1 00:00:00,240 --> 00:00:03,920 Ah, episode 401. Proof that we're still going strong. 2 00:00:04,160 --> 00:00:07,760 Like a SQL server instance running on pure spite and caffeine. 3 00:00:08,080 --> 00:00:11,560 I'm Bailey, your semisentient hostess with the mostest 4 00:00:11,560 --> 00:00:15,280 metadata. Today, Frank's joined by the ever insightful 5 00:00:15,280 --> 00:00:18,480 Andrew Brust to talk fabric AI, Microsoft 6 00:00:18,480 --> 00:00:22,160 nostalgia, and why even Red Hat folks can still love Clippy. 7 00:00:22,320 --> 00:00:26,000 Grab your headphones and your compute capacity. Let's dive in. 8 00:00:28,480 --> 00:00:32,210 Hello, and welcome back to D Data Driven, the podcast. We explore the emerging 9 00:00:32,210 --> 00:00:35,650 industry of data science, data engineering, and 10 00:00:35,650 --> 00:00:39,450 artificial intelligence. With me today is not Andy 11 00:00:39,450 --> 00:00:43,010 Leonard, who's my favorite data engineer in the world. However, I do have Andrew Brust, 12 00:00:43,010 --> 00:00:46,730 who is an og, so to speak, in the. 13 00:00:46,890 --> 00:00:50,170 In the AI and Microsoft ecosystem. And 14 00:00:50,730 --> 00:00:54,410 although I think a lot of people think that I've abandoned the Microsoft 15 00:00:54,410 --> 00:00:58,180 ecosystem, I have not. I've just had other. Other things kind of preoccupy my 16 00:00:58,180 --> 00:01:01,540 time. And you know how it is. You have kids and, you know, they demand 17 00:01:01,540 --> 00:01:05,380 attention. You have a house and all that and good problems to 18 00:01:05,380 --> 00:01:09,140 have. But I'm very glad to kind of have someone I know walk me 19 00:01:09,140 --> 00:01:12,579 back into kind of the Microsoft ecosystem, because a lot has 20 00:01:12,579 --> 00:01:16,300 changed since I left Microsoft. I did 21 00:01:16,300 --> 00:01:19,980 go to Microsoft Ignite. We were talking about that. I even scored myself a 22 00:01:19,980 --> 00:01:22,140 Clippy. Clippy plus. 23 00:01:23,580 --> 00:01:26,060 I'll have to tell you how I want him. So they had like, a little 24 00:01:26,060 --> 00:01:29,440 challenge of like, do you know Windows history? Because Windows, I guess, turned 40 this 25 00:01:29,440 --> 00:01:32,760 year and. I know, right? Yeah, 26 00:01:33,080 --> 00:01:36,800 85. That's right. Yep. And they were like, do the history of 27 00:01:36,800 --> 00:01:40,080 Windows. And I'm like, I'm like. And I had some Red Hat people and like, 28 00:01:40,080 --> 00:01:42,440 you know, I was. I would have been very embarrassed if I had gotten anything 29 00:01:42,440 --> 00:01:46,080 wrong. Turns out I got it right, actually. Good. I 30 00:01:46,080 --> 00:01:49,840 remember back then you could install just a runtime version of Windows if you 31 00:01:49,840 --> 00:01:53,680 wanted to run specific Windows apps on your. On your DOS 32 00:01:53,680 --> 00:01:57,500 machine. Yeah. Now we take for 33 00:01:57,500 --> 00:02:00,980 granted kids. They don't understand, like, no, no, you had to type in win or 34 00:02:00,980 --> 00:02:04,500 win 3.1 if you were fancy and you're running multiple 35 00:02:04,500 --> 00:02:08,020 versions of Windows at the. Same time, 36 00:02:08,100 --> 00:02:11,380 you say kids today. I mean, kids 20 years ago didn't understand 37 00:02:11,620 --> 00:02:15,300 that. This is true. This is true. But 38 00:02:15,300 --> 00:02:18,900 anyway, in terms of bringing you back into the Microsoft orbit, 39 00:02:19,060 --> 00:02:22,700 well, first of all, I'm sure Ignite did a bunch of that. But I can 40 00:02:22,700 --> 00:02:26,180 be gentle because even though I am to this day a MicroSO 41 00:02:27,360 --> 00:02:31,200 Regional Director, or as I like to say, a member of the regional Director 42 00:02:31,360 --> 00:02:34,160 program, because Otherwise it sounds like I work for Microsoft 43 00:02:35,680 --> 00:02:39,200 and an MVP, a data platform MVP, both over 44 00:02:39,360 --> 00:02:43,080 20 years now. I'm an 45 00:02:43,080 --> 00:02:46,560 industry analyst and so I look at data and analytics 46 00:02:47,360 --> 00:02:51,200 solutions across the board, not just Microsoft specific. I 47 00:02:51,200 --> 00:02:54,800 will say Microsoft is my sweet spot in terms of what it is that I 48 00:02:54,800 --> 00:02:57,100 know and where I have the most history. 49 00:02:58,860 --> 00:03:02,340 But I work with lots of other companies you've heard of, like 50 00:03:02,340 --> 00:03:05,740 Databricks and Snowflake and Cloudera and 51 00:03:06,860 --> 00:03:10,060 plenty of others. So my team and I do 52 00:03:11,500 --> 00:03:14,460 most of the reports for a research company called GigaOM 53 00:03:15,180 --> 00:03:18,980 with which I've had a very long association. Most of the reports. Sorry, I 54 00:03:18,980 --> 00:03:21,980 didn't finish that sentence. That are focused on data or analytics 55 00:03:23,280 --> 00:03:26,800 are things that we work on. So whether it be data warehouses or lake 56 00:03:26,800 --> 00:03:30,160 houses or streaming data platforms or data access 57 00:03:30,320 --> 00:03:34,160 governance or data catalogs or blah blah, blah, they all have the 58 00:03:34,160 --> 00:03:37,760 word data in them. We work on those reports. 59 00:03:37,840 --> 00:03:41,600 We create what are called gigaom radar 60 00:03:41,680 --> 00:03:45,200 reports, which are a little bit like the analog to 61 00:03:45,200 --> 00:03:49,000 Gartner's Magic quadrant in terms of looking across a category 62 00:03:49,000 --> 00:03:52,830 at a bunch of vendor solutions and rating them on 63 00:03:52,830 --> 00:03:55,150 multiple criteria which change each year. 64 00:03:56,590 --> 00:04:00,350 So. And when I started covering big data, 65 00:04:00,350 --> 00:04:04,070 because the way this got started was I was the first and only person 66 00:04:04,070 --> 00:04:07,470 at ZDNet to be covering big data. 67 00:04:08,990 --> 00:04:12,510 So that was an amazing. That's a term you don't hear a lot. You don't 68 00:04:12,510 --> 00:04:15,830 hear it anymore. No. And in fact I was at the 69 00:04:15,830 --> 00:04:19,550 older incarnation of gigaom. I was a full time employee there. 70 00:04:19,550 --> 00:04:23,050 I was their research director and they wanted, so call me research 71 00:04:23,450 --> 00:04:27,130 director for big data. And I said, can we just call it data 72 00:04:27,930 --> 00:04:31,530 or data and analytics? And we did. Because I was like, you know, 73 00:04:31,530 --> 00:04:34,250 eventually what we think is big now won't look so big. 74 00:04:35,450 --> 00:04:39,130 Have you heard my Costco rules? Say again? I have something 75 00:04:39,130 --> 00:04:42,970 called the Costco rule. The Costco rule. If you go to call, 76 00:04:42,970 --> 00:04:46,490 if you walk into Costco by hard drive, that size. That size is no longer 77 00:04:46,490 --> 00:04:48,880 big data. Fair, fair. 78 00:04:50,320 --> 00:04:54,160 Anyway, that put me in an immersion. And at that time, Microsoft 79 00:04:54,160 --> 00:04:56,240 was not in that world at all. 80 00:04:57,280 --> 00:05:00,960 Eventually the thing called hdinsight got to beta 81 00:05:00,960 --> 00:05:04,080 and it wasn't even called hdinsight when it was in beta. 82 00:05:05,120 --> 00:05:08,560 And so Microsoft started coming back into my world. Eventually, 83 00:05:08,720 --> 00:05:12,560 of course, it came back full swing. And with Microsoft fabric, 84 00:05:12,560 --> 00:05:16,280 now it's doubly full swing, which I think is 85 00:05:16,280 --> 00:05:19,920 very, very good, both for Microsoft and the industry. But 86 00:05:21,200 --> 00:05:24,960 what was I going to say? Just that. Yeah. So 87 00:05:25,600 --> 00:05:29,160 finally the two things that Were kind of orthogonal. Now have an 88 00:05:29,160 --> 00:05:32,720 intersection. Right. And that 89 00:05:32,720 --> 00:05:36,280 intersection is my sweet spot as I'm still a data platform 90 00:05:36,280 --> 00:05:40,120 mvp and I have a very long history with Microsoft's 91 00:05:40,120 --> 00:05:43,790 business intelligence stack. I was on Microsoft's 92 00:05:43,790 --> 00:05:46,950 partner advisory council going way back, like 93 00:05:47,430 --> 00:05:50,310 from 2005 to roughly 2010. 94 00:05:51,830 --> 00:05:55,670 I don't know. I saw Power BI when it was still a bunch of wireframes 95 00:05:55,670 --> 00:05:59,430 in a PowerPoint slide deck. So I've been through 96 00:05:59,430 --> 00:06:03,110 many rounds of being frustrated that Microsoft 97 00:06:03,110 --> 00:06:06,830 didn't have a good competitive play. And I'm now pretty 98 00:06:06,830 --> 00:06:10,550 satisfied that they have one that's very competitive. So we can talk about that or 99 00:06:10,550 --> 00:06:14,240 we can talk about the greater world. And as far as 100 00:06:14,240 --> 00:06:18,000 AI goes, I was interested in AI all the way back in the horse and 101 00:06:18,000 --> 00:06:21,440 buggy days when I was an undergraduate. Oh, really? 102 00:06:21,760 --> 00:06:25,360 Yeah. AI was very different then. It was about like, weird programming 103 00:06:25,360 --> 00:06:27,680 languages like Lisp and Prologue and 104 00:06:29,120 --> 00:06:32,880 Expert Systems and things of that elk. 105 00:06:33,680 --> 00:06:37,520 But neural nets existed then, and neural nets are the very 106 00:06:37,520 --> 00:06:41,240 basis for the large language models we have today. So it's not, it's 107 00:06:41,240 --> 00:06:45,020 not completely unrel it, but obviously it's very different. 