1 00:00:01,080 --> 00:00:04,799 How'd you like to listen to dot net rocks with no ads? Easy? 2 00:00:05,360 --> 00:00:09,480 Become a patron for just five dollars a month. You get access to a 3 00:00:09,480 --> 00:00:14,240 private RSS feed where all the shows have no ads. Twenty dollars a month. 4 00:00:14,279 --> 00:00:18,440 We'll get you that and a special dot net Rocks patron mug. Sign 5 00:00:18,519 --> 00:00:36,560 up now at Patreon dot dot NetRocks dot com. Hey guess what it's dot 6 00:00:36,600 --> 00:00:41,600 net Rocks Portugal Edition. I'm Carl Franklin. We're at NDC Porto. Yeah, 7 00:00:41,840 --> 00:00:45,479 and just like last time, when we're in here, the sun is 8 00:00:45,479 --> 00:00:49,799 out, yes, when we walk side to the hotel. Poor. Yeah. 9 00:00:49,840 --> 00:00:53,439 I got to tell you that I was here the first day I was 10 00:00:53,439 --> 00:00:56,640 here, yesterday, I was walking around. I got an umbrella, like 11 00:00:56,679 --> 00:01:00,719 a four euro umbrella. Boy took get outside and the wind and it went 12 00:01:02,679 --> 00:01:07,599 and it turned into like a big mangle of metal and broken four year four 13 00:01:07,680 --> 00:01:12,239 euro pinata. Yeah, wind pinata. I gave it instantly smashed one minute 14 00:01:12,280 --> 00:01:17,920 for each euro. That's how long it lasted. And then I got smart 15 00:01:17,959 --> 00:01:23,159 and bought a ten euro umbrella which was quite a bit more sturdy and worked 16 00:01:23,200 --> 00:01:26,319 really well. And feeling all smugs. Somebody in the hotel speaker said, 17 00:01:26,359 --> 00:01:33,480 you know, the hotel will just give you an umbrella. Thanks. I 18 00:01:33,480 --> 00:01:36,719 feel smart now, Yeah, I feel smart now. Hey, let's roll 19 00:01:36,719 --> 00:01:47,719 the music for better no framework, all right. So at breakfast the other 20 00:01:47,760 --> 00:01:51,920 day I was asking my friend Richard Campbell's HoloLens. I haven't heard anything about 21 00:01:52,040 --> 00:01:53,719 Microsoft HoloLens in a long time, and He's gone, I don't know. 22 00:01:55,760 --> 00:01:59,920 So of course I went and did what most people do. I Google it. 23 00:02:00,239 --> 00:02:06,680 Banged it enough Google bang And so it turns out there's a story from 24 00:02:06,840 --> 00:02:15,240 on MSFT like on Microsoft dot com from October eleventh. Microsoft releases Windows Holographic 25 00:02:15,879 --> 00:02:21,759 twenty three h two update. Nice, here's what's new, And so I 26 00:02:21,879 --> 00:02:28,599 guess there's uh. You know, they've got their military version close to deployment 27 00:02:28,680 --> 00:02:30,280 now, and you can see in this article that their people in the army 28 00:02:30,319 --> 00:02:36,120 weren't wearing it, and you know it's been a big project for them. 29 00:02:36,520 --> 00:02:42,840 Yeah, and it's a big, multi billion dollar contract they have with the 30 00:02:43,000 --> 00:02:49,159 army, and so hopefully that will turn into good stuff for us in the 31 00:02:49,240 --> 00:02:53,199 future. Yeah. I presume this is just Windows ten, but the holographic 32 00:02:53,360 --> 00:02:55,960 edition, so Windows eleven. Yeah, I think it's still ten. That 33 00:02:57,479 --> 00:03:00,039 they're using for the Hololends too. I don't know they've actually moved up to 34 00:03:00,080 --> 00:03:04,360 the new Okay, which why you know they because Windows ten is getting a 35 00:03:04,400 --> 00:03:07,199 twenty three h two. So this is the equivalent, just for the holographic 36 00:03:07,360 --> 00:03:09,240 edition, right right, right right? Okay? Well, anyway, so 37 00:03:09,400 --> 00:03:15,520 you can click on the HoloLens to release notes from October tenth, twenty twenty 38 00:03:15,599 --> 00:03:21,080 three and read all about it. So there's no real product as far as 39 00:03:21,159 --> 00:03:23,919 I know yet. You can still buy a HoloLens two, but you can 40 00:03:23,960 --> 00:03:27,960 get a Halo Lends two. It's four grand and it's been years. Yeah, 41 00:03:28,000 --> 00:03:30,159 it's been years, and I don't see you hardware on the immediate horizon. 42 00:03:30,240 --> 00:03:34,159 But you know, they all they've they've basically been talking about we're waiting 43 00:03:34,319 --> 00:03:38,159 for the right chipset, like the right REV to come along. It gives 44 00:03:38,240 --> 00:03:39,960 us enough battery life and enough things that it's gonna be worth making the next 45 00:03:40,000 --> 00:03:45,080 one. Right. Meanwhile, Meta Quest three, yeah, is five hundred 46 00:03:45,080 --> 00:03:49,280 dollars. Yeah, although it's a VR headset, not a VR, not 47 00:03:49,400 --> 00:03:53,680 an AR. Yes, but everybody says it's a lot of fun. Yeah. 48 00:03:53,960 --> 00:03:55,319 I've got a quest too. I have to decide whether I want to 49 00:03:55,360 --> 00:03:59,120 upgrade a quest three? Yeah, me too. Cool. Anyway, that's 50 00:03:59,159 --> 00:04:01,360 what I'm thinking about today. Who's talking to us today right well, knowing 51 00:04:01,360 --> 00:04:04,840 we're going to talk about sustainable development. Yes. I dug back into the 52 00:04:05,000 --> 00:04:11,280 archives and I found a show from twenty eleven, episode seven twenty five, 53 00:04:11,639 --> 00:04:16,240 so twelve hundred episodes ago. In December twenty eleven. I think I was 54 00:04:16,360 --> 00:04:21,360 fourteen years old at that time. I don't think that's true. We talked 55 00:04:21,399 --> 00:04:28,040 to Kathy Malone, who was a tech ed speaker about being a green developer. 56 00:04:28,959 --> 00:04:31,959 Great conversation, you know, forty five minutes long. And know she's 57 00:04:31,959 --> 00:04:35,399 gotten here, there and everywhere, and admittedly quite a while ago. And 58 00:04:35,439 --> 00:04:39,480 there was a comment on the show at the time, eleven years ago. 59 00:04:39,560 --> 00:04:44,439 This is from Timothy Clinky who said, in our company, telling a developer 60 00:04:44,480 --> 00:04:47,639 to make their code more environmentally friendly means they need to make it more green 61 00:04:48,079 --> 00:04:51,920 by adding comments they come up green in they tell a sense you know, 62 00:04:53,160 --> 00:04:56,480 Oh, come on, it's what it does. It really does that. 63 00:04:57,160 --> 00:04:59,839 I thought it was funny, Tim, I thought it was real funny. 64 00:05:00,160 --> 00:05:01,639 I did not. It's pretty funny. Okay, Well, you say, 65 00:05:01,920 --> 00:05:04,160 Tim, I'm going to send you a copy of music Code by whether Carl 66 00:05:04,199 --> 00:05:08,720 thought it was funny or not. It made me laugh. So and if 67 00:05:08,759 --> 00:05:11,439 you'd like a copy of music Code by write a comment on the website at 68 00:05:11,480 --> 00:05:14,399 don at Rocks dot com or on the facebooks. We publish every show there 69 00:05:14,680 --> 00:05:15,839 and if you comment there and a read on the show, we'll send you 70 00:05:15,879 --> 00:05:18,959 a copy of music Code by write a green comment, please very green. 71 00:05:19,040 --> 00:05:21,839 You know, if it's only put the double slash on it, it'll be 72 00:05:21,879 --> 00:05:27,519 Greening'll be fine. Yeah, And you can definitely follow us on Twitter if 73 00:05:27,560 --> 00:05:30,079 you want to. But the cool kids are over at Macedon. I'm at 74 00:05:30,160 --> 00:05:34,319 Carl Franklin at Techhub dot social and I'm Rich Campbell at Macedon dot social. 