760 know what to test and also how to test. 438 00:21:52,000 --> 00:21:54,120 I feel like it's going to come more towards the forefront. 439 00:21:54,120 --> 00:21:56,880 Some people are already front loading that testing part and 440 00:21:56,880 --> 00:21:59,240 then just generating based on the tests and the 441 00:21:59,240 --> 00:22:00,720 specifications. It's also a way of doing it. 442 00:22:00,760 --> 00:22:05,560 Yeah, until the point where we really don't have to write that 443 00:22:05,560 --> 00:22:08,000 code anymore. Like if you look at open spec. 444 00:22:08,040 --> 00:22:11,120 Yeah, that's pretty amazing. What I've learned from a 445 00:22:11,120 --> 00:22:14,600 colleague has been demonstrating that during an AI evening we had 446 00:22:15,000 --> 00:22:17,760 is that it's really great for starting something up, but at 447 00:22:17,760 --> 00:22:20,560 some point still you need to take that code and treat it as 448 00:22:20,560 --> 00:22:22,960 yourself. Which by the way is how I treat 449 00:22:22,960 --> 00:22:26,960 all AI generated code. But I'm pretty sure I mean like 450 00:22:26,960 --> 00:22:30,880 what is it 2026 now? When was when became ChatGPT 451 00:22:30,880 --> 00:22:33,320 popular? I even lost track like. 2-3 end 452 00:22:33,320 --> 00:22:36,080 of 2023 ish I think. It's only two years ago, two 453 00:22:36,080 --> 00:22:37,080 years. Ago end of 2020. 454 00:22:37,080 --> 00:22:39,440 Three and just imagine where we're now it's quick. 455 00:22:39,640 --> 00:22:42,680 It's, and I'm not sure if you talk to, I intended to talk 456 00:22:42,680 --> 00:22:46,320 during hello stuff anger conference by Richard Campbell 457 00:22:46,320 --> 00:22:50,680 from Doner rocks and he kind of told us that like the technology 458 00:22:50,680 --> 00:22:53,640 for AI has kind of reached its potential and the, and the fact 459 00:22:53,840 --> 00:22:55,280 no, not they didn't reach its potential. 460 00:22:55,280 --> 00:23:00,200 It reached the roof. Like they can't really, they can 461 00:23:00,200 --> 00:23:03,480 still optimize the GPT and the, the, the machine learning and 462 00:23:03,480 --> 00:23:08,120 the, the generative AI technology, but there's no, they 463 00:23:08,120 --> 00:23:11,760 really have to make some kind of paradigm shift to move it 464 00:23:11,760 --> 00:23:13,960 further. That's his statement right now. 465 00:23:13,960 --> 00:23:16,320 And you still see improvements, but it's not this big anymore. 466 00:23:16,600 --> 00:23:19,640 It's a tooling around that tries to, you know, it becomes faster 467 00:23:19,640 --> 00:23:22,920 and more powerful, But I don't actually understand how AI 468 00:23:22,920 --> 00:23:24,520 works. Honestly, I have no clue. 469 00:23:25,320 --> 00:23:28,560 I mean I tried to deep dive but then I was like am I doing this 470 00:23:28,560 --> 00:23:30,880 out of curiosity? Out of interest, what problem am 471 00:23:30,880 --> 00:23:32,840 I solving with this? Is just knowledge to have the 472 00:23:32,840 --> 00:23:35,520 knowledge. I see the tools, I can 473 00:23:35,520 --> 00:23:38,080 experiment. I love experimenting in personal 474 00:23:38,080 --> 00:23:41,800 projects, but also if I can in production specifically. 475 00:23:42,160 --> 00:23:45,560 I've had cases where I went fully AI assisted development 476 00:23:45,560 --> 00:23:47,800 workflow and then all of a sudden my context got 477 00:23:47,800 --> 00:23:50,000 compressed. And they always say context is 478 00:23:50,000 --> 00:23:51,520 like milk, never let it compress. 479 00:23:51,520 --> 00:23:54,320 And I, I know why because it compressed the context and it 480 00:23:54,320 --> 00:23:57,120 was like, oh, we have changes, let's merge them and let's just 481 00:23:57,160 --> 00:23:58,720 push them. And I was on the main branch and 482 00:23:58,720 --> 00:24:01,480 I was actually experimenting. So that was not a good idea. 483 00:24:01,880 --> 00:24:03,240 Yeah, we didn't have the safety net. 484 00:24:03,240 --> 00:24:06,400 So I, I've, I've made my mistakes and I've, I have my 485 00:24:06,400 --> 00:24:08,560 learnings, but that's mainly through doing. 486 00:24:08,880 --> 00:24:11,600 I'm having a lot of fun though. Like I was in a role responsible 487 00:24:11,600 --> 00:24:14,560 for product like you mentioned, I need to be hands on to have 488 00:24:14,560 --> 00:24:16,440 things to talk about on this podcast. 489 00:24:16,440 --> 00:24:19,400 And I had a lot of FOMO because things accelerated and I think 490 00:24:19,400 --> 00:24:22,560 they're still accelerating. So yeah, if you're not in an 491 00:24:22,560 --> 00:24:24,560 environment where you can experiment, where you can 492 00:24:24,560 --> 00:24:27,800 actually figure out what works for you in this newer way of 493 00:24:27,800 --> 00:24:31,240 working, it's going to be hard to kind of catch up at some 494 00:24:31,240 --> 00:24:33,920 point. Yeah, kind of similar experience 495 00:24:33,960 --> 00:24:36,600 I had like I think yesterday or the day before I was, I was 496 00:24:36,600 --> 00:24:40,120 using I think clothes within writer to make some code 497 00:24:40,120 --> 00:24:42,760 changes. And I also made some manual code 498 00:24:42,760 --> 00:24:44,600 changes. And I I keep pushing them. 499 00:24:44,680 --> 00:24:47,800 Like I use a couple of git aliases to quickly push 500 00:24:47,800 --> 00:24:50,240 temporary commits to my brand so I don't lose anything. 501 00:24:50,480 --> 00:24:54,200 So whenever I use AII, first commit my changes, push them so 502 00:24:54,200 --> 00:24:58,320 I can roll back and I made some changes, continue to make some 503 00:24:58,320 --> 00:25:00,280 changes. And at some point I was looking 504 00:25:00,280 --> 00:25:02,720 at the actual production code. Well, not production code, but 505 00:25:02,720 --> 00:25:04,280 the, the, the real code, not the tests. 506 00:25:04,960 --> 00:25:07,000 And I was, I was like, Hey, this looks like my previous 507 00:25:07,000 --> 00:25:07,960 implementation. What happened? 508 00:25:07,960 --> 00:25:12,240 And apparently for some reason, Junie or I think it was Claude 509 00:25:12,240 --> 00:25:15,160 made some changes and, and did some of my changes. 510 00:25:15,480 --> 00:25:18,920 And I was like, am I now like I was, I was challenging myself or 511 00:25:18,920 --> 00:25:20,760 doubting myself like did I make a mistake? 512 00:25:21,040 --> 00:25:22,640 Did I excellently forget to push it? 513 00:25:23,240 --> 00:25:26,240 And then I used my knowledge of a git and did a Git revlog. 