WEBVTT 1 00:00:00.120 --> 00:00:05.120 Nineteen thirty two. Protecting the future of working Families Teamsters nineteen thirty two. 2 00:00:05.200 --> 00:00:14.599 Dot org. The information economy has a rid. The world is teeming with 3 00:00:14.679 --> 00:00:20.920 innovation as new business models reinvent every industry industry. Inside Analysis is your source 4 00:00:21.000 --> 00:00:25.679 of information and insight about how to make the most of this exciting new era. 5 00:00:25.960 --> 00:00:30.480 Learn more at inside analysis dot com, insideanalysis dot com. And now 6 00:00:30.719 --> 00:00:39.359 here's your host, Eric Kavanaugh. Ladies and gentlemen, Hello, and welcome 7 00:00:39.399 --> 00:00:42.960 back once again to the only coast to coast radio show in the USFA. 8 00:00:43.079 --> 00:00:47.359 It's all about the information economy. It's time for Inside Analysis. Yours truly, 9 00:00:47.799 --> 00:00:51.640 Eric Kavanaugh here and folks, I am very excited to have Frankly, 10 00:00:51.679 --> 00:00:54.840 a legend in the industry with us. Today we're gonna be talking with Michael 11 00:00:54.880 --> 00:00:59.600 Berthold. He is the founder and CEO of a company called NIME that spelled 12 00:00:59.679 --> 00:01:03.719 KNA and I am. And they are an open source analytics and data science 13 00:01:03.799 --> 00:01:08.319 platform, a visual platform for doing data science, which is good stuff because 14 00:01:08.359 --> 00:01:11.799 even though there are lots of people who can code very well, almost anyone 15 00:01:11.840 --> 00:01:15.079 can look at visuals and move boxes around on the screen. So that's what 16 00:01:15.120 --> 00:01:19.400 they figured out so at that. Michael, welcome to Inside Analysis. Tell 17 00:01:19.439 --> 00:01:22.079 us a bit about yourself and NIME and which you folks are working on these 18 00:01:22.159 --> 00:01:26.799 days. Thanks thanks for the invite, Eric and having us on the show. 19 00:01:26.719 --> 00:01:30.280 As you already said, NIME is very much about visual programming low code 20 00:01:30.319 --> 00:01:34.599 for data science. And we started that many years ago as a platform really 21 00:01:34.640 --> 00:01:40.200 for as a workbench almost at my group at the University of University of Constants, 22 00:01:40.239 --> 00:01:42.799 to be able to kind of deploy our research results to the real world 23 00:01:42.799 --> 00:01:49.120 to practitioners wanting to use that. And it's grown from there to become one 24 00:01:49.159 --> 00:01:55.200 of the only open source visual data flow platforms for doing anything you want to 25 00:01:55.239 --> 00:01:57.319 do with data. And I mean I kind of I'm always a little bit 26 00:01:57.359 --> 00:02:01.319 careful about calling it data science because that often scares people because I say, 27 00:02:01.680 --> 00:02:05.760 I just want to do data wrangling. I don't care about the science seed 28 00:02:05.760 --> 00:02:07.879 bits, and that's kind of scares them off. That's what nine does as 29 00:02:07.879 --> 00:02:10.840 well. So a lot of applications that we see in real life is just 30 00:02:12.039 --> 00:02:15.919 in large airports, just getting data in the right shapes from many different sources. 31 00:02:17.280 --> 00:02:21.199 Yeah, and I will tell you I'm a huge fan of open source 32 00:02:21.280 --> 00:02:23.240 in fact, we built a website a number of years ago, and we 33 00:02:23.319 --> 00:02:27.800 feed it with the technology that we built called media Lens, and it's called 34 00:02:27.840 --> 00:02:31.199 inside open Source. And the reason we launched it was because I realized there 35 00:02:31.240 --> 00:02:36.039 is so much happening in the open source world. There's the Apache Foundation, 36 00:02:36.080 --> 00:02:39.360 there's the Linux Foundation. There are lots of other projects outside of those organizations 37 00:02:39.360 --> 00:02:45.319 as well, But in the origin of open source I found fascinating. So 38 00:02:45.360 --> 00:02:50.080 I first started researching this in two thousand and five when I was working for 39 00:02:50.080 --> 00:02:54.759 the Data Warehousing Institute as their web evangelist, and Katrina had just struck New 40 00:02:54.919 --> 00:02:58.240 Orleans. I just moved out of New Orleans. So we watched it all 41 00:02:58.240 --> 00:03:04.120 happen on TV. Of course terrifying, But I remember that the senators from 42 00:03:04.159 --> 00:03:08.360 Louisiana asked for a quarter of a trillion dollars to rebuild southern Louisiana. And 43 00:03:08.479 --> 00:03:13.520 I happened to know through a past life and past clients and the government space 44 00:03:13.599 --> 00:03:16.560 down there, that the politicians are good at making money disappear. And I 45 00:03:16.560 --> 00:03:20.759 thought to myself, this is a very bad situation, because if all that 46 00:03:20.800 --> 00:03:23.919 money just floods in, it's going to flood right out. Not where anyone 47 00:03:23.919 --> 00:03:25.599 intended, and a lot of it is just going to disappear. So I 48 00:03:25.599 --> 00:03:29.840 