75 00:05:34,480 --> 00:05:39,240 Send us a two We will read them. And with that, let me 76 00:05:39,399 --> 00:05:50,920 introduce the one the only Leah Close. Yes, let's go okay. Leah 77 00:05:51,439 --> 00:05:56,399 is a passionate technologist that loves to get people together to solve problems. With 78 00:05:56,560 --> 00:06:00,199 about ten years of working as a software engineer, Leah has been work working 79 00:06:00,279 --> 00:06:06,319 in London as a c sharp back end engineer delivering value to fund apps clients 80 00:06:06,639 --> 00:06:13,480 for the last five years. Fund Apps plural fundaps is clients for the last 81 00:06:13,519 --> 00:06:16,720 five years. Yes, yes, and we're talking about sustainable development. But 82 00:06:16,839 --> 00:06:20,680 first, welcome to dot NetRocks. Thank you very much. It's very sure 83 00:06:20,720 --> 00:06:25,839 to have you here and such a cool topic. Yes, and I really 84 00:06:25,920 --> 00:06:29,079 liked that we talked about this, you know, twelve years ago, because 85 00:06:29,120 --> 00:06:31,879 I've got to hope things have progressed. So when you say sustainable development, 86 00:06:32,519 --> 00:06:39,439 what do you mean? Yeah, I think sustainable development changed over the years. 87 00:06:40,560 --> 00:06:43,519 To me, it means I only buy a laptop every other year, 88 00:06:43,800 --> 00:06:46,480 nice, not every year. Yeah, I mean when we talk about sustainable 89 00:06:46,480 --> 00:06:49,199 there are like three aspects, like if you read online if you google it, 90 00:06:50,160 --> 00:06:56,199 economical, social, and environmental. Obviously, because you're here talking about 91 00:06:56,240 --> 00:07:01,920 green we are talking a bit more about environmental impact. I think, you 92 00:07:02,000 --> 00:07:06,279 know, like last decades or even more, it was how do you make 93 00:07:06,360 --> 00:07:14,279 developers code most sustainably, not create tech? That and all of that to 94 00:07:14,399 --> 00:07:17,399 that To me, that that says maintainable rather than which is a kind of 95 00:07:17,439 --> 00:07:21,519 sustainability exactly. But it's a different I would say, it's a different kind 96 00:07:21,560 --> 00:07:27,160 of sustainability. It's not environmental, right, sustainability, Like they're different aspects. 97 00:07:27,199 --> 00:07:31,920 Sure, without and can I stay the obvious like planet is fucked. 98 00:07:32,279 --> 00:07:39,959 So we are more thinking about environmental aspects nowadays. Don't use AI, don't 99 00:07:40,040 --> 00:07:45,319 use the cloud and listen, the planet's fine. It's all of us that 100 00:07:45,439 --> 00:07:47,680 have got the problem. Like the planet's going to make it. We may, 101 00:07:47,920 --> 00:07:54,000 we may be burned off, still be here. You know, when 102 00:07:54,040 --> 00:07:57,560 I do the climate change talk for the high schoolers, you talk about the 103 00:07:57,800 --> 00:08:01,720 in the e scene, which is like twenty four million years ago, carbon 104 00:08:01,759 --> 00:08:05,600 dioxide was at eleven hundred parts per million, right as opposed to that right 105 00:08:05,639 --> 00:08:09,279 now for eighteen it's like planet was here. Heck, it was just a 106 00:08:09,279 --> 00:08:13,319 big old pile algae, that's all. You know, Antarctica had probably had 107 00:08:15,199 --> 00:08:20,000 trees growing on it. So it is we are. The troubles we're having 108 00:08:20,399 --> 00:08:26,759 are impacting humans being able to function here exactly. So all of the fires 109 00:08:26,000 --> 00:08:31,399 and the floods and sea levels rising, and that's just direct impact, like 110 00:08:31,600 --> 00:08:35,120 people not being able to live on this planet. Right there is also like 111 00:08:35,240 --> 00:08:41,080 all the health issues and what's causing that the subtler parts yea of all these 112 00:08:41,120 --> 00:08:46,879 things. So you basically have three categories. But I'm immediately thinking there's got 113 00:08:46,000 --> 00:08:50,159 to be hundreds of issues and topics that we could talk about, right, 114 00:08:50,320 --> 00:08:54,080 Yeah, it's got to be hundreds, and I'm sure there is. But 115 00:08:54,399 --> 00:09:01,320 as software engineers and software professionals, we can talk about cool and what we 116 00:09:01,480 --> 00:09:07,759 can do. And I learned yesterday that twenty to twenty five percent of electricity 117 00:09:07,879 --> 00:09:13,480 usage can be attributed to a digital sector. Sure, And I mean that's 118 00:09:13,559 --> 00:09:18,440 why we are talking about it even more, and that's why it is more 119 00:09:18,519 --> 00:09:22,600 important. And even Richard, you had to talk about future of electricity, 120 00:09:24,240 --> 00:09:28,679 of energy, actually the future of energy. Yeah, and our move towards 121 00:09:28,080 --> 00:09:35,919 using less carbon emitting energy sources which certainly helps matters. Yes, but it 122 00:09:35,159 --> 00:09:39,320 is interesting in the news, especially around the large language models, which are 123 00:09:39,519 --> 00:09:43,360 very compute intensive. Now there's folks talking about, well, just how much 124 00:09:43,440 --> 00:09:48,120 energy is that? Is it really necessary? Also the cooling parts of it, 125 00:09:48,440 --> 00:09:54,919 you're complaining the cloud vendors and Microsoft's among them, have come up with 126 00:09:54,000 --> 00:10:00,559 strategies to make very inexpensive cooling systems for computers, but they're fairly water intensive, 127 00:10:01,759 --> 00:10:05,120 and so the you know the sort of reality of you're taking fresh water 128 00:10:05,399 --> 00:10:09,559 and you're essentially missing it into these data centers to cool them. What could 129 00:10:09,600 --> 00:10:13,519 go wrong? Well, they tell you what the cost of it. I 130 00:10:13,720 --> 00:10:20,000 built an I nine machine in my house and it has a giant heat sting 131 00:10:20,399 --> 00:10:22,879 on the CPU. And when I say giant, like I can I can 132 00:10:24,000 --> 00:10:26,679 barely get the case cover on it, right, that's how big it is. 133 00:10:26,759 --> 00:10:30,759 It's probably a half a foot square, but it's probably a couple of 134 00:10:30,840 --> 00:10:35,720 pounds of cover. Yeah right, yeah, But it's really quiet and it's 135 00:10:35,879 --> 00:10:39,879 really cool, right, and it doesn't require water cooling, and it's an 136 00:10:39,919 --> 00:10:43,919 nine nine like it's a beefy machine. Yeah. And the graphics also, 137 00:10:43,200 --> 00:10:46,360 yeah, the modern video cards. You look at the family assemblies that are 138 00:10:46,399 --> 00:10:50,240 on those. That's not trivial either. The real question is if you pinned 139 00:10:50,279 --> 00:10:54,159 it, like if you had that process or working flat out, how warm 140 00:10:54,240 --> 00:10:56,879 would it get? I mean, you know question. We know that cooler 141 00:10:56,919 --> 00:11:01,799 would disperse that heat. But that's just means it's in the room and it's 142 00:11:01,840 --> 00:11:03,159 a let's say say it's a client machine, right, yeah, it's not 143 00:11:03,240 --> 00:11:09,240 a server machine serving up lots of CPU. Yeah. But I wouldn't want 144 00:11:09,360 --> 00:11:13,799 my I wouldn't want my data center to be completely reliant on solar energy either, 145 00:11:15,039 --> 00:11:20,519 because okay, my servers down, yeah, I mean they use a 146 00:11:20,519 --> 00:11:24,399 lot of hydrological power like one. Since I would think that in general, 147 00:11:24,480 --> 00:11:30,799 the cloud movement has meant that a relatively few number of people are responsible for 148 00:11:30,879 --> 00:11:35,960 a tremendous amount of compute availability, and when green requirements are put on them, 149 00:11:37,000 --> 00:11:39,360 they have a better chance of implementing it than we do if everybody was 150 00:11:39,399 --> 00:11:43,679 building their own data center. Yeah, I mean it's tier if numbers right, 151 00:11:43,960 --> 00:11:48,840 But like economy of scale, if we all use cloud providers, surely 152 00:11:50,519 --> 00:11:54,600 we can push them to do something better and in a better way. Plus 153 00:11:54,759 --> 00:11:58,519 we can all reuse these resources instead of every one of us having a server 154 00:11:58,679 --> 00:12:03,559 at home. Yeah, running ten percent utilization. I think it's also that's 155 00:12:03,679 --> 00:12:07,960 true. I think it's also making us lazy. Yeah, well, I 156 00:12:07,120 --> 00:12:11,159 don't know that we've ever prioritized this. Like if you think about the priorities 157 00:12:11,200 --> 00:12:16,399 of a software developer, it's like one, make it work. Two, 158 00:12:16,440 --> 00:12:22,279 make it work fast right, Three make it work so that somebody else could 159 00:12:22,399 --> 00:12:26,759 possibly fix it. Like, where is any other consideration at that point? 