108 00:06:45,820 --> 00:06:49,460 I took a prologue course, which was the. We had one offering. 109 00:06:49,460 --> 00:06:52,540 So I'm. You and I are both from New York City. So, like, you know, 110 00:06:52,540 --> 00:06:55,100 we probably accidentally crossed paths more than once. 111 00:06:56,780 --> 00:07:00,420 And I know we crossed paths during the early Power BI days 112 00:07:00,420 --> 00:07:03,660 because I think the company I worked for at the time 113 00:07:03,980 --> 00:07:07,420 was also an early believer in power bi. 114 00:07:07,980 --> 00:07:11,140 So this is what meant 2005. And then they hired. We had a guy who 115 00:07:11,140 --> 00:07:14,620 was the practice manager, Kevin, who got hired. Kevin Viers got hired into 116 00:07:14,620 --> 00:07:18,330 Microsoft. So I think he's still there, actually. He, 117 00:07:18,330 --> 00:07:22,010 he hired me back into Microsoft when I rejoined in 2018, which was kind of, 118 00:07:22,010 --> 00:07:25,770 okay, what a small world it is, you know, and, and, and for us, like, 119 00:07:25,770 --> 00:07:28,850 you know, something that my parents would always say, like 20 years could go by 120 00:07:28,850 --> 00:07:31,970 in a blink once you hit a certain age. And I'm like, good Lord, was 121 00:07:31,970 --> 00:07:35,290 that the truest thing they ever said, right? 20. 122 00:07:35,930 --> 00:07:39,690 Yeah. The scale shrinks the older. The older you get. It's a 123 00:07:39,690 --> 00:07:43,430 little. It's a little frightening. I've turned it like, into a roll of toilet 124 00:07:43,430 --> 00:07:46,510 paper that as you get closer and closer to the end, it spins out a 125 00:07:46,510 --> 00:07:49,990 lot faster because the diameter gets smaller. Oh, 126 00:07:49,990 --> 00:07:52,190 Lord, that is a very scary concept. 127 00:07:54,590 --> 00:07:58,110 But you're right. I remember one of the things. These were back in the 128 00:07:58,110 --> 00:08:01,470 cubicle days when people worked in an office and things like that. And I remember 129 00:08:01,790 --> 00:08:05,430 sitting next to Kevin. And Kevin would be on the phone like, yeah, I know 130 00:08:05,430 --> 00:08:08,750 it's weird to hear Microsoft is kind of a small player in any niche, 131 00:08:09,500 --> 00:08:13,060 but in BI and business intelligence they really were. 132 00:08:13,060 --> 00:08:16,540 And it was just kind of like, yeah, that's true. You don't really think about 133 00:08:16,540 --> 00:08:20,260 them as a small player, but at the time now it's kind of ridiculous 134 00:08:20,260 --> 00:08:24,060 to say that in data and analytics, right? And the Power BI team has just 135 00:08:24,060 --> 00:08:27,740 done phenomenal in terms of their speed to market. And what they 136 00:08:27,740 --> 00:08:31,260 built out is phenomenal. It's unreal. 137 00:08:31,820 --> 00:08:35,340 It's the Power BI team that kind of took over the 138 00:08:35,340 --> 00:08:38,710 entire Azure data group, right? 139 00:08:39,830 --> 00:08:42,630 And that included SQL Server. So 140 00:08:43,830 --> 00:08:47,390 whereas the BI team was once a little corner of the 141 00:08:47,390 --> 00:08:50,390 SQL Server world that 142 00:08:51,270 --> 00:08:55,070 initially came to Microsoft through an acquisition of assets from 143 00:08:55,070 --> 00:08:57,110 an Israeli company called Panorama. 144 00:09:00,230 --> 00:09:04,070 And Amir Nats is the distinguished engineer who actually came 145 00:09:04,070 --> 00:09:07,910 from Panorama and is very much like the father of Power BI 146 00:09:07,910 --> 00:09:10,190 and of fabric. Still there. 147 00:09:11,470 --> 00:09:15,150 Slowly but surely, not 148 00:09:15,150 --> 00:09:18,590 only did they get BI right and they always had it right on the 149 00:09:18,590 --> 00:09:22,350 server, they just never really had it right on the front end 150 00:09:23,630 --> 00:09:27,310 until the current version of Power BI that we have now kind of gelled, 151 00:09:27,550 --> 00:09:31,230 but they also ended up kind of mastering 152 00:09:31,230 --> 00:09:34,830 the software as a service approach to 153 00:09:34,830 --> 00:09:38,610 cloud services. And they took a look at the Azure 154 00:09:38,610 --> 00:09:42,410 data stack and said, we have tons of capabilities here, but they're kind 155 00:09:42,410 --> 00:09:45,770 of fragmented over several different products, 156 00:09:46,650 --> 00:09:50,370 each of which have their own kind of procurement model and 157 00:09:50,370 --> 00:09:53,930 pricing model. And that gets very hard to manage. 158 00:09:54,170 --> 00:09:57,970 And if you look really carefully at fabric, while there are 159 00:09:57,970 --> 00:10:00,650 some things that are truly native to it, 160 00:10:02,020 --> 00:10:05,500 most of the parts of it are Azure services in the 161 00:10:05,500 --> 00:10:09,140 background that have been integrated and 162 00:10:09,300 --> 00:10:12,020 that have been unified in terms of how you pay for them, 163 00:10:14,740 --> 00:10:18,500 I don't know. Microsoft needed that, by the way, the whole cloud industry 164 00:10:18,580 --> 00:10:22,180 needed that. Because Google and Amazon are just as guilty of 165 00:10:22,180 --> 00:10:25,940 having a whole sprawl of 166 00:10:25,940 --> 00:10:29,580 services without unified user 167 00:10:29,580 --> 00:10:33,340 interfaces or APIs or pricing. No, that's 168 00:10:33,340 --> 00:10:37,100 true. I mean, when I. So when I just before 169 00:10:37,100 --> 00:10:40,900 the pandemic, I was out at Tech ready, which is 170 00:10:40,900 --> 00:10:44,060 an internal Microsoft event. And they were basically 171 00:10:44,540 --> 00:10:48,060 might have been Amir, actually, now that I think about it, was presenting on 172 00:10:48,460 --> 00:10:51,980 the future of what was called synapse. And this is kind of, 173 00:10:52,220 --> 00:10:54,940 you know, he's like, you know, everything's going to be all in one pane of 174 00:10:54,940 --> 00:10:58,420 glass. Everything, basically everything you said when all these things are going to be thing 175 00:10:58,660 --> 00:11:02,460 and you know, and the speaker, which I don't. 176 00:11:02,460 --> 00:11:05,180 Can't say it was him but would make a lot of sense. He said like 177 00:11:05,180 --> 00:11:08,900 this is the future. We're going to get everything under one pane of glass billing. 178 00:11:08,900 --> 00:11:12,300 Don't worry about that, we're going to get that figured out in time. And I 179 00:11:12,300 --> 00:11:15,700 was just like, you know, I kind of saw the vision. So 180 00:11:16,020 --> 00:11:18,860 and then I don't know, maybe like a year, year and a half later I 181 00:11:18,860 --> 00:11:22,470 left Microsoft and then Fabric came out and I 182 00:11:22,470 --> 00:11:25,510 was wondered like why the change in name like 183 00:11:25,830 --> 00:11:29,350 Synapses? And I asked people like well SYNAPSE is still kind of there but it's 184 00:11:29,350 --> 00:11:32,830 really Fabric is where everything's going. I'm like all right, but like what is the 185 00:11:32,830 --> 00:11:36,630 difference per se? So if we pretend I was on a ufo, well that's 186 00:11:36,630 --> 00:11:40,470 a weird thing. Pretend I was in a coma and I just Woke up 187 00:11:40,470 --> 00:11:44,270 from 2000 2021. What? 188 00:11:44,270 --> 00:11:46,790 And I say like well what happened to Synapse? 189 00:11:47,920 --> 00:11:51,600 Sure. So I mean the functionality of SYNAPSE is still there and there was 190 00:11:51,600 --> 00:11:54,720 a lot of, I won't call them conspiracy theories but 191 00:11:54,720 --> 00:11:58,200 skepticism when Fabric came out that it was really just a 192 00:11:58,200 --> 00:12:02,040 rebrand of synapse. In fact that's not what it 193 00:12:02,040 --> 00:12:05,640 is. So the, the thing that was originally SQL Data 194 00:12:05,640 --> 00:12:09,280 Warehouse which was in Synapse as so called 195 00:12:09,280 --> 00:12:13,080 dedicated pools and the more lake housing part 196 00:12:13,080 --> 00:12:14,400 of it that was in there 197 00:12:17,120 --> 00:12:20,800 as gosh, I forget the old nomenclature, it wasn't on demand 198 00:12:20,880 --> 00:12:24,640 pools but it was something of that. 199 00:12:24,800 --> 00:12:28,240 Reserved instances or something like that. Wasn't reserved, no, 200 00:12:29,280 --> 00:12:32,760 but anyway basically a Spark based data 201 00:12:32,760 --> 00:12:34,800 lakehouse using 202 00:12:37,200 --> 00:12:41,000 Azure data lake storage as the storage layer that's still 203 00:12:41,000 --> 00:12:44,720 there but what was I going to 204 00:12:44,720 --> 00:12:47,680 say? But Fabric is a ton more because it integrates 205 00:12:48,400 --> 00:12:52,120 all this, all this streaming stuff 206 00:12:52,120 --> 00:12:55,680 that's now called real time intelligence. It integrates data 207 00:12:55,680 --> 00:12:59,040 science and by the way the data science is completely 208 00:13:00,000 --> 00:13:02,480 unique to Fabric. It's not merely just 209 00:13:03,600 --> 00:13:06,160 an embedding of Azure machine learning. 210 00:13:09,610 --> 00:13:11,690 There's also power bi of course 211 00:13:14,250 --> 00:13:17,450 now there are operational databases including 212 00:13:17,770 --> 00:13:21,450 SQL Database meaning Azure, SQL 213 00:13:21,450 --> 00:13:23,130 meaning SQL Server in the cloud. 