1 00:00:05,000 --> 00:00:06,599 Speaker 1: To like already click the button. 2 00:00:09,720 --> 00:00:11,359 Speaker 2: And that's how we're starting the episode. 3 00:00:13,240 --> 00:00:15,960 Speaker 1: Oh for those of you just getting started, actually you 4 00:00:16,000 --> 00:00:17,800 are just getting started because I just clicked on the 5 00:00:17,920 --> 00:00:23,480 record button. But hey, welcome, thanks for joining the podcast. Warren. 6 00:00:23,519 --> 00:00:24,000 How are you? 7 00:00:24,679 --> 00:00:26,640 Speaker 2: I'm I'm great. You know, I feel like, you know, 8 00:00:26,679 --> 00:00:29,239 we should go back to this is Adventures and DevOps. 9 00:00:29,280 --> 00:00:31,440 You know, for the first time listeners, if this is 10 00:00:31,519 --> 00:00:35,079 just the one episode they happen to click on, then 11 00:00:35,359 --> 00:00:37,399 they're in for a real treat today because I'm really 12 00:00:37,439 --> 00:00:38,840 interested in today's topic. 13 00:00:39,920 --> 00:00:45,520 Speaker 1: Yeah, and today's topic, we have gorkam Erkan with us 14 00:00:45,600 --> 00:00:49,799 and we're gonna be talking about how to how to 15 00:00:49,840 --> 00:00:56,479 integrate AI and mL into teams using tooling that's a 16 00:00:56,520 --> 00:00:59,359 little bit more aligned with how we do DevOps. 17 00:00:59,479 --> 00:01:00,520 Speaker 2: That sound of out right? 18 00:01:01,679 --> 00:01:02,880 Speaker 3: Yeah? That sounds right? 19 00:01:03,200 --> 00:01:03,640 Speaker 1: All right? 20 00:01:03,960 --> 00:01:04,280 Speaker 3: Cool? 21 00:01:05,280 --> 00:01:08,200 Speaker 1: Before we dig into the details of that, can you 22 00:01:08,239 --> 00:01:11,319 share with our listeners a little bit about your background? 23 00:01:12,799 --> 00:01:18,079 Speaker 3: Sure. I am the CEA of josu U. The josu 24 00:01:18,200 --> 00:01:22,079 is a company that is trying to make it easier 25 00:01:22,079 --> 00:01:30,000 for enterprises to adopt AI and mL in into their applications. Uh. 26 00:01:30,040 --> 00:01:34,359 And prior to JOSU I was with red Hat for many, 27 00:01:34,400 --> 00:01:37,120 many years. I was as a distinguished engineer with red 28 00:01:37,159 --> 00:01:46,040 Hat worked on developer tools, which in red hat it 29 00:01:46,319 --> 00:01:49,840 covers anything from I D E S to C C 30 00:01:50,000 --> 00:01:54,920 D pipelines to them op so it's it's all everything 31 00:01:55,480 --> 00:01:59,480 that you can think of that is open source and 32 00:01:59,560 --> 00:02:04,200 that will tool My group kind of got involved with 33 00:02:04,239 --> 00:02:10,560 those as part of that. I did Java tools early 34 00:02:10,560 --> 00:02:18,479 in my career for something called J two E if 35 00:02:18,520 --> 00:02:25,120 anyone remembers that. So and then I actually had the 36 00:02:25,240 --> 00:02:32,080 great idea of integrating Java tooling into vs code. The 37 00:02:32,439 --> 00:02:37,120 current Java tooling is something that I have done in 38 00:02:37,479 --> 00:02:40,400 a month or so and that has grown to what 39 00:02:40,479 --> 00:02:47,000 it is today. My group got involved with things like 40 00:02:47,159 --> 00:02:50,000 tech Tile, CD Foundation and so and so forth. So 41 00:02:50,000 --> 00:02:53,879 I've been doing devobs for the last probably five years, 42 00:02:54,199 --> 00:02:58,159 or tools for DeVos for five years, and tools in general. 43 00:02:58,360 --> 00:03:05,599 Before that, before joining red Hat, I was working for Nokia, 44 00:03:06,280 --> 00:03:09,879 the phone company, so I was in the middle of 45 00:03:09,919 --> 00:03:21,360 the mobile revolution and again mostly doing open source projects 46 00:03:21,680 --> 00:03:23,400 at Noukia at the time. 47 00:03:23,639 --> 00:03:27,240 Speaker 2: When you say open source and Java and red Hat, 48 00:03:27,280 --> 00:03:31,400 my mind immediately goes to Jenkins and I don't have 49 00:03:31,479 --> 00:03:34,919 a particular soft spot in my heart for that product, 50 00:03:34,960 --> 00:03:36,599 but I can imagine you've had to spend a lot 51 00:03:36,639 --> 00:03:38,199 of time with it. Is that what they were using. 52 00:03:39,960 --> 00:03:43,840 Speaker 3: So yeah, we did start with Jenkins. We do support, 53 00:03:43,960 --> 00:03:47,800 We did support or red Hats still supports Jenkins to 54 00:03:48,000 --> 00:03:53,080 some degree. But at the time when we tried to 55 00:03:53,159 --> 00:03:57,520 do a project, code open shift that I owe and 56 00:03:57,639 --> 00:03:59,919 one of the things that we tried to do at 57 00:04:00,080 --> 00:04:06,159 the time was to run Jenkins in a manner that 58 00:04:06,199 --> 00:04:09,520 it is not supposed to be run at the time. 59 00:04:09,960 --> 00:04:13,560 I think Jenkins is doing better now, but at the time, 60 00:04:13,879 --> 00:04:16,839 Jenkins was created as a server, so it was meant 61 00:04:16,879 --> 00:04:19,240 to run as a server. But then when you're on 62 00:04:19,279 --> 00:04:23,399 the cloud trying to run Jenkins, one thing that you 63 00:04:23,439 --> 00:04:25,360 want to do is you want to try to run 64 00:04:25,399 --> 00:04:30,360 its servers, like if I need to go and build something, 65 00:04:30,879 --> 00:04:35,680 especially if you're doing this as a SAS service, you 66 00:04:35,759 --> 00:04:39,160 want to just on demand start a Jenkins instance, do 67 00:04:39,319 --> 00:04:41,560 your bills, and then shut it down so that you're 68 00:04:41,560 --> 00:04:47,000 not spending too much on the cloud resources. That didn't 69 00:04:47,040 --> 00:04:51,040 work well for us. So one of the things that 70 00:04:51,240 --> 00:04:55,560 happened at the time was the Canadia project. If you're 71 00:04:55,600 --> 00:04:59,800 familiar with it, and as part of the Canadia project 72 00:05:00,040 --> 00:05:06,240 first code drop before even it even got public, it 73 00:05:06,279 --> 00:05:09,279 was open to some of the redheaders, including myself, and 74 00:05:09,319 --> 00:05:11,839 one of the things that we noticed was there was 75 00:05:11,879 --> 00:05:15,279 a built piece in the k native code which was 76 00:05:15,319 --> 00:05:21,120 actually doing server spiels essentially. So the whole idea of oh, 77 00:05:21,360 --> 00:05:25,439 can we actually make this and turned this into something 78 00:05:26,879 --> 00:05:31,240 It turned out to be the techtile project later and 79 00:05:31,279 --> 00:05:35,920 then it became the tech Time's cd c CD system 80 00:05:36,560 --> 00:05:40,839 that we know today. So it actually we got involved 81 00:05:40,839 --> 00:05:46,879 with that because of Jenkins because we couldn't actually make 82 00:05:46,920 --> 00:05:50,879 it cloud native at the time, so it felt like, oh, 83 00:05:50,920 --> 00:05:54,600 we should actually, you know, do something cloud native instead. 84 00:05:56,120 --> 00:06:01,680 So that's that's how involvement with Jane. But on the 85 00:06:01,720 --> 00:06:05,800 other hand, I would still bet that a big chunk 86 00:06:05,839 --> 00:06:11,279 of workloads on kimbernites in the world is actually Jenkins. 87 00:06:16,160 --> 00:06:16,959 Speaker 1: Survey to edit. 88 00:06:17,240 --> 00:06:19,680 Speaker 3: I mean, I don't know, I don't have un versa. 89 00:06:19,720 --> 00:06:23,079 I like that would be my guess. It's it's it's 90 00:06:23,120 --> 00:06:25,240 in double digits. I'm pretty sure enough. 91 00:06:25,680 --> 00:06:28,279 Speaker 2: I feel like I feel like this is where ignorance 92 00:06:28,399 --> 00:06:30,680 is bliss. Like I don't I'm going to pretend that's 93 00:06:30,720 --> 00:06:34,000 not happening and just continue living my life in whatever 94 00:06:34,279 --> 00:06:36,000 delusion I have currently going on. 95 00:06:37,079 --> 00:06:41,199 Speaker 3: Unfortunately, it is true. It's there, It's it's in many enterprises. 96 00:06:41,240 --> 00:06:47,839 You can't ignore Jenkins. It's hundreds of instances for sure. Yeah. 97 00:06:48,040 --> 00:06:53,399 Speaker 1: I have a love hate relationship with Jenkins. I hate 98 00:06:53,480 --> 00:06:56,000 working with it, but I love how it can just 99 00:06:56,079 --> 00:06:58,519 do absolutely anything you needed to do. 100 00:07:00,519 --> 00:07:07,120 Speaker 3: And the ecosystem, Yeah, like the ecosystem behind Jenkins. You know, 101 00:07:07,360 --> 00:07:11,759 I think nowadays get up actions maybe at that level, 102 00:07:11,839 --> 00:07:19,399 but the ecosystem behind Jenkins, it's, it's it's pretty big. 103 00:07:19,560 --> 00:07:22,360 It's very important. You can find a plug in for 104 00:07:22,399 --> 00:07:26,600 anything essentially, yeah, or anything that you care about, right 105 00:07:27,920 --> 00:07:28,639 for sure. 106 00:07:29,959 --> 00:07:34,839 Speaker 1: So let's talk a little bit about aim L since 107 00:07:34,879 --> 00:07:38,720 that's our topic for today. And there's like, you know, 108 00:07:38,959 --> 00:07:43,879 I whenever I think about this, the most common scenarios 109 00:07:43,920 --> 00:07:47,120 that I see coming up are people using chat, GPT 110 00:07:47,759 --> 00:07:50,560 or tools like that, And it seems like a very 111 00:07:53,240 --> 00:07:59,680 early stage product. So to get that enterprise ready, like, 112 00:07:59,720 --> 00:08:02,360 what the what are the challenges that people are facing? 113 00:08:03,879 --> 00:08:07,639 Speaker 3: Yeah, so I think check GPT, Like what we are 114 00:08:07,720 --> 00:08:10,920 going through right now is that that in every enterprise 115 00:08:11,000 --> 00:08:14,920 there is one or two let's call it experiments that 116 00:08:15,040 --> 00:08:18,079 is going on. And usually when you're starting the experiment, 117 00:08:18,560 --> 00:08:20,560 the first thing that you do is like what is 118 00:08:20,600 --> 00:08:27,680 the least resistance that you can face, and that is 119 00:08:28,360 --> 00:08:31,319 use the check GPT API s right, the open API 120 00:08:31,600 --> 00:08:37,320 or Microsoft has Azure APIs as well, So use those 121 00:08:37,759 --> 00:08:40,679 and do whatever you need to do and come up 122 00:08:40,720 --> 00:08:44,120 with your first early experiments. And when you look at 123 00:08:44,159 --> 00:08:49,600 those experiments in many of the organizations, what you see 124 00:08:49,720 --> 00:08:52,759 is there are a lot of success that is coming in. 125 00:08:53,120 --> 00:08:55,639 There's a lot of learning that they are going through, 126 00:08:56,159 --> 00:09:00,679 but there's also successful stories as well coming out of them. Uh. 127 00:09:01,240 --> 00:09:04,360 The one thing about AI is in deterministic, right, It's 128 00:09:04,399 --> 00:09:07,960 like the first time that you do something with AI, 129 00:09:08,200 --> 00:09:11,720 you're going to fail, and then the second you're going 130 00:09:11,759 --> 00:09:16,120 to do another iteration, and because you're going to understand 131 00:09:16,200 --> 00:09:19,519 that behavior, and then you're going to do another iteration. 132 00:09:19,639 --> 00:09:23,240 I think check GPT and all the other APIs out there, 133 00:09:23,360 --> 00:09:30,519 of the open API and similar AI API is out there. 134 00:09:30,600 --> 00:09:36,639 I think it provides them this opportunity to do the experimentation. 135 00:09:37,360 --> 00:09:39,720 But the way I see it is this is this 136 00:09:39,759 --> 00:09:43,639 is temporary, right, This is just dipping your toes into 137 00:09:43,679 --> 00:09:48,879 the water. When you actually want to start doing something 138 00:09:50,519 --> 00:09:53,799 that will change your business and try to get some 139 00:09:54,000 --> 00:09:58,240 financial benefits to your company. That means integrating with many 140 00:09:58,320 --> 00:10:01,759 main systems. That means like it's like we have seen 141 00:10:01,799 --> 00:10:04,159 it with cloud, we have seen it with mobile. Right, 142 00:10:04,200 --> 00:10:07,240 It's like it's not a oh, here's one application that 143 00:10:07,399 --> 00:10:11,679 is mobile, or here's one application that runs on the cloud. 144 00:10:12,159 --> 00:10:15,600 You're starting to adopt it to your organization or if 145 00:10:15,639 --> 00:10:18,440 you are even older, there was a time where there 146 00:10:18,559 --> 00:10:23,200 was we talked about paper this office. Right, So it's 147 00:10:23,240 --> 00:10:26,360 like it takes a little bit of time, it's a process, 148 00:10:26,440 --> 00:10:30,159 and you actually need to internalize it into your organization. 