160 00:12:26,840 --> 00:12:31,039 Make it available all the time. Yeah, yeah, I don't remember these 161 00:12:31,120 --> 00:12:35,480 times because probably I was not born or I was just a kid. But 162 00:12:37,240 --> 00:12:41,840 I mean, once upon a time, memory and CPU and hardware were very 163 00:12:41,919 --> 00:12:48,240 expensive and constraint and huge. Yeah, it was a massive expense. So 164 00:12:48,399 --> 00:12:52,080 like you did need to consider how you code it was it needed to be 165 00:12:52,240 --> 00:12:56,720 super efficient. It need to be for computers. That's how we got solid 166 00:12:56,759 --> 00:13:01,879 principles, like we were writing from machines to be efficient well because we had 167 00:13:01,919 --> 00:13:05,360 to, because we had to exactly. We also limited the feature stat accordingly, 168 00:13:05,799 --> 00:13:09,159 like yes, once upon a time you could not waste cycles on a 169 00:13:09,279 --> 00:13:13,759 guy, much less a touchscreen. Yeah, you know that's a lot of 170 00:13:13,799 --> 00:13:18,759 extra compute. Yeah. We move now into a world of memory is cheap, 171 00:13:18,840 --> 00:13:22,200 CPUs, cheap, storage is infinite. Storage is infinite. You can 172 00:13:22,279 --> 00:13:26,159 scale these servers, you can add more nodes, you can add more memory 173 00:13:26,200 --> 00:13:31,440 and CPU on fly. It's very easy. Yeah. Yeah. And everything 174 00:13:31,559 --> 00:13:35,320 was about performance, like everything was like I wanted them immediately. Yeah, 175 00:13:35,600 --> 00:13:37,159 and that's a great version of performance. Not even the code is fast, 176 00:13:37,240 --> 00:13:41,279 but you got the code written fast, like the performance of get the feature 177 00:13:41,320 --> 00:13:43,759 out. Get the feature out, but also run it fast. But not 178 00:13:43,960 --> 00:13:48,679 by writing the code performance all scale it and like it's going to be one 179 00:13:48,799 --> 00:13:52,879 hundred nodes and it's going to be super quick. I will fix this with 180 00:13:52,000 --> 00:13:58,600 the giant compute button exactly. So do you see your role as shining light 181 00:13:58,679 --> 00:14:01,840 on these inefficiencies and saying, you know, hey, we have to get 182 00:14:01,879 --> 00:14:07,320 better at this or do you have solutions to these problems that can be implemented 183 00:14:07,360 --> 00:14:13,840 today. Yes and no. Well it's a complicated question, but I do 184 00:14:13,039 --> 00:14:20,519 think we need to change something. It's either for the planet because we see 185 00:14:20,559 --> 00:14:28,200 what's happening, we can get smarter at it. But it's also if we 186 00:14:28,399 --> 00:14:33,360 don't do something as software engineers, our boss is going to come after us 187 00:14:33,399 --> 00:14:37,519 and it's going to say you need to write it in this way because either 188 00:14:37,679 --> 00:14:41,840 government told me that I need to, or because we can't sign any more 189 00:14:41,919 --> 00:14:45,960 clients because they are all asking for all of these checks that we are thinking 190 00:14:46,039 --> 00:14:50,759 about these things. I mean even in Fundos now, all of our prospects 191 00:14:50,799 --> 00:14:54,480 are looking for these vendor checks and like are you doing this? Are you 192 00:14:54,559 --> 00:14:58,840 doing that? And the reason why we get the signature, well not only 193 00:14:58,960 --> 00:15:03,200 because of that, but it's because we are thinking about these things and at 194 00:15:03,200 --> 00:15:09,519 the moment they are optional. They're not there if you would, you know, 195 00:15:09,799 --> 00:15:13,240 if companies thinking about it and thinking about the environment. But in my 196 00:15:13,399 --> 00:15:18,679 opinion. Soon this is going to be a must. It's going to be 197 00:15:18,759 --> 00:15:20,480 a mandatory. Yeah. Well, and I wonder if it just won't become 198 00:15:20,519 --> 00:15:24,799 table sticks. Like Microsoft and all of these tech companies have been talking a 199 00:15:24,879 --> 00:15:30,559 lot about zero carbon and zero impact technology centers. Then you know, they 200 00:15:30,639 --> 00:15:33,960 help check those boxes off for you. And one could argue that some of 201 00:15:35,039 --> 00:15:39,320 this is quote unquote greenwashing, like are they really doing anything about it? 202 00:15:39,879 --> 00:15:41,039 But it also makes it easy for us. It's like, hey, just 203 00:15:41,120 --> 00:15:45,360 write your code, don't worry, we'll make sure it has no impact on 204 00:15:45,440 --> 00:15:48,879 the planet. Right, But they also bill us by the minute for the 205 00:15:48,000 --> 00:15:50,240 compute. Like there is an angle on this. It says, if I 206 00:15:50,320 --> 00:15:54,960 write my code more efficiently, it will actually cost me less. You know, 207 00:15:56,039 --> 00:15:58,559 when you own the computer and you don't fully utilize it anyway, you 208 00:15:58,600 --> 00:16:02,320 don't care. You have all the computing and you have enough compute for what 209 00:16:02,360 --> 00:16:06,799 you're doing. But if you're charged by the transaction and by the compute load 210 00:16:07,480 --> 00:16:10,600 writing more efficient code, that's money in the bank. So get up, 211 00:16:10,679 --> 00:16:15,120 Copilot, I'll go. Let's say while true and get up COPI Mmmm, 212 00:16:15,320 --> 00:16:18,879 you can't do that. That's a tight loop. Yeah, I mean we 213 00:16:18,360 --> 00:16:22,720 do have servilis at the moment. Like all of these cloud providers, you 214 00:16:22,799 --> 00:16:29,159 can spin up. You can spin up containers when you need them. Then 215 00:16:29,200 --> 00:16:32,799 you can scale them down and shut them down. You can turn on these 216 00:16:32,960 --> 00:16:37,399 lambders. I mean even us when we code, we have the massive problems 217 00:16:37,440 --> 00:16:41,679 of things taking too long. Sure, and you develop something for a client 218 00:16:41,720 --> 00:16:45,679 that is very simple, things take I don't know, five minutes, and 219 00:16:45,759 --> 00:16:48,919 then it scales up. You get a more complex client takes two and a 220 00:16:49,000 --> 00:16:52,679 half hours. For let's say a badge process. You remove a too string 221 00:16:53,679 --> 00:16:57,399 that it's done tens of millions of times. It saves half an hour. 222 00:16:57,600 --> 00:17:02,799 Yeah, no kidding, scale effect there. You are not paying for that 223 00:17:03,080 --> 00:17:08,160 half an hour of these I don't know three hundred containers being on, especially 224 00:17:08,160 --> 00:17:11,240 if there's a big pipeline running like you probably like fewer instances too, Like, 225 00:17:11,440 --> 00:17:15,759 yeah, that is money you're not spending. Yeah do it well. 226 00:17:15,880 --> 00:17:22,519 I think there is a direct link from how much you're paying something to how 227 00:17:22,640 --> 00:17:29,519 much energy that use. Sure, Like you're basically eliminating this infrastructure that you 228 00:17:29,640 --> 00:17:36,680 don't need to run. So there are a couple of things you can just 229 00:17:37,359 --> 00:17:41,039 run these machines when you need them when you're running something instead of constantly having 230 00:17:41,119 --> 00:17:48,319 them on. So elasticity elasticity sufficient exactly. That's one thing. The other 231 00:17:48,400 --> 00:17:52,480 thing is write more performing code so you don't need to have them on for 232 00:17:52,599 --> 00:17:55,599 an hour. Maybe you need them just for half an hour, and that 233 00:17:55,880 --> 00:17:59,720 is good for the environment. You're assuming less resources and happens to save you 234 00:17:59,759 --> 00:18:06,480 money exactly. But if you perform the same amount of CPU processing that you 235 00:18:06,559 --> 00:18:07,960 would do in a half an hour, if you do that in five minutes, 236 00:18:08,000 --> 00:18:12,039 you're jacking up the CPU. Are you really saving anything? Maybe? 