514 00:25:26,640 --> 00:25:29,160 I started to look at all the commits that were made and I 515 00:25:29,160 --> 00:25:32,920 discovered that two commits before I use claw to make some 516 00:25:32,920 --> 00:25:35,800 changes, pushed it without looking and it actually undid 517 00:25:35,800 --> 00:25:38,560 some of my changes. So you really have to be careful 518 00:25:38,560 --> 00:25:41,960 sometimes and using a branch, a feature branch and pull requests 519 00:25:42,240 --> 00:25:45,480 really help, especially because you can go to git up or if 520 00:25:45,480 --> 00:25:48,160 you're stuck on Azure DevOps and go back and see what you what 521 00:25:48,160 --> 00:25:50,200 you force pushed all the time because I use a lot of force 522 00:25:50,200 --> 00:25:53,000 pushing. And the other thing you said 523 00:25:53,000 --> 00:25:57,600 about experimenting, I'm not the I'm, I don't like experimenting 524 00:25:57,600 --> 00:26:00,280 in the sense that I just start doing something. 525 00:26:01,000 --> 00:26:03,400 You could say I use my open source projects to experiment, 526 00:26:03,400 --> 00:26:05,360 but I'm not going to build something and throw it away. 527 00:26:05,360 --> 00:26:08,240 I don't like that. But I had a discussion or not a 528 00:26:08,240 --> 00:26:11,560 discussion. Like I was talking with one of 529 00:26:11,560 --> 00:26:14,960 my managers at Aviva and he was asking like, yeah, I get all 530 00:26:14,960 --> 00:26:18,680 these questions that people want to use AI and do we want, don't 531 00:26:18,680 --> 00:26:21,040 we want to standardise or something or make a decision 532 00:26:21,040 --> 00:26:24,080 which one we use? And I kind of said like, don't 533 00:26:24,080 --> 00:26:26,960 do that. Let people use whatever AI tool 534 00:26:26,960 --> 00:26:30,760 they want because it changes all the time. 535 00:26:31,280 --> 00:26:34,280 And there may be tools that we didn't know about and can be 536 00:26:34,280 --> 00:26:36,400 very valuable. Let them experiment. 537 00:26:36,400 --> 00:26:40,160 We just track, you know, who's using which and and regularly 538 00:26:40,160 --> 00:26:43,120 chip check up like, OK, are you still using this tool? 539 00:26:43,120 --> 00:26:45,920 Should we extend the license? Because it's not about the cost. 540 00:26:46,440 --> 00:26:49,120 You will earn back that money in in minutes. 541 00:26:49,120 --> 00:26:52,800 You know, I also had some clients like I think at that 542 00:26:52,800 --> 00:26:55,160 time where I tried to decide should we use copilot? 543 00:26:55,160 --> 00:26:57,840 Do we even want to do that? Yeah, but we have 300 544 00:26:57,840 --> 00:26:58,960 developers. Yeah. 545 00:26:58,960 --> 00:27:02,320 So 300 developers times $15.00 a month. 546 00:27:02,920 --> 00:27:06,040 That sounds like a lot of money. If you, if you think like that, 547 00:27:06,840 --> 00:27:11,600 it's $15.00 a month, you earn that back in 5 minutes and 10 548 00:27:11,600 --> 00:27:13,920 fold because it will give you so much power. 549 00:27:14,720 --> 00:27:17,080 But then it's the other side. I think we didn't talk about 550 00:27:17,080 --> 00:27:19,560 that. There's also a lot of developers 551 00:27:19,560 --> 00:27:22,280 junior, I would say aspiring developers, junior sounds so 552 00:27:22,280 --> 00:27:25,760 negative. Inexperienced developers that 553 00:27:25,800 --> 00:27:29,360 either are too inexperienced to, to. 554 00:27:30,200 --> 00:27:32,080 I would say there's 22 versions of them. 555 00:27:32,440 --> 00:27:36,240 You have the people that don't want to use AI because they're 556 00:27:36,240 --> 00:27:39,160 afraid they lose the capability to learn themselves. 557 00:27:39,160 --> 00:27:42,080 They kind of see that. I mean, if you experienced 558 00:27:42,080 --> 00:27:44,840 enough, you've you, you went through that process, you build 559 00:27:44,840 --> 00:27:48,600 the skills on how to find something using Google, but 560 00:27:48,600 --> 00:27:51,680 they, they don't want to use AI because they want to force 561 00:27:51,680 --> 00:27:54,560 themselves to do the investigation and the 562 00:27:54,560 --> 00:27:57,360 experimenting themselves. They do want to make mistakes. 563 00:27:57,640 --> 00:27:59,920 Go to Stack Overflow, although there's not much happening 564 00:27:59,920 --> 00:28:02,080 there. Go to Google, read 565 00:28:02,080 --> 00:28:05,040 documentation, try to interpret them themselves because they're 566 00:28:05,040 --> 00:28:08,360 afraid if they don't, they lose these, these learning skills, 567 00:28:08,360 --> 00:28:10,680 the cognitive skills. And I read somewhere that 568 00:28:10,680 --> 00:28:14,040 there's actually that the current generation has less 569 00:28:14,040 --> 00:28:17,160 cognitive skills than the previous generation because 570 00:28:17,160 --> 00:28:19,520 they're so grown up with, with Chachi Petit and everything. 571 00:28:21,200 --> 00:28:26,760 And you have like, relatively inexperienced developers that 572 00:28:26,880 --> 00:28:31,600 are using AI everywhere they can, but they miss the skills or 573 00:28:31,600 --> 00:28:35,800 the experience to judge whether what the AI generated actually 574 00:28:35,800 --> 00:28:38,280 makes sense. And I've seen that. 575 00:28:38,280 --> 00:28:42,000 I've seen developers creating a pool quest with some changes. 576 00:28:42,600 --> 00:28:45,240 And I was looking at it and I thought, like, does that even 577 00:28:45,240 --> 00:28:46,920 work? And he said, yeah, the test 578 00:28:46,920 --> 00:28:49,200 work, but do you understand how it works? 579 00:28:50,120 --> 00:28:52,520 No. So you used AI. 580 00:28:52,760 --> 00:28:55,160 Yes, I did. So you don't actually know what 581 00:28:55,160 --> 00:28:58,160 this code is doing. Well, I know that know what it 582 00:28:58,160 --> 00:29:00,320 tries to do, but I don't understand what these lines of 583 00:29:00,320 --> 00:29:02,320 code are doing. And that's the point where I 584 00:29:02,320 --> 00:29:08,760 said, OK, dude, I really support that you use AI, but use it in a 585 00:29:08,760 --> 00:29:12,440 way that I don't see it, you know, kind of saying like, like, 586 00:29:12,440 --> 00:29:14,880 I don't care if you don't practice test driven development 587 00:29:15,640 --> 00:29:18,720 really properly, but make sure that the code is actually 588 00:29:18,720 --> 00:29:22,320 designed in a way that I would if I practice DDD because I, I, 589 00:29:22,400 --> 00:29:25,640 I, I, I'm a big fan of that. As long as you you can use 590 00:29:25,640 --> 00:29:28,160 whatever code you, whatever tools you want, whatever 591 00:29:28,160 --> 00:29:31,320 practice you want, as long as the code is really the way we 592 00:29:31,320 --> 00:29:32,840 want to have it, then we can maintain it. 593 00:29:33,160 --> 00:29:37,000 I think this is very interesting because in my, at least if I 594 00:29:37,000 --> 00:29:39,880 look at myself, my educational career was always OK, you go and 595 00:29:39,880 --> 00:29:41,400 you do the research yourself, right? 