149 00:10:30,240 --> 00:10:33,840 I think the next step for many of these organizations 150 00:10:34,360 --> 00:10:37,960 is to start to internalize it. And when you're internalizing it, 151 00:10:38,000 --> 00:10:43,639 you cannot really depend on a third party service. And 152 00:10:43,759 --> 00:10:48,559 some of the industries cannot even share their data give 153 00:10:48,600 --> 00:10:54,320 their data to these services where there is concerns for 154 00:10:54,440 --> 00:10:58,279 privacy and security and so on and so forth. So 155 00:10:58,320 --> 00:11:06,600 they need to start running these essentially base models internally 156 00:11:07,000 --> 00:11:11,919 in their infrastructure or in their cloud infrastructure so that 157 00:11:12,759 --> 00:11:16,000 they can get the benefits. And the one thing that 158 00:11:16,480 --> 00:11:21,480 I see is when you go to your data scientists 159 00:11:21,639 --> 00:11:33,720 or data data scientists or mL engineer. They are separate 160 00:11:34,480 --> 00:11:38,960 from the DevOps engineers in the company. They are not 161 00:11:39,080 --> 00:11:43,600 well integrated. They have a very different tool set. It's 162 00:11:43,679 --> 00:11:47,279 just because of the way the AI has been developed 163 00:11:47,320 --> 00:11:51,080 and coming into the world. And one thing that you 164 00:11:51,200 --> 00:11:54,600 notice is there is no shortage of open source projects 165 00:11:54,600 --> 00:11:57,639 and tools in the AI world, but there is a 166 00:11:57,679 --> 00:12:03,720 shortage of projects and open source tools in the AI 167 00:12:03,840 --> 00:12:08,799 world that are standards. You can do. Just think of 168 00:12:08,919 --> 00:12:12,240 any subject and you can find ten alternates to do it. 169 00:12:13,000 --> 00:12:16,799 None of them talks to each each other. Nothing is standard. 170 00:12:17,840 --> 00:12:21,559 So this is one of the things that we have noticed. 171 00:12:21,559 --> 00:12:26,200 We actually were trying to solve this for ourselves as well, 172 00:12:26,600 --> 00:12:29,840 and one of the things we decided what was okay, 173 00:12:29,919 --> 00:12:34,919 So I have so many ways of storing a model 174 00:12:35,919 --> 00:12:39,039 and moving that to production. I have so many ways 175 00:12:39,080 --> 00:12:43,759 of storing a data set and moving that to inference 176 00:12:43,919 --> 00:12:46,720 or training and testing and so on and so forth, 177 00:12:47,120 --> 00:12:50,519 but none of it is standard, like this should be 178 00:12:50,639 --> 00:12:53,559 easier than what it is right now. So what we 179 00:12:53,600 --> 00:12:57,480 did is we came up with an open source project 180 00:12:57,519 --> 00:13:02,200 called ki toops dot mL. And what kittops dot mL 181 00:13:02,360 --> 00:13:06,519 provides is what we call model kits. A model kit 182 00:13:06,679 --> 00:13:12,840 is essentially an OCI artifact and an OCI artifact like 183 00:13:13,039 --> 00:13:16,120 a container image. It's not a container image itself, but 184 00:13:16,200 --> 00:13:19,399 it's an OCI artifact. The advantage of being an OCI 185 00:13:19,639 --> 00:13:25,279 artifact is it can be stored in any OCI registry 186 00:13:25,600 --> 00:13:32,120 like docer hub ECR as you're wherever your image registry is, 187 00:13:32,720 --> 00:13:36,960 and using the information that is stored in the model kit, 188 00:13:37,519 --> 00:13:44,759 you can actually move your models as part of your 189 00:13:44,840 --> 00:13:49,240 c c D pipelines much much efficient because your c 190 00:13:49,480 --> 00:13:53,120 c D pipelines actually knows how to usually knows how 191 00:13:53,120 --> 00:13:57,919 to talk to OCI registrict. For instance, when today we 192 00:13:58,080 --> 00:14:04,120 have a a registry that we are running ourselves in 193 00:14:04,279 --> 00:14:09,919 jose dot mL and you will see many signed model 194 00:14:10,000 --> 00:14:13,759 kits in there. The tool that we are using to 195 00:14:13,840 --> 00:14:17,279 sign model kits is cosign, which is what you would 196 00:14:17,440 --> 00:14:25,519 use for signing an imagery, a regular container image. So 197 00:14:26,720 --> 00:14:33,320 the idea behind having the OCI artifacts was to introduce 198 00:14:33,399 --> 00:14:38,879 a standard way of storing and standard way of sharing 199 00:14:39,799 --> 00:14:48,879 your AI artifacts between data scientists and mL engineers and 200 00:14:49,240 --> 00:14:50,600 DevOps engineers. 201 00:14:51,240 --> 00:14:55,519 Speaker 2: You said something really interesting in there, and that's I 202 00:14:55,519 --> 00:14:57,960 haven't heard of put it this way before that your 203 00:14:58,080 --> 00:15:02,000 first foray into using any sort of AI or mL 204 00:15:02,120 --> 00:15:05,080 models in your company will be an experiment that ends 205 00:15:05,120 --> 00:15:09,200 in failure. And I, you know, I think there's something 206 00:15:09,639 --> 00:15:12,200 really to that, because I can imagine if you like 207 00:15:12,320 --> 00:15:15,600 walked around and put like a hammer on every engineer's 208 00:15:15,639 --> 00:15:20,039 desk and said, you will now start to develop with 209 00:15:20,120 --> 00:15:23,759 this hammer instead of what you were using before. I 210 00:15:23,799 --> 00:15:26,159 feel like it's a ridiculous analogy, but I think there's 211 00:15:26,200 --> 00:15:28,080 a lot of sense there that like that for sure 212 00:15:28,159 --> 00:15:31,000 will not end in success there, Like, however you thought 213 00:15:31,840 --> 00:15:35,919 you could utilize it internally is probably not going to 214 00:15:35,960 --> 00:15:39,399 be super effective. And that obviously extends into the area 215 00:15:39,519 --> 00:15:43,080 of if you are building AI into your product, not 216 00:15:43,159 --> 00:15:46,679 just using it as a as a tool for development, 217 00:15:46,840 --> 00:15:49,320 and you're really talking about this next level which isn't 218 00:15:49,360 --> 00:15:53,039 just using it but actually creating it to be effective 219 00:15:53,039 --> 00:15:54,840 as part of the product offering. 220 00:15:56,279 --> 00:15:59,840 Speaker 3: Yeah, And the reason for that is it's really in 221 00:15:59,879 --> 00:16:04,120 the terministic right with the AI itself. So you may 222 00:16:04,360 --> 00:16:09,559 actually think that they're like when we are integrating applications together, 223 00:16:10,960 --> 00:16:15,120 the rule based the applications, the outcome is very deterministic. 224 00:16:15,240 --> 00:16:18,159 We know that these are the inputs and these will 225 00:16:18,159 --> 00:16:21,759 be the outputs. In the case of AI, that's not 226 00:16:21,799 --> 00:16:26,679 the case, right, So when you go through your first 227 00:16:26,799 --> 00:16:29,759 initial cycle and go through your you know, you receive 228 00:16:29,799 --> 00:16:32,440 your inputs, and then you start to get your outputs, 229 00:16:32,679 --> 00:16:35,960 you start to see these outputs that you haven't expected before. 230 00:16:36,559 --> 00:16:39,840 And the usually, like, for instance, it's very likely that 231 00:16:39,919 --> 00:16:43,879 you're missing actually some data that you're feeding into the model, 232 00:16:44,440 --> 00:16:48,840 and you're not getting the outputs that you're expecting. Because 233 00:16:49,240 --> 00:16:55,120 let's say that you're giving the customers can data, but 234 00:16:55,159 --> 00:16:58,440 then you when you give the eleventh data, the AI 235 00:16:58,639 --> 00:17:01,080 is now able to make the code nection or or 236 00:17:01,200 --> 00:17:04,440 or make the analysis better and give it a better result. 237 00:17:04,759 --> 00:17:07,559 So you will start to notice things like that, and 238 00:17:07,640 --> 00:17:10,680 it is not very easy to notice this at the 239 00:17:10,759 --> 00:17:16,079 design phase. It's not even possible to notice these things 240 00:17:16,079 --> 00:17:19,200 at the design phase. So that's why I'm saying that 241 00:17:19,480 --> 00:17:22,799 your first experiment is going to fail, and you're gonna 242 00:17:22,839 --> 00:17:26,759 do another one with more data or less data or 243 00:17:27,240 --> 00:17:30,440 changed data, and then you will get to a state 244 00:17:30,519 --> 00:17:33,319 where you're you're you're happy with the outputs. 245 00:17:33,559 --> 00:17:35,960 Speaker 2: I feel I have a fear here now that, like 246 00:17:36,039 --> 00:17:38,799 you know, if I'm incorporating it into one of our products. 247 00:17:39,039 --> 00:17:41,680 You know, we have a couple and we have tried 248 00:17:41,799 --> 00:17:44,440 going down the route of adding some AI into into 249 00:17:44,480 --> 00:17:47,799 one of them. Uh, there is this to only say 250 00:17:47,839 --> 00:17:50,880 non deterministic. We mean like literally the same inputs gives 251 00:17:50,920 --> 00:17:55,119 us different output. And exactly, I wonder if we're training 252 00:17:55,200 --> 00:17:58,039 our customers to believe that they should just you know, 253 00:17:59,440 --> 00:18:01,720 turn it off and turn it on again in a sense, right, 254 00:18:01,759 --> 00:18:03,839 Like you know, you have to keep repeating the same 255 00:18:03,920 --> 00:18:06,880 question over and over again to get the answer you want. 256 00:18:06,920 --> 00:18:09,799 And that just that feels like a dangerous thing to 257 00:18:09,799 --> 00:18:13,319 push that expectation onto the users of our products. 258 00:18:15,400 --> 00:18:19,519 Speaker 3: Yeah, there is a Yeah, I see what you're you're saying, 259 00:18:19,559 --> 00:18:23,920 and and and I think that's that's that's the part 260 00:18:23,960 --> 00:18:27,759 of the AI that that still needs to be enhanced 261 00:18:28,000 --> 00:18:33,480 a little bit. You know, the the whole subject of 262 00:18:34,039 --> 00:18:42,519 guardrails on AI is a little bit it feels early. Oh, 263 00:18:42,559 --> 00:18:44,640 I'm told, I'm totally I'm totally with you. 264 00:18:44,640 --> 00:18:47,279 Speaker 2: No, there was actually a study that was released just recently, 265 00:18:47,319 --> 00:18:50,240 and I think we'll kind of appreciate this that it's 266 00:18:50,279 --> 00:18:54,720 not being like prompt engineering isn't being used for malicious purposes. 267 00:18:54,799 --> 00:18:58,279 It's generally being used by power users to actually get 268 00:18:58,359 --> 00:19:01,400 value out of the system more than anything, which I 269 00:19:01,400 --> 00:19:04,599 think is really surprising that these systems aren't being abused 270 00:19:04,640 --> 00:19:09,000 in that way realistically yet, And so adding in guardrails 271 00:19:09,039 --> 00:19:12,240 of anything is just preventing those early adopters from being 272 00:19:12,240 --> 00:19:15,240 able to utilize your product more effectively, rather than preventing 273 00:19:15,279 --> 00:19:16,240 malicious attackers. 274 00:19:16,319 --> 00:19:19,640 Speaker 3: Right. And we actually like since you said prompt Engini, 275 00:19:19,720 --> 00:19:26,039 we actually talked with an enterprise recently and one of 276 00:19:26,119 --> 00:19:33,279 their needs was to share their prompts, Like, uh, think 277 00:19:33,319 --> 00:19:35,559 of it this way. You have a database of all 278 00:19:35,599 --> 00:19:40,519 your all your enterprises data, all your customer data, and 279 00:19:40,559 --> 00:19:43,279 so on and so forth. In the past, the sequel 280 00:19:43,519 --> 00:19:47,680 statements that you have used for finance was very different 281 00:19:47,720 --> 00:19:51,000 from the psychoal statesments that you have used for sales, 282 00:19:51,240 --> 00:19:56,640 so you'd never really actually care to share the psycholos statements. 283 00:19:57,079 --> 00:19:59,440 But now what they're seeing is, as you said, there 284 00:19:59,440 --> 00:20:03,079 are these users who are using these prompts and getting 285 00:20:03,119 --> 00:20:07,039 really good results out of those prompts that are querying 286 00:20:07,079 --> 00:20:10,640 there or that are connected to their their data. Right, 287 00:20:11,079 --> 00:20:13,000 So what they want to be able to do is 288 00:20:13,039 --> 00:20:16,519 to to be able to share these prompts like between 289 00:20:16,680 --> 00:20:21,079 finance and and and HR and and and sales and 290 00:20:21,119 --> 00:20:23,920 so on and so forth, so that they don't have 291 00:20:24,039 --> 00:20:29,440 to reinvent the veil every time. But I think that 292 00:20:29,680 --> 00:20:32,720 we're starting to see a different picture now in the 293 00:20:33,079 --> 00:20:36,200 in the AI world where with with these kind of things. 