237 00:18:12,519 --> 00:18:17,319 Yeah. I think that's just what I'm saying is the developers need to think 238 00:18:17,400 --> 00:18:22,559 about not just going faster. But it's really kind of a compute cycle problem. 239 00:18:22,359 --> 00:18:29,160 I like your lambda scenario because typically when I see a spikelad like that, 240 00:18:29,240 --> 00:18:30,480 I've got a big old queue of calls, like there's a bunch of 241 00:18:30,559 --> 00:18:33,559 work that's just come in. We've got a whole bunch of stuff to run, 242 00:18:33,319 --> 00:18:37,720 and it is containers under the hood. We all know that, right, 243 00:18:37,119 --> 00:18:40,359 we just don't have to deal with it lights them up, and to 244 00:18:40,440 --> 00:18:41,440 pump through it as quickly as possible, it's going to light up a lot 245 00:18:41,519 --> 00:18:45,039 of them, but if you've made them more efficient, you'll literally light fewer. 246 00:18:45,960 --> 00:18:48,319 So you know, at the in the end, you're going to see 247 00:18:48,359 --> 00:18:52,319 how many instances it was the same workload, but how many parallel instances did 248 00:18:52,359 --> 00:18:56,599 it need to execute it? And because it was finishing faster, it won't 249 00:18:56,680 --> 00:19:00,599 as lit as many the same one will be reused more like there's real savings 250 00:19:00,680 --> 00:19:04,160 to be had there. I guess that's for a puzzle for everybody to individually 251 00:19:04,240 --> 00:19:07,920 figure out, sure, like how to make things more efficient. Well, 252 00:19:07,119 --> 00:19:11,279 and we played these map reduce games before where it's like, hey, I 253 00:19:11,319 --> 00:19:14,400 can run map reduce on one machine. Don't do that, it's dumb, 254 00:19:14,880 --> 00:19:17,319 but you'll, you know, and it'll take an hour. I can run 255 00:19:17,440 --> 00:19:23,759 on six machines and it'll take twelve minutes. I can run it on sixty 256 00:19:23,839 --> 00:19:27,880 machines and it'll take four minutes. So you know, so what's the environmental 257 00:19:27,920 --> 00:19:32,640 impact run? You're exactly right, like one would argue that six machines was 258 00:19:32,720 --> 00:19:37,519 kind of the efficient middle point where it took less time and the overhead of 259 00:19:37,599 --> 00:19:41,839 the mapping part and the recombining part wasn't as high as over distributing it, 260 00:19:42,319 --> 00:19:45,960 where you know, we may be able to cut the time in half, 261 00:19:45,000 --> 00:19:48,640 but we up the load by five times to make it work. I don't 262 00:19:48,640 --> 00:19:52,160 know if you've thought about this lab, but I know Richard has. Do 263 00:19:52,240 --> 00:19:59,680 you think quantum computing is going to change the equation in terms of sustainability? 264 00:20:00,359 --> 00:20:06,920 And you know in our data centers? Can you think of a quantum data 265 00:20:06,960 --> 00:20:11,480 center what that would look like. I don't know that much about quantum computing, 266 00:20:11,519 --> 00:20:18,240 but but there is an interesting example. My brother, he's a physicist 267 00:20:18,480 --> 00:20:22,200 and he works with I think solid states physics and kind of combination with chemistry 268 00:20:22,240 --> 00:20:26,920 and finding new materials, and this was an interesting thing. I was talking 269 00:20:26,960 --> 00:20:30,640 to him the other day, and what he's trying to do is use machine 270 00:20:30,720 --> 00:20:37,920 learning to find these different to write programs that find these different combinations of materials 271 00:20:38,720 --> 00:20:44,640 that would otherwise be done with quantum computing, which takes a lot of power, 272 00:20:44,680 --> 00:20:48,960 which takes a lot of energy, so mostly in the cooling. Yeah. 273 00:20:49,319 --> 00:20:53,720 So I can't say anything about quantum computers, like how efficient they are? 274 00:20:53,839 --> 00:20:56,519 What are dytaing that there? Are? They exist? Yet? Do 275 00:20:56,640 --> 00:21:00,640 they? I think they kind of do, but they're with a limitations. 276 00:21:00,680 --> 00:21:03,160 I mean, I always look at quantum computers as a class of supercomputer, 277 00:21:03,839 --> 00:21:10,039 which is to say, complex problems that take a long time to stage, 278 00:21:10,720 --> 00:21:12,519 potentially take a long time to process, and then you sort of have to 279 00:21:12,599 --> 00:21:18,119 act on it. They tend to be very deterministic problems like computing airflow over 280 00:21:18,200 --> 00:21:22,680 a wing, those kinds of massively complex computers. General compute, you don't 281 00:21:22,720 --> 00:21:26,759 think that. I don't think there's a generic compute scenario here, just because 282 00:21:26,759 --> 00:21:30,200 it's just the nature of the kind of problem space that it's good at that 283 00:21:30,359 --> 00:21:33,920 you wouldn't use it for anything else. And again, we may not be 284 00:21:34,000 --> 00:21:41,200 able to see that today, just because they are bespoke machines cooled liquid helium 285 00:21:41,279 --> 00:21:45,640 with people in white lab coats taking care of them. And that might transform 286 00:21:45,680 --> 00:21:48,759 at some point, but not anytime soon. Are there any other technologies that 287 00:21:49,400 --> 00:21:55,319 either you can think of coming on the horizon that might make this make the 288 00:21:55,440 --> 00:21:57,920 data center problem a little more efficient? Well, I think a big one, 289 00:21:59,039 --> 00:22:00,680 and I'm going to I'm already to take from your show, No go 290 00:22:00,759 --> 00:22:04,759 ahead, but I'm learning something from an old school hardware geek. As you 291 00:22:04,960 --> 00:22:11,400 know, the move to arm compute on the server is a simplification of CPU 292 00:22:11,559 --> 00:22:18,799 architecture a big simplification of CPU architecture, so reduced instructions, it's only a 293 00:22:18,839 --> 00:22:22,920 reduced instruction set, it's fewer electrons need to move around to get the same 294 00:22:23,079 --> 00:22:29,839 result. The downside of the Intel architecture that lives in a lot of servers 295 00:22:30,319 --> 00:22:33,759 is that it's persistence the nineteen seventies has just been added to and while it's 296 00:22:33,799 --> 00:22:40,559 performant, it is complicated, and so it takes. While it's fast and 297 00:22:40,680 --> 00:22:42,759 we can do the results, it's just not what we do today with the 298 00:22:42,839 --> 00:22:45,759 resource that we have. It's jumping forward through more hoops than it needs. 299 00:22:45,839 --> 00:22:48,559 It's a little brute force, isn't it. It's just because it's old, 300 00:22:48,960 --> 00:22:53,599 because compatibility was more important, and so maintain that cambility as we kind of 301 00:22:53,680 --> 00:22:59,240 dragged baggage from the eighties with us all the way into the twenty twenties. 