596 00:29:41,600 --> 00:29:43,600 Because if you ask me and I give you the answer, it's not going 597 00:29:43,600 --> 00:29:46,800 to stick. Well, I did that and then I went 598 00:29:46,880 --> 00:29:50,200 into kind of adulthood and in my first job and they said, well, 599 00:29:50,680 --> 00:29:52,320 do the research and definitely do that. 600 00:29:52,320 --> 00:29:54,960 But also if you're stuck and if you feel like you need a quicker 601 00:29:54,960 --> 00:29:58,200 answer, we are here to be productive as a team. 602 00:29:58,200 --> 00:30:00,320 So if I can give you the answer instead of you researching and 603 00:30:00,320 --> 00:30:04,080 being on an island by yourself for a day, ask and I'll, I'll, I 604 00:30:04,080 --> 00:30:07,360 will brainstorm. So I asked, I learned how to ask 605 00:30:07,360 --> 00:30:10,360 and I think the answer stuck with me because the answers made 606 00:30:10,360 --> 00:30:12,000 sense. And if I didn't understand, I 607 00:30:12,000 --> 00:30:15,000 would ask again and I would go deeper with colleagues and with 608 00:30:15,000 --> 00:30:17,760 my team specifically. I learned a lot through that 609 00:30:17,800 --> 00:30:20,880 through just reflecting with people, bouncing off ideas, 610 00:30:21,080 --> 00:30:23,800 genuinely being curious and being like, just can you explain 611 00:30:23,920 --> 00:30:27,360 exactly XY and Z? And I feel like I can't revert 612 00:30:27,360 --> 00:30:28,960 time. I can't change how I learned, 613 00:30:28,960 --> 00:30:32,480 but if I were to have CHA GPT right now, also from a coding 614 00:30:32,480 --> 00:30:35,760 perspective, if I have a specific language, if there are 615 00:30:35,760 --> 00:30:38,720 conventions that I don't understand, I ask and I go 616 00:30:38,720 --> 00:30:42,480 through line by line, not every single line because I feel like 617 00:30:42,480 --> 00:30:44,800 some stuff it's fine. If I don't understand this bit, 618 00:30:44,800 --> 00:30:46,480 it's going to be OK. In the end. 619 00:30:46,480 --> 00:30:48,840 I feel like it makes sense, but it's because you already have a 620 00:30:48,840 --> 00:30:50,800 mental model. I feel like you can do two 621 00:30:50,800 --> 00:30:52,320 things right? You can be the engineer that 622 00:30:52,320 --> 00:30:55,040 copy paste from stack overflow pushes it to production, it 623 00:30:55,040 --> 00:30:58,440 works, it's OK. Same with CHA GPT, it generated, 624 00:30:58,440 --> 00:31:01,640 it pushes, it's, it's OK. Or you can be the person that 625 00:31:01,640 --> 00:31:05,080 leverages whatever you have now. And that curiosity kind of drive 626 00:31:05,080 --> 00:31:09,600 you in doing AQ and a with yourself with an AI tool that 627 00:31:09,600 --> 00:31:12,640 can actually explain and reason and go deep and be like what 628 00:31:12,640 --> 00:31:14,080 these are the architecture conventions. 629 00:31:14,080 --> 00:31:17,120 And this is actually where it, it doesn't follow these 630 00:31:17,120 --> 00:31:19,640 standards. And then you can ask, I think 631 00:31:19,640 --> 00:31:21,600 your colleagues or your team, does this make sense? 632 00:31:21,600 --> 00:31:23,920 Or is it actually a conscious choice? 633 00:31:23,920 --> 00:31:27,080 And as long as you have a curious mind, as long as you try 634 00:31:27,080 --> 00:31:30,160 and maintain your critical thinking, I think you can speed 635 00:31:30,160 --> 00:31:34,160 up very quickly. You can, I'm, I'm trying to 636 00:31:34,160 --> 00:31:37,680 think like in the past. I, I sometimes had to, well, I 637 00:31:37,720 --> 00:31:41,440 often worked with, again, aspiring developers and you 638 00:31:41,440 --> 00:31:44,720 typically saw scale. You have people that asked every 639 00:31:44,720 --> 00:31:46,240 question. They were asking questions all 640 00:31:46,240 --> 00:31:48,440 the time. And you have people that were 641 00:31:48,440 --> 00:31:51,720 afraid to ask questions. And I can also imagine that the 642 00:31:51,720 --> 00:31:55,160 people who are very hesitant in bothering the team, even though 643 00:31:55,160 --> 00:31:58,080 that's the whole point of the team, we'll probably go to chat 644 00:31:58,080 --> 00:32:01,920 Chiapati instead of even asking. I'm not sure if that will 645 00:32:01,920 --> 00:32:04,800 actually give you a good or bad result because I do it myself. 646 00:32:04,800 --> 00:32:08,120 I tend to like, I think a year or two years ago, a year ago, I 647 00:32:08,120 --> 00:32:12,680 think I was wondering between Azure Function Apps or Azure 648 00:32:12,680 --> 00:32:16,640 Container Apps and if you go on the Internet and try to find the 649 00:32:16,640 --> 00:32:19,760 answers, you typically would go to stacco flow because this is 650 00:32:19,760 --> 00:32:22,400 where the descriptions happen. Maybe in Reddit these days. 651 00:32:22,840 --> 00:32:26,960 Now I just ask Chetty Petty and it will give me a nice breakdown 652 00:32:26,960 --> 00:32:29,720 between the options and the pros and cons and I can continue 653 00:32:29,720 --> 00:32:33,000 asking like what is acceptable. Another example, like a couple 654 00:32:33,000 --> 00:32:35,600 of weeks ago, we were were wondering like, OK, we need to 655 00:32:35,600 --> 00:32:40,640 capture some guidelines on how fast our website should respond. 656 00:32:40,640 --> 00:32:43,640 What is acceptable, what's not? In the past, I would Google. 657 00:32:43,880 --> 00:32:47,080 Now, I just asked ChatGPT and say, hey, do you know any 658 00:32:47,080 --> 00:32:50,120 industry except the standards for defining how fast the 659 00:32:50,120 --> 00:32:53,880 website should respond? And it gives me a beautiful 660 00:32:53,880 --> 00:32:55,880 answer. And it can even tell me like, 661 00:32:56,160 --> 00:32:59,640 oh, you want some examples of websites that document this or, 662 00:32:59,920 --> 00:33:02,440 or sometimes there's documentation that's not 663 00:33:02,440 --> 00:33:05,200 properly. There's no, there's no industry 664 00:33:05,200 --> 00:33:08,080 standard, but it has kind of, you know, yeah. 665 00:33:08,400 --> 00:33:12,440 What I was trying to say is that ChatGPT or GPT or AI is capable 666 00:33:12,720 --> 00:33:16,320 of interpreting different versions of the truth and try to 667 00:33:16,320 --> 00:33:19,160 combine that in a certain way, which is a nice way because 668 00:33:19,160 --> 00:33:21,600 sometimes you do want to hear different stories or different 669 00:33:21,840 --> 00:33:25,440 perspectives of a problem and get the combined result. 