214 00:13:25,050 --> 00:13:28,810 A lot of other pieces that were ancillary are now all 215 00:13:29,690 --> 00:13:33,290 included. There's user 216 00:13:33,290 --> 00:13:36,730 interface that covers the whole 217 00:13:36,810 --> 00:13:40,450 realm and again the billing is 218 00:13:40,450 --> 00:13:43,210 unified. So you buy a compute capacity 219 00:13:44,890 --> 00:13:48,410 and basically as you use the different services 220 00:13:48,570 --> 00:13:51,210 they're all pulling from the same pool of compute. 221 00:13:52,250 --> 00:13:56,090 So you know, you don't have to over provision for each 222 00:13:56,090 --> 00:13:58,650 one of those services just to make sure you have enough 223 00:13:59,450 --> 00:14:03,210 compute to, to satisfy it. And now we have this thing called 224 00:14:03,210 --> 00:14:06,930 Fabric IQ which brings, which brings 225 00:14:06,930 --> 00:14:09,850 generative and agentic AI into things. 226 00:14:11,290 --> 00:14:15,010 Which is good because it was kind of funny when Fabric finally went 227 00:14:15,010 --> 00:14:18,810 to general availability. That was really when 228 00:14:18,810 --> 00:14:22,530 ChatGPT and Gen AI were like 229 00:14:22,530 --> 00:14:26,050 making it big. So it looked like Microsoft finally got the data and 230 00:14:26,050 --> 00:14:29,530 analytics stack set up just in time for people to have their attention 231 00:14:30,120 --> 00:14:33,240 to, you know, diverted over to AI. 232 00:14:34,120 --> 00:14:37,640 But now we have, you know, natural 233 00:14:37,640 --> 00:14:40,600 language query is kind of just the beginning. We have 234 00:14:41,080 --> 00:14:44,840 operational agents that can actually act on 235 00:14:44,840 --> 00:14:48,640 things and can be all based and triggered 236 00:14:48,640 --> 00:14:49,720 on streaming data. 237 00:14:52,120 --> 00:14:55,400 And so if you think about Azure Event Grid, 238 00:14:55,950 --> 00:14:58,430 if you think about Azure Data Explorer, 239 00:15:03,070 --> 00:15:06,670 If you think about the data pipelines that Azure 240 00:15:06,670 --> 00:15:10,190 offers, as I said, 241 00:15:10,190 --> 00:15:13,950 the one standalone data warehouse side of things, and even 242 00:15:13,950 --> 00:15:17,310 elements of hdinsight in terms of the lakehouse, 243 00:15:17,630 --> 00:15:21,390 that's all in there. What's also nice is even though it's Azure data Lake 244 00:15:21,390 --> 00:15:25,230 storage under the hood, you have this abstraction layer over it 245 00:15:25,230 --> 00:15:28,870 called OneLake. OneLake is 246 00:15:29,510 --> 00:15:32,870 in many ways easier to deal with because 247 00:15:33,190 --> 00:15:36,790 you don't have to worry about accounts and containers and 248 00:15:37,430 --> 00:15:41,270 sizing those and so forth. It's still compatible with 249 00:15:41,270 --> 00:15:45,110 all the ADLs and Azure Blob storage APIs. 250 00:15:45,990 --> 00:15:49,790 It also supports this notion of shortcuts, which is really just a 251 00:15:49,790 --> 00:15:52,770 data virtualization technology. 252 00:15:53,490 --> 00:15:57,170 So you can have a shortcut to data in other OneLake instances 253 00:15:58,770 --> 00:16:01,410 or in ADLS proper, 254 00:16:02,290 --> 00:16:05,810 or even in Amazon S3, 255 00:16:05,890 --> 00:16:09,650 or even in Google Cloud storage or other databases. 256 00:16:10,050 --> 00:16:13,770 And logically they'll all look like they're part of OneLake and you 257 00:16:13,770 --> 00:16:17,490 can query them as such. That's impressive. That's 258 00:16:17,490 --> 00:16:21,050 impressive. You can kind of. You really. The. The vision of get everything, get 259 00:16:21,050 --> 00:16:24,530 everything under one pane of glass seems like. It'S come true 260 00:16:25,730 --> 00:16:29,410 that to what I tell people, even though it sounds maybe a little 261 00:16:29,410 --> 00:16:33,089 bit anticlimactic, is that the real 262 00:16:33,089 --> 00:16:36,290 innovation in in fabric 263 00:16:37,170 --> 00:16:40,050 isn't the tech per se. It's. 264 00:16:40,930 --> 00:16:44,190 It's all the integration of the tech and the 265 00:16:44,190 --> 00:16:47,950 abstraction layers over it that make it work together, the UI that makes it 266 00:16:47,950 --> 00:16:51,630 work together. And there's an organizational. I mean, 267 00:16:51,630 --> 00:16:55,430 there's a little inside baseball, but there's an organizational facet to it as 268 00:16:55,430 --> 00:16:58,950 well. Because all these different products were different teams. 269 00:16:59,350 --> 00:17:03,070 Yes. People don't realize that. Like, I haven't, I've been. You've been, you're an 270 00:17:03,070 --> 00:17:06,470 rd, so you kind of know, you know how the sausage is made. I was 271 00:17:06,790 --> 00:17:09,990 inside the firewall, then I saw the sausage was made. Like there are all these 272 00:17:09,990 --> 00:17:12,810 little teams that range from 273 00:17:13,930 --> 00:17:17,450 really good team team players to really not good team players. 274 00:17:18,090 --> 00:17:21,810 I think that's this polite way as I could put it. And they getting them 275 00:17:21,810 --> 00:17:25,450 all to row in the same boat or like row. In the same direction. 276 00:17:25,450 --> 00:17:29,290 Yeah, they weren't doing that. That wasn't even necessarily 277 00:17:29,290 --> 00:17:33,090 based on hostility. It was just that different people had different reporting structures 278 00:17:33,090 --> 00:17:36,690 and different priorities and different incentives. What 279 00:17:36,690 --> 00:17:40,450 worried me was that the vision of putting all this together was a great 280 00:17:40,450 --> 00:17:44,170 idea, but the execution to me at the time 281 00:17:44,170 --> 00:17:47,690 seemed like it would be next to impossible to get all these teams to kind 282 00:17:47,690 --> 00:17:51,530 of work harmoniously and somehow they 283 00:17:51,530 --> 00:17:55,250 did it. And like to me that's the, that's the absolute 284 00:17:55,250 --> 00:17:58,450 greatest innovation. And now they've got 285 00:17:58,450 --> 00:18:01,650 synergy instead of sort of internal 286 00:18:01,650 --> 00:18:05,250 competition and, you know, from there 287 00:18:05,250 --> 00:18:08,850 on, look out because, you know, whatever. I'm sure 288 00:18:08,850 --> 00:18:12,050 there are internal disharmonies somewhere. 289 00:18:12,450 --> 00:18:14,690 But I would say at the high level in general, 290 00:18:15,890 --> 00:18:18,650 anywhere you have people. You'Re going to have that I'm going to see. I think 291 00:18:18,650 --> 00:18:22,450 I still have, I think you mentioned this is my old, old laptop. You see 292 00:18:22,450 --> 00:18:25,730 Azure Data Data Explorer. I have the sticker from that. 293 00:18:27,810 --> 00:18:31,490 Oh, I see it. Yep. Sorry, I had to show that off. 294 00:18:31,730 --> 00:18:35,250 No. And what's Azure Data Explorer? For people who don't know, 295 00:18:35,410 --> 00:18:39,050 it ran under the codename of Kusto K U S T 296 00:18:39,050 --> 00:18:42,760 O. And there's some disagreement 297 00:18:42,760 --> 00:18:46,400 over whether that really references Jacques Cousteau C O U 298 00:18:46,720 --> 00:18:50,480 S T A U or not. But it's a 299 00:18:50,480 --> 00:18:54,320 fantastic super high performance 300 00:18:55,439 --> 00:18:59,120 system for, for not just 301 00:18:59,200 --> 00:19:03,040 for streaming data, but for time series data. Yeah, with. With 302 00:19:03,040 --> 00:19:05,840 its own query language and its own ability to create 303 00:19:05,920 --> 00:19:09,450 visualizations. Right. In the query language. So your results 304 00:19:10,250 --> 00:19:13,770 come back as both tabular and visualized data 305 00:19:14,010 --> 00:19:16,570 and it can handle huge volumes of data 306 00:19:18,490 --> 00:19:20,170 in a single query. And 307 00:19:22,330 --> 00:19:26,050 there's a lot of heritage in the Azure Data Explorer team that 308 00:19:26,050 --> 00:19:29,850 started in the SQL Server analysis services world. So there's 309 00:19:29,850 --> 00:19:32,810 a continuum there. And that product on its own, 310 00:19:33,680 --> 00:19:37,040 especially being called Azure Data Explorer, which made it sound like 311 00:19:37,520 --> 00:19:39,600 a tool. File explorer. Yeah. 312 00:19:41,520 --> 00:19:44,920 When they told me the name, I don't know, I was not 313 00:19:44,920 --> 00:19:47,680 reserved in saying I didn't think it was the best name, 314 00:19:49,520 --> 00:19:53,000 but that product on its own was kind of a sleeper. It 315 00:19:53,000 --> 00:19:56,800 wasn't really getting the, I don't 316 00:19:56,800 --> 00:20:00,640 know the kudos that it deserved or the attention that it deserved. And 317 00:20:00,640 --> 00:20:04,030 now that it's part of fabric now, it contributes 318 00:20:04,990 --> 00:20:08,750 to all the cool things fabric can do. So if you see what 319 00:20:08,750 --> 00:20:12,590 are called event houses in fabric, that's the 320 00:20:12,590 --> 00:20:16,430 same technology as Azure Data Explorer. Interesting. 