302 00:22:59,319 --> 00:23:03,440 Now talk about my latest MacBook. Yeah, the M ones and the M 303 00:23:03,519 --> 00:23:07,920 twos being some of the finest machines ever made. So when you say and 304 00:23:07,119 --> 00:23:14,559 the battery lasts forever and and well, and that chip is an incredibly large 305 00:23:14,720 --> 00:23:18,720 chip, it's not actually that complex a chip. Because the ARM architecture simpler, 306 00:23:18,799 --> 00:23:25,160 they have room to include a GPU, include an NPU, include more 307 00:23:25,240 --> 00:23:29,359 memory, all things that make it perform better. So we're starting to see 308 00:23:29,359 --> 00:23:33,519 a push towards ARMS insert in the server clausets and in the cloud. And 309 00:23:33,599 --> 00:23:37,440 I think you're going to talk about double digit decreases in energy consumption for the 310 00:23:37,559 --> 00:23:42,000 same results. Wow. And I think that's a good point. Like Pentium 311 00:23:42,119 --> 00:23:48,960 is now talking like working on chips that are low consumption. They're no longer 312 00:23:48,039 --> 00:23:52,480 saying when marketing, like even Apple with their chips, they no longer saying 313 00:23:52,559 --> 00:23:56,119 this is super fast and performance, like how fast do we need? Like 314 00:23:56,720 --> 00:24:00,200 you know, like we're it's fast enough. We are now in a world 315 00:24:00,200 --> 00:24:04,680 where it's like this is more sustainable battery will like it will use less energy. 316 00:24:06,160 --> 00:24:10,240 So it's not just us like Apple, like all of these companies are 317 00:24:10,279 --> 00:24:14,119 also marketing on this. When you're right that like ARMS pitch has always been 318 00:24:14,200 --> 00:24:17,880 batter efficiency, those M ones and M two's all day battery in your laptop 319 00:24:18,079 --> 00:24:21,039 and it's not a big battery. The idea that you put it into server 320 00:24:21,200 --> 00:24:26,200 because literally be a less kilootts is just another level of the same thinking. 321 00:24:26,440 --> 00:24:30,680 So we should be able to get there MEAs you're cool well, inevitably because 322 00:24:30,720 --> 00:24:34,680 you're literally moving the electrons around less, the pipelines aren't as complicated as not 323 00:24:34,720 --> 00:24:38,240 as many instruction sets, and in the end, all that heat comes from 324 00:24:38,359 --> 00:24:42,160 vibrating electrons. Well, there you go, So we don't have to fix 325 00:24:42,200 --> 00:24:48,160 our code. We're fine. Just you know, dot net CLR will automatically 326 00:24:48,319 --> 00:24:52,319 compile onto ARM and we're going to get the computer results you want. I 327 00:24:52,359 --> 00:24:56,200 think there is also another lens to it, like we are using all of 328 00:24:56,279 --> 00:25:02,319 these programs hopefully to bring some value. So what problems are they're solving, 329 00:25:02,400 --> 00:25:06,000 Like when we're talking about machine learning, when we are talking about quantum computing, 330 00:25:06,400 --> 00:25:10,000 when we are talking about any software that we develop, what or even 331 00:25:10,079 --> 00:25:17,839 AI and crypto like non crypto like by blockchain, and like this distributed system 332 00:25:17,960 --> 00:25:26,079 is like what problems are we solving? How much energy they use might be 333 00:25:26,240 --> 00:25:30,319 okay if we are thinking about the problems that we are solving, and maybe 334 00:25:30,480 --> 00:25:34,680 those problems would be using even more energy if you're not doing this. So 335 00:25:36,319 --> 00:25:38,119 I think we need to kind of look at the bigger picture, not in 336 00:25:38,240 --> 00:25:42,960 the like in smaller isolation, this is using x y z amount of energy. 337 00:25:44,599 --> 00:25:47,759 It's like, what is that replacing. Yeah, it's and the trick 338 00:25:47,880 --> 00:25:51,279 is a comparison because they're not going to be easy comparison. Now, guys, 339 00:25:51,440 --> 00:25:55,200 I'm going to ask you to pause a moment for these very important messages. 340 00:25:59,559 --> 00:26:03,079 And we're it's Dot at Rocks. I'm Carl Franklin, that's my friend 341 00:26:03,160 --> 00:26:07,240 Richard Campbell, and we're talking to Leah Maladiger. Yes, I got it 342 00:26:07,319 --> 00:26:15,000 right again about sustainable computing. And like you know, there's so many topics 343 00:26:15,079 --> 00:26:18,480 subtopics under here, but we're going for the big one, which is data 344 00:26:18,559 --> 00:26:21,720 centers, or at least that's what we have been talking about. Yes, 345 00:26:22,079 --> 00:26:26,519 writting more efficient code is important too, but after the data center problem, 346 00:26:26,880 --> 00:26:30,839 what's number two on your list? Well, it's in his data centers where 347 00:26:30,880 --> 00:26:37,440 we run our code. So these hyper scalers like AWS, Google, like 348 00:26:37,720 --> 00:26:44,200 Microsoft. I think it's it's really just just thinking about it, Like we 349 00:26:44,319 --> 00:26:47,039 need to start talking about this. I wonder if we can start getting the 350 00:26:47,640 --> 00:26:52,680 instrumentation where I am talking about, you know, watts per transaction, we 351 00:26:52,799 --> 00:26:57,759 can also be talking about that greenwashing that you mentioned earlier, Richard, and 352 00:26:57,839 --> 00:27:00,240 I think that's a big pain. I think with anything that we do, 353 00:27:00,400 --> 00:27:06,920 that's a big pain, like because people stop trusting the numbers and their impact. 354 00:27:07,559 --> 00:27:12,200 If all of these companies are scamming in a way, let's go with 355 00:27:12,279 --> 00:27:18,160 that way, because then you stop trusting it. And even if okay, 356 00:27:18,400 --> 00:27:21,839 if we're thinking about where we run our software, how we write our software, 357 00:27:21,880 --> 00:27:26,039 and how performing it is, we can't reduce it to zero, right 358 00:27:26,000 --> 00:27:32,559 like we would stop living or we would not write anything, and that's the 359 00:27:32,599 --> 00:27:34,839 best thing for planet. But I don't know if that's true. But yeah, 360 00:27:36,000 --> 00:27:38,079 okay, but you know, like we can't do that. So like 361 00:27:38,480 --> 00:27:42,119 it's also offsetting, and there is a lot of greenwashing there as well. 362 00:27:42,759 --> 00:27:51,279 So I think our next step is as well, kind of removing the fog 363 00:27:51,440 --> 00:27:56,599 around all of that or trying to figure out creating more transparency, more transparency 364 00:27:56,680 --> 00:28:02,279 maybe standards. Like it's electricity, right, like we know how to calculate 365 00:28:02,319 --> 00:28:07,920 the electricity, but looking at the environmental impact of it given electrical sources challenging. 366 00:28:07,799 --> 00:28:11,440 It depends on what you're using. I mean, I've had a chance 367 00:28:11,480 --> 00:28:15,920 to tour the Quincy Data Center in eastern Washington, which you're really test sites 368 00:28:17,079 --> 00:28:19,720 for Azure, but they're powered by hydro electric power, which is about as 369 00:28:19,759 --> 00:28:22,960 green as you're going to get. You know, it's not like hydraulic power 370 00:28:23,000 --> 00:28:26,839 doesn't have environmental consequences. You do flood a valley and a bunch of plant 371 00:28:26,920 --> 00:28:30,400 life goes underwater, and other things like, there are consequences. Yeah, 372 00:28:30,400 --> 00:28:34,440 whatever happened to that whole? You know, let's put a data center underwater. 