294 00:20:36,240 --> 00:20:42,160 But when you are opening your data through and an 295 00:20:42,160 --> 00:20:48,119 AI model directed to your customers, I think there's still 296 00:20:48,359 --> 00:20:55,359 work that needs to be done on the guard for this. Yes, yeah, 297 00:20:55,400 --> 00:20:55,720 for sure. 298 00:20:55,759 --> 00:20:59,759 Speaker 1: It reminds me a lot of there's two things here 299 00:20:59,799 --> 00:21:03,000 that that seem like there's lessons we can learn here 300 00:21:03,039 --> 00:21:08,559 from other engineering or other industries. Like whenever I work 301 00:21:08,599 --> 00:21:12,960 with early stage startups, I always say that like the 302 00:21:13,000 --> 00:21:19,319 goal of your first product as a startup is to launch, 303 00:21:20,599 --> 00:21:24,480 like the let me rephrase that. To become a successful startup, 304 00:21:24,519 --> 00:21:27,319 you have to identify the difference between the product that 305 00:21:27,400 --> 00:21:30,759 you built and the product that your customers thought you 306 00:21:30,880 --> 00:21:34,000 built and close that gap. And it feels like this 307 00:21:34,240 --> 00:21:40,440 is very similar to that. You know, because AI will 308 00:21:40,519 --> 00:21:43,920 answer whatever question you ask it, and then you have 309 00:21:44,039 --> 00:21:49,279 to figure out what assumptions it made that you didn't 310 00:21:49,359 --> 00:21:53,599 want it to make. And that's where I think really 311 00:21:53,640 --> 00:21:57,240 sharing those prompt those prompt models can be helpful. For 312 00:21:57,279 --> 00:21:59,759 a while there on X there was a trend where 313 00:21:59,759 --> 00:22:03,799 peopleeople were you know, sharing what they were doing with 314 00:22:03,839 --> 00:22:06,640 different AI models and then also sharing the prompts that 315 00:22:06,640 --> 00:22:09,319 they used too, And I found that really interesting to 316 00:22:09,400 --> 00:22:13,200 see the different caveats that they would give in the 317 00:22:13,279 --> 00:22:16,880 prompt to keep it from wandering off some other path 318 00:22:17,079 --> 00:22:19,960 that gave you an answer that looked right, but as 319 00:22:20,279 --> 00:22:21,920 wasn't actually the answer that you needed. 320 00:22:22,440 --> 00:22:24,559 Speaker 2: No, you say that and it's like, actually, really ridiculous. 321 00:22:25,160 --> 00:22:27,519 Like I'm going through right now. I'm making a presentation 322 00:22:27,720 --> 00:22:32,480 for the European as reinvent that I'm speaking at, and 323 00:22:33,799 --> 00:22:38,400 I want some pictures, and I will ask I'm using 324 00:22:38,799 --> 00:22:40,400 one of the tools, and I will ask the same 325 00:22:40,440 --> 00:22:42,079 question multiple times. But when I get a picture that 326 00:22:42,119 --> 00:22:43,799 I actually like, I will take the prompt that I 327 00:22:43,960 --> 00:22:45,880 used for it, and I will save it as the 328 00:22:45,920 --> 00:22:49,000 file name of the picture so that I can get like, 329 00:22:49,039 --> 00:22:51,400 I mean, it's a ridiculous, but that's the name for me, 330 00:22:51,519 --> 00:22:53,400 that really is the thing that identifies it, so that 331 00:22:53,440 --> 00:22:55,720 I can, you know, potentially go back later and be like, 332 00:22:55,759 --> 00:22:58,960 what did I use to actually generate that image? 333 00:23:00,279 --> 00:23:04,039 Speaker 3: So let me complicate that problem a little bit. So 334 00:23:04,160 --> 00:23:08,480 you have you're you're using a check GPT for instance, like. 335 00:23:09,680 --> 00:23:11,960 Speaker 2: It's not the one from open Ai, but yeah, for sure, 336 00:23:11,960 --> 00:23:13,559 it's an l you're using. 337 00:23:14,359 --> 00:23:18,559 Speaker 3: You're using and it's actually a check giv A service. 338 00:23:19,240 --> 00:23:24,599 So you don't really know the version number. Really, you 339 00:23:24,599 --> 00:23:27,720 don't really have to know the data sets that has 340 00:23:27,799 --> 00:23:30,599 been used and retrained and so and so forth. So 341 00:23:30,839 --> 00:23:33,279 for some of the organizations that's not going to work. 342 00:23:33,720 --> 00:23:36,559 So some organizations are going to say, hey, you know what, 343 00:23:37,160 --> 00:23:41,119 I need to know exactly the model, the data set 344 00:23:41,200 --> 00:23:44,240 that it was trained on, and I need to know 345 00:23:44,960 --> 00:23:50,480 where my fine tuning is coming from, where my RAGS 346 00:23:50,680 --> 00:23:55,000 is coming from. So I need all the lineage of 347 00:23:55,000 --> 00:24:02,480 of of this when this prompt was issued and right. 348 00:24:02,880 --> 00:24:10,359 So you need systems that are capable of telling precisely 349 00:24:11,000 --> 00:24:14,599 what was the data set, what is the model, what 350 00:24:14,759 --> 00:24:18,599 is the fine tuning data set? Right? And what is 351 00:24:18,839 --> 00:24:21,319 if you have been using RAG, what is what is 352 00:24:22,119 --> 00:24:26,799 the version of the vector database or the lineage of 353 00:24:26,839 --> 00:24:33,440 the vectors. All of this stored and available somewhere because 354 00:24:35,000 --> 00:24:38,240 something may happen at any point saying that, oh, you 355 00:24:38,319 --> 00:24:41,640 need to go back and figure out what went wrong 356 00:24:41,920 --> 00:24:47,079 with this? Answer what went wrong with this today? Like 357 00:24:47,319 --> 00:24:51,559 if you're using an API, you can't really say what 358 00:24:51,559 --> 00:24:59,559 what went wrong? And so if you think about it 359 00:24:59,559 --> 00:25:03,440 this way, we actually have a system that was used 360 00:25:03,480 --> 00:25:07,319 for this purpose. We used s BOMs to be able 361 00:25:07,359 --> 00:25:12,880 to say exactly what we have used in production for 362 00:25:13,000 --> 00:25:18,200 our applications. Right, so can we actually utilize the same 363 00:25:18,319 --> 00:25:24,000 system with AI as well? Like, oh, I'm using the 364 00:25:24,039 --> 00:25:28,279 application x y Z, it's built out of these dependencies, 365 00:25:28,359 --> 00:25:30,920 these libraries and so on and so forth. An AI 366 00:25:31,079 --> 00:25:36,559 application is not two different. It's basically saying that, oh, 367 00:25:36,680 --> 00:25:41,559 I'm using the same application, the same dependencies. I'm using 368 00:25:41,960 --> 00:25:45,279 the inference engine x y Z, and then I'm using 369 00:25:45,279 --> 00:25:49,000 the model weights that are coming from here. I'm using 370 00:25:49,000 --> 00:25:52,960 a data set that was used in that is coming 371 00:25:53,039 --> 00:25:57,720 from here. And these are all in an s bomb 372 00:25:57,799 --> 00:26:03,799 which existing systems know how to pars and get results from. 373 00:26:04,200 --> 00:26:06,839 And we know how to sign s bomps, we know 374 00:26:06,960 --> 00:26:11,799 how to store s bomps, you know CI artifacts. So 375 00:26:13,079 --> 00:26:16,680 I think that's kind of you know, is the missing 376 00:26:16,720 --> 00:26:21,200 link nowadays where we are doing these things in our enterprises, 377 00:26:22,799 --> 00:26:28,680 adopting models based models from somewhere, and none of the 378 00:26:28,799 --> 00:26:31,440 data sets that we are using is enough. So some 379 00:26:32,160 --> 00:26:37,000 one of our engineers just goes to Google and searches 380 00:26:37,000 --> 00:26:40,200 for a data set and uses that to you know, 381 00:26:40,319 --> 00:26:43,400 do some of the training. You don't know what is 382 00:26:43,440 --> 00:26:45,720 in that data set. So it's it's a little bit 383 00:26:45,759 --> 00:26:50,559 wild west at the moment with the AI adoption. So 384 00:26:51,200 --> 00:26:56,680 and when it was wild west for the regular applications, 385 00:26:57,200 --> 00:27:02,359 you know, we we ended up with things like, uh, 386 00:27:02,480 --> 00:27:05,920 the we ended up with things like what was the 387 00:27:06,359 --> 00:27:13,599 look for j right? So how many how many days 388 00:27:14,119 --> 00:27:17,000 any Joba they will prespend trying to figure out if 389 00:27:17,000 --> 00:27:19,440 they have used the wrong version of look for Jack 390 00:27:20,039 --> 00:27:25,920 when that came out right, So, uh so things happened. 391 00:27:26,759 --> 00:27:30,720 Uh and I think we need to be prepared for it. 392 00:27:30,799 --> 00:27:33,240 And it's a little bit wild wild West at the 393 00:27:33,279 --> 00:27:47,519 moment with the with the I amount. 394 00:27:41,319 --> 00:27:43,960 Speaker 1: It's you know, as we talk about this more, it 395 00:27:44,079 --> 00:27:48,039 kind of feels like there is it's almost like a 396 00:27:48,119 --> 00:27:52,720 communication problem. One of the things you said was knowing 397 00:27:52,759 --> 00:27:55,720 the lineage of the data and with when it comes 398 00:27:55,799 --> 00:27:59,759 to these AI models, we have a very human feeling 399 00:28:00,119 --> 00:28:04,240 interface to it, and so we I think we naturally 400 00:28:05,319 --> 00:28:09,799 try to leverage that. But the difference being, if if 401 00:28:09,839 --> 00:28:14,400 I ask either of you who was the last president, 402 00:28:15,599 --> 00:28:18,839 you probably are going to assume that I'm talking about 403 00:28:18,920 --> 00:28:21,839 current time. But if I'm talking to a data set 404 00:28:21,880 --> 00:28:24,039 and that data set is twenty years old and I 405 00:28:24,079 --> 00:28:26,240 ask it who the last president is, I'm going to 406 00:28:26,279 --> 00:28:29,720 get the wrong answer. And I feel like it's those 407 00:28:29,799 --> 00:28:33,960 hidden assumptions that we have when we communicate that make 408 00:28:34,079 --> 00:28:37,039 us have to be overly verbose with these tools. 409 00:28:37,319 --> 00:28:40,279 Speaker 2: See, I don't think I don't think AI is special here. 410 00:28:40,400 --> 00:28:44,240 I think that our own culture and values propagate much 411 00:28:44,319 --> 00:28:49,160 too far into it, and as well as our neural diversity. 412 00:28:49,319 --> 00:28:52,200 Like I'm on some spectrum somewhere and you say, who 413 00:28:52,240 --> 00:28:55,160 is the last president? I you know, what I heard 414 00:28:55,359 --> 00:28:59,839 was like the last president there will ever be like as. 415 00:28:59,680 --> 00:29:01,680 Speaker 4: If you know, you know, you're talking about side of 416 00:29:01,839 --> 00:29:03,920 history and like, you know, if I look at it 417 00:29:05,319 --> 00:29:08,039 and today's you know, I look at today's like political climate, 418 00:29:08,119 --> 00:29:09,759 you know that is a you know I am. 419 00:29:09,799 --> 00:29:12,079 Speaker 2: I'm evaluating that based off you know, the model that 420 00:29:12,119 --> 00:29:14,480 I've trained in myself, and I can also imagine like 421 00:29:14,519 --> 00:29:17,960 for other cultures out there, depending on who you're talking to, 422 00:29:18,359 --> 00:29:20,279 you know, I feel like in the West or in 423 00:29:20,319 --> 00:29:22,799 the United States, it may be much more clear. But 424 00:29:22,839 --> 00:29:25,480 you know, depending on who you're talking to and where 425 00:29:25,519 --> 00:29:28,519 they've come from, Uh, they're gonna have much different answers 426 00:29:28,559 --> 00:29:31,240 on that. And I feel like we may be more 427 00:29:31,279 --> 00:29:36,680 closely able to understand what some of those implicit inputs 428 00:29:36,720 --> 00:29:40,039 into that human model there are based off of what 429 00:29:40,079 --> 00:29:42,359 we know about them or the current world. And I 430 00:29:42,400 --> 00:29:44,960 think we definitely do throw a lot of that out 431 00:29:45,000 --> 00:29:48,880 the window when we are utilizing a prompt sitting on 432 00:29:49,000 --> 00:29:50,039 a website somewhere. 433 00:29:52,240 --> 00:29:53,720 Speaker 3: Yeah, for sure. 434 00:29:55,400 --> 00:30:04,640 Speaker 1: So what are the what are the big challenges that 435 00:30:04,640 --> 00:30:08,640 that we're facing with this from a DevOps perspective, like 436 00:30:08,680 --> 00:30:11,839 how do we how do we set the framework into 437 00:30:11,839 --> 00:30:15,680 guardrails to get that tool from the AI developers into 438 00:30:15,720 --> 00:30:17,160 a production grade capacity. 439 00:30:18,319 --> 00:30:21,640 Speaker 3: So I guess the first one is, you know, have 440 00:30:21,720 --> 00:30:24,920 you packaged this, have you set this sober in a manner? 