670 00:33:26,320 --> 00:33:28,600 And it's ridiculous. But I use it every day, many 671 00:33:28,600 --> 00:33:33,440 times a day, even when I was trying to buy a new car. 672 00:33:33,520 --> 00:33:37,240 Get feedback on what are common problems with that car or or you 673 00:33:37,240 --> 00:33:39,840 know what is better or what what do people say about it, about 674 00:33:39,840 --> 00:33:43,120 this particular type of car. I used everywhere now. 675 00:33:43,240 --> 00:33:45,800 Yeah, I'd love to get your perspective on this because I 676 00:33:45,800 --> 00:33:49,160 did AQ and a episode recently and someone said, OK, if you 677 00:33:49,160 --> 00:33:53,640 were to be responsible for actually building up a person 678 00:33:53,640 --> 00:33:55,400 that wants to be a software engineer, they start from 679 00:33:55,400 --> 00:33:59,160 scratch, what would you focus on specifically with what you know 680 00:33:59,160 --> 00:34:02,400 nowadays? That's a really good question. 681 00:34:02,560 --> 00:34:03,600 Sorry. That's a really good. 682 00:34:04,520 --> 00:34:08,280 I don't think I answered that in a good way because this is one 683 00:34:08,280 --> 00:34:11,159 that can go quite in a lot of directions to be honest. 684 00:34:12,080 --> 00:34:14,800 This happens like this particular situation happens a 685 00:34:14,800 --> 00:34:18,520 lot because I have a couple of pretty young developers in my 686 00:34:18,520 --> 00:34:23,400 team. In the end, I think it's still 687 00:34:23,400 --> 00:34:29,360 the way it happened in the past. It's letting them fix box in an 688 00:34:29,360 --> 00:34:34,040 existing code base, letting them pick up small improvements where 689 00:34:34,040 --> 00:34:36,719 they touch a big chunk of the code base and learn about that 690 00:34:36,719 --> 00:34:40,320 code base, let them ask questions. 691 00:34:41,320 --> 00:34:43,400 And one of the things I do, and I've been doing this for many 692 00:34:43,400 --> 00:34:48,080 years, when they create a pool request, I, I review it and I 693 00:34:48,080 --> 00:34:51,320 use emojis in my poetry, not because I need to express my 694 00:34:51,520 --> 00:34:58,760 emotions, but what I'm trying to do is I tried to emphasize how 695 00:34:58,760 --> 00:35:00,960 they should interpret my, my code review comments. 696 00:35:01,240 --> 00:35:05,400 So I use, for example, a pickaxe to say I'm, I'm nitpicking. 697 00:35:05,400 --> 00:35:09,080 And especially in the beginning and with new, new joiners, I do 698 00:35:09,080 --> 00:35:11,480 that more, I do that more extensively, even though I 699 00:35:11,480 --> 00:35:15,200 shouldn't be nitpicking, but I, it's my way of saying, OK, this 700 00:35:15,200 --> 00:35:18,360 is OK. This can be done better for 701 00:35:18,360 --> 00:35:20,880 business reasons. And then I use different things. 702 00:35:20,880 --> 00:35:25,320 I use thinking person to say, OK, you know what? 703 00:35:25,800 --> 00:35:28,040 This is probably OK. There's different ways of doing 704 00:35:28,040 --> 00:35:29,400 it. Let's discuss that. 705 00:35:29,840 --> 00:35:32,880 And then I use that opportunity. They probably call me or we at 706 00:35:32,880 --> 00:35:35,320 the office or we discussed that and I tried to explain them the 707 00:35:35,320 --> 00:35:39,160 pros and cons of that solution and make them thinking about it, 708 00:35:39,200 --> 00:35:42,240 make them think about it. The other things are sometimes I 709 00:35:42,240 --> 00:35:45,640 use a camp campground icon, you know, to leave the campground 710 00:35:45,640 --> 00:35:49,440 cleaner, which also is for them an opportunity to understand, 711 00:35:49,440 --> 00:35:53,280 OK, so my code is OK, but there's, there's ways to make 712 00:35:53,280 --> 00:35:55,720 that better. And I think that's, that's 713 00:35:55,720 --> 00:35:57,760 already a learning exercise by itself. 714 00:35:58,440 --> 00:36:02,320 And if they're stuck on a particular solution, sometimes 715 00:36:02,320 --> 00:36:05,280 even though I know the answer, I ask them, look, no, maybe you 716 00:36:05,280 --> 00:36:08,720 should ask Cha Chi Patee or or Juni or whatever AI you're using 717 00:36:09,040 --> 00:36:11,080 to see what they come up, come up with and use that as a 718 00:36:11,080 --> 00:36:13,520 discussion point. You know, they do that and we 719 00:36:13,520 --> 00:36:15,040 have actually something to discuss about. 720 00:36:15,280 --> 00:36:16,280 Interesting. Yeah. 721 00:36:16,720 --> 00:36:20,280 Because especially if you talk about architecting the system, 722 00:36:20,520 --> 00:36:22,280 what kind of abstraction do we need? 723 00:36:22,440 --> 00:36:26,200 Why would we use the dependency inversion principle in this 724 00:36:26,200 --> 00:36:28,160 solution to decouple something from each other? 725 00:36:28,520 --> 00:36:30,280 That is not something an AI does for you. 726 00:36:30,720 --> 00:36:35,400 That requires like a holistic view on what you're trying to 727 00:36:35,400 --> 00:36:37,960 do. And I don't think an AI can do 728 00:36:37,960 --> 00:36:39,920 that right now. This is actually the value. 729 00:36:40,560 --> 00:36:42,800 And as long as we need to maintain the code, we need to 730 00:36:42,800 --> 00:36:45,760 think about coupling A cohesion and that kind of things. 731 00:36:45,760 --> 00:36:46,840 We have still to think about that. 732 00:36:46,920 --> 00:36:49,880 Yeah, yeah. The thing you triggered me when 733 00:36:49,880 --> 00:36:52,600 you said, OK, some people will ask ChatGPT instead of asking a 734 00:36:52,600 --> 00:36:54,280 colleague. For me, there are a lot of 735 00:36:54,280 --> 00:36:58,000 things that my opinion is of myself, I should know these 736 00:36:58,000 --> 00:37:01,440 things and I just don't. And then I can ask a colleague, 737 00:37:01,440 --> 00:37:02,600 right? I, I feel he's a bit 738 00:37:02,600 --> 00:37:04,560 embarrassing and probably from the other side, it's not 739 00:37:04,560 --> 00:37:06,120 embarrassing at all. This is really just this 740 00:37:06,120 --> 00:37:10,320 imposter syndrome feeling. Or I ask and I use a tool to 741 00:37:10,320 --> 00:37:14,200 kind of better figure out what is actually going on before I 742 00:37:14,200 --> 00:37:16,840 ask a question because I will still want kind of human 743 00:37:16,840 --> 00:37:20,600 feedback, a human in the loop in the end for my decision making 744 00:37:20,600 --> 00:37:22,480 power. In essence, I feel like we have 745 00:37:22,480 --> 00:37:25,560 a tool that can give us a lot of input with regards to what is 746 00:37:25,560 --> 00:37:28,080 there and also how historically things have built. 