373 00:28:36,519 --> 00:28:40,119 You know what? Funny thing speaking of green washing or whatever, I 374 00:28:40,200 --> 00:28:44,440 talked. I talked to Mark Rasinovich about that on run As a while back 375 00:28:44,920 --> 00:28:48,039 in the build time frame, and we talked about Project Natick, which was 376 00:28:48,160 --> 00:28:53,680 this test of putting essentially racks in a metal cylinder under the ocean. To 377 00:28:55,039 --> 00:28:57,519 do they take a shipping container, put all the servers in it, have 378 00:28:57,680 --> 00:29:00,559 it water cooled, boom, it's a it's a pressure vessel, so it's 379 00:29:00,640 --> 00:29:03,599 cylindrical, and they tested in the orkneys and so forth. But in the 380 00:29:03,720 --> 00:29:10,799 end it came down to data sovereignty controls, like just being able to protect 381 00:29:10,839 --> 00:29:15,720 the machines, so uh, like who owns it? Well, not only 382 00:29:15,799 --> 00:29:18,640 who owns it, who has access to it? Like one of the problems 383 00:29:18,680 --> 00:29:22,039 with it being underwater is you don't know if people are approaching it. We've 384 00:29:22,079 --> 00:29:26,319 noticed, but we're living in a world lately was a few bad state actors 385 00:29:26,359 --> 00:29:30,000 out there. This is exactly what Sinners was talking about. It's like, 386 00:29:30,759 --> 00:29:34,680 you know, we like, we like our big buildings made of concrete with 387 00:29:34,839 --> 00:29:40,799 big fences around them, and and you know, strict controls on access, 388 00:29:41,119 --> 00:29:42,559 and you can't do that when it's sitting at the bottom of bey in a 389 00:29:42,720 --> 00:29:48,279 city. So they in the end it became improuctial. The only thing they 390 00:29:48,319 --> 00:29:49,680 show that one of the things that came out of that was really interesting is 391 00:29:49,720 --> 00:29:56,440 that the hardware lasted longer when the environment was completely undisturbed, that the machine 392 00:29:56,480 --> 00:30:03,519 failure rates were actually lower because has nobody walked by them because they were in 393 00:30:03,559 --> 00:30:07,400 a middle cylinder run of the ocean. So did you ever hear about Subsea 394 00:30:07,559 --> 00:30:11,680 Cloud. That was another company that planned to have a commercially available under sea 395 00:30:11,799 --> 00:30:17,400 data center opering off the U operating off the US coast before the end of 396 00:30:17,440 --> 00:30:19,359 twenty twenty two. Yeah, I don't think it made it. That didn't 397 00:30:19,400 --> 00:30:22,920 happen apparently, No, an, I think you get back to the same 398 00:30:22,960 --> 00:30:26,200 old problem. It costs more and it's questionable safety. Yeah, right, 399 00:30:26,440 --> 00:30:30,799 Like the ocean is a dangerous place, sure, and in a secretive place, 400 00:30:30,880 --> 00:30:33,799 So I think it's problematic. Overall, there was a bunch of benefits 401 00:30:33,799 --> 00:30:37,720 to it, without a doubt. I mean it was the logical thing was 402 00:30:37,720 --> 00:30:40,079 you were going to be able to put them closer to cities where the latensy 403 00:30:40,160 --> 00:30:44,640 be low in a space that it would be inexpensive. Yeah, but there 404 00:30:44,720 --> 00:30:47,640 are unusual regulations when you get into the water that are hard to deal with. 405 00:30:47,839 --> 00:30:49,559 Yeah, and most importantly, you don't have direct control over everything. 406 00:30:49,640 --> 00:30:52,279 Well, probably the worst place to put a data center is above ground, 407 00:30:52,359 --> 00:30:55,279 isn't that? Well? Where else are you going to put it underground? 408 00:30:55,480 --> 00:30:59,240 Yeah? Yeah, as long as you can control the space around it, 409 00:30:59,319 --> 00:31:02,400 right, and they certainly you get more heat management that way. I think 410 00:31:02,440 --> 00:31:06,440 you see a lot of organizations talking a lot about where they're getting power from 411 00:31:06,640 --> 00:31:11,759 and how they're cooling, I mean, the two big energy consumers, and 412 00:31:11,240 --> 00:31:17,559 for the most part, like I want to this seems to be the logical 413 00:31:17,599 --> 00:31:22,559 thing for us to do is we pick the provider whose greens, whose green 414 00:31:22,599 --> 00:31:26,599 approaches are appropriate. I suspect they're all going to be the same ultimately, 415 00:31:27,079 --> 00:31:33,200 and they all have a certain amount of green washing in them. I just 416 00:31:33,799 --> 00:31:36,920 and we need to fight against it. Like I'm not saying it's acceptable, 417 00:31:37,319 --> 00:31:41,000 just that's like part of our responsibility going you know this part where you're buying 418 00:31:41,039 --> 00:31:42,279 all these carbon offsets because he didn't actually say what you're going to do, 419 00:31:42,359 --> 00:31:47,119 which was have no carbon output on this, like stop it, stop emitting 420 00:31:47,319 --> 00:31:51,759 carbon on it. But that's the thing, like, as we care more 421 00:31:52,119 --> 00:31:55,799 about this, as we care more about the environment, we are choosing those 422 00:31:55,880 --> 00:32:00,640 providers yea. And at the end we'll stop choosing those that we don't trust 423 00:32:01,319 --> 00:32:05,440 or that are not doing a good thing. Whatever we define that that could 424 00:32:05,519 --> 00:32:09,079 think is, or whatever the values are and standards are. I think we'll 425 00:32:09,119 --> 00:32:13,960 get better. I think we just at the moment trying a lot of different 426 00:32:14,039 --> 00:32:19,160 things, and people are putting the numbers or ratings that that would make them 427 00:32:19,200 --> 00:32:22,920 look better. I think that's important. It is interesting to weigh the price 428 00:32:23,200 --> 00:32:29,839 of building ultra efficient code that will consume as little energy as possible versus the 429 00:32:29,920 --> 00:32:34,839 maintainability of that code. Yes, my experience with making highly optimized code is 430 00:32:34,920 --> 00:32:39,240 it's harder to take care of the simplest code is the best code for maintainability, 431 00:32:39,519 --> 00:32:43,480 even if it is going to consume more energy In the process. There 432 00:32:43,559 --> 00:32:46,359 is always a balance, right, I think we sway too much in the 433 00:32:46,519 --> 00:32:50,240 direction of we want to scale, we want to face. We don't care 434 00:32:50,279 --> 00:32:53,000 about how we write the code. I mean, honestly, I don't think 435 00:32:53,039 --> 00:32:57,119 people are thinking about this for the most part. Yeah, I get that, 436 00:32:57,480 --> 00:33:01,759 And I think the first steps are to start talking about it, get 437 00:33:01,839 --> 00:33:07,359 the data on like what are you doing where you are? Because most of 438 00:33:07,480 --> 00:33:12,039 this is we can do loads of stuff, and I think we should do 439 00:33:12,200 --> 00:33:15,240 loads of stuff now at the beginning. When we don't know what's the right 440 00:33:15,319 --> 00:33:20,359 thing, we should experiment, we should learn from it. But it's mostly 441 00:33:20,440 --> 00:33:22,359 to get the data. I know what brings the most value. Yeah, 442 00:33:22,480 --> 00:33:27,599 Starting to measure, just measure, Yeah, I mean data and metrics. 443 00:33:28,039 --> 00:33:32,279 It's another thing to add to these storages that don't impact the environment in a 444 00:33:32,359 --> 00:33:37,480 good way. Ye. Measuring is also a consumption of exactly, but at 445 00:33:37,599 --> 00:33:43,000 least we can make an informed decision on what we are doing. The thing 446 00:33:43,200 --> 00:33:49,319 that we need to be mindful about is that we shouldn't wait to get something. 447 00:33:49,440 --> 00:33:52,000 We should all start thinking about it and do something about it. Because 448 00:33:52,200 --> 00:33:57,079 I don't think there's going to be a perfect solution, but nobody's perfect. 449 00:33:57,160 --> 00:34:00,799 No solution will be perfect. It's a moving target, moving it target. 450 00:34:00,839 --> 00:34:05,680 But everything is. This is a complex topic that we need to solve and 451 00:34:05,759 --> 00:34:08,719 at some point we will have better data, we will have more information on 452 00:34:09,320 --> 00:34:14,320 what is, you know, like going to make it all better or which 453 00:34:14,360 --> 00:34:16,000 direction we want to go in, and there's going to be a couple of 454 00:34:16,079 --> 00:34:21,760 options. But at the moment, let's just trial. Let's do something. 