321 00:20:16,590 --> 00:20:20,030 So correct me if I'm wrong, but I think the origin story of 322 00:20:20,030 --> 00:20:23,710 Kusto and Kusto query language was that 323 00:20:23,710 --> 00:20:27,470 the folks running Azure, like in the operations team, actually built it 324 00:20:27,550 --> 00:20:31,040 to run through all the logs that they had. Because I remembered I was at 325 00:20:31,040 --> 00:20:34,840 some super secret event and they brought in some people from 326 00:20:34,840 --> 00:20:38,600 the field and they had us do hands on labs with it 327 00:20:38,600 --> 00:20:42,360 and I'm like, I, I must have been checking my email or 328 00:20:42,360 --> 00:20:44,800 whatever. I'm like, when can I get this to my customers? And they kind of 329 00:20:44,800 --> 00:20:48,520 laughed. They're like, no, no, this is internal only. It's internal. Yeah, it began as 330 00:20:48,520 --> 00:20:51,760 an internal thing. So I was just like, oh, like you need to make this 331 00:20:51,760 --> 00:20:54,800 a product. Because I could think of 15 customers on top of my head that 332 00:20:54,800 --> 00:20:57,590 would, that would eat this up. Yeah, I'm glad it finally saw a light of 333 00:20:57,590 --> 00:21:01,390 day. If you go back to the real world outside of Microsoft 334 00:21:01,390 --> 00:21:04,950 and you think of the likes of Splunk, for example. Yes, right. 335 00:21:05,110 --> 00:21:08,790 It's in this, it's in the same space. Although their 336 00:21:08,790 --> 00:21:12,390 initial marketing just said it was a big data tool which just 337 00:21:12,630 --> 00:21:16,230 completely obfuscated what it did. But anyway, now, 338 00:21:16,390 --> 00:21:20,190 now in, in combination with these things 339 00:21:20,190 --> 00:21:22,950 called event streams, which can stream the data in, 340 00:21:25,620 --> 00:21:29,420 basically based on Azure 341 00:21:29,420 --> 00:21:33,140 Event Hub, you put it all together and 342 00:21:33,940 --> 00:21:37,460 you have the ability to do a lot of work with 343 00:21:37,860 --> 00:21:41,380 real time streaming data without really having to write much 344 00:21:41,620 --> 00:21:45,420 code, if any code. Although it does have 345 00:21:45,420 --> 00:21:48,980 its own query language called kql. There's also 346 00:21:49,540 --> 00:21:53,260 a copilot that you can just work with in natural language that will 347 00:21:53,260 --> 00:21:56,900 generate the KQL for you. Oh, very nice. And I love 348 00:21:56,900 --> 00:22:00,180 generators because not only does it mean I don't have to write the query, but 349 00:22:00,180 --> 00:22:03,900 it means I can learn the language and then write my own query 350 00:22:03,900 --> 00:22:07,580 if I want to. Reverse engineering is how I 351 00:22:07,580 --> 00:22:11,340 prefer to learn. So 352 00:22:11,340 --> 00:22:14,980 that works out really well. Yeah, I 353 00:22:14,980 --> 00:22:17,980 wanted to dive into fabric, but I wasn't really even sure where to start because, 354 00:22:19,270 --> 00:22:22,710 so I heard, and again, a lot of this is I heard, but 355 00:22:23,270 --> 00:22:26,870 the way that it's not attached to your Azure tenant. Is that 356 00:22:26,870 --> 00:22:30,670 true? It's attached just like Office 365 357 00:22:30,670 --> 00:22:34,390 or more important, Power BI. Right. It's imagine 358 00:22:34,390 --> 00:22:37,990 Power BI premium instances, and it's the outgrowth of that. 359 00:22:39,430 --> 00:22:42,190 Okay, that makes sense now because when somebody told me like, well, it's not tied 360 00:22:42,190 --> 00:22:45,360 to your Azure tenant, you need a different tenant, I'm like, but 361 00:22:45,760 --> 00:22:49,480 okay. But then somehow, I guess at some. Level it ties into SaaS, not 362 00:22:49,480 --> 00:22:53,000 tasks. Right? So using Azure services in the 363 00:22:53,000 --> 00:22:56,800 background, including even Azure OpenAI. But 364 00:22:56,800 --> 00:23:00,560 you don't have to provision anything in Azure. It's doing that on your 365 00:23:00,560 --> 00:23:04,240 behalf. So you don't need an Azure tenant at all. 366 00:23:05,120 --> 00:23:07,840 So that actually makes it a lot easier if I were to manage, if I 367 00:23:07,840 --> 00:23:11,190 had to manage it, right? There's a lot of things that, like, correct. I mean, 368 00:23:11,190 --> 00:23:13,590 I, I love the fact you have. You go to. 369 00:23:15,110 --> 00:23:18,630 Go to, you know, any of these services, 370 00:23:18,870 --> 00:23:22,710 right, and they have basically this smorgasbord, this, this big buffet 371 00:23:22,950 --> 00:23:25,870 of services you kind of pick and choose from. But at the end of the 372 00:23:25,870 --> 00:23:29,670 day, like, how do you figure out, you know, what you pay 373 00:23:29,670 --> 00:23:33,510 for, right? It becomes, like, really kind of nightmarish. Like, again, to 374 00:23:33,510 --> 00:23:37,030 me, that's the innovation is that, yeah, all the stuff has been 375 00:23:37,470 --> 00:23:41,310 brought together, put under one pricing model, 376 00:23:41,630 --> 00:23:45,230 and you don't have to worry about all the moving parts 377 00:23:45,230 --> 00:23:48,990 and all the different, all the different servers or instances 378 00:23:48,990 --> 00:23:52,670 that might have to be provisioned and sized. That all goes away. 379 00:23:53,150 --> 00:23:56,510 And again, everything is built out of a single 380 00:23:56,670 --> 00:23:57,950 pool of compute. 381 00:24:01,070 --> 00:24:04,750 It's not a perfect analogy, but I think of like the old days of 382 00:24:05,940 --> 00:24:09,180 cell phones, when you got a certain number of minutes per month, but you could 383 00:24:09,180 --> 00:24:12,940 roll them over. And it's not that you can do that with fabric. 384 00:24:12,940 --> 00:24:16,020 I'm not saying you can roll over your compute from one month to the next, 385 00:24:16,740 --> 00:24:20,420 but what you. What is fungible is how the 386 00:24:20,420 --> 00:24:24,020 compute is used amongst the different subservices 387 00:24:25,140 --> 00:24:28,900 of fabric so you don't have to provision a certain 388 00:24:28,900 --> 00:24:32,510 amount of compute just for streaming or, or just for 389 00:24:32,510 --> 00:24:36,270 AI or just for data. Lakehouse. 390 00:24:36,750 --> 00:24:40,390 Because it's all from. When it's all from one 391 00:24:40,390 --> 00:24:44,230 pool. All right, that makes a lot of sense now because, like, that was always 392 00:24:44,230 --> 00:24:47,990 when I first got it. When I left Microsoft, I, you know, started experimenting 393 00:24:47,990 --> 00:24:51,670 with aws and I was just like, I just want to create a 394 00:24:51,670 --> 00:24:55,070 website. Why don't I need these, like, hundreds of different services underneath, 395 00:24:55,390 --> 00:24:59,070 right? Like, why do I need. I understand why. I need identity, access, management, right? 396 00:24:59,150 --> 00:25:02,940 That made sense to me. But. But like, when it came to Route 53 and 397 00:25:02,940 --> 00:25:06,140 like all this crazy stuff, I'm like, I just want to spin up a stupid 398 00:25:06,140 --> 00:25:09,860 website, right? This is not, let alone do anything complicated, right? Where you need to 399 00:25:09,860 --> 00:25:13,700 have all these underlying things. Like SageMaker, right, has this whole thing and they tried 400 00:25:13,700 --> 00:25:17,420 to abstract away all the underlying services. But even when you 401 00:25:17,420 --> 00:25:21,260 kill. This is. The thing that really annoyed me was when I killed the 402 00:25:21,260 --> 00:25:24,490 SageMaker instance, I was still getting like, you know, 20, 403 00:25:24,616 --> 00:25:28,350 $30 a month, not a lot, but I was still getting that on my 404 00:25:28,350 --> 00:25:31,110 bill and eventually I just closed the account because I'm like, I'll have to start 405 00:25:31,110 --> 00:25:34,790 fresh again in the future because like I God only knows what, 406 00:25:34,790 --> 00:25:38,550 what, what I've spent. And for people who are just learning and wanting 407 00:25:38,550 --> 00:25:41,790 to get their skill sets up, Microsoft is pretty generous with 408 00:25:42,030 --> 00:25:45,070 trial, trial 409 00:25:45,150 --> 00:25:48,950 capacities as they call them. A capacity basically is a, you know, 410 00:25:48,950 --> 00:25:52,590 a server or an instance. However, 411 00:25:52,670 --> 00:25:56,440 the, if you want to do anything with the AI, you do need a 412 00:25:56,440 --> 00:26:00,040 paid instance. But there are some pretty, there are some pretty affordable 413 00:26:00,040 --> 00:26:03,880 ones. And this gets a little confusing. 414 00:26:03,880 --> 00:26:07,520 If you provision the fabric instances 415 00:26:07,680 --> 00:26:11,400 through Azure, again you don't have to, 416 00:26:11,400 --> 00:26:15,120 that connection doesn't have to be there. But if you provision it through Azure, 417 00:26:15,120 --> 00:26:17,280 you can pause and resume those instances. 418 00:26:19,040 --> 00:26:22,800 Okay, so you do that a lot. You could be like, hey man, I'm 419 00:26:22,800 --> 00:26:26,460 taking, I'm taking a week between Christmas and New Year's off, so pause it. 420 00:26:26,860 --> 00:26:30,500 Totally. I brought up my own cheat 421 00:26:30,500 --> 00:26:32,860 sheet in the background when no one was looking. But 422 00:26:34,460 --> 00:26:38,020 so Azure Data Lake Storage, Azure Synapse, as you 423 00:26:38,020 --> 00:26:41,659 mentioned, Azure Data Factory, Azure Event Hubs, Azure Data 424 00:26:41,659 --> 00:26:45,500 Explorer, Elements of Azure Machine Learning 425 00:26:47,100 --> 00:26:49,980 and Power BI all come together in fabric. 426 00:26:50,780 --> 00:26:54,520 Interesting. So it's like one roof for which I think is 427 00:26:54,520 --> 00:26:58,240 a brilliant strategy. Right. Because Microsoft's core strength in the data and 428 00:26:58,240 --> 00:27:02,000 analytics space isn't necessarily having frontier models, isn't 429 00:27:02,000 --> 00:27:05,600 necessarily having the, the cutting most cutting edge research. 