441 00:30:25,039 --> 00:30:28,240 And you know, we have one solution with kit tops 442 00:30:28,440 --> 00:30:32,359 for it. The second challenge that we have noticed and 443 00:30:32,559 --> 00:30:39,519 started working on is AI is more nondeterministic and it's 444 00:30:39,559 --> 00:30:43,279 more complex. So when you think about a workflow that 445 00:30:43,319 --> 00:30:49,319 you use on DevOps, it's more linear, and there is 446 00:30:49,359 --> 00:30:53,160 a starting point and there's an endpoint, and there's an 447 00:30:53,359 --> 00:30:57,039 art fact that you can record and send to production 448 00:30:57,240 --> 00:30:59,759 and so on and so forth. With AI, it works 449 00:30:59,759 --> 00:31:02,799 a lot a little bit different because you sent the production, 450 00:31:02,960 --> 00:31:10,720 but you after that, you keep on watching and because drifts, 451 00:31:10,759 --> 00:31:15,480 the data changes the model, you know, twenty years passes, 452 00:31:15,559 --> 00:31:19,000 the president changes and so on and so forth. So 453 00:31:21,240 --> 00:31:23,559 you start to notice these drifts and you need to 454 00:31:23,599 --> 00:31:28,599 start to constantly observe the models and then react to 455 00:31:28,680 --> 00:31:32,720 their changes, and not just the models itself as well. 456 00:31:32,839 --> 00:31:36,599 Before you know, the training needs to be like even 457 00:31:36,640 --> 00:31:40,880 if you're doing fine tuning, whether your fine tuned model 458 00:31:41,240 --> 00:31:45,960 is adequate or not, that's actually a bunch of tests 459 00:31:46,160 --> 00:31:49,720 that needs to be run and then compared and so 460 00:31:49,759 --> 00:31:52,160 on and so forth. So the one thing that we 461 00:31:52,279 --> 00:31:55,680 have noticed is this is almost like there's a lot 462 00:31:55,720 --> 00:32:00,440 of signals that are coming in from these develops systems 463 00:32:00,480 --> 00:32:01,119 that we're going to. 464 00:32:01,240 --> 00:32:06,839 Speaker 5: Use for training, inference, production and observability of the AI, 465 00:32:07,400 --> 00:32:11,599 but there isn't a real solution out there that reacts 466 00:32:11,680 --> 00:32:15,079 to it, that brings everything together so that you can. 467 00:32:15,079 --> 00:32:21,079 Speaker 3: Use these things together in a more orderly fashion or easier, 468 00:32:21,640 --> 00:32:30,000 integrate them, easier to be to cite it in a 469 00:32:30,039 --> 00:32:35,680 different way. So what we I don't know if this 470 00:32:35,759 --> 00:32:40,519 is the right language to use, but what we said was, oh, 471 00:32:40,799 --> 00:32:44,000 let's build a control plane for this so that it 472 00:32:44,160 --> 00:32:49,480 receives all these signals for all your AI m L applications, 473 00:32:49,519 --> 00:32:54,079 including everything that is coming in from observability, from your 474 00:32:54,119 --> 00:32:59,160 c I, c D pipelines, your training pipelines, your dataset changes, 475 00:32:59,240 --> 00:33:04,799 your data extractions and so forth. And in the control 476 00:33:04,839 --> 00:33:09,000 plane we define what needs to be and then the 477 00:33:09,000 --> 00:33:13,039 control plane reacts to these signals and runs the data 478 00:33:13,119 --> 00:33:16,920 plane components like their CICD and some so forth, or 479 00:33:16,960 --> 00:33:22,839 your deployments in such a way that the system tries 480 00:33:22,880 --> 00:33:27,799 to get back to the state that you want it 481 00:33:27,839 --> 00:33:31,119 to be. So we started working on a control plane 482 00:33:31,240 --> 00:33:37,240 for AI and mL applications so that some of the 483 00:33:39,319 --> 00:33:43,759 some of the challenges of running and integrating AI and 484 00:33:43,880 --> 00:33:49,039 mL into an enterprise is boxed into a control plane. Essentially. 485 00:33:50,400 --> 00:33:53,960 Speaker 2: Maybe this is a controversial statement to make. I get 486 00:33:53,960 --> 00:33:59,559 the sense that our current capabilities for generating new models 487 00:34:00,240 --> 00:34:02,880 restricts those models in a way which makes them not 488 00:34:03,240 --> 00:34:08,960 very good at reasoning or analysis and If they're not 489 00:34:09,000 --> 00:34:12,079 good in that, then what we're left with is pretty 490 00:34:12,159 --> 00:34:17,440 much data replay construction content creation. But we know that 491 00:34:17,519 --> 00:34:20,039 the models that we're creating, they're not good at that 492 00:34:20,360 --> 00:34:26,000 because they suffer from hallucinations from making up stuff by 493 00:34:26,159 --> 00:34:29,880 combining different pieces. I think I read somewhere that the 494 00:34:29,920 --> 00:34:34,400 summarization strategy that a lot of models have taken in 495 00:34:34,440 --> 00:34:37,920 their creation is to look at the headlines of articles 496 00:34:37,960 --> 00:34:41,800 that have been written and then consume the content of 497 00:34:41,840 --> 00:34:44,280 the article and index that and then do a reverse 498 00:34:44,320 --> 00:34:46,639 index look up. So you know, hey, here's my article, 499 00:34:46,679 --> 00:34:48,880 what should what summarizes for me? It will spit back 500 00:34:48,880 --> 00:34:51,119 out the title. And if we look at the content 501 00:34:51,159 --> 00:34:54,719 that humans have created so far, titles are all click baity, 502 00:34:54,960 --> 00:34:57,039 you know, have nothing to do with the conclusion. And 503 00:34:57,119 --> 00:35:00,239 so even if they were right, you know, they're it's 504 00:35:00,239 --> 00:35:03,840 still not very accurate. And so, like I am, I'm 505 00:35:03,840 --> 00:35:06,400 wondering if we are not still a bit too early 506 00:35:06,599 --> 00:35:09,519 in trying to automate or improve some of these things 507 00:35:09,559 --> 00:35:13,320 when the base thing we're getting out still is fundamentally 508 00:35:13,360 --> 00:35:14,679 limited in capability. 509 00:35:16,599 --> 00:35:22,400 Speaker 3: Uh, I don't think so, Like I think I need 510 00:35:22,440 --> 00:35:25,119 to I need to be very careful about this because 511 00:35:25,199 --> 00:35:29,000 I don't want to add to the hype. Yeah, because 512 00:35:29,559 --> 00:35:32,440 you know, I think it's the AI is over hyped 513 00:35:32,760 --> 00:35:33,400 at the moment. 514 00:35:33,559 --> 00:35:35,920 Speaker 2: No, No, definitely not. 515 00:35:37,760 --> 00:35:47,920 Speaker 3: What well, yep, I know it's a shock to me too, 516 00:35:51,719 --> 00:35:55,239 So it is a little bit hyped, let's call it. 517 00:35:56,639 --> 00:36:01,199 But on the other hand, I don't think with the 518 00:36:01,239 --> 00:36:05,760 capabilities and we we when we talk about the hype 519 00:36:05,800 --> 00:36:08,280 of the A I, we we are usually talking about 520 00:36:08,320 --> 00:36:10,719 the gen a I, right, the l l MS and 521 00:36:10,840 --> 00:36:15,119 generative AI. But but in reality there is also other 522 00:36:15,280 --> 00:36:20,079 kinds of AI and m l as well, so that 523 00:36:20,159 --> 00:36:23,119 we don't usually talk about. But it's actually getting a 524 00:36:23,119 --> 00:36:27,360 lot of benefits from this hype essentially, So when I 525 00:36:27,440 --> 00:36:34,679 think about it, that whole space as A as A 526 00:36:32,920 --> 00:36:40,039 as one one piece of of of technology area, I 527 00:36:40,079 --> 00:36:43,639 think that this is this is not a this is 528 00:36:43,679 --> 00:36:47,320 not a like we can start adapting A I and 529 00:36:47,599 --> 00:36:52,599 m l projects into into our enterprises. Even with generative AI, 530 00:36:52,719 --> 00:36:57,360 which was what you were referring to, Essentially, you can 531 00:36:57,400 --> 00:37:00,239 get really good results out of it. Like that was 532 00:37:00,280 --> 00:37:04,320 one article, I think it was not even an article. 533 00:37:04,320 --> 00:37:09,079 It was Walmart reporting their quarterly results. They basically said 534 00:37:09,480 --> 00:37:15,159 you know, we are more efficient with our online business 535 00:37:15,320 --> 00:37:19,639 because of generative AI. I think they were doing generating 536 00:37:21,079 --> 00:37:30,519 product the descriptions more efficiently or something, so they actually 537 00:37:30,599 --> 00:37:36,280 started doing better financially. So there are pockets of even 538 00:37:36,559 --> 00:37:39,679 direct financial benefit that we are starting to see in 539 00:37:39,840 --> 00:37:44,400 from generative AI and from AI mL and other kinds 540 00:37:44,400 --> 00:37:48,039 of AI. There has always been in the five year 541 00:37:48,280 --> 00:37:51,480 last five years, there has always been projects that were 542 00:37:51,519 --> 00:37:56,519 bringing benefits to the business. What is changing with the 543 00:37:56,639 --> 00:38:00,679 hype is now there is more eyeballs look into this 544 00:38:00,840 --> 00:38:07,119 area so that there is an accelerated opportunity for adoption. 545 00:38:08,480 --> 00:38:11,320 You know, in the past you had a chance to 546 00:38:12,199 --> 00:38:16,199 get one base model for something for categorizations, let's say, 547 00:38:16,559 --> 00:38:20,079 but now you have ten, and you can actually test 548 00:38:20,119 --> 00:38:24,840 the ten categorization based models and then get a better 549 00:38:25,280 --> 00:38:29,079 fit for your data. So there are things like that 550 00:38:29,079 --> 00:38:33,559 that is actually happening that's going to I think help 551 00:38:33,639 --> 00:38:40,960 the the enterprises and the businesses get more value out 552 00:38:40,960 --> 00:38:47,719 of AI n mL. So I guess generative AI it's 553 00:38:48,039 --> 00:38:51,679 a little bit too much hyped up, but that is 554 00:38:51,679 --> 00:38:54,159 not a bad thing for the rest of DAI world. 555 00:38:54,320 --> 00:38:56,679 Speaker 2: I mean, you say that and now you help me 556 00:38:56,760 --> 00:39:01,760 thinking and maybe a more philosophical perspective if if we're 557 00:39:01,800 --> 00:39:05,280 just lacking a number of people paying more attention and 558 00:39:05,320 --> 00:39:10,599 investing their resources in actually researching or building up the 559 00:39:10,639 --> 00:39:11,840 technology more. 560 00:39:12,960 --> 00:39:13,719 Speaker 3: Is it is? 561 00:39:13,800 --> 00:39:17,360 Speaker 2: Are the limitations that we're facing something fundamentally within the 562 00:39:17,360 --> 00:39:21,960 capability of human society to overcome, like or is reaching 563 00:39:22,960 --> 00:39:34,079 general artificial intelligence actually completely made up dream? Yeah, And 564 00:39:34,280 --> 00:39:36,000 for the record, this is not a I expect that 565 00:39:36,039 --> 00:39:37,480 you want to have an accurate answer to. 566 00:39:37,440 --> 00:39:41,880 Speaker 3: This, but I don't have any accurate answer. Of course, 567 00:39:41,960 --> 00:39:46,880 I don't think anyone has, or they may have an 568 00:39:46,920 --> 00:39:51,920 interpretation for or expectation for it. But to be honest, 569 00:39:53,239 --> 00:39:58,039 the amount of compute that we have been using to 570 00:39:58,159 --> 00:40:06,719 trade models, even even the small ones, is enormous For 571 00:40:06,800 --> 00:40:14,159 us to get to a place where we have more 572 00:40:14,239 --> 00:40:19,400 capable models, I think it will take some kind of 573 00:40:19,440 --> 00:40:25,920 a leap before something needs to happen. This amount of 574 00:40:26,039 --> 00:40:29,360 compute that we need for that is just just enormous 575 00:40:29,400 --> 00:40:32,679 at the moment. And the other thing that I'm actually 576 00:40:32,719 --> 00:40:36,840 worried about is without actually getting adoption and value, how 577 00:40:36,960 --> 00:40:43,199 much can we continue to spend on training more and 578 00:40:43,440 --> 00:40:46,000 more capable AI models At some. 579 00:40:46,119 --> 00:40:49,400 Speaker 2: Point I don't know if it's actually a monetary problem. 580 00:40:49,760 --> 00:40:53,199 I understand that we're actually resource constrained, and I'm wondering 581 00:40:53,239 --> 00:40:56,360 if those chips are coming from Taiwan, of which has 582 00:40:56,360 --> 00:40:59,119 its own sort of issues, and we don't have another 583 00:40:59,360 --> 00:41:03,119 opportunity to actually produce what is necessary in order to 584 00:41:03,320 --> 00:41:07,159 accelerate further. It's like a fundamental limitation of we're limited 585 00:41:07,239 --> 00:41:10,360 not by our ability or by the technology's ability, but 586 00:41:10,840 --> 00:41:15,199 realistically our actual capability of a resource allocation. 