455 00:34:22,480 --> 00:34:25,559 I mean, I think the easiest thing I can get by a senior leadership 456 00:34:25,800 --> 00:34:30,159 is this, we have a SaaS offering, and I can reduce the cost 457 00:34:30,199 --> 00:34:32,079 of as offering a SaaS offering for the same set of customers. So effectively, 458 00:34:32,119 --> 00:34:36,480 we're going to make more money off of its existing customers by making this 459 00:34:36,559 --> 00:34:40,199 code more efficient. And if you want to then do a little green marketing 460 00:34:40,280 --> 00:34:44,360 on the back of we make our code more efficient to reduce our impact on 461 00:34:44,440 --> 00:34:47,880 the planet. That fine you know up to you, Yes, there is 462 00:34:49,719 --> 00:34:57,320 that thing where we need to be able to know how to present this, 463 00:34:57,880 --> 00:35:01,360 how which value this brings. There is also that thing you want to work 464 00:35:01,400 --> 00:35:05,639 for the company that wants to do good and if they are not, they 465 00:35:05,639 --> 00:35:09,760 don't understand what that means. You want to explain them. But if they 466 00:35:09,800 --> 00:35:14,400 don't get it, and if they don't want to do it, find another 467 00:35:14,599 --> 00:35:17,039 job. Like we are spending. Are we going to get to that point 468 00:35:17,079 --> 00:35:21,039 where it's like I can't work here because you don't care about these things? 469 00:35:21,119 --> 00:35:23,280 I mean, that's interesting to me. It is, no, if it's 470 00:35:23,360 --> 00:35:29,559 true, but well, there is an interesting not interesting. There is a 471 00:35:30,440 --> 00:35:36,679 phrase that I was reminded by recently by my CEO, and it was every 472 00:35:36,840 --> 00:35:39,960 dollar you choose to spend is a vote for the future you want to live 473 00:35:40,000 --> 00:35:45,440 in. Exactly. I think that's fair. And arguably the pandemic has made 474 00:35:45,519 --> 00:35:51,079 us more aware of the culture we want to be in, and a culture 475 00:35:51,119 --> 00:35:54,320 of an organization that takes environmental impact seriously is maybe a culture you want to 476 00:35:54,360 --> 00:35:58,199 be a part of. So that does affect your choice of where you want 477 00:35:58,239 --> 00:36:01,639 to work. Yes, I think we shape it right. I think we 478 00:36:01,920 --> 00:36:09,159 choose and if our employees are not willing or don't care about it, they 479 00:36:09,280 --> 00:36:13,360 need to start because people will not want to live, we will not want 480 00:36:13,440 --> 00:36:16,039 to work there. No, it's a signal about culture. It's a signal, 481 00:36:16,280 --> 00:36:21,280 and it's a singal we want to do something right and Okay, That's 482 00:36:21,360 --> 00:36:25,119 one go is like they will not be able to get software engineers good software. 483 00:36:25,519 --> 00:36:29,239 The other thing is like these vendors like they'll not be able to get 484 00:36:29,239 --> 00:36:32,280 the clients, or the government will say like they will be they will start 485 00:36:32,440 --> 00:36:38,320 charging you on your environmental impacts. That's the other thing. It's interesting to 486 00:36:38,320 --> 00:36:42,760 get to a point where it's like, can we get ratings on our software 487 00:36:43,320 --> 00:36:45,440 for environmental impact? I don't think that idea. It doesn't exist at this 488 00:36:45,519 --> 00:36:50,960 point obviously, but you know, we've certainly talked about other rating systems for 489 00:36:51,159 --> 00:36:54,320 quality. Yeah, and so you know, could there be one that measures 490 00:36:54,360 --> 00:36:59,800 this as well? I mean absolutely, you know, it's interesting just how 491 00:37:00,800 --> 00:37:05,840 unmeasured our industry is that And again I think the clouds so sort of work 492 00:37:05,880 --> 00:37:10,639 around to this is the byproduct of cloud is and now we've now concentrated that 493 00:37:10,719 --> 00:37:15,559 responsibility down to a few organizations that we can press against for better or worse. 494 00:37:16,400 --> 00:37:22,760 I think there's a client conversation here too. I saw the EUS talking 495 00:37:22,800 --> 00:37:28,320 more about the maintainability of smartphones. You already said this earli earlier. This 496 00:37:28,480 --> 00:37:30,480 is like do you really need more compute? It's like do I need another 497 00:37:30,599 --> 00:37:35,639 camera in my phone? I think three is enough. Right. The main 498 00:37:35,679 --> 00:37:37,719 thing, the main reason on replaced phones these days is because the battery is 499 00:37:37,760 --> 00:37:42,639 starting to fail because batteries aren't replaceable. And now you see the EU coming 500 00:37:42,719 --> 00:37:46,400 up with going back to we used to have changeable batteries, right, and 501 00:37:47,239 --> 00:37:52,559 does it could you tolerate a millimeter thicker on your phone? But every few 502 00:37:52,639 --> 00:37:53,519 years now you're going to be able to take the old battery out, have 503 00:37:53,599 --> 00:37:55,760 it recycled, and get a new battery for your phone and can use it 504 00:37:55,800 --> 00:38:00,519 for longer. Well, certainly Apple doesn't want that do that because they can't 505 00:38:00,559 --> 00:38:04,760 sell you a new phone. Well, and I'm not just gonna lay that 506 00:38:04,800 --> 00:38:07,800 at the feet of Apple any All manufacturers are in the business of selling you 507 00:38:07,920 --> 00:38:13,840 do devices, Samson. But the cadence is decreasing, right even you know, 508 00:38:14,000 --> 00:38:17,599 people aren't getting every new iPhone and maybe they're getting every other. Certainly 509 00:38:17,880 --> 00:38:23,440 over an enterprise land we're talking about keeping workstations. Used to be two years, 510 00:38:23,679 --> 00:38:27,559 then it was four years. Now we're talking five and six years. 511 00:38:28,079 --> 00:38:34,360 Because a they're good enough, like it's not like the employees productivity is impaired. 512 00:38:34,719 --> 00:38:37,719 The only reason that at five or six years we start wanting to swap 513 00:38:37,800 --> 00:38:40,480 out the machines is we start having more problems with them. I had my 514 00:38:40,639 --> 00:38:45,840 last workstation for over ten years, Yeah, over ten and this new one 515 00:38:45,960 --> 00:38:49,119 is gonna last me a long time too. They either the last set of 516 00:38:49,199 --> 00:38:51,800 servers I had when I had a server closet, because I don't have one 517 00:38:51,800 --> 00:38:54,239 anymore, and I'm not unhappy about that. That's greenwashing right there, right 518 00:38:54,280 --> 00:38:59,920 there. But that set of servers. Twice I changed out all the hard 519 00:39:00,039 --> 00:39:04,119 drives on all the fans in it. Like the motherboards were good enough and 520 00:39:04,320 --> 00:39:07,000 there's nothing wrong with the chassis, but eventually the bearings on the fans start 521 00:39:07,039 --> 00:39:09,880 to go, and that will kill a machine as it stops pumping air through. 522 00:39:10,000 --> 00:39:13,679 Well. Well, I have thirty two solar panels on my roof, 523 00:39:13,800 --> 00:39:17,280 so I feel pretty smug right now. But again, we're thinking about in 524 00:39:17,400 --> 00:39:22,840 terms of when we reduce impact, when we maintain equipment for longer, and 525 00:39:22,920 --> 00:39:28,400 there does seem to be a general sense that a lot of these devices are 526 00:39:28,400 --> 00:39:30,480 sufficiently performed. We don't need to replace them to get anything from that, 527 00:39:31,039 --> 00:39:35,400 but they should be more repaarable. Yes, I really hope we're going in 528 00:39:35,480 --> 00:39:38,440 that direction because I mean we are a consumer, how do you say, 529 00:39:38,519 --> 00:39:44,800 like consumed? Yes, yeah, we have been trained by a system that 530 00:39:45,039 --> 00:39:49,360 profits from us consuming constantly. Exactly if we're consuming constantly, we want everything 531 00:39:49,480 --> 00:39:53,800 Like you go on Amazon and you order things immediately, clothes, gadgets, 532 00:39:54,360 --> 00:39:59,199 everything is I just want it now. It doesn't need to be good quality 533 00:39:59,320 --> 00:40:00,880 because I'm going to to be good. I'm going to get bored of it 534 00:40:01,000 --> 00:40:07,239 in the month. And this is the thing, like these companies that that 535 00:40:07,840 --> 00:40:12,800 produce these things that are not incentivized to make it no last longer, and 536 00:40:12,920 --> 00:40:15,440 that's what you get. You end up with like EU regulations saying, well, 537 00:40:15,639 --> 00:40:21,679 they forced iPhone to use the USB C port and it's not like they 538 00:40:21,719 --> 00:40:23,760 couldn't do it, and there was much acclaim from everyone. And now they've 539 00:40:23,800 --> 00:40:29,119 also said stop shipping chargers with the phone because we've got enough chargers, like 540 00:40:29,440 --> 00:40:34,480 I had enough cables at home. Story for you. So my second marriage 541 00:40:35,880 --> 00:40:39,599 I inherited two step daughters and they're wonderful bonus daughters. But at one point 542 00:40:39,639 --> 00:40:45,280 one of them came and said, Mom, are we broke? And she 543 00:40:45,400 --> 00:40:50,159 goes, no, why, It says, because I haven't seen any Amazon 544 00:40:50,280 --> 00:40:59,039 boxes on the front porch in like two weeks. Yeah, cold border something. 