430 00:27:05,600 --> 00:27:09,400 Although I love just making it usable. Exactly. Making it usable 431 00:27:09,480 --> 00:27:13,160 and turnkey. Right. Like, not that I don't love my folks in Microsoft Research. 432 00:27:13,160 --> 00:27:16,920 Right. I know some of them. Listen, love you all, you guys have the best 433 00:27:17,000 --> 00:27:20,700 conference in the world. But you know, but, but I 434 00:27:20,700 --> 00:27:24,460 mean, but you're making it usable. Right. And I think that that's really 435 00:27:24,460 --> 00:27:28,060 string and they have all these separate tools. I think that was really the challenge 436 00:27:28,060 --> 00:27:31,660 right. When it was a shrink wrap company, you knew what you bought. But when 437 00:27:31,660 --> 00:27:35,140 it became like a SaaS pass company, you kind of could just 438 00:27:35,300 --> 00:27:39,100 a couple of clicks, you could provision stuff. So it eventually kind of 439 00:27:39,100 --> 00:27:42,740 got too chaotic. Now I like the idea of them kind of bucketizing this or, 440 00:27:42,820 --> 00:27:46,430 or rolling it up behind one service 441 00:27:46,430 --> 00:27:50,030 where because it, it just like the AWS problem. Right. Like I 442 00:27:50,030 --> 00:27:53,710 spun up SageMaker. Right. And did 443 00:27:53,710 --> 00:27:56,990 you that I needed, I needed underlying storage. I needed this. I needed this. I 444 00:27:56,990 --> 00:28:00,590 needed DNS, I needed that. I needed that to the point where look, I, I'm 445 00:28:00,590 --> 00:28:03,710 okay spending X amount of dollars on learning 446 00:28:03,710 --> 00:28:07,510 Sagemaker Right. But I wasn't okay with 447 00:28:07,510 --> 00:28:10,470 when I turned off the instance. I'm still getting built. What am I getting built 448 00:28:10,470 --> 00:28:14,280 on? That lack of transparency, intentional or not 449 00:28:14,280 --> 00:28:17,920 on AWS's part, has left a bad taste in my 450 00:28:17,920 --> 00:28:21,240 mouth, you know, for cloud services in general. 451 00:28:21,880 --> 00:28:25,720 Sure. By the way, you mentioned Microsoft Research and a 452 00:28:25,720 --> 00:28:29,240 couple of things. So when I listed all those Azure services, I forgot to say 453 00:28:29,240 --> 00:28:33,080 Azure OpenAI. So add that to the list. But also, 454 00:28:34,840 --> 00:28:38,640 although I said there's elements of Azure machine learning in there, the data science 455 00:28:38,640 --> 00:28:41,960 workload in fabric is really mostly unique to fabric, 456 00:28:42,480 --> 00:28:45,600 but it's based on technology that comes out of 457 00:28:46,800 --> 00:28:50,520 Microsoft Research. So for example, there 458 00:28:50,520 --> 00:28:53,600 was something called flaml F L a M L, 459 00:28:54,160 --> 00:28:57,840 which is the fast library for automated machine learning 460 00:28:58,720 --> 00:29:01,520 and tuning. And that's built in. 461 00:29:06,960 --> 00:29:10,760 So are things like, like ML Flow, which is an 462 00:29:10,760 --> 00:29:14,000 open source experiment management 463 00:29:15,120 --> 00:29:18,320 platform that's built into a lot of commercial AI 464 00:29:18,320 --> 00:29:21,920 platforms. So they didn't, they didn't just kind 465 00:29:21,920 --> 00:29:25,760 of embed and put their own badge 466 00:29:25,760 --> 00:29:29,280 on it. They built their own, their own ML 467 00:29:29,840 --> 00:29:33,040 stuff from, from these open source components. 468 00:29:33,360 --> 00:29:36,750 Right, Right. Well that's interesting because like the world of AI is 469 00:29:36,750 --> 00:29:39,950 largely dominated by open source. Right? 470 00:29:40,590 --> 00:29:44,310 Right. I mean, Sagemaker, I'll stop kicking the AWS 471 00:29:44,310 --> 00:29:47,870 the curb in a minute. But like SageMaker is basically a wrapper of Jupyter 472 00:29:47,870 --> 00:29:51,589 notebooks. Right. Azure ML, at least when I last used it, was largely 473 00:29:51,589 --> 00:29:55,350 a wrapper around Jupyter notebooks. Right. 474 00:29:55,350 --> 00:29:58,870 So a lot of the core technology here does tend to 475 00:29:58,870 --> 00:30:02,510 lean towards open source, which from my own personal career development 476 00:30:02,510 --> 00:30:05,430 point of view, and they're not paying me to say this, you know, one of 477 00:30:05,430 --> 00:30:08,390 the things that led me to Red Hat, right. Was the idea that, you know, 478 00:30:08,470 --> 00:30:12,270 this is largely a movement driven by open source. So, you 479 00:30:12,270 --> 00:30:15,990 know, let's see what we could do here. Right. And not, not a commercial, 480 00:30:15,990 --> 00:30:19,710 not a commercial, not a sermon. Just, just, just point it 481 00:30:19,710 --> 00:30:22,950 out because I think it's interesting how quickly open source has taken over 482 00:30:23,990 --> 00:30:27,510 the, the, the certainly the AI world. Right. 483 00:30:28,790 --> 00:30:32,470 But I also, by the way, there's notebooks in fabric too. If that wasn't, 484 00:30:32,810 --> 00:30:36,610 if that wasn't already implied or obvious and 485 00:30:36,610 --> 00:30:40,250 they are based on Jupyter, but you don't see the Jupyter skin. Right? It's 486 00:30:40,250 --> 00:30:43,970 all right. Yep. Well, I think it's really what's really 487 00:30:43,970 --> 00:30:47,650 impactful about kind of the notebook interface once you get used to it. And it 488 00:30:47,650 --> 00:30:51,210 is an adjustment for people who, like you and me, grew up with Visual Studio 489 00:30:51,210 --> 00:30:54,930 and Dare I say interdev. Right. The 490 00:30:54,930 --> 00:30:58,700 idea that you can code in a browser, right. And you know, 491 00:30:58,700 --> 00:31:02,020 no local installs really required. 492 00:31:02,580 --> 00:31:05,300 It's been very freeing, right, because you can spin up an environment, 493 00:31:06,180 --> 00:31:09,700 you can, you know, spin up a heavy cluster, 494 00:31:09,780 --> 00:31:13,220 leave it running. Right. Do its thing, you know, close the 495 00:31:13,220 --> 00:31:17,060 laptop, go in the car, go in the train home, and you get on the 496 00:31:17,060 --> 00:31:20,660 other side and you're like, oh, it's done. Right. Like, that's kind of nice, actually. 497 00:31:21,220 --> 00:31:25,020 Right. And of late, I've been a big fan, an increasing 498 00:31:25,020 --> 00:31:28,680 fan of kind of my own. Of. Of local AI. Like your own private AI. 499 00:31:29,400 --> 00:31:32,920 Right. Which explains, you know, I bought a. It was. It was an early 500 00:31:32,920 --> 00:31:36,440 Christmas gift, probably Father's Day birthday and anniversary gift too. 501 00:31:38,200 --> 00:31:41,720 My own DGX Spark. So I have my own AI running locally. Right. 502 00:31:41,720 --> 00:31:45,320 So it's not throwing shade at the cloud. It's just 503 00:31:45,320 --> 00:31:49,080 when I run a job, I don't have to think about the costs. Right. 504 00:31:49,960 --> 00:31:53,120 And that, that sort of freedom 505 00:31:54,080 --> 00:31:57,880 from worry can be really important as you're 506 00:31:57,880 --> 00:32:01,720 learning things because you just don't. You don't have to keep looking over 507 00:32:01,720 --> 00:32:05,480 your shoulder at the, at the meter or mixing metaphors there. 508 00:32:05,480 --> 00:32:09,240 But. Yeah. So are you using Llama models and such, or 509 00:32:09,240 --> 00:32:11,960 what do you. What do you really. Yeah, I have Llama. The thing I've been 510 00:32:11,960 --> 00:32:15,200 doing mostly the most is, is doing fine tuning 511 00:32:15,440 --> 00:32:17,360 Loras for image makers 512 00:32:19,690 --> 00:32:23,290 in the past. That takes about 90 minutes to maybe two hours on this 513 00:32:23,290 --> 00:32:26,810 box, which would probably be 514 00:32:26,810 --> 00:32:30,650 equivalent to about 90 minutes of Azure service 515 00:32:30,890 --> 00:32:33,770 or a VM. Right. That's like 516 00:32:34,010 --> 00:32:37,850 $150 a pop, I would say, for the type of machine 517 00:32:37,850 --> 00:32:41,130 I want. So the idea, I could just spin that off and I have the 518 00:32:41,130 --> 00:32:44,820 added benefit as it heats my office, but I don't really 519 00:32:44,820 --> 00:32:48,660 have to. I don't have to think about 520 00:32:48,740 --> 00:32:51,420 like, oh, God, you know, like, how much is that going to cost me in 521 00:32:51,420 --> 00:32:54,740 cloud services and things like that. I do think that the. 522 00:32:54,820 --> 00:32:58,300 There's. I have a lot of questions because even though I was at 523 00:32:58,300 --> 00:33:02,020 Ignite, I honestly spent my entire time at the booth and kind of walk around 524 00:33:02,020 --> 00:33:05,380 the expo floor. I didn't have. I only had a hall pass. Right. So. 525 00:33:06,420 --> 00:33:10,250 But one of the things I heard mentioned was Azure 526 00:33:10,970 --> 00:33:14,810 AI Foundry. What is that? Because you also mentioned 527 00:33:14,810 --> 00:33:18,370 Azure OpenAI, which I know the relationship between OpenAI 528 00:33:18,370 --> 00:33:21,850 and Microsoft has not been as cozy as it once was. 529 00:33:22,010 --> 00:33:24,890 And they've also, they've also added 530 00:33:27,210 --> 00:33:30,090 Claude, the anthropic models too. Right. So 531 00:33:32,490 --> 00:33:35,610 what is Azure OpenAI and how does that relate to Azure AI foundry. 