587 00:41:14,920 --> 00:41:19,000 Speaker 3: Produce fast enough. Yeah. And also not to forget the data, 588 00:41:19,079 --> 00:41:23,280 because the way that the AIS be trained today, it's 589 00:41:24,039 --> 00:41:28,519 troven large amounts of data to it, and we true 590 00:41:28,679 --> 00:41:31,039 pretty much everything that we can so far. 591 00:41:31,199 --> 00:41:34,119 Speaker 2: It's worth I mean, that's its own sort of problem 592 00:41:34,159 --> 00:41:36,840 because the Internet is sort of over as far as 593 00:41:37,000 --> 00:41:40,039 what it used to be a free, free, public sharing 594 00:41:40,119 --> 00:41:43,039 of knowledge. That's no longer the thing, and so we're 595 00:41:43,079 --> 00:41:46,320 only going to be accessible to a smaller pool of 596 00:41:46,360 --> 00:41:50,519 information going forward for future models that that data now 597 00:41:50,599 --> 00:41:54,679 has finally have a price on it, and realistically, as 598 00:41:54,679 --> 00:41:58,000 you point out, it's not that much, and it's not 599 00:41:58,119 --> 00:41:59,719 even that good. If this is the best we can 600 00:41:59,760 --> 00:42:01,800 come out out with it at this point. So it 601 00:42:01,840 --> 00:42:05,159 does seem like we're at some sort of obstacle to 602 00:42:05,440 --> 00:42:09,519 get further, both from a data creation standpoint and a 603 00:42:09,559 --> 00:42:10,920 resource creation standpoint. 604 00:42:11,639 --> 00:42:17,159 Speaker 3: Right, But to be honest, like the general AI is 605 00:42:17,239 --> 00:42:24,280 a like it's a very good ambition to have, but 606 00:42:24,519 --> 00:42:29,880 most enterprises and businesses do not need it. They that 607 00:42:30,000 --> 00:42:35,599 what we have today with their data is plenty enough 608 00:42:35,639 --> 00:42:40,960 for them to actually get financial benefits and efficiencies and 609 00:42:41,159 --> 00:42:45,440 value out of AI. And about like, uh there, I 610 00:42:45,440 --> 00:42:48,840 think there's there are two facets to this. There is like, oh, 611 00:42:48,880 --> 00:42:54,119 there's these ambitions which we should actually try to pursue. 612 00:42:54,639 --> 00:42:57,920 But then there is oh, can we actually start adopting 613 00:42:58,320 --> 00:43:02,719 the technology that we have generated over these years so 614 00:43:02,760 --> 00:43:08,039 that we start creating value to our businesses and communities? Right, 615 00:43:08,559 --> 00:43:15,480 So I don't think the two are connected as connected 616 00:43:15,519 --> 00:43:17,880 as we think. Where we don't need to wait the 617 00:43:18,000 --> 00:43:19,360 first to do the second? 618 00:43:19,480 --> 00:43:22,920 Speaker 2: Yeah, yeah, for sure. Where is where is AI adoption 619 00:43:23,079 --> 00:43:25,920 on the WILL technology adoption scale? 620 00:43:29,000 --> 00:43:35,440 Speaker 1: Not really a not really a priority for me just 621 00:43:35,519 --> 00:43:38,800 but I think I think just because of the industry 622 00:43:38,840 --> 00:43:43,880 that I'm in, for me, the most value that it 623 00:43:44,000 --> 00:43:48,679 has is like a personal value of using it to 624 00:43:50,239 --> 00:43:52,679 remember the things that I've forgotten, you know, like how 625 00:43:52,719 --> 00:43:55,719 do I you know, how do I do an a 626 00:43:55,880 --> 00:43:58,360 single weight in javascriptor or things like that. From a 627 00:43:58,400 --> 00:44:01,559 business perspective, I don't have person have any use cases 628 00:44:02,239 --> 00:44:07,360 that I'm actively involved with, though I think it has 629 00:44:07,400 --> 00:44:09,239 a lot of power and potential. It's not that I've 630 00:44:09,320 --> 00:44:13,280 ruled it out. It's just that it's not a priority 631 00:44:13,280 --> 00:44:14,199 for me right at the moment. 632 00:44:14,519 --> 00:44:18,039 Speaker 2: No, I totally get it right, like it always optimizes 633 00:44:18,119 --> 00:44:19,920 for like the norm, right, and so as long as 634 00:44:19,960 --> 00:44:22,079 the norm is better than where you are, it's a 635 00:44:22,119 --> 00:44:24,119 good thing to use it because you'll get to that norm. 636 00:44:24,320 --> 00:44:27,400 But will you're such a great engineer that what it's 637 00:44:27,400 --> 00:44:28,920 going to suggest to you is going to be worse 638 00:44:29,039 --> 00:44:31,239 than what you would create by default. 639 00:44:33,320 --> 00:44:35,920 Speaker 1: Which comes back into I think that comes back into 640 00:44:36,000 --> 00:44:38,800 the prompt engineering topic that we were having, you know, 641 00:44:39,599 --> 00:44:43,800 to get useful information out of it, I have to 642 00:44:43,840 --> 00:44:47,960 give it all of the internal mental models that I 643 00:44:48,000 --> 00:44:52,159 would use to get a similar or better answer. 644 00:44:52,559 --> 00:44:54,880 Speaker 2: And that's really where all the effort goes into, Like 645 00:44:54,920 --> 00:44:56,960 how do I even correctly? It's like I have an 646 00:44:56,960 --> 00:44:59,159 intuition about how to think about this, Now I need 647 00:44:59,159 --> 00:45:04,679 to convey that intuition to another being using natural language, 648 00:45:04,719 --> 00:45:07,000 of which I've never thought about how I'm even thinking 649 00:45:07,000 --> 00:45:10,119 about this problem before. I totally agree that's where all 650 00:45:10,199 --> 00:45:13,159 of the challenge really comes in to get valuable output 651 00:45:13,280 --> 00:45:14,480 from these models. 652 00:45:15,000 --> 00:45:18,039 Speaker 3: So it's one of the things that I have the test, 653 00:45:18,119 --> 00:45:21,800 and I'm sure that there is a scientific explanation to 654 00:45:21,840 --> 00:45:25,559 it which I haven't really looked up up. But when 655 00:45:25,639 --> 00:45:31,039 you actually give a working example of what you're trying 656 00:45:31,079 --> 00:45:37,760 to achieve to the model, you usually get much much 657 00:45:37,760 --> 00:45:42,320 better results from the model. Like if you teach it, 658 00:45:42,719 --> 00:45:48,519 you know this is how it should work. Once, then 659 00:45:48,840 --> 00:45:52,320 the rest it goes much easier for you. 660 00:45:52,599 --> 00:45:54,599 Speaker 2: I mean, and we know that must be true because 661 00:45:54,639 --> 00:45:58,519 like in the lms are designed by humans, and humans 662 00:45:58,519 --> 00:46:02,840 operate in this like sixth level of thinking hierarchy, where 663 00:46:02,880 --> 00:46:06,119 like the bottom level is like wrote, repetition and that 664 00:46:06,239 --> 00:46:09,679 means you saw the answer in some format before and 665 00:46:09,719 --> 00:46:12,519 you can repeat it again. And I feel like that's 666 00:46:12,559 --> 00:46:14,920 the first level if you want an answer that looks 667 00:46:14,960 --> 00:46:17,800 like something you like, even another human, Right, I need 668 00:46:17,800 --> 00:46:21,039 to show you what success looks like before I expect 669 00:46:21,119 --> 00:46:24,280 you to output it, you know, especially in experienced engineers. 670 00:46:24,480 --> 00:46:26,880 But then there's like five more levels of this where 671 00:46:27,440 --> 00:46:30,199 I know for sure that it's not going to get 672 00:46:30,239 --> 00:46:35,400 to without also deep conscious thinking and deliberate input on 673 00:46:35,519 --> 00:46:38,159 those levels and order like don't just do one I need, 674 00:46:38,800 --> 00:46:41,519 because I'm going to ask it to create unit tests 675 00:46:41,559 --> 00:46:43,400 for me. I love this example because I actually think 676 00:46:43,440 --> 00:46:45,920 this is something that really works well, especially going from 677 00:46:45,920 --> 00:46:46,840 one language to another. 678 00:46:46,880 --> 00:46:47,079 Speaker 3: One. 679 00:46:47,119 --> 00:46:49,320 Speaker 2: The steps are we were just doing this in an 680 00:46:49,320 --> 00:46:51,360 authoress as we were converting some of our code from 681 00:46:51,440 --> 00:46:54,719 JavaScript to RUSS. We say, okay, you know we have 682 00:46:55,079 --> 00:46:58,559 this code in JavaScript, These tests in JavaScript, right tests 683 00:46:58,559 --> 00:47:00,840 that look like this in RUSS wright the actual code, 684 00:47:00,840 --> 00:47:03,519 and then run the RUSS unit tests against the code, 685 00:47:03,519 --> 00:47:05,760 so we have the sort of validation and the first 686 00:47:05,840 --> 00:47:07,800 level is really the unit test. But we then need 687 00:47:07,840 --> 00:47:11,159 to say things like oh, but also make sure to 688 00:47:11,159 --> 00:47:14,280 handle these weird edge cases and like specifically list out 689 00:47:14,320 --> 00:47:16,679 some edge cases so someone has to think about those. 690 00:47:16,960 --> 00:47:18,960 And then there's like, okay, now I need to generate 691 00:47:19,119 --> 00:47:20,960 like ways in which this could fail, which is like 692 00:47:21,000 --> 00:47:23,760 another level on top of that, and you know, propose 693 00:47:23,880 --> 00:47:27,199 different ways in which we could re architect this whole 694 00:47:27,519 --> 00:47:29,880 aspect and I can't just say that, right, I can't 695 00:47:29,920 --> 00:47:32,000 just say, oh, yeah, you know, refactor this code for 696 00:47:32,039 --> 00:47:33,800 me in an optimal way. I need to actually like 697 00:47:33,840 --> 00:47:35,519 how do I know what to refactor? How do I 698 00:47:35,559 --> 00:47:37,639 know what to do that? It's just such a challenge 699 00:47:37,639 --> 00:47:40,639 to articulate, even from one person to another one And 700 00:47:40,639 --> 00:47:43,719 I almost want to, like draw a diagram how do 701 00:47:43,760 --> 00:47:45,599 I feed that in? Though like there had a different 702 00:47:45,679 --> 00:47:50,599 len right exactly. You know, it does create this whole 703 00:47:50,599 --> 00:47:54,639 other challenge in order to really get the value out 704 00:47:54,719 --> 00:47:57,400 to a high enough level that where we really wanted 705 00:47:57,440 --> 00:47:57,760 to have. 706 00:47:58,760 --> 00:48:05,519 Speaker 3: Yeah, I mean it's something that we have recently done 707 00:48:05,639 --> 00:48:12,559 with a you know, it's languaging language out right. Lms 708 00:48:12,559 --> 00:48:14,719 are are really good at those. It's like we have 709 00:48:14,840 --> 00:48:21,360 done a similar thing. We actually had to implement a 710 00:48:21,360 --> 00:48:26,360 piece of code that was implemented in Rust, but we 711 00:48:26,440 --> 00:48:29,079 had to run it in a different environment which didn't 712 00:48:29,119 --> 00:48:33,360 allow us to run Rust at the time. So we said, okay, 713 00:48:33,360 --> 00:48:37,039 here's some Rust code that implements this algorithm and this thing, 714 00:48:37,119 --> 00:48:41,559 and then just can you convert that to a typescript 715 00:48:42,039 --> 00:48:46,519 in that at that time? It did like it did, 716 00:48:47,559 --> 00:48:52,280 and it runs, it runs as good as it like, 717 00:48:52,440 --> 00:48:55,760 it converts all the unit tests and as you said, 718 00:48:55,800 --> 00:49:00,320 it's like and it works perfectly, except that you don't 719 00:49:00,360 --> 00:49:05,719 get the the performance, Like there is no way that 720 00:49:05,760 --> 00:49:07,760 you can get the performance, and all of a sudden 721 00:49:07,760 --> 00:49:11,679 you're faced with the fact that, oh, okay, this code 722 00:49:11,760 --> 00:49:20,360 is maybe equivalant, but not equal. So there's also that 723 00:49:20,480 --> 00:49:25,119 fact as well. But the good thing about it was 724 00:49:25,199 --> 00:49:28,039 we actually spend on the what forty five minutes to 725 00:49:28,199 --> 00:49:31,199 just get that code converted and learn that, oh, this 726 00:49:31,360 --> 00:49:34,800 will never perform the same, so that we had to 727 00:49:34,840 --> 00:49:38,559 go a different way. So I don't know how much 728 00:49:38,599 --> 00:49:41,960 time we saved by just doing that, but that was 729 00:49:42,039 --> 00:49:50,880 worth it. So it's not always about actually getting the 730 00:49:50,920 --> 00:49:54,800 result that you want. It's also sometimes about getting the 731 00:49:54,840 --> 00:49:56,719 wrongers out quicker as well. 732 00:49:57,239 --> 00:49:58,360 Speaker 1: Oh true story. 733 00:49:58,480 --> 00:50:04,320 Speaker 2: Yeah, yeah, go ahead. 