545 00:41:02,519 --> 00:41:12,360 Yeah. We are just used to having everything now and the world is 546 00:41:12,400 --> 00:41:15,760 getting complicated, you know, like all of this stuff, all of these 547 00:41:15,880 --> 00:41:20,960 frameworks or everything, you know, like run this shiplogs there. I want 548 00:41:21,000 --> 00:41:24,840 to filter on this. We talked about observability yesterday. That's more data. 549 00:41:25,000 --> 00:41:29,519 That's more you know, Like, I'm not saying any of that is bad. 550 00:41:29,840 --> 00:41:34,199 I'm just saying what is the value that that's bringing? What are problems 551 00:41:34,239 --> 00:41:37,039 are we solving with that? But you hit an interesting point. It's like 552 00:41:37,119 --> 00:41:40,039 there's an awful lot of consumption going on without assessing the value from it exactly. 553 00:41:40,079 --> 00:41:43,159 It's like, if you're going to do that level of logging, tell 554 00:41:43,199 --> 00:41:45,159 me you look at the lot. Yeah, why are you using them? 555 00:41:45,440 --> 00:41:49,119 Is it? Just like first of all, like do you need it? 556 00:41:49,239 --> 00:41:52,800 What's the problem you're trying to solve? The other then is like what's the 557 00:41:52,880 --> 00:41:55,280 tool for the job? And this is the thing? Like is then the 558 00:41:55,400 --> 00:42:00,119 machine learning? Is it? This AI? Is it? Can you run 559 00:42:00,239 --> 00:42:05,800 this during the night when energy consumption is lower? Can you run it in 560 00:42:05,880 --> 00:42:09,760 a region where we have greener energy? Sure? Yeah, that's an interesting 561 00:42:09,800 --> 00:42:15,239 idea that we shift workloads, time and geography for lower impact. Again, 562 00:42:15,280 --> 00:42:17,519 I don't know how much consideration that is right there, but it's an interesting 563 00:42:17,599 --> 00:42:22,400 question to ask, like these are not those are not stunningly hard problems. 564 00:42:22,880 --> 00:42:24,599 No, if you had the numbers in front of you, Yeah, I 565 00:42:24,679 --> 00:42:29,880 think there are different ways. It's just we do need to start talking about 566 00:42:29,960 --> 00:42:34,079 them. Sure, we need to start making it easier for people to do 567 00:42:34,239 --> 00:42:38,760 them. Yeah, make the path the path of least resistance is also a 568 00:42:38,840 --> 00:42:44,679 path of optimal outcome. Yeah, I buy it. It's an interesting conversation 569 00:42:44,800 --> 00:42:47,119 to have and to kick upstairs. Yes, you know, to take to 570 00:42:47,280 --> 00:42:51,480 leadership and say where does this fit into our conversations? Like how are we 571 00:42:51,559 --> 00:42:55,920 thinking about this? And again I'm as someone who's had to convince leaders to 572 00:42:57,000 --> 00:43:00,599 do things for many years. It's like, come with a dollar for They 573 00:43:00,719 --> 00:43:05,599 understand that absolutely. I mean, you need to speak with the language of 574 00:43:05,760 --> 00:43:08,280 the people you're talking to, So you need to know your audience. So 575 00:43:08,480 --> 00:43:13,119 what triggers I can make the CFO happy and the CEO happy at the same 576 00:43:13,239 --> 00:43:17,119 time. I've been I can really tell you that, like infant ofs, 577 00:43:17,280 --> 00:43:24,679 we made this platform greener by accidents at early stages and this was just because 578 00:43:24,719 --> 00:43:30,440 we were trying to not just because one of our problems was we couldn't scale 579 00:43:30,920 --> 00:43:36,159 with the client sizes and the hardware is just too expensive. So by trying 580 00:43:36,199 --> 00:43:42,159 to reduce the cost, we actually made the platform greener. Scale it up, 581 00:43:42,199 --> 00:43:45,480 scale it down when you're not using it, distributed in a different way, 582 00:43:45,119 --> 00:43:50,800 like those different things. But that's the language that our CFO understands. 583 00:43:51,000 --> 00:43:55,679 It's cheaper, it was cheap, it should be cheaper. That's awesome. 584 00:43:55,960 --> 00:44:00,199 So what's next for you? What's in your inbox? What's next for me? 585 00:44:00,519 --> 00:44:04,840 Well, with fund ups, I'm lucky to work in a company that 586 00:44:04,960 --> 00:44:07,760 actually cares about this, that does want to make a good impact and not 587 00:44:08,119 --> 00:44:15,599 just profit. And we just rearc that our platform, yeah, to be 588 00:44:15,679 --> 00:44:22,320 able to support our clients, but also make it cheaper and greener. The 589 00:44:22,440 --> 00:44:28,079 next thing is for us are making that code more performance, figuring the ways 590 00:44:28,199 --> 00:44:35,920 to maybe use those machines that have lower energy consumption arm processors. Yes, 591 00:44:36,119 --> 00:44:39,159 as I actually come online, but it's an interesting question to kick to a 592 00:44:39,239 --> 00:44:43,360 Microsoft to say, tell me which is your greenest data center, Like do 593 00:44:43,440 --> 00:44:45,400 you ue the track of that where the power comes from, how the cooling 594 00:44:45,519 --> 00:44:50,639 is done. There was a talk yesterday and this guy showed, like in 595 00:44:50,840 --> 00:44:57,800 parasites this much impact on the environment, like in Ireland, the data centers 596 00:44:57,840 --> 00:45:01,039 have this much impact. In Germany. You could be choosing and scaling up 597 00:45:01,079 --> 00:45:07,440 and scaling down servers depending where the energy is greener, where the consumption is 598 00:45:07,599 --> 00:45:13,159 better or lower. Like if you look at Canada like fires and you need 599 00:45:13,280 --> 00:45:16,400 water and we are cooling down these data centers with this enormous amount of water, 600 00:45:17,159 --> 00:45:20,840 if you can kind of switch that somewhere else now and then you know, 601 00:45:21,039 --> 00:45:23,880 like I think we can be smooth smarter about this. You can push 602 00:45:24,000 --> 00:45:27,679 for value in that. Yeah, I like that. Leah, thanks so 603 00:45:27,800 --> 00:45:30,679 much for joining us. It's been a pleasure talking to you. Thank you 604 00:45:30,880 --> 00:45:34,920 very much for Yeah, I'm making this a very enjoyable talk, all right, 605 00:45:35,519 --> 00:46:00,880 and we'll see you next time on dot net rocks. Dot net Rocks 606 00:46:00,960 --> 00:46:05,519 is brought to you by Franklin's Net and produced by Pop Studios, a full 607 00:46:05,599 --> 00:46:09,920 service audio, video and post production facility located physically in New London, Connecticut, 608 00:46:10,239 --> 00:46:15,639 and of course in the cloud online at pwop dot com. Visit our 609 00:46:15,679 --> 00:46:19,960 website at d O T N E T r o c k s dot com 610 00:46:20,480 --> 00:46:24,239 for RSS feeds, downloads, mobile apps, comments, and access to the 611 00:46:24,320 --> 00:46:29,920 full archives going back to show number one, recorded in September two thousand and 612 00:46:30,000 --> 00:46:32,400 two. And make sure you check out our sponsors. They keep us in 613 00:46:32,480 --> 00:46:38,039 business. Now go write some code, See you next time you got Jamal vanst