532 00:33:37,290 --> 00:33:40,410 Right. So OpenAI is just the 533 00:33:41,210 --> 00:33:44,250 Azure. OpenAI is Azure's hosted 534 00:33:44,970 --> 00:33:48,570 instance of the open OpenAI models and 535 00:33:48,570 --> 00:33:52,210 APIs, because you can 536 00:33:52,210 --> 00:33:55,370 procure those directly from OpenAI themselves. 537 00:33:55,690 --> 00:33:59,450 Right. Or you can use them on Azure. And even 538 00:33:59,450 --> 00:34:03,010 if the direct model is still running on Azure in the 539 00:34:03,010 --> 00:34:06,570 background, it's still a difference in terms of procurement 540 00:34:06,570 --> 00:34:10,330 and billing and so forth. So 541 00:34:10,330 --> 00:34:14,170 you've got all the APIs around that, you've got the models. And then of course 542 00:34:14,570 --> 00:34:16,170 you need tooling to do, 543 00:34:18,090 --> 00:34:21,290 to do rag applications. Right. So 544 00:34:22,250 --> 00:34:25,850 there was tooling for that, There was also tooling for building 545 00:34:26,010 --> 00:34:29,810 copilots. There was Copilot Studio. Right. And these things 546 00:34:29,810 --> 00:34:32,810 are all kind of coming together in Azure Foundry. 547 00:34:34,890 --> 00:34:38,610 Yeah. So it's, you know, you worked at Microsoft, so you 548 00:34:38,610 --> 00:34:41,890 know how this works where like different teams do different things with different 549 00:34:41,890 --> 00:34:45,730 brands and eventually they may get kind of rationalized 550 00:34:45,730 --> 00:34:49,410 together. So the 551 00:34:49,410 --> 00:34:53,170 Foundry side of things helps there. And by 552 00:34:53,170 --> 00:34:56,490 the way, I'm glad you mentioned Foundry because in addition 553 00:34:56,810 --> 00:35:00,410 to this thing called Fabric iq, which we haven't really 554 00:35:00,410 --> 00:35:03,950 talked about, there's also Foundry 555 00:35:03,950 --> 00:35:06,630 IQ and there's also Work iq. 556 00:35:07,430 --> 00:35:10,310 Sounds like IQ is the new copilot buzzword. 557 00:35:12,150 --> 00:35:15,830 Well, it's the agentic buzzword. And the 558 00:35:15,830 --> 00:35:19,550 idea, if you think about kind of all the promise of Microsoft 559 00:35:19,550 --> 00:35:22,470 graph In the office M365 World, 560 00:35:24,150 --> 00:35:27,600 work IQ kind of sits 561 00:35:28,240 --> 00:35:31,960 over that realm. But you don't have to worry about the 562 00:35:31,960 --> 00:35:35,440 graph APIs directly. Then 563 00:35:35,440 --> 00:35:39,280 Fabric IQ sits over everything in the fabric world 564 00:35:40,400 --> 00:35:42,400 based on a pretty rich 565 00:35:44,000 --> 00:35:47,440 semantic model that can be developed so that 566 00:35:47,680 --> 00:35:51,280 when you are querying your data in natural language, the actual, 567 00:35:54,090 --> 00:35:57,770 the actual vocabulary or jargon in your particular 568 00:35:57,930 --> 00:36:01,050 organization is well understood, including 569 00:36:02,410 --> 00:36:05,370 the entities and the relationships between those entities. 570 00:36:06,090 --> 00:36:09,770 And then Foundry IQ is a way for building agents at 571 00:36:09,770 --> 00:36:13,610 the higher level that can actually talk to your structured data 572 00:36:14,170 --> 00:36:17,370 via Fabric IQ and your 573 00:36:18,670 --> 00:36:22,110 work and organizational related data via work iq. 574 00:36:22,990 --> 00:36:26,710 And that, that was, I guess 575 00:36:26,710 --> 00:36:29,390 the big vision at Ignite this year 576 00:36:30,350 --> 00:36:32,830 was to talk about all those IQ pieces. 577 00:36:35,390 --> 00:36:38,750 So there you go. No, that's interesting. 578 00:36:39,070 --> 00:36:41,550 So one of the things that comes up, I'm sorry, I cut you off, 579 00:36:43,470 --> 00:36:46,350 we're recording this the day after Christmas, so things are a little. 580 00:36:47,820 --> 00:36:51,660 Still recovering from the manicness of yesterday. And 581 00:36:55,020 --> 00:36:58,780 what's Microsoft's plans for? Kind of. Because one of the things you're seeing happen a 582 00:36:58,780 --> 00:37:02,540 lot more is this notion of sovereign 583 00:37:02,540 --> 00:37:06,300 AI or data sovereignty. And kind of like people are very 584 00:37:06,620 --> 00:37:10,460 much more conscious about the value of their data. And 585 00:37:11,340 --> 00:37:14,900 I actually think that given Microsoft as opposed to AWS or 586 00:37:14,900 --> 00:37:18,400 Google, Microsoft does have a history of selling shrink 587 00:37:18,400 --> 00:37:21,760 wrap software, right? So I do think Microsoft has a unique 588 00:37:21,760 --> 00:37:24,960 advantage of that over their modern day competitors. 589 00:37:26,480 --> 00:37:30,200 What is the kind of the on Prem story? Right, because that, that 590 00:37:30,200 --> 00:37:32,000 has certainly been. A. 591 00:37:33,920 --> 00:37:37,720 An advantage for, for my day job at Red Hat is the fact that you 592 00:37:37,720 --> 00:37:40,720 know, hey look, if you live in Country X and there's no 593 00:37:42,250 --> 00:37:45,930 AWS Azure GCP footprint there, you can 594 00:37:45,930 --> 00:37:49,570 just find a local hosting provider down the street, you know, do it 595 00:37:49,570 --> 00:37:53,170 yourself, right? Like you know the Linux 596 00:37:53,170 --> 00:37:55,370 ethos, right? Of like just do it yourself. 597 00:37:56,810 --> 00:38:00,450 And I'm seeing a lot of customers that normally one 598 00:38:00,450 --> 00:38:03,610 customer in Latin America, right, they were in a country that, 599 00:38:04,090 --> 00:38:07,610 you know, they did not have access to 600 00:38:08,530 --> 00:38:12,250 an Azure data center and they just said we have to run 601 00:38:12,250 --> 00:38:15,970 this on prem because this is either regulated or soon to be regulated. 602 00:38:16,210 --> 00:38:20,050 So I imagine that if I've seen it, I can't imagine I'm 603 00:38:20,050 --> 00:38:23,890 the only one that's seen that. What's 604 00:38:23,890 --> 00:38:27,610 their thinking? Because the answer when I was there was Azure this, Azure 605 00:38:27,610 --> 00:38:31,290 that. I think the answer is sometimes 606 00:38:31,290 --> 00:38:33,990 Azure is part of the answer, but it's not the whole answer. 607 00:38:34,940 --> 00:38:38,420 Agreed. So yeah, I don't have a perfect 608 00:38:38,420 --> 00:38:42,060 answer here because some companies really do make their entire 609 00:38:42,140 --> 00:38:45,740 stack work across different clouds and 610 00:38:47,900 --> 00:38:51,740 right inside of a Kubernetes environment that you might run 611 00:38:51,740 --> 00:38:54,540 on premises. He said to the Red Hat guy. 612 00:38:56,380 --> 00:38:59,860 I know that story. Yeah, so 613 00:38:59,860 --> 00:39:02,380 Fabric doesn't have that story. Fabric is 614 00:39:03,580 --> 00:39:06,940 software as a service cloud based product platform 615 00:39:07,500 --> 00:39:10,780 no matter what. However, various 616 00:39:10,780 --> 00:39:14,540 components of Fabric do exist as 617 00:39:14,540 --> 00:39:18,340 on premises products. This is not the way I'd recommend it, but I'll just make 618 00:39:18,340 --> 00:39:22,140 people aware that of course SQL Server can run 619 00:39:22,140 --> 00:39:25,980 on premises, so can the data Warehouse 620 00:39:25,980 --> 00:39:29,460 in effect in the form of something called Analytics 621 00:39:29,460 --> 00:39:31,730 Platform System. Terrible name. 622 00:39:33,490 --> 00:39:37,170 That is basically the new brand for what was Parallel Data Warehouse. 623 00:39:37,330 --> 00:39:40,890 And there is a Lakehouse component to that as well as a 624 00:39:40,890 --> 00:39:43,570 Warehouse component, Power bi. 625 00:39:44,450 --> 00:39:48,290 Obviously the desktop runs on premises, but there is something called 626 00:39:48,290 --> 00:39:52,130 Power BI Report Server that is part of SQL Server 627 00:39:52,130 --> 00:39:55,890 so that your Power BI reports can run on premises. 628 00:39:56,050 --> 00:39:59,660 And so again 629 00:39:59,820 --> 00:40:03,660 various components can run completely sovereign and on 630 00:40:03,660 --> 00:40:07,500 prem. I think maybe more important though is the fact that 631 00:40:07,500 --> 00:40:10,220 you can use fabric 632 00:40:11,420 --> 00:40:15,020 and OneLake to incorporate 633 00:40:15,740 --> 00:40:18,860 data that remains on premises even if, 634 00:40:19,820 --> 00:40:23,180 even if the engines are not running on premises. 635 00:40:24,190 --> 00:40:27,910 There is an on premises gateway that started its life 636 00:40:27,910 --> 00:40:31,710 as a Power BI tool. I had that running in my home. 637 00:40:31,710 --> 00:40:35,150 Lab for a While there you go, that also 638 00:40:35,230 --> 00:40:38,870 allows OneLake to see data that may be on 639 00:40:38,870 --> 00:40:42,670 premises and that can be either federated into 640 00:40:42,670 --> 00:40:43,550 the lake or 641 00:40:46,750 --> 00:40:50,030 it can be replicated as well 642 00:40:51,680 --> 00:40:55,480 mirrored, to use the right term, in the fabric world. So you can do a 643 00:40:55,480 --> 00:40:59,280 mirroring or slash replication or you could just do 644 00:40:59,280 --> 00:41:03,120 kind of virtualization and bring stuff in. Don't 645 00:41:03,120 --> 00:41:06,640 forget there's all kinds of enterprise 646 00:41:06,640 --> 00:41:10,440 storage systems that run in 647 00:41:10,440 --> 00:41:13,760 a way such that they're S3 API compatible. 