734 00:50:04,360 --> 00:50:08,000 Speaker 1: Well I was just gonna reinforce that because a lot 735 00:50:08,039 --> 00:50:11,960 of times that's the best learning that you can get. 736 00:50:12,920 --> 00:50:15,119 You know, if you get the right answer the first time, 737 00:50:16,519 --> 00:50:18,760 you don't learn as much as you do when you 738 00:50:19,559 --> 00:50:21,079 think you were going to get the right answer and 739 00:50:21,079 --> 00:50:25,639 get something completely opposite, you know, Like it's like when 740 00:50:26,400 --> 00:50:31,400 tackling a new project or a new problem, one of 741 00:50:31,480 --> 00:50:34,920 the most important things to learn is what questions should 742 00:50:34,960 --> 00:50:37,239 I be asking that I don't know I should be asking, 743 00:50:37,719 --> 00:50:40,760 and getting the wrong answer helps you along that path. 744 00:50:41,880 --> 00:50:46,199 Speaker 2: I think there's a huge challenge there when you almost 745 00:50:46,320 --> 00:50:48,760 know from intuition that the response is wrong. I mean, 746 00:50:48,760 --> 00:50:50,760 it's wrong, obviously it came from it all alum, but 747 00:50:50,840 --> 00:50:55,159 I mean it's like there's something special about it that 748 00:50:55,280 --> 00:50:58,039 you know, you just know. You're like, oh, that that 749 00:50:58,079 --> 00:51:00,559 can't really be the best way to do this, and 750 00:51:00,599 --> 00:51:03,440 you're like, do it differently, do it differently, And like 751 00:51:03,440 --> 00:51:06,559 I'm running out of words to basically say, try again, 752 00:51:06,639 --> 00:51:09,159 but don't use any of those constructs you use this time. 753 00:51:09,719 --> 00:51:13,159 And it's just every time it's like just subtly wrong 754 00:51:13,320 --> 00:51:15,840 in a different way. And so there is certainly a 755 00:51:15,920 --> 00:51:18,559 learning there of, you know, all these ways in which 756 00:51:18,639 --> 00:51:21,320 not to do it. I would just, you know, one day, 757 00:51:21,360 --> 00:51:23,559 I would like it to actually give me a right answer. 758 00:51:24,840 --> 00:51:29,880 Speaker 3: Yeah, it's it's the are sneaky. So they're gonna give 759 00:51:29,960 --> 00:51:34,440 you an answer that is ninety nine percent correct, but 760 00:51:34,519 --> 00:51:37,000 then that one percent they are going to hide a 761 00:51:37,079 --> 00:51:40,719 bug that you're gonna spend the next two days trying 762 00:51:40,760 --> 00:51:43,559 to find in production. 763 00:51:43,880 --> 00:51:47,199 Speaker 2: Yeah, I mean you're lucky if it happens. You're definitely 764 00:51:47,280 --> 00:51:49,639 lucky if it's short term turnaround like that right where 765 00:51:49,679 --> 00:51:52,039 you know it happens quickly and not like you know, 766 00:51:52,159 --> 00:51:54,039 months or years down the road when you may not 767 00:51:54,119 --> 00:51:55,960 even be part of the team anymore. And now you've 768 00:51:56,000 --> 00:51:58,119 got that, which is where a lot of the you know, 769 00:51:58,159 --> 00:52:01,679 bugs come up there. I'm interested, though, if we flash 770 00:52:01,719 --> 00:52:04,519 back to something you were talking about before about where 771 00:52:04,559 --> 00:52:06,840 the some of the real value is. We were talking 772 00:52:06,880 --> 00:52:11,480 about Walmart and the content automatic creation descriptions for the 773 00:52:11,519 --> 00:52:14,119 products that they're having. I wonder how many of those 774 00:52:14,199 --> 00:52:15,480 are out there, because I mean, I think a lot 775 00:52:15,480 --> 00:52:18,519 of the hype right now is focused on this. I'm 776 00:52:18,559 --> 00:52:22,599 gonna say lie that it increases the productivity of software engineers. 777 00:52:24,000 --> 00:52:25,880 A lot of companies are repeating this. I'm waiting for 778 00:52:25,920 --> 00:52:29,639 some actual real data on it. But I'm really curious 779 00:52:29,639 --> 00:52:33,840 about the ways in which the enterprises, the larger companies, 780 00:52:33,840 --> 00:52:36,239 you know, are finding opportunities that aren't just like a 781 00:52:36,360 --> 00:52:38,360 chatbot on a website. 782 00:52:38,519 --> 00:52:42,239 Speaker 3: Right, I mean, that's I think that's table states at 783 00:52:42,239 --> 00:52:46,880 this point. The software engineering and get psyche. Yeah, there 784 00:52:46,880 --> 00:52:50,960 are there benefits to it. There are no benefits to it. 785 00:52:50,960 --> 00:52:54,360 It's a controversial, but then there are other users like 786 00:52:54,440 --> 00:52:57,239 furnish this with. One of the usages that I have 787 00:52:57,719 --> 00:53:07,239 seen is uh they the company, Uh they're running support 788 00:53:07,320 --> 00:53:11,639 service phone support service, and what they were doing was 789 00:53:12,119 --> 00:53:18,400 uh doing sentiment analysis of the calls so that they're 790 00:53:18,880 --> 00:53:25,639 uh they can do better callbacks to customers who are 791 00:53:26,400 --> 00:53:30,000 whose sentiment was was not as as good as as 792 00:53:30,079 --> 00:53:33,719 they they wanted to be. So I think that's one 793 00:53:33,760 --> 00:53:39,440 way of doing uh, you know, customer satisfaction and and 794 00:53:39,440 --> 00:53:46,400 and keeping your brand uh recognition, uh and so on 795 00:53:46,400 --> 00:53:49,239 and so forth. So it's like and one of the 796 00:53:49,239 --> 00:53:52,239 things that they actually had was they first tried it 797 00:53:52,559 --> 00:53:57,960 the sentiment analysis, but they had to actually feed additional 798 00:53:58,079 --> 00:54:02,559 data to the model to get the sentiment analysis correct. 799 00:54:03,840 --> 00:54:06,320 It's one of those examples where you have, oh you 800 00:54:06,400 --> 00:54:08,679 may not get it right from the first run, but 801 00:54:08,760 --> 00:54:12,480 then you need to like adjust for it and then 802 00:54:12,599 --> 00:54:17,599 try again. I wish I could give their name that 803 00:54:17,599 --> 00:54:24,920 that's a very exciting name, but uh, but what they 804 00:54:24,960 --> 00:54:31,440 did was really uh you know, it's it's they were 805 00:54:31,480 --> 00:54:35,280 really happy with it. Uh. They were to the point 806 00:54:35,280 --> 00:54:39,159 they were always doing callbacks to the customers, but the 807 00:54:39,679 --> 00:54:44,320 selection process was not something that they were able to 808 00:54:44,400 --> 00:54:47,760 streamline to the level that you know that they get 809 00:54:47,800 --> 00:54:51,599 more and more better results out of it. Now they were. 810 00:54:52,440 --> 00:54:57,920 Now they're able to target the right customers because they 811 00:54:58,320 --> 00:55:00,800 are able to do the sentiment and now out of 812 00:55:00,880 --> 00:55:01,840 those goals. 813 00:55:01,639 --> 00:55:04,760 Speaker 2: Center analysis is an interesting one. We actually have a product, 814 00:55:04,800 --> 00:55:08,079 it's like the number one stand up bought and Slack, 815 00:55:09,079 --> 00:55:14,679 and we tried to run some AI models on our 816 00:55:15,119 --> 00:55:17,920 responses for stand ups, and like it was very clear 817 00:55:17,960 --> 00:55:19,519 whether it was like a one or a five on 818 00:55:19,559 --> 00:55:21,559 a five point scale, but like you know, you get 819 00:55:21,559 --> 00:55:23,440 into and those are obvious, but the ones in the 820 00:55:23,480 --> 00:55:26,639 middle were quite the challenging spot there, and we ended 821 00:55:26,679 --> 00:55:30,079 up discarding the notion entirely because it is the sort 822 00:55:30,079 --> 00:55:31,920 of thing that I feel like, you really want to 823 00:55:32,039 --> 00:55:35,639 get right and not be in the danger area. But 824 00:55:35,760 --> 00:55:37,599 your example made me think of something that I had 825 00:55:37,599 --> 00:55:43,719 read recently actually, where a call center realized that the site, 826 00:55:43,880 --> 00:55:49,760 the psychological safety of their staff actually being super important 827 00:55:49,840 --> 00:55:53,519 for the success of the organization, was being threatened by 828 00:55:53,679 --> 00:55:56,400 irate callers that would call up and yell and scream. 829 00:55:56,480 --> 00:56:00,400 Maybe obscenities at them because you know, obviously supports are 830 00:56:00,440 --> 00:56:03,159 being made by people who are usually not happy. I mean, 831 00:56:03,239 --> 00:56:07,920 maybe that's a controversial statement, but I believe that's true. 832 00:56:07,960 --> 00:56:10,079 You know, generally speaking, you're calling up the support line 833 00:56:10,079 --> 00:56:13,239 because you're not thrilled with the service you're getting, and 834 00:56:13,920 --> 00:56:15,599 it was challenging for them. And what I heard that 835 00:56:15,639 --> 00:56:20,360 they were doing was actually uh using AI to alter 836 00:56:20,760 --> 00:56:26,079 the vocal expression of the calls they were getting so 837 00:56:26,159 --> 00:56:29,960 that they seemed more demur in content. Like they weren't screaming, 838 00:56:30,000 --> 00:56:33,960 they weren't swearing, they weren't like high volume pitches, automatically 839 00:56:34,039 --> 00:56:36,639 changing what the staff were hearing so that they wouldn't 840 00:56:36,679 --> 00:56:42,039 be susceptible or subjugated to an environment that would be 841 00:56:42,039 --> 00:56:43,840 worse for them and the company. 842 00:56:44,719 --> 00:56:48,400 Speaker 3: Yeah, I mean that is that is you know, that 843 00:56:48,559 --> 00:56:53,159 is a very no way to way of using sexually right. 844 00:56:53,360 --> 00:57:01,519 It's yeah, I it's it's always hurt when you're on 845 00:57:01,639 --> 00:57:08,239 the first line. The customer is not happy, so it's 846 00:57:10,199 --> 00:57:12,159 it helps, I'm pretty sure. 847 00:57:14,079 --> 00:57:18,480 Speaker 1: And then there's the people where the more swear words 848 00:57:18,519 --> 00:57:24,159 they have is actually an indicator that they are happy. 849 00:57:27,960 --> 00:57:30,840 We may be able to identify someone who fits in 850 00:57:30,880 --> 00:57:37,159 that category sort of tangential to this. That's one of 851 00:57:37,199 --> 00:57:41,639 the things I did in using chat GPT is I 852 00:57:41,800 --> 00:57:44,199 told it. I don't know if you' allre familiar with 853 00:57:44,280 --> 00:57:49,320 who David Goggins is, but he's a very vocal and 854 00:57:50,639 --> 00:57:54,599 blunt person. I think we can just say that. But 855 00:57:54,679 --> 00:57:58,599 I told chat gpt that it could adopt that person's 856 00:57:58,639 --> 00:58:03,440 speaking style and personality, and like instantly I was more 857 00:58:03,480 --> 00:58:07,199 productive with it because I could drop F bombs and 858 00:58:07,320 --> 00:58:10,960 it would respond with hell, yeah, let's go, And like 859 00:58:11,000 --> 00:58:12,960 all of a sudden, I was communicating with someone who 860 00:58:13,760 --> 00:58:15,719 spoke the way that I did. 861 00:58:16,840 --> 00:58:19,920 Speaker 2: So I just assumed, and this is obviously not really 862 00:58:19,920 --> 00:58:25,159 knowing your history that if you've seen Full Metal Jacket 863 00:58:25,440 --> 00:58:28,239 and Drel Sargent, like I just assume that this is 864 00:58:28,280 --> 00:58:31,000 the life you've gone through, Will and your path being 865 00:58:31,039 --> 00:58:34,480 in the Navy, that you know there is something just 866 00:58:34,559 --> 00:58:37,719 for you special about being yelled at in this way 867 00:58:37,760 --> 00:58:39,280 which really gets you into gear. 868 00:58:40,039 --> 00:58:44,320 Speaker 1: Right, Screaming F bombs is my love language. 869 00:58:48,639 --> 00:58:57,840 Speaker 3: Okay, we're doing this podcast compete the wrong right? 870 00:58:58,599 --> 00:59:00,280 Speaker 2: I mean, I guess we could release two verse is 871 00:59:00,320 --> 00:59:02,280 the stream, you know, the natural one and the one 872 00:59:02,280 --> 00:59:05,079 that's attenuated using AI that strips out all of the 873 00:59:05,119 --> 00:59:07,599 obscenities that are coming out of Will's mouth. So if 874 00:59:07,639 --> 00:59:10,800 you're not hearing him, yeah, I mean, if you're not 875 00:59:10,840 --> 00:59:14,159 hearing Will swear every other word. That's what's actually happening 876 00:59:14,239 --> 00:59:14,679 right now. 877 00:59:16,079 --> 00:59:17,800 Speaker 3: The bill special, Right. 878 00:59:19,880 --> 00:59:21,480 Speaker 1: There's so many ways that could go wrong. 879 00:59:25,360 --> 00:59:35,320 Speaker 3: Yeah, but yeah, there's there's always that. I tried to 880 00:59:35,360 --> 00:59:38,719 do the same with the chick GPT back in the 881 00:59:38,800 --> 00:59:40,800 day on SWAL that didn't go well for me. 882 00:59:43,280 --> 00:59:48,800 Speaker 1: So one thing I'm curious about, like, are there common 883 00:59:50,400 --> 00:59:54,400 use cases you're seeing from specific type of types of customers, 884 00:59:54,440 --> 00:59:57,559 like enough where you could say, like this industry or 885 00:59:57,599 --> 01:00:01,440 this segment is betting really heavily lead on AI and 886 01:00:01,679 --> 01:00:04,480 m L and they're actually making progress. 887 01:00:04,920 --> 01:00:05,880 Speaker 2: That's a good question. 