747 00:37:28,520 --> 00:37:31,320 Sometimes if you are in a project that is quite far along, 748 00:37:31,320 --> 00:37:33,920 there is no single person that knows the context and the 749 00:37:33,920 --> 00:37:36,320 history of the code base. But now you have a tool that can 750 00:37:36,320 --> 00:37:38,720 actually facilitate that information, that distribution 751 00:37:38,720 --> 00:37:42,280 of information to you and it can do it quite well, which is going 752 00:37:42,280 --> 00:37:45,360 to be quite interesting for like modernizing software in essence. 753 00:37:45,880 --> 00:37:48,960 But we are still, we still have to take ownership, we still have 754 00:37:48,960 --> 00:37:53,040 to do the decision making and we still have to align on things as 755 00:37:53,040 --> 00:37:56,560 a team, I would say. Yes, but the tool can help you 756 00:37:58,440 --> 00:38:00,560 understand a complicated code base. 757 00:38:01,000 --> 00:38:04,040 In fact, I use that also like especially when you have a lot 758 00:38:04,040 --> 00:38:06,280 of micro service architecture, you try to figure out like how 759 00:38:06,280 --> 00:38:08,440 does this information flow from one surface to the other? 760 00:38:08,800 --> 00:38:11,040 You can do that yourself. Go through the abstractions. 761 00:38:11,040 --> 00:38:12,920 Oh, I'm doing an HTP endpoint call here. 762 00:38:13,120 --> 00:38:16,280 Where's the the handler on? Where's that HTPAPI on the other 763 00:38:16,280 --> 00:38:20,000 side? I use AI for that, but 764 00:38:20,000 --> 00:38:24,200 understanding the history and the reasons behind that, that is 765 00:38:24,200 --> 00:38:26,120 the source control if you're lucky. 766 00:38:26,720 --> 00:38:29,520 And that's also one of the reasons why I still are very 767 00:38:29,600 --> 00:38:32,840 obnoxious about making sure that people create pull requests that 768 00:38:32,840 --> 00:38:37,960 describe not just what it tries to do, but what is what's the 769 00:38:37,960 --> 00:38:41,160 reason for that. And I still do that for commits, 770 00:38:41,240 --> 00:38:42,600 for pull requests, for everything. 771 00:38:42,840 --> 00:38:45,520 It should capture what you're trying to do, not how you did 772 00:38:45,520 --> 00:38:47,800 it, because I can see that because that information is 773 00:38:47,800 --> 00:38:49,920 lost. It's not it's not on a code bit. 774 00:38:50,080 --> 00:38:53,440 Maybe by using comments you can I even do that and documenting 775 00:38:53,440 --> 00:38:56,160 my code still properly 9 out of time. 776 00:38:56,160 --> 00:38:58,600 I do that using AI, but it's really good. 777 00:38:58,640 --> 00:39:01,080 So yeah. And by the way, that's funny 778 00:39:01,080 --> 00:39:04,600 because I'm actually surprised. I use I use writer. 779 00:39:04,600 --> 00:39:09,160 As I said, I often use AI or writer to generate description 780 00:39:09,160 --> 00:39:12,320 of my pool of my commit because I still have a client that uses 781 00:39:12,320 --> 00:39:13,880 Azure DevOps. I'm a big fan of it. 782 00:39:14,200 --> 00:39:18,440 I'm not a fan of it that I said, but I I let the AI generate the 783 00:39:18,440 --> 00:39:21,680 description and I'm often surprised that it somehow 784 00:39:21,680 --> 00:39:25,480 figured out what I'm trying to do so just didn't explain what 785 00:39:25,480 --> 00:39:28,920 files are changed or but it actually sees the value of that 786 00:39:28,920 --> 00:39:31,000 and can and can write that in a functional way. 787 00:39:31,240 --> 00:39:33,200 That's really solid. I don't even understand how it 788 00:39:33,200 --> 00:39:36,560 does that, but I take that, that description and I put it in the 789 00:39:36,640 --> 00:39:38,800 pull across description and sometimes tweak a little bit, 790 00:39:39,520 --> 00:39:42,920 but it's better than Copilot, Git of Copilot, it's own summary 791 00:39:42,920 --> 00:39:44,840 functionality because that's very technical. 792 00:39:45,200 --> 00:39:48,240 But for some reason Claude and Junie can actually do that for 793 00:39:48,240 --> 00:39:50,080 me. Maybe because it looks at the 794 00:39:50,080 --> 00:39:53,480 history of the codebase. I don't know what we were 795 00:39:53,480 --> 00:39:54,160 saying. Oh yeah. 796 00:39:54,520 --> 00:39:59,280 So it's still valuable to make sure that your your git history 797 00:39:59,680 --> 00:40:02,520 captures the essence the the the the purpose of what you're 798 00:40:02,520 --> 00:40:03,800 doing. Yeah, absolutely. 799 00:40:04,120 --> 00:40:07,360 I saw GitHub, they do this quite extensively. 800 00:40:07,360 --> 00:40:09,960 They make business decision records, so not architectural 801 00:40:09,960 --> 00:40:13,040 decision records because they want business outcomes. 802 00:40:13,040 --> 00:40:15,720 It can be cost savings, it could be new revenue streams, 803 00:40:16,160 --> 00:40:19,960 different or or magnitudes of scale, for example, that can 804 00:40:19,960 --> 00:40:22,760 reach the business outcomes and they capture that in quite an 805 00:40:22,760 --> 00:40:26,440 extensive documentation. If you already do AD Rs and if 806 00:40:26,440 --> 00:40:30,440 you really work on historically capturing your context through 807 00:40:30,440 --> 00:40:33,080 commits, through pull requests, through ticketing systems, 808 00:40:33,400 --> 00:40:37,280 decision making records, I feel like you will be, I think 809 00:40:37,280 --> 00:40:40,680 investing in the future because that context is quite useful for 810 00:40:40,680 --> 00:40:44,120 humans and it's hard for a human to go through all the context, 811 00:40:44,120 --> 00:40:46,680 But humans now have tools to actually go through all the 812 00:40:46,680 --> 00:40:48,600 context. So the richer the context is, 813 00:40:48,880 --> 00:40:51,440 the better we can make decisions and kind of understand 814 00:40:51,440 --> 00:40:54,640 historically what has happened, which I think is really cool to 815 00:40:54,640 --> 00:40:56,280 be honest. The companies that have invested 816 00:40:56,280 --> 00:41:00,400 in that, they will be leaps and bounds ahead compared to the 817 00:41:00,400 --> 00:41:02,600 companies that don't invest in that agree. 