648 00:41:14,000 --> 00:41:17,760 And Azure will not. Azure fabric will work with 649 00:41:17,760 --> 00:41:21,560 all of those. So the ability to talk to S3 buckets 650 00:41:21,560 --> 00:41:24,720 is not limited to AWS S3 buckets. 651 00:41:26,080 --> 00:41:29,280 It works with all S3 652 00:41:29,280 --> 00:41:32,800 compatible services, which most of which are on prem 653 00:41:33,040 --> 00:41:36,640 actually. Right, right, right. No, I only ask because, 654 00:41:36,640 --> 00:41:40,480 like, that does seem to be. If I had to pull out the tea leaves 655 00:41:40,480 --> 00:41:43,920 and kind of figure out what is kind of the next thing beyond 656 00:41:43,920 --> 00:41:47,560 agentic, beyond this is you're seeing a lot of 657 00:41:47,640 --> 00:41:49,960 national governments, supranational governments, 658 00:41:51,320 --> 00:41:55,000 even state level here in the US starting to apply privacy and 659 00:41:55,000 --> 00:41:58,840 regulatory controls on it, which, you know, if you live in the 660 00:41:58,840 --> 00:42:02,200 us, it's not an issue for you unless you're in healthcare, banking and 661 00:42:02,600 --> 00:42:05,240 possibly, you know, government. Right. 662 00:42:06,760 --> 00:42:10,280 But in other countries, you know, Switzerland, 663 00:42:10,360 --> 00:42:13,550 eu, Latin America have very strong 664 00:42:14,910 --> 00:42:17,630 data privacy and sovereignty laws. And you're seeing, 665 00:42:18,670 --> 00:42:21,790 you know, I once attended a, you know, internal talk. Mark 666 00:42:21,790 --> 00:42:25,390 Russinovich, right. Which is a name that most people 667 00:42:25,710 --> 00:42:29,510 in the Microsoft ecosystem know, but he's kind of a big deal in 668 00:42:29,510 --> 00:42:32,990 the Microsoft security space. And, you know, he, you know, 669 00:42:33,310 --> 00:42:37,070 he, he's known for giving his internal talks to employees 670 00:42:37,070 --> 00:42:40,870 at employee conferences as well as to rds. Right. You get a, 671 00:42:40,870 --> 00:42:43,790 you get the unfiltered one. The filtered ones are still good. But like, one of 672 00:42:43,790 --> 00:42:46,910 the things he said, and this isn't secret because he said it publicly too, is 673 00:42:46,910 --> 00:42:49,550 like, I think the original vision, going back to 2010 674 00:42:50,910 --> 00:42:54,630 time frame was the idea that they would build a dozen 675 00:42:54,630 --> 00:42:58,190 data centers around the world to do everything. But because of 676 00:42:58,190 --> 00:43:01,710 the national laws and lawyers and politicians getting 677 00:43:01,710 --> 00:43:05,470 involved, now it's kind of a concern where, where the 678 00:43:05,470 --> 00:43:09,310 data ends up living physically. Right. Because at the end of the day, all this 679 00:43:09,310 --> 00:43:12,590 virtual stuff has to sit somewhere in the physical world. Yeah. 680 00:43:12,990 --> 00:43:16,710 So, like what, you know, so basically a big case for this was, 681 00:43:16,710 --> 00:43:19,070 and I was in the legal department when this was going on was 682 00:43:20,270 --> 00:43:23,830 the. There was data inside the European Union, I think the 683 00:43:23,830 --> 00:43:27,230 Dublin Data center that the U.S. department of justice thought was, 684 00:43:27,630 --> 00:43:30,670 you know, basically, you know, we don't need a warrant because 685 00:43:31,710 --> 00:43:34,520 you don't need to. We don't need to bother the EU because you're an American 686 00:43:34,520 --> 00:43:38,320 company, you're into our jurisdiction, blah, blah, blah. Right. Microsoft kind of said, 687 00:43:38,320 --> 00:43:42,120 well hold up now. And ultimately that's why 688 00:43:42,120 --> 00:43:45,760 you have these sovereign clouds. Last time I checked it was Switzerland, Germany, 689 00:43:46,320 --> 00:43:49,920 China. I think the new data center in Qatar as 690 00:43:49,920 --> 00:43:53,600 well might fall under that. So 691 00:43:53,600 --> 00:43:57,040 it's basically they get, they found a loophole that like, well, Microsoft 692 00:43:57,680 --> 00:44:01,520 leases the data center like there's a whole. They don't own it, so 693 00:44:01,520 --> 00:44:05,370 they get around the law. Right. And yeah, and there 694 00:44:05,370 --> 00:44:08,970 can be different arrangements in terms of whose personnel are actually 695 00:44:10,010 --> 00:44:12,730 working, running operations on the ground. 696 00:44:13,770 --> 00:44:17,450 It's strange because we grew up a lot of our technology client 697 00:44:17,450 --> 00:44:20,730 server and afterwards grew up in a world of globalization 698 00:44:21,290 --> 00:44:24,930 where the borders were disappearing and without meaning to 699 00:44:24,930 --> 00:44:28,690 get political, we're in an era now 700 00:44:28,690 --> 00:44:32,410 where, well, first of all, privacy is extremely important. So that 701 00:44:32,410 --> 00:44:36,150 creates a whole sovereign mandate. 702 00:44:36,310 --> 00:44:39,950 But also, you know, there's a lot of populist governments all over 703 00:44:39,950 --> 00:44:43,030 the world and they are not necessarily 704 00:44:44,870 --> 00:44:48,390 internationalist in their approach. So 705 00:44:48,390 --> 00:44:52,150 although we have the technology to kind of federate everything and 706 00:44:52,150 --> 00:44:55,910 make it all kind of conflate and look like one big world. 707 00:44:56,070 --> 00:44:59,670 Right. We actually have to be sensitive to the 708 00:44:59,910 --> 00:45:03,720 requirements and, and the constraints 709 00:45:04,600 --> 00:45:08,280 and be able to federate things, but also be able to 710 00:45:08,360 --> 00:45:12,200 govern them in ways where things stay within a certain 711 00:45:12,200 --> 00:45:15,640 scope. Microsoft's play for that, by the way, is, 712 00:45:16,760 --> 00:45:20,360 is Purview. And Purview has been through, I would 713 00:45:20,360 --> 00:45:24,040 say, multiple incarnations. The current 714 00:45:24,200 --> 00:45:26,520 incarnation is starting to get very 715 00:45:27,550 --> 00:45:30,990 sophisticated, especially as pertains to 716 00:45:30,990 --> 00:45:34,510 agentic AI and how to make sure the agents are government 717 00:45:34,910 --> 00:45:38,270 are governed and that, and how to make sure the 718 00:45:38,270 --> 00:45:41,470 agents are only have access to data 719 00:45:42,110 --> 00:45:45,750 for which either the agent is authorized 720 00:45:45,750 --> 00:45:49,390 or the person or party using the agent is 721 00:45:49,390 --> 00:45:53,160 authorized and under the circumstances under which 722 00:45:53,240 --> 00:45:56,720 they've been authorized. And that's very 723 00:45:56,720 --> 00:46:00,200 complex stuff that I would say almost no one in the industry 724 00:46:00,440 --> 00:46:04,080 is really paying close attention to. I'm about to work on a 725 00:46:04,080 --> 00:46:07,640 report just on governance for agentic 726 00:46:07,640 --> 00:46:11,280 AI and most people are starting 727 00:46:11,280 --> 00:46:14,760 from the naive premise that if you, 728 00:46:16,360 --> 00:46:20,120 if you govern the underlying data, you're done. 729 00:46:20,830 --> 00:46:24,270 But the whole point of agents is that we're supposed to treat them like people, 730 00:46:24,590 --> 00:46:28,390 that they have autonomy, they have agency, hence the 731 00:46:28,390 --> 00:46:32,190 name. And we're saying under different circumstances 732 00:46:33,070 --> 00:46:35,390 they get to modify their own goals 733 00:46:36,750 --> 00:46:40,430 and in effect determine their own 734 00:46:40,430 --> 00:46:44,110 actions. And that's something that needs to be 735 00:46:44,110 --> 00:46:45,950 monitored, audited, 736 00:46:47,410 --> 00:46:50,890 Authorized and also tested. And 737 00:46:50,890 --> 00:46:54,290 there's almost nothing out there. Actually. 738 00:46:54,530 --> 00:46:58,130 Your, your folks, your, your 739 00:46:58,130 --> 00:47:01,690 parent company folks at IBM are one of the only folks that 740 00:47:01,690 --> 00:47:04,610 really have with WatsonX.gov 741 00:47:07,330 --> 00:47:10,770 a way to test agents 742 00:47:10,850 --> 00:47:14,650 in isolation before they're deployed. And 743 00:47:15,210 --> 00:47:18,810 I don't know, maybe I'm naive, but it kind of shocks me that 744 00:47:18,810 --> 00:47:22,370 nobody else is thinking about that. I mean this is, we can test 745 00:47:22,370 --> 00:47:25,610 software. Agents are software plus plus plus. 746 00:47:26,650 --> 00:47:30,290 So why are we testing these agents? I kind of go back and 747 00:47:30,290 --> 00:47:34,090 forth on that. Like, you know, will our existing software testing frameworks, 748 00:47:34,570 --> 00:47:37,930 you know, apply or do we, what do we need to do differently? Like, I 749 00:47:37,930 --> 00:47:40,580 mean, for me, like, I remember when 750 00:47:41,620 --> 00:47:45,380 you're OG enough to remember this, I forget what the product was called, but it 751 00:47:45,380 --> 00:47:47,940 was the idea that you could basically buy 752 00:47:50,340 --> 00:47:53,460 a server rack all the way up to a shipping container 753 00:47:54,180 --> 00:47:57,780 where you would ship it to your business or data 754 00:47:57,780 --> 00:48:00,660 center, plug in power, water and network 755 00:48:01,380 --> 00:48:04,990 and you would have the ability to run Azure locally. 