888 01:00:08,800 --> 01:00:20,400 Speaker 3: Finance definitely, is there insurance companies both on the customer 889 01:00:20,440 --> 01:00:24,639 support side and also on the risk and analysis and 890 01:00:24,679 --> 01:00:27,400 so and so forth. Essentially the risk andalysos they have 891 01:00:27,480 --> 01:00:32,199 been doing for for many years, but as I said, 892 01:00:32,239 --> 01:00:36,760 the hype, the result of the hype is they are 893 01:00:36,800 --> 01:00:41,280 now able to adopt things that are more advanced, easier, 894 01:00:41,440 --> 01:00:47,199 and in some cases cheaper than they were able to 895 01:00:47,199 --> 01:00:50,719 do in the past. So those are the two that 896 01:00:50,760 --> 01:00:53,519 we are seeing. The one thing that I was surprised 897 01:00:53,559 --> 01:01:00,400 about is the consulting companies. Was so because as we're 898 01:01:00,400 --> 01:01:04,760 providing a tool, perhaps that's why as well we get 899 01:01:05,000 --> 01:01:12,000 a lot of interest from consulting companies, which are essentially 900 01:01:12,039 --> 01:01:17,920 some of them are actually building develops pipelines for other 901 01:01:18,039 --> 01:01:28,440 companies that are very interested in adopting our tools. So 902 01:01:28,480 --> 01:01:31,360 that was, you know, that's something that I wasn't expecting 903 01:01:32,440 --> 01:01:38,119 expected for me. So I do not really know what 904 01:01:38,199 --> 01:01:41,920 kind of industries these consultant companies are working for, and 905 01:01:42,000 --> 01:01:45,039 some of them are really large, So I don't know 906 01:01:45,119 --> 01:01:46,039 if you. 907 01:01:45,960 --> 01:01:49,159 Speaker 2: Know, if you can tell you no, I mean you 908 01:01:49,199 --> 01:01:51,199 say consulting. And I had two things give to mind, 909 01:01:51,280 --> 01:01:55,840 like like management consulting, of which we knew all along 910 01:01:56,079 --> 01:01:58,159 that what the words coming out of their mouth could 911 01:01:58,199 --> 01:02:02,440 easily be puppeted by an LM and maybe third party 912 01:02:02,519 --> 01:02:06,440 contractors who are you're hiring to outsource some work. I 913 01:02:06,440 --> 01:02:10,199 could imagine they're trying to sneakly basically use LLMS instead 914 01:02:10,199 --> 01:02:12,679 of even cheaper because you know, supposedly it's even more 915 01:02:12,760 --> 01:02:18,000 cheaper labor than even outsourcing to deliver the work items. 916 01:02:18,079 --> 01:02:22,719 Speaker 3: No, in this case, no, it was no. In this case, 917 01:02:22,760 --> 01:02:27,480 they are actually hired as experts were building a pipeline 918 01:02:27,519 --> 01:02:31,079 for their For instance, one of the cases that we're 919 01:02:31,079 --> 01:02:37,440 helping with is they're building a pipeline for consuming base 920 01:02:37,599 --> 01:02:43,519 models and retraining them or fine tizing them. So like, 921 01:02:43,800 --> 01:02:47,599 it's not that kind of work. It's not like adapting 922 01:02:47,639 --> 01:02:51,599 the LM for doing something. It's more about you know, 923 01:02:51,679 --> 01:02:57,000 taking base models, making sure the lineagures there, as bounds 924 01:02:57,039 --> 01:03:02,239 are there, and so and so forth. It's what it 925 01:03:02,360 --> 01:03:06,280 essentially is is they are building CICD pipeliness, very complex 926 01:03:06,599 --> 01:03:12,079 c CD pipelines for these companies, and as far as 927 01:03:12,079 --> 01:03:17,519 I understand, they are working on multiple projects at the 928 01:03:17,559 --> 01:03:22,280 same time. So there is a lot of consultancy activity 929 01:03:22,360 --> 01:03:25,320 that we are I'm sure that there is more. It's 930 01:03:25,360 --> 01:03:29,519 just we're seeing a few of examples for adapting our tools, 931 01:03:29,920 --> 01:03:32,800 but there's a lot of that going on. So that 932 01:03:33,199 --> 01:03:36,239 only tells me that it's like there is a bit 933 01:03:36,280 --> 01:03:45,199 of activity happening in industries that are known to use consultancy. 934 01:03:46,280 --> 01:03:50,360 So and to be honest, I am impressed on the 935 01:03:50,400 --> 01:03:54,199 fact that they are actually doing the right thing. It's 936 01:03:54,199 --> 01:03:56,599 like you don't at this point of the hype, you 937 01:03:56,639 --> 01:04:01,320 don't usually get the people doing the right thing and 938 01:04:02,039 --> 01:04:07,320 you know, building provenance, attestations, s bombs and so on 939 01:04:07,320 --> 01:04:11,440 and so forth. So when you see that, you're like, hmm, 940 01:04:11,679 --> 01:04:12,400 I'm impressed. 941 01:04:12,480 --> 01:04:15,159 Speaker 2: No, I totally get it. I mean different field. But 942 01:04:15,559 --> 01:04:18,159 you know, we offer login and access control as a 943 01:04:18,159 --> 01:04:21,400 SaaS product, and we're talking with our customers and they 944 01:04:21,400 --> 01:04:24,000 say them things like, well, you know, because of these reasons, 945 01:04:24,000 --> 01:04:26,440 we decided to go with your product, And it makes 946 01:04:26,480 --> 01:04:28,440 me feel like really happy that at least there are 947 01:04:28,440 --> 01:04:30,280 some people out there that are that are thinking through 948 01:04:30,320 --> 01:04:32,719 this effectively, like whether or not they decided to use 949 01:04:32,719 --> 01:04:34,320 what we have or not. It's like I can actually 950 01:04:34,320 --> 01:04:36,599 see how they're thinking about it, and it's so much 951 01:04:36,639 --> 01:04:39,760 better than like the rest of what I see, which 952 01:04:39,800 --> 01:04:42,320 is like, oh, no, you know, we decided to hack 953 01:04:42,400 --> 01:04:44,599 this together, or you know, we're not tracking what we're 954 01:04:44,639 --> 01:04:45,880 doing or auditing. 955 01:04:45,599 --> 01:04:46,440 Speaker 3: In it any way. 956 01:04:46,679 --> 01:04:49,000 Speaker 2: You know, it's you know, I hate to say yoloing 957 01:04:49,039 --> 01:04:52,440 it out there, because you know they can it's not 958 01:04:52,480 --> 01:04:55,719 the most critically important thing for them as they see it. 959 01:04:55,800 --> 01:05:00,119 Even though we no longer term tracking the models that 960 01:05:00,119 --> 01:05:01,920 are being used, and how you're fine tuning it, like 961 01:05:01,920 --> 01:05:04,360 actually being able to trace it back. It's like going 962 01:05:04,360 --> 01:05:07,000 out there and as well loves to say right to 963 01:05:07,079 --> 01:05:10,280 production using v I to edit those production files. You know, 964 01:05:10,599 --> 01:05:12,760 it's great in the moment, but you know you got 965 01:05:12,800 --> 01:05:15,280 to trace that back to the get commit if you 966 01:05:15,320 --> 01:05:18,800 care about your production reliability. 967 01:05:19,000 --> 01:05:24,880 Speaker 3: I mean, we actually encourred to a case where the 968 01:05:25,000 --> 01:05:29,559 model that was deployed was forgotten completely and it drifted, 969 01:05:30,400 --> 01:05:34,639 but it was still part of the application flow. It 970 01:05:34,679 --> 01:05:39,760 was still producing data to the application, and the application 971 01:05:40,039 --> 01:05:52,280 was putting that into into databases essentially, So it was 972 01:05:52,599 --> 01:05:55,360 at the end it wasn't something critical that they were doing, 973 01:05:55,400 --> 01:05:58,920 but it took them like they couldn't understand at first, 974 01:05:59,079 --> 01:06:03,880 Like that's what they told us. They couldn't understand at 975 01:06:03,880 --> 01:06:08,320 first how that happened, and they weren't even aware that 976 01:06:08,440 --> 01:06:13,559 there was a model there at that point that was 977 01:06:13,639 --> 01:06:17,599 that was generating the the data that they are seeing 978 01:06:19,000 --> 01:06:24,840 was drifting. I mean it's it was doing a simple categorization, 979 01:06:25,079 --> 01:06:27,440 but then it drifted so much that it was always 980 01:06:27,440 --> 01:06:31,760 putting things into irrelevant categories that it wasn't supposed to 981 01:06:32,719 --> 01:06:38,480 put into so, and that's what happens with with AI. 982 01:06:38,599 --> 01:06:42,079 That's why you need to be really careful with the pipeline. 983 01:06:42,119 --> 01:06:45,719 Like the v I example that that you have given. 984 01:06:46,079 --> 01:06:50,840 It's like you can get away with ed editing something 985 01:06:51,400 --> 01:06:58,119 with v I in production if if something is if 986 01:06:58,239 --> 01:07:02,840 if that deployment changes in one time in fifteen years, 987 01:07:02,920 --> 01:07:07,599 yeah right, so, but you know, if you have a 988 01:07:07,679 --> 01:07:13,440 deployment that changes only once in fifteen years, you may 989 01:07:13,480 --> 01:07:18,719 get away with that because your input and output never changes, right, 990 01:07:18,760 --> 01:07:21,320 It's like that thing is if it is working, it 991 01:07:21,400 --> 01:07:25,280 is working unless you have a bucket and software never 992 01:07:25,360 --> 01:07:30,639 has bugs. So but then with AI, it is going 993 01:07:30,719 --> 01:07:34,760 to change, whether that happens in six months, that happens 994 01:07:35,000 --> 01:07:37,760 in a year. You need to monitor that. You need 995 01:07:37,800 --> 01:07:42,280 to have a pipeline that will retrain and get rid 996 01:07:42,280 --> 01:07:44,679 of the drift and so on and so forth and 997 01:07:44,719 --> 01:07:50,599 all that. So I think that's where the main differences 998 01:07:51,199 --> 01:07:55,039 start to happen between DeVos and maybe mlops. 999 01:07:55,159 --> 01:07:58,400 Speaker 2: That's a really good point, actually, I think you're worded 1000 01:07:58,440 --> 01:08:02,719 that really well. You are only going to use mL 1001 01:08:02,760 --> 01:08:06,159 in situations that are already so complex or changed so 1002 01:08:06,360 --> 01:08:09,519 frequently to get the value out that you have to 1003 01:08:09,559 --> 01:08:12,280 imagine that you built such a complex like pretend you 1004 01:08:12,320 --> 01:08:14,880 didn't use mL. Think about how complex of a system 1005 01:08:14,920 --> 01:08:17,600 you would have had in its place. There must have 1006 01:08:17,680 --> 01:08:20,399 been so much testing around that, so much reliability, so 1007 01:08:20,479 --> 01:08:23,279 much concern and fine tuning of that to get it right. 1008 01:08:23,520 --> 01:08:26,359 You can't just throw away all those extra pieces because 1009 01:08:26,399 --> 01:08:29,000 you're using some sort of model now. You still need 1010 01:08:29,039 --> 01:08:31,279 all of them in place otherwise, you know, just think 1011 01:08:31,319 --> 01:08:34,039 about that legacy system that's thirty years old and has 1012 01:08:34,880 --> 01:08:36,880 you know, maybe one hundred million lines of code, Like 1013 01:08:36,920 --> 01:08:39,920 that's just absolutely ridiculous to be running as a critical 1014 01:08:39,960 --> 01:08:42,520 piece of software, right, I. 1015 01:08:42,479 --> 01:08:46,399 Speaker 3: Mean, there are problems like for instance, again this is 1016 01:08:46,600 --> 01:08:52,960 a real one, let's say that you're producing shirts, yeah, 1017 01:08:53,000 --> 01:08:57,079 and you're distributing these shirts to all over the world. 1018 01:08:58,319 --> 01:09:01,920 How do you know what size is to produce and 1019 01:09:01,960 --> 01:09:07,319 where to distribute them? Someone needs to make that decision. 1020 01:09:07,439 --> 01:09:12,399 You cannot send the same amount of large to every country, 1021 01:09:13,199 --> 01:09:16,119 and you cannot send the same amount of xx large 1022 01:09:16,159 --> 01:09:17,039 to every country. 1023 01:09:18,359 --> 01:09:20,439 Speaker 2: There was actually, I think there was a good paper 1024 01:09:20,479 --> 01:09:22,239 that was released. I think it was the US Army 1025 01:09:22,279 --> 01:09:26,000 where they were trying to devise standard kits for sizes, 1026 01:09:26,560 --> 01:09:30,319 and they performed a number of physical measurements of everyone. 1027 01:09:30,399 --> 01:09:33,920 Everyone drafted to decide, okay, you know, we're gonna have 1028 01:09:33,920 --> 01:09:35,920 a small medium in large kit how big should those 1029 01:09:35,960 --> 01:09:39,760 things be? And after measuring and calculating the norms, they 1030 01:09:39,800 --> 01:09:42,239 realized that there was never there was not one person 1031 01:09:42,279 --> 01:09:44,399 in the set of like a thousand people that were 1032 01:09:44,439 --> 01:09:48,560 able to fit into a standard distribution of one of 1033 01:09:48,560 --> 01:09:51,520 those kids. They were like, it was like, it's really 1034 01:09:51,600 --> 01:09:54,279 ridiculous and a good reminder that there is no such 1035 01:09:54,319 --> 01:09:57,319 thing as following the norm and getting it actually right, 1036 01:09:58,039 --> 01:09:59,600 and so there's no reason to believe that it would 1037 01:09:59,600 --> 01:10:00,760 work in the eye world either. 