818 00:41:03,080 --> 00:41:06,200 And, and, and you're preaching to the choir here because this 819 00:41:06,200 --> 00:41:08,920 is exactly what I do with my clients quite often. 820 00:41:09,360 --> 00:41:11,800 I try to convince them to adopt Confluence, Atlassian 821 00:41:11,800 --> 00:41:13,720 Confluence, because it's a beautiful documentation 822 00:41:13,720 --> 00:41:16,160 platform. And we use that to create 823 00:41:16,480 --> 00:41:20,240 basically decision records, which is simple pages with which 824 00:41:20,240 --> 00:41:23,120 demonstrate or show you some of the history behind decisions, 825 00:41:23,120 --> 00:41:26,280 like the pros and cons. And for any non trivial 826 00:41:26,280 --> 00:41:29,880 decision, we capture that because you make so many 827 00:41:29,880 --> 00:41:32,960 decisions. And I've, I've been at certain 828 00:41:32,960 --> 00:41:35,960 clients for a long time and what you, what happened at some 829 00:41:35,960 --> 00:41:39,280 point, somebody new joins, maybe a new senior developer, maybe a 830 00:41:39,280 --> 00:41:42,280 new architect, and they may challenge certain decisions that 831 00:41:42,280 --> 00:41:44,920 I made in the past. And I make sure that I've 832 00:41:44,920 --> 00:41:48,400 captured exactly what were the reasons at that time to adopt A 833 00:41:48,400 --> 00:41:52,080 particular strategy. Also demonstrate even simple 834 00:41:52,080 --> 00:41:54,960 tables, you know, with all the pros and cons and who were 835 00:41:54,960 --> 00:41:59,080 involved in the decision and what other options did we reject 836 00:41:59,080 --> 00:42:02,720 and for what reasons? I used why statements, why dash 837 00:42:02,720 --> 00:42:06,120 statements, which is a nice way of doing that because I've, I've 838 00:42:06,120 --> 00:42:10,080 had the pain of having to deal with that situation and I do it 839 00:42:10,080 --> 00:42:11,840 everywhere. And most teams are like, why do 840 00:42:11,840 --> 00:42:14,920 we need to capture that? And when you see after a year or 841 00:42:14,920 --> 00:42:17,800 so, they started to see the failure like, oh crap, we 842 00:42:17,800 --> 00:42:21,400 actually wrote that down. And in Confluence, you also have 843 00:42:21,400 --> 00:42:24,360 like, AI, it's called Rovo. I think you can now just ask 844 00:42:24,360 --> 00:42:28,200 like, hey, can you figure out why we decided to go for Azure 845 00:42:28,200 --> 00:42:31,520 Container Apps, for example, as we'll find the decision page and 846 00:42:31,520 --> 00:42:36,480 tell you, yeah, because of these reasons or, yeah, why did we use 847 00:42:36,480 --> 00:42:40,960 private endpoints Azure instead of going through application API 848 00:42:40,960 --> 00:42:42,160 management or something like that? 849 00:42:42,840 --> 00:42:44,520 Those are decisions that you want to capture. 850 00:42:44,520 --> 00:42:47,200 They're not even a code. They happen outside the code 851 00:42:47,200 --> 00:42:50,360 base. That's maybe the flip side, but 852 00:42:50,360 --> 00:42:53,760 I don't like writing a markdown in GitHub just for capturing 853 00:42:53,760 --> 00:42:55,680 functional or technical decisions. 854 00:42:55,920 --> 00:42:58,520 I think it's cumbersome. Have you ever seen those 855 00:42:58,800 --> 00:43:03,360 practices of capturing decisions linked to a sunken cost where we 856 00:43:03,360 --> 00:43:06,280 say, OK, we've made the trade-offs we chose for XY and Z 857 00:43:06,280 --> 00:43:09,760 in this context, and that's why we're sticking to it rather than 858 00:43:09,760 --> 00:43:12,240 looking at what is now and actually maybe pivoting or 859 00:43:12,240 --> 00:43:14,960 shifting because of going through those motions of 860 00:43:14,960 --> 00:43:16,640 actually capturing those decisions? 861 00:43:16,920 --> 00:43:19,160 Yeah. In fact the example that I just 862 00:43:19,160 --> 00:43:21,160 gave you about the new architect, I've seen that it 863 00:43:21,160 --> 00:43:23,920 happened a couple of times, like I've been for one particular 864 00:43:23,920 --> 00:43:26,000 client for a longer period of time, couple of years. 865 00:43:26,720 --> 00:43:31,000 And like what happened a couple of times, management changes and 866 00:43:31,000 --> 00:43:34,760 then management tend to talk to everybody, like trying to 867 00:43:34,840 --> 00:43:39,440 understand the the department and they may talk to people that 868 00:43:39,480 --> 00:43:42,320 are pretty young but influential, which didn't 869 00:43:42,320 --> 00:43:44,840 actually understand some of the decisions I made in the past. 870 00:43:45,320 --> 00:43:47,720 I started to talk to a manager and then a manager at some point 871 00:43:47,720 --> 00:43:51,360 will come to me like, hey, I've heard that we're using this in 872 00:43:51,360 --> 00:43:54,160 this architecture style, which the manager doesn't understand 873 00:43:54,160 --> 00:43:56,560 at all, but it doesn't matter. And I heard it's causing a lot 874 00:43:56,560 --> 00:44:00,080 of pain, you know, and then because I can go back and 875 00:44:00,080 --> 00:44:05,120 demonstrate all that I can kind of circumfent somebody will 876 00:44:05,120 --> 00:44:06,600 decide. Now we have to remove that 877 00:44:06,600 --> 00:44:08,800 because we need to use a simple architecture. 878 00:44:08,800 --> 00:44:13,760 Even though in all my I, I can honestly look at my heart and 879 00:44:13,760 --> 00:44:17,240 say, like, I made that decision based on the situation at that 880 00:44:17,240 --> 00:44:20,320 point of time with all the knowledge that I have, if the 881 00:44:20,320 --> 00:44:22,920 situation has changed, this case, the situation didn't 882 00:44:22,920 --> 00:44:26,000 change and there's no alternative, there's no reason 883 00:44:26,000 --> 00:44:28,960 to change that. But I know that because I'm an 884 00:44:28,960 --> 00:44:30,440 architect, I have the experience. 885 00:44:30,840 --> 00:44:33,600 A manager doesn't have that. They just listen to, you know, 886 00:44:33,600 --> 00:44:36,320 hearsay. So in that case has saved me a 887 00:44:36,320 --> 00:44:37,680 couple of times. Yeah, Yeah. 888 00:44:37,680 --> 00:44:39,480 I think that's super solid, one of the. 889 00:44:39,480 --> 00:44:42,040 Last thoughts I wanted to share with you because I like your 890 00:44:42,040 --> 00:44:44,560 perspective on this has to do with open source. 