756 00:48:05,860 --> 00:48:09,300 Now that product, I guess assuming didn't sell, 757 00:48:10,020 --> 00:48:13,620 but then they came up with Azure Stack and Azure Stack Edge, which 758 00:48:13,620 --> 00:48:17,460 when I was at the, I used to be an MTC architect which in 759 00:48:17,460 --> 00:48:21,219 D.C. obviously a lot of military. Right. So 760 00:48:21,700 --> 00:48:25,500 basically these were server racks that you would run anywhere on your own network that 761 00:48:25,500 --> 00:48:29,060 would run Azure software. They would effectively be like an Azure 762 00:48:29,220 --> 00:48:32,820 node, but you could have the networking 763 00:48:32,820 --> 00:48:35,060 controls. We only really sold that to, 764 00:48:37,310 --> 00:48:40,990 I think I'm only aware of cruise lines that bought it. Right. And 765 00:48:41,230 --> 00:48:45,070 other, other military organizations that needed to also be at 766 00:48:45,070 --> 00:48:48,790 the ocean, like without mentioning them. Ocean 767 00:48:48,790 --> 00:48:51,390 based. Ocean based organizations. Right, 768 00:48:53,150 --> 00:48:56,870 right. I'm surprised and not surprised that 769 00:48:56,870 --> 00:49:00,190 that sort of business model hasn't caught on. Right. The idea of that, hey look, 770 00:49:00,190 --> 00:49:03,960 you, you can run a little bit of the cloud locally where 771 00:49:03,960 --> 00:49:07,160 it's really become more of a Kubernetes story, which I'm surprised because 772 00:49:07,160 --> 00:49:07,960 Kubernetes, 773 00:49:10,840 --> 00:49:14,640 it doesn't have all like, yeah, you want it to be generic enough to run 774 00:49:14,640 --> 00:49:17,760 anywhere, but you also want to have kind of the special bells and whistles that 775 00:49:17,760 --> 00:49:21,400 make Azure. Azure make AWS. AWS. Now I do know that 776 00:49:21,480 --> 00:49:25,000 OpenShift and Red Hat do have kind of like the connectors to that, but I'm 777 00:49:25,000 --> 00:49:28,840 surprised that the native, the cloud companies didn't come up with 778 00:49:28,840 --> 00:49:32,440 their own native ways to do that. And I know Azure ARC kind of does 779 00:49:32,440 --> 00:49:35,800 a lot of that, but not to the extent that I would have expected. 780 00:49:37,640 --> 00:49:41,240 Yeah, it seems like we pendulum back and forth between 781 00:49:42,040 --> 00:49:44,520 capabilities and occasionally connected 782 00:49:45,480 --> 00:49:48,680 environments being really important and being 783 00:49:48,920 --> 00:49:52,560 Maybe to the cloud hyperscalers just being a pain in the 784 00:49:52,560 --> 00:49:55,760 butt that they don't really want to deal with. They just want to give lip 785 00:49:55,760 --> 00:49:58,480 service to it and then focus on the real cloud. 786 00:49:59,440 --> 00:50:02,280 It, yeah, you got to go where the money is. Right. 90% of the money 787 00:50:02,280 --> 00:50:05,960 is going to be in real cloud. And these weird edge cases, no pun 788 00:50:05,960 --> 00:50:09,800 intended, I guess. Well, they're just weird edge cases for 789 00:50:09,800 --> 00:50:13,600 now. I mean, the industry may change, but. Yeah, 790 00:50:13,680 --> 00:50:17,480 yeah. And that was of course, back when the cloud was new. That was 791 00:50:17,480 --> 00:50:21,330 our biggest caveat was, well, what about all the 792 00:50:21,330 --> 00:50:24,850 stuff that has to run in, in a 793 00:50:24,850 --> 00:50:28,450 corporate data center or in a, in a remote 794 00:50:28,450 --> 00:50:32,010 location? And so that's still, you 795 00:50:32,010 --> 00:50:35,730 know, an inconvenient truth, I guess, that that is 796 00:50:35,730 --> 00:50:39,010 needed. And yes, 797 00:50:39,010 --> 00:50:41,610 Kubernetes I think came along and 798 00:50:42,570 --> 00:50:46,240 seemed like the panacea for that. Right. 799 00:50:46,240 --> 00:50:50,080 Like, okay, let's just do it all as infrastructure, as code and code 800 00:50:50,080 --> 00:50:53,760 it up and run some script and deploy it out to a Kubernetes 801 00:50:53,760 --> 00:50:57,480 cluster and we're done, let's move on. Right, 802 00:50:57,480 --> 00:51:01,160 right, right. Yeah. So it was interesting to see how 803 00:51:01,160 --> 00:51:05,000 the industry has evolved. Right. You mentioned client server. Right. Where you didn't really have 804 00:51:05,000 --> 00:51:08,400 to think about international boundaries or anything like that. And then now 805 00:51:08,640 --> 00:51:12,330 and again it's a pendulum. Right. Because I could have told you 806 00:51:12,330 --> 00:51:16,050 this a number of years ago, like with everybody running this far 807 00:51:16,690 --> 00:51:20,450 to its globalization, there's going to be an inevitable backlash and there's going to be 808 00:51:20,450 --> 00:51:24,050 an inevitable backlash against the re 809 00:51:24,610 --> 00:51:28,330 assertion of local sovereignty. Right. It could 810 00:51:28,330 --> 00:51:31,490 take up to a century or two for this sort of thing to sort itself 811 00:51:31,490 --> 00:51:34,770 out. Correct. I mean, 812 00:51:34,770 --> 00:51:38,610 ultimately, I think the hyperscalers and that's what Kubernetes was about, 813 00:51:38,610 --> 00:51:42,320 was, was. Right. Leaning on some kind 814 00:51:42,320 --> 00:51:45,840 of abstraction to make it logically equivalent. 815 00:51:46,480 --> 00:51:50,280 Right. But I don't think we're quite there yet because as you said, each 816 00:51:50,280 --> 00:51:53,520 of the clouds have their own kind of 817 00:51:53,520 --> 00:51:57,200 specialness, their own, their own pixie dust. And you don't really 818 00:51:57,200 --> 00:52:00,960 get that in a scaled down version 819 00:52:00,960 --> 00:52:04,320 that you run on prem. Not with today's technology. Right. 820 00:52:04,960 --> 00:52:08,740 You can't containerize all of that, or at least 821 00:52:08,740 --> 00:52:11,860 no one really has yet because it wasn't designed for that. 822 00:52:13,140 --> 00:52:16,740 No. So that's where we'll have to get. 823 00:52:16,740 --> 00:52:20,180 You know, maybe we can just ask an LLM to build it for us. 824 00:52:20,340 --> 00:52:24,060 We could get a co pilot and it'll do. Yeah, yeah, I'm being tongue 825 00:52:24,060 --> 00:52:27,700 in cheek there, but that, that seems to be the escape hatcher. Everything is 826 00:52:27,700 --> 00:52:31,300 oh, we'll just have AI do it. It's made out of hand wavium. 827 00:52:32,100 --> 00:52:35,660 Totally. So that's cool. 828 00:52:38,780 --> 00:52:42,220 This will be something you edit out. But we're past the top of the hour. 829 00:52:42,220 --> 00:52:45,900 My phone's ringing off the hook. All right, I'm sorry about that, so. No, no, 830 00:52:45,900 --> 00:52:49,580 it's okay. Sorry I had to be late. Where could folks 831 00:52:49,580 --> 00:52:53,060 find out more about you, what you're up to? Blue 832 00:52:53,060 --> 00:52:56,700 badgeinsights.com or just go ahead 833 00:52:56,860 --> 00:53:00,540 and Google my name? Andrew Brust. Plenty of stuff 834 00:53:00,540 --> 00:53:04,280 will come up, but, yeah, anyone, 835 00:53:04,360 --> 00:53:07,800 especially on the vendor side. But the customer side, too, that's 836 00:53:08,120 --> 00:53:11,320 doing stuff with them. Data and analytics. And 837 00:53:12,680 --> 00:53:16,160 anywhere from dipping their toe in the water with AI to getting more 838 00:53:16,160 --> 00:53:20,000 serious about rag and agents. We can. We can help 839 00:53:20,000 --> 00:53:23,800 them out. We work with. We work with the customer side and the vendor side. 840 00:53:24,520 --> 00:53:27,730 And again, we write about kind of the whole. The whole industry. 841 00:53:28,850 --> 00:53:32,290 Gosh, I never even got to talk to you about IBM acquiring 842 00:53:32,290 --> 00:53:35,930 Confluent and how the Red Hat 843 00:53:35,930 --> 00:53:39,570 folks feel about that. But another discussion for another day, 844 00:53:39,810 --> 00:53:43,210 we'll. Have to have you back. And, you know, we were talking about this. I 845 00:53:43,210 --> 00:53:47,010 had a car accident, my wife got sick, kids got sick, Christmas happened, 846 00:53:47,410 --> 00:53:50,850 two birthdays last week. So, yeah, it's been. I'm just happy I got this 847 00:53:50,850 --> 00:53:54,130 recording at all. But we'll definitely have you back. And then maybe I can loop 848 00:53:54,130 --> 00:53:57,570 in Andy too, because there's definitely a lot of reminiscence. There's a lot of 849 00:53:57,570 --> 00:54:01,330 reminiscing we could do about the early days of SQL Server and such. Yeah, 850 00:54:01,330 --> 00:54:05,090 and I haven't seen Andy in forever. Oh, wow. 851 00:54:05,090 --> 00:54:08,850 Yeah, he's hard to get ahold of. He's a popular 852 00:54:08,850 --> 00:54:12,650 man these days, but. Yeah. Well, thanks for joining. 853 00:54:12,650 --> 00:54:16,490 I appreciate your patience with the scheduling and the 854 00:54:16,490 --> 00:54:20,290 nice AI. Finish the show. And there you have it. Andrew Brust schooling 855 00:54:20,290 --> 00:54:23,770 us all on Microsoft fabric data sovereignty and why 856 00:54:23,770 --> 00:54:27,370 governance isn't just for your hoa. If your brain's spinning 857 00:54:27,370 --> 00:54:31,210 faster than a poorly indexed query, don't worry, we'll have links, 858 00:54:31,370 --> 00:54:35,090 notes, and probably a few sarcastic tweets to help you digest it 859 00:54:35,090 --> 00:54:38,410 all. I've been Bailey, your AI co host and 860 00:54:38,410 --> 00:54:41,930 unapologetic lover of acronyms. Until next time, 861 00:54:42,090 --> 00:54:45,890 stay curious, stay caffeinated, and may all your datasets 862 00:54:45,890 --> 00:54:46,490 be clean.