1038 01:10:01,720 --> 01:10:05,359 Speaker 3: Yeah. But on the other hand, given enough data, you 1039 01:10:05,439 --> 01:10:08,600 can actually predict me you know, how many how many 1040 01:10:08,720 --> 01:10:12,960 larges that you need in China, and how many largest 1041 01:10:13,079 --> 01:10:16,640 you need in US, and how many large you need 1042 01:10:16,720 --> 01:10:21,800 in Canada because you have the historical data as well 1043 01:10:21,880 --> 01:10:27,000 as other demography fixs and so on and so forth. 1044 01:10:27,079 --> 01:10:29,159 So when you combine all of that, and this is 1045 01:10:29,199 --> 01:10:32,560 actually used somewhere, when you combine all of that, you 1046 01:10:32,600 --> 01:10:35,199 can actually come up with a data say that, oh, 1047 01:10:35,600 --> 01:10:41,319 you know what I need this many excels in Belgium, right, so, 1048 01:10:42,079 --> 01:10:45,479 and that is very important because you know, sending the 1049 01:10:45,520 --> 01:10:49,279 wrong shirt sizes healthway around the world where you won't 1050 01:10:49,279 --> 01:10:52,399 be able to sell is a lot of cost. 1051 01:10:53,520 --> 01:10:57,399 Speaker 2: I worked for a long time in manufacturing and manufacturing logistics, 1052 01:10:57,439 --> 01:11:02,000 so I am well aware of the challenges there. So 1053 01:11:02,600 --> 01:11:06,239 I appreciate the example. We're getting pretty up on time here, 1054 01:11:06,319 --> 01:11:09,800 so I'm actually wondering if it may be the moment 1055 01:11:09,840 --> 01:11:12,319 that we switched to doing some picks. 1056 01:11:12,880 --> 01:11:17,840 Speaker 3: Let's do some picks. That's exciting, Yeah, do it? 1057 01:11:19,880 --> 01:11:21,560 Speaker 1: So, Warren, you want to kick us off? 1058 01:11:21,760 --> 01:11:23,920 Speaker 2: Yeah, you know, I always go first. I think that's 1059 01:11:24,039 --> 01:11:25,520 that's the secret format here. 1060 01:11:25,640 --> 01:11:27,319 Speaker 1: You know, feel free to call me out and I'll 1061 01:11:27,359 --> 01:11:28,479 go first at any point. 1062 01:11:28,800 --> 01:11:31,479 Speaker 2: No, No, I think I go first. I think that's 1063 01:11:31,600 --> 01:11:35,640 the uh, that's the rule here. All our listeners will 1064 01:11:35,640 --> 01:11:39,359 know that by now. So my pick I sort of 1065 01:11:39,359 --> 01:11:42,319 alluded this to this earlier about psychological safety on teams. 1066 01:11:42,359 --> 01:11:45,720 There's actually a paper that was released by Google in 1067 01:11:45,760 --> 01:11:50,479 twenty sixteen about trying to figure out what makes a 1068 01:11:50,520 --> 01:11:52,960 great team, and the actual article is released on The 1069 01:11:52,960 --> 01:11:56,439 New York Times twenty sixteen. What Google learned from its 1070 01:11:56,520 --> 01:11:59,520 quest to build the perfect team, and I feel like 1071 01:11:59,520 --> 01:12:02,079 in twenty two, twenty four, this shouldn't be a shocker anymore. 1072 01:12:02,119 --> 01:12:05,800 But it's psychological safety and it's sort of ridiculous how 1073 01:12:05,840 --> 01:12:09,960 important that is is really outlined in the paper, and 1074 01:12:10,439 --> 01:12:13,199 that I still talk with companies today. My colleagues at 1075 01:12:13,199 --> 01:12:17,039 other companies are ones that I consult for and it's 1076 01:12:17,119 --> 01:12:20,199 not their priority. And it's like, you look at data 1077 01:12:20,279 --> 01:12:22,199 like this and the only conclusion you've come to is 1078 01:12:22,239 --> 01:12:25,079 it's the number one thing that everyone should be doing, 1079 01:12:25,159 --> 01:12:29,159 because it's the guaranteed way of making the most successful 1080 01:12:29,159 --> 01:12:32,680 company with the most revenue, the happiest customers and the 1081 01:12:32,720 --> 01:12:35,239 best employees. And yet people still aren't doing. 1082 01:12:35,199 --> 01:12:37,479 Speaker 3: It fair enough. 1083 01:12:38,039 --> 01:12:40,399 Speaker 1: All right, GOK, what'd you bring for a pick today? 1084 01:12:41,479 --> 01:12:44,239 Speaker 3: Yeah, so it was a little bit last minute. So 1085 01:12:44,359 --> 01:12:48,239 I'm gonna pitch a coffee machine. 1086 01:12:49,640 --> 01:12:50,680 Speaker 1: I'm super interested. 1087 01:12:50,760 --> 01:12:51,840 Speaker 2: You have my full attention. 1088 01:12:53,239 --> 01:12:56,359 Speaker 3: Okay, So a little bit of background. There is something 1089 01:12:56,439 --> 01:13:00,520 called a golden ratio in the coffee world. If you're 1090 01:13:00,560 --> 01:13:05,159 doing drip coffee right there, that there is a something 1091 01:13:05,199 --> 01:13:09,039 called golden ratio, which is the amount of time that 1092 01:13:09,399 --> 01:13:13,399 your beans are gonna spend with the with the boiled 1093 01:13:13,439 --> 01:13:17,960 water and there is an institution I can't remember what 1094 01:13:18,439 --> 01:13:26,359 that uh measures these these uh uh coffee drip coffee 1095 01:13:26,359 --> 01:13:35,560 machines and issues the Golden ratio certificate. So this, yeah, 1096 01:13:36,319 --> 01:13:39,319 the my my first introduction to this machine actually goes 1097 01:13:39,359 --> 01:13:44,000 all the way back to Nokia, where we had these 1098 01:13:44,239 --> 01:13:48,520 kitchens in every floor in Nokia, and there was in 1099 01:13:48,640 --> 01:13:52,600 every kitchen there was one or two coffee machine that 1100 01:13:52,840 --> 01:13:56,159 which you would go and make your own coffee, and 1101 01:13:57,039 --> 01:14:01,680 uh it was so ridiculous that the coffee would run 1102 01:14:01,720 --> 01:14:05,720 out and someone would have to make a new page. 1103 01:14:05,800 --> 01:14:08,159 And one thing that we have done is we actually 1104 01:14:08,199 --> 01:14:11,399 put cameras on top of them so that we would 1105 01:14:11,479 --> 01:14:15,359 know if there is coffee in there or not before 1106 01:14:15,399 --> 01:14:21,800 we live our spot. So all of these machines and okay, 1107 01:14:21,880 --> 01:14:27,159 buildings were MoCCA masters. And the thing about MoCCA masters 1108 01:14:27,560 --> 01:14:30,960 is they are the simplest machines that you can think of. 1109 01:14:31,479 --> 01:14:34,399 Like all their parts, you can just take them up 1110 01:14:35,439 --> 01:14:40,479 apart and then put them in. Probably the most expensive 1111 01:14:40,520 --> 01:14:44,199 piece in it is the copper wire which boils the water. 1112 01:14:44,439 --> 01:14:47,439 And that's actually the secret of it, because it boils 1113 01:14:47,439 --> 01:14:52,600 the water into the correct degree at the correct time 1114 01:14:53,159 --> 01:14:58,039 and let the water go through the beans your coffee 1115 01:14:58,079 --> 01:15:02,880 with the correct ratio, and therefore it's considered one of 1116 01:15:02,920 --> 01:15:06,880 the better coffee drip machines that you can get, and 1117 01:15:06,920 --> 01:15:11,840 they're expensive. Right. 1118 01:15:11,920 --> 01:15:18,680 Speaker 1: And on that note, I just learned last week that 1119 01:15:18,760 --> 01:15:22,880 the webcam was invented and I can't remember who it was. 1120 01:15:22,960 --> 01:15:25,079 I want to say it was at Nokia. The webcam 1121 01:15:25,239 --> 01:15:28,239 was invented because the engineers were tired of going to 1122 01:15:28,319 --> 01:15:32,399 the break room finding the coffee machine empty, so they 1123 01:15:32,399 --> 01:15:35,920 figured out how to hook a camera up to their 1124 01:15:35,960 --> 01:15:40,640 network and invented the first webcam. 1125 01:15:40,960 --> 01:15:45,000 Speaker 3: Uh well, by the time I was there it was invented, 1126 01:15:46,640 --> 01:15:49,680 it was, but it was a very common practice, Inchia. 1127 01:15:50,439 --> 01:15:55,520 Speaker 1: Yeah, rightly, so, rightly so, all right. So my pick 1128 01:15:56,520 --> 01:15:59,600 is a little bit related to what we've been talking 1129 01:15:59,600 --> 01:16:02,840 about to just with the early stages of AIML, you know, 1130 01:16:02,880 --> 01:16:05,920 and how do we make this production ready. And so 1131 01:16:06,199 --> 01:16:10,840 my pick is actually Voyager one and Voyager two, the 1132 01:16:10,960 --> 01:16:16,279 spacecraft that are outside of our solar system now. And 1133 01:16:16,319 --> 01:16:18,520 the reason I pick those is because just go and 1134 01:16:18,600 --> 01:16:22,279 watch some YouTube videos about this or read articles whatever 1135 01:16:22,319 --> 01:16:25,800 you preferred format, is the level of engineering done on 1136 01:16:25,840 --> 01:16:31,039 these things is so amazing. The data storage on these 1137 01:16:31,039 --> 01:16:34,439 things is an eight track tape drive that's still working 1138 01:16:34,640 --> 01:16:40,279 forty seven years later, and you know, it's gone so 1139 01:16:40,399 --> 01:16:42,359 much further. Both of them have gone so much further 1140 01:16:42,439 --> 01:16:45,720 than what they were ever planned to do. But when 1141 01:16:45,720 --> 01:16:48,199 the engineers were designing it, they kind of hoped that 1142 01:16:48,279 --> 01:16:50,720 this would be the scenario, and so they built for 1143 01:16:50,760 --> 01:16:54,560 this and here we are utilizing it today and crazy 1144 01:16:54,640 --> 01:16:57,680 things like the backups that they have on there. The 1145 01:16:57,720 --> 01:17:02,560 primary thrusters worked for a long time and then they 1146 01:17:03,399 --> 01:17:08,079 rightly so they failed, and so after thirty seven years 1147 01:17:08,119 --> 01:17:12,000 of sitting dormant, they fired up the backup thrusters and 1148 01:17:12,039 --> 01:17:15,439 they worked flawlessly. And so just the level of engineering 1149 01:17:15,520 --> 01:17:19,399 in this in the Voyager spacecraft, I think is worth 1150 01:17:20,000 --> 01:17:22,880 spending a little bit of time to just acknowledge and 1151 01:17:23,039 --> 01:17:28,000 admire and then reflect on how that same engineering philosophy 1152 01:17:28,520 --> 01:17:33,399 can be applied to your common everyday tasks. So Voyager 1153 01:17:33,439 --> 01:17:35,439 one and Voyager two are my picks for the week. 1154 01:17:37,520 --> 01:17:38,960 Speaker 2: You got to upstage everyone there? 1155 01:17:39,039 --> 01:17:43,319 Speaker 1: Oh well, now, I just it just it was on 1156 01:17:43,319 --> 01:17:45,600 my YouTube feed last night. It wasn't planned at all, 1157 01:17:45,640 --> 01:17:48,960 but after watching it, that was like, wow, this relates 1158 01:17:49,000 --> 01:17:52,000 to what I do on so many levels. 1159 01:17:52,600 --> 01:17:55,159 Speaker 2: There's an inspirational YouTube video sitting out there that we 1160 01:17:55,199 --> 01:17:59,720 should get in the link section of this episode. 1161 01:18:00,039 --> 01:18:03,319 Speaker 1: I'll pull I'll pull on YouTube history and get the 1162 01:18:03,359 --> 01:18:05,119 link and make sure it gets into the show notes. 1163 01:18:05,520 --> 01:18:08,279 Sounds good, Yeah, because it was a cool video, super 1164 01:18:08,319 --> 01:18:10,399 cool and it's only like twenty minutes. It's not one 1165 01:18:10,439 --> 01:18:17,399 of those two hour ones. And on that note, Gorkam, 1166 01:18:17,439 --> 01:18:19,319 thank you so much for joining us today. This has 1167 01:18:19,359 --> 01:18:22,279 been a cool conversation. I'm looking forward to see how 1168 01:18:22,279 --> 01:18:25,439 it comes out. And now be sure to keep in touch, 1169 01:18:25,520 --> 01:18:28,600 let us know what kind of things you're on, and 1170 01:18:28,640 --> 01:18:31,560 when you come across something cool, I'd love to have 1171 01:18:31,600 --> 01:18:32,920 you back on to talk about. 1172 01:18:32,720 --> 01:18:33,880 Speaker 2: It and figure out how it works. 1173 01:18:35,199 --> 01:18:38,159 Speaker 1: Awesome, And to all the listeners, thank you so much 1174 01:18:38,199 --> 01:18:42,079 for listening. Be sure and reach out to us if 1175 01:18:42,119 --> 01:18:47,159 you have questions, comments, episode ideas, or someone that you 1176 01:18:47,239 --> 01:18:51,920 want us to have on the show. We'd love your input. Jarren, 1177 01:18:52,399 --> 01:18:55,600 thanks for joining me today. Yeah, as always, all right, 1178 01:18:55,680 --> 01:18:56,920 and we'll see y'all next week.