891 00:44:44,920 --> 00:44:47,240 It was a few weeks back, I think I read this article about 892 00:44:47,240 --> 00:44:50,600 Tailwind and the team that supports Tailwind, they had to 893 00:44:50,600 --> 00:44:54,560 lay off 75% of their team. Now 75% sounds big, but it's 894 00:44:54,720 --> 00:44:57,600 three out of four people, which is still a lot in essence, but 895 00:44:57,600 --> 00:45:00,240 it's because their business model didn't add up anymore till 896 00:45:00,240 --> 00:45:02,920 when it's there, it's free, it's open source and they have this 897 00:45:02,920 --> 00:45:05,440 premium features where certain components and component 898 00:45:05,440 --> 00:45:08,400 libraries that they've made, which are really cool, you can 899 00:45:08,400 --> 00:45:10,480 pay for them and you get those out-of-the-box rather than 900 00:45:10,480 --> 00:45:13,320 creating them yourself. Well, now with AI assisted 901 00:45:13,320 --> 00:45:16,280 development tooling, you can basically say I want this 902 00:45:16,280 --> 00:45:18,600 component that is in their premium functionality. 903 00:45:18,800 --> 00:45:20,960 So you don't have to pay, but you still get the benefits 904 00:45:21,360 --> 00:45:22,840 because everything is open source. 905 00:45:22,840 --> 00:45:26,200 This is now a possibility, but it kind of brings at risk 906 00:45:26,640 --> 00:45:29,880 certain options for making code open source. 907 00:45:29,880 --> 00:45:32,560 Because if there's no business model in the end and you still 908 00:45:32,560 --> 00:45:36,840 need to have a maintenance and a team behind it, it's kind of a 909 00:45:36,840 --> 00:45:39,520 chicken and egg problem, right? If there's no money in it, then 910 00:45:39,520 --> 00:45:42,480 people might lose the incentive to then make things open source, 911 00:45:42,840 --> 00:45:46,360 Which is a shame because I, I really loved that part of this 912 00:45:46,400 --> 00:45:49,200 community and kind of being a software engineer part of 913 00:45:49,200 --> 00:45:51,040 things. Yeah, I definitely have thought 914 00:45:51,040 --> 00:45:53,840 about it and talked. About it this last year, given 915 00:45:53,920 --> 00:45:56,760 that I actually commercialized an open source project, which by 916 00:45:56,760 --> 00:45:59,600 the way wasn't planned, it just happened was an opportunity. 917 00:46:01,160 --> 00:46:05,280 I've always said, even way before I was doing that if you 918 00:46:05,280 --> 00:46:08,040 use open source, it doesn't mean you get anything. 919 00:46:08,200 --> 00:46:12,480 Well, how do I phrase this? There's no guarantee that that 920 00:46:12,480 --> 00:46:15,360 code will last. And that's actually what 921 00:46:15,360 --> 00:46:17,280 happened. A lot of people felt entitled or 922 00:46:17,280 --> 00:46:20,240 feel entitled that if you use, if they use your open source 923 00:46:20,240 --> 00:46:24,920 library, they feel like they kind of expect that this will be 924 00:46:24,920 --> 00:46:26,600 maintained. Definitely like a service 925 00:46:26,960 --> 00:46:29,160 almost. And what I often. 926 00:46:29,160 --> 00:46:35,200 Do as part of an architecture role is write down, write down 927 00:46:35,200 --> 00:46:37,520 guidelines for open source consumption. 928 00:46:38,000 --> 00:46:40,760 And what I usually no what I always do is write down 929 00:46:40,760 --> 00:46:44,560 something that states something in the matter of in the style 930 00:46:44,560 --> 00:46:47,920 of. If you use open source, look at 931 00:46:47,920 --> 00:46:53,160 the quality of that open source library, because if the person 932 00:46:53,160 --> 00:46:56,320 or group of people behind that will no longer support it, you 933 00:46:56,320 --> 00:46:59,320 have to support it. So just like AI code is your 934 00:46:59,320 --> 00:47:03,040 code, if you use a, if you use open source code, it may 935 00:47:03,040 --> 00:47:05,560 eventually become your own code that you have to maintain. 936 00:47:06,200 --> 00:47:09,680 So there are things like, OK, is it just one man or one person 937 00:47:09,680 --> 00:47:12,120 that maintains it, or is there an entire community behind that? 938 00:47:12,520 --> 00:47:13,800 It's not. It doesn't mean that you can't 939 00:47:13,800 --> 00:47:15,280 use it just means that you have to think about it. 940 00:47:15,720 --> 00:47:18,640 Is it something that you use in your build pipeline or unit 941 00:47:18,640 --> 00:47:21,480 tests? Because then it's easier, you 942 00:47:21,480 --> 00:47:23,920 can be more lenient. Or is it something you use in 943 00:47:23,920 --> 00:47:26,080 production? Because if it if it stops being 944 00:47:26,080 --> 00:47:28,640 maintained, you have to do it yourself and it might be very 945 00:47:28,640 --> 00:47:31,000 complicated. What's the quality of that code? 946 00:47:31,000 --> 00:47:34,080 Does it have automatic tests? Does it have documentation? 947 00:47:34,080 --> 00:47:39,200 Is the, does it use all kinds of code checking tools to make sure 948 00:47:39,200 --> 00:47:42,720 that code is, is at the same level that you as a company or 949 00:47:42,720 --> 00:47:46,560 team tries to uphold? And there's a bunch of more like 950 00:47:46,840 --> 00:47:49,080 do they use a a proper branching strategy? 951 00:47:49,360 --> 00:47:52,680 What's the code coverage? Do they proper read means? 952 00:47:52,680 --> 00:47:55,440 Do they do release modes, release notes, stuff like that? 953 00:47:55,840 --> 00:47:59,320 That kind of, it's not a single rule that says you can't use it, 954 00:47:59,600 --> 00:48:02,840 but it's to make you really aware that it is in the end your 955 00:48:02,840 --> 00:48:06,320 code. If you take that approach, it's 956 00:48:06,320 --> 00:48:08,480 easier to to take something and and support it. 957 00:48:08,960 --> 00:48:11,600 Because if, if, if the original developer decides to stop 958 00:48:11,600 --> 00:48:14,720 maintaining it, yeah, then you fork the repostory and maintain 959 00:48:14,720 --> 00:48:16,840 it yourself. You could have also decided to 960 00:48:16,840 --> 00:48:19,040 say, OK, I'm going to write that code yourself. 961 00:48:19,720 --> 00:48:22,920 I think in the end that is still much more expensive than take 962 00:48:22,920 --> 00:48:24,680 something from the open source community. 963 00:48:25,080 --> 00:48:28,280 We quite, quite a lot of people underestimate how much 964 00:48:28,280 --> 00:48:31,080 complexity is behind that. I know that I've experienced 965 00:48:31,080 --> 00:48:35,200 this first hand where people said like, oh, I don't, I don't 966 00:48:35,200 --> 00:48:36,840 need these stupid fluent assertions. 967 00:48:37,080 --> 00:48:39,280 I can build this in a couple of days yourself. 968 00:48:39,520 --> 00:48:42,600 Sure. I've spent 15 years perfecting 969 00:48:42,600 --> 00:48:48,160 it, you know, and even I I built four or five new open source 970 00:48:48,160 --> 00:48:50,800 projects this year and there's a lot of energy. 971 00:48:50,840 --> 00:48:54,280 It still takes me weeks of energy to improve things. 972 00:48:54,280 --> 00:48:57,120 Even in fluent session eight, I was improving some kind of 973 00:48:57,560 --> 00:48:59,600 reporting thing. When you compare collections. 974 00:48:59,600 --> 00:49:02,840 I spent weeks on trying to find a good algorithm and 975 00:49:02,840 --> 00:49:04,960 implementing in a way that actually makes it natural. 976 00:49:05,160 --> 00:49:08,360 And I still use AI. People underestimate the amount 977 00:49:08,360 --> 00:49:09,960 of work. I think Duanda is a good 978 00:49:09,960 --> 00:49:11,960 example. Duanda has this AI. 979 00:49:11,960 --> 00:49:15,680 Sorry, this authentication authorization library used to be 980 00:49:15,680 --> 00:49:18,120 called Identity Server. A lot of people use that. 981 00:49:18,120 --> 00:49:21,200 And when they became commercial, there was no discussion whether 982 00:49:21,200 --> 00:49:23,320 the people want to pay for that because they realized how 983 00:49:23,320 --> 00:49:28,800 complex something like that is. I think that kind of is an 984 00:49:28,800 --> 00:49:30,680 answer to you. I think one of your points. 985 00:49:30,920 --> 00:49:34,360 So in the end, use open source because it's probably cheaper 986 00:49:34,440 --> 00:49:38,560 for you to start using it. And if it really dies, maintain 987 00:49:38,560 --> 00:49:41,080 it yourself. And if that's too expensive, 988 00:49:41,360 --> 00:49:42,960 then probably find something else. 989 00:49:42,960 --> 00:49:45,720 But probably it was too complicated from the start and 990 00:49:45,720 --> 00:49:47,600 doing it yourself would never have been feasible. 991 00:49:48,920 --> 00:49:51,360 Now I lost track of my thoughts. No worries if people are 992 00:49:51,360 --> 00:49:54,080 wondering, should I? Build something that is a 993 00:49:54,080 --> 00:49:56,200 building block or I have something, Should I open source 994 00:49:56,200 --> 00:49:57,680 it? You're a person that open source 995 00:49:57,680 --> 00:50:00,560 a lot, why should they do that? What would you say to them? 996 00:50:00,920 --> 00:50:03,160 It's fun. It is. 997 00:50:03,200 --> 00:50:05,800 A nice way to demonstrate what you can. 998 00:50:06,640 --> 00:50:10,520 It's super cool if you see other people appreciate that, you 999 00:50:10,520 --> 00:50:12,440 know? Yeah. 1000 00:50:12,760 --> 00:50:18,560 And if you're lucky, people start adopting it and enjoy it. 1001 00:50:18,560 --> 00:50:21,160 And when that happens, and I had that with Mokli yesterday, the 1002 00:50:21,160 --> 00:50:24,200 first time one of my colleagues, new colleague, actually enjoyed 1003 00:50:24,200 --> 00:50:26,240 it so much, he thought like, OK, we found a bug. 1004 00:50:26,800 --> 00:50:28,120 You know what? I'm going to file a bug. 1005 00:50:28,120 --> 00:50:31,040 And I said, like, you're open to create a pull request and fix 1006 00:50:31,040 --> 00:50:32,960 it. It will automatically list you 1007 00:50:32,960 --> 00:50:34,600 as a contributor. And he said, yeah, I will do 1008 00:50:34,600 --> 00:50:36,240 that. And he created a pull request. 1009 00:50:36,240 --> 00:50:38,560 And that for me is always valuable. 1010 00:50:39,160 --> 00:50:41,440 And if I look at all the other open source projects, especially 1011 00:50:41,440 --> 00:50:43,560 through the assertions, I learned a lot from that. 1012 00:50:44,120 --> 00:50:46,840 There were so many improvements. People contributed at some 1013 00:50:46,840 --> 00:50:50,800 point, ideas of using tools or methods or something like that 1014 00:50:50,800 --> 00:50:54,200 that I didn't know about. I now have another library which 1015 00:50:54,200 --> 00:50:58,320 is called the.net Library Starter Kit, which I used to 1016 00:50:58,320 --> 00:51:01,760 combine everything I've learned about building packages into 1017 00:51:01,760 --> 00:51:04,720 likeyoucansay.net new and then it will generate an entire 1018 00:51:04,720 --> 00:51:07,160 solution. You can use it for open source, 1019 00:51:07,160 --> 00:51:10,000 for inner source, for GitHub, Azure DevOps capturing 1020 00:51:10,000 --> 00:51:12,800 everything I learned. And people contribute to that by 1021 00:51:12,800 --> 00:51:16,080 saying, hey, did you know about, I think it's called annotation 1022 00:51:16,080 --> 00:51:19,000 or not annotation attestation in GitHub. 1023 00:51:19,000 --> 00:51:21,760 I never heard about that. And you learn from that or 1024 00:51:22,280 --> 00:51:23,760 somebody else said, hey, maybe you can. 1025 00:51:23,760 --> 00:51:25,720 I have another library called Package Card, which is for 1026 00:51:25,720 --> 00:51:28,000 scanning your licences and open source packages. 1027 00:51:28,520 --> 00:51:32,000 And maybe you can add some additional statistics about like 1028 00:51:32,000 --> 00:51:33,920 the maintainability of an open source library. 1029 00:51:33,920 --> 00:51:35,120 And I thought like, hey, that's interesting. 1030 00:51:35,360 --> 00:51:37,720 I didn't even know about this service that I can use for that. 1031 00:51:38,240 --> 00:51:42,600 It is almost like you give, but you get so much back from it. 1032 00:51:43,520 --> 00:51:47,360 Knowledge, experience, new tools, new insights, visibility, 1033 00:51:47,360 --> 00:51:50,320 of course, and that is very valuable. 1034 00:51:50,440 --> 00:51:52,520 Yeah, I love that. Thank you so much for. 1035 00:51:52,520 --> 00:51:54,600 Coming on, Dennis, and sharing your perspective, this was a 1036 00:51:54,600 --> 00:51:55,680 blast. Yeah. 1037 00:51:56,040 --> 00:51:59,160 Last time was four years ago. Next time, hopefully it's not 1038 00:51:59,160 --> 00:52:02,320 four years I lost even I lost track already. 4 times. 1039 00:52:02,520 --> 00:52:04,280 Yeah. If you're still with us, let me. 1040 00:52:04,280 --> 00:52:06,400 Know in the comments section what you thought of this episode 1041 00:52:06,400 --> 00:52:07,400 and we'll see you in the next one.