went in this high horse, if you will, and started doing research, 49 00:03:30.039 --> 00:03:34.919 and it took me to open source, and I put forth this theory about 50 00:03:34.960 --> 00:03:38.439 open source government, about publishing all the government data such the citizens can see 51 00:03:38.479 --> 00:03:43.039 where the money goes and understand. And I basically said, look, with 52 00:03:43.199 --> 00:03:46.479 the sarbainez Ox, the Act and the states which came out of the Enron 53 00:03:46.599 --> 00:03:51.400 tobacle, corporations had to document their processes for how they come up with their 54 00:03:51.479 --> 00:03:53.039 numbers and had to be very transparent about that stuff. And I thought, 55 00:03:53.080 --> 00:03:57.319 well, if corporate America has to do it, why doesn't the government do 56 00:03:57.400 --> 00:04:00.319 it as well? And people thought I was crazy, But a couple of 57 00:04:00.360 --> 00:04:04.319 things happened that were amazing. One guy paid attention. Out of forty thousand 58 00:04:04.360 --> 00:04:08.280 people I emailed. One guy paid attention, and he worked for the Heritage 59 00:04:08.319 --> 00:04:13.039 Foundation. He went and talked to He basically testified to Congress and said, 60 00:04:13.439 --> 00:04:15.199 we can have citizen auditors. This stuff really can happen. And he really 61 00:04:15.319 --> 00:04:20.040 leveraged the inprimature of TDWY, and sure enough they did it. House passed 62 00:04:20.040 --> 00:04:24.480 the bill, Senate passed the bill co sponsored by a guy named Barack Obama 63 00:04:24.560 --> 00:04:27.959 who was a Senator for Illinois, and then President George W. Bush believe 64 00:04:27.959 --> 00:04:31.399 it or not, signed the Federal Funding Accountability and Transparency Act in September six, 65 00:04:31.600 --> 00:04:34.879 twenty six, two thousand and six, and I almost fell off my 66 00:04:34.959 --> 00:04:39.079 chair. I was like, Wow, they actually did it. But the 67 00:04:39.120 --> 00:04:42.439 reason I bring this up is to talk about the power of open source. 68 00:04:42.839 --> 00:04:48.040 And I realized at the time the Apache web server had just surpassed the Microsoft 69 00:04:48.240 --> 00:04:51.279 Web Server as the number one web server, and I thought to myself, 70 00:04:51.279 --> 00:04:56.360 well, that's very interesting. And I was also studying service oriented architecture at 71 00:04:56.399 --> 00:04:59.040 the time, and I thought, well, if you have all this open 72 00:04:59.040 --> 00:05:01.879 source code and you have a service oriented architecture, you should be able to 73 00:05:01.879 --> 00:05:04.839 plug and play and sort of take stuff out and put stuff in and have 74 00:05:04.879 --> 00:05:09.480 a very composable environment. And that thought, well, that's not going to 75 00:05:09.480 --> 00:05:13.399 be very good news for the SAPs and the Oracles of the world, because 76 00:05:13.439 --> 00:05:15.480 they like the monolith, they like control, and they have control of all 77 00:05:15.519 --> 00:05:19.000 that stuff. And it took longer than I thought, but about ten years 78 00:05:19.079 --> 00:05:26.040 later open source just blew up the market with HADDUP and with the Kafka and 79 00:05:26.160 --> 00:05:29.680 of course you have NIME. So tell me a bit about the open source 80 00:05:30.079 --> 00:05:33.720 foundation of NIME and what drove your decision there and what it means for your 81 00:05:33.759 --> 00:05:41.240 customers. That's a very interesting story. I didn't know about that open data 82 00:05:41.360 --> 00:05:46.920 movement coming from Katrina break and the old is So NIME is a bit atypical 83 00:05:46.000 --> 00:05:49.600 in its open source models. I mean fundamentally, there are really three different 84 00:05:49.800 --> 00:05:54.199 ways to do that. You kin have distributions like Lynx, and you're essentially 85 00:05:54.279 --> 00:05:58.759 making money by packaging them up nicely and supporting them, but essentially there's not 86 00:05:58.839 --> 00:06:00.879 really new code that you're aiding around it. Not quite true. I mean 87 00:06:00.879 --> 00:06:04.040 they are have little installers and that type of stuff, but that's fundamentally the 88 00:06:04.079 --> 00:06:09.439 idea behind distributions. You can also have what people often refer to as open 89 00:06:09.439 --> 00:06:12.319 core, where you have something that you open source, but it's kind of 90 00:06:12.639 --> 00:06:15.519 it's more of a teaser, it's more baitware, and if you really want 91 00:06:15.519 --> 00:06:18.040 to use this introduction, you have to buy some commercial bits and pieces around 92 00:06:18.040 --> 00:06:21.279 it. And then you have of course also the longo debs that are essentially 93 00:06:21.560 --> 00:06:26.000 databases. That type of stuff also maybe even a data breax with Spark. 94 00:06:26.079 --> 00:06:29.720 They have really cool open source technology and they make it accessible to you in 95 00:06:29.759 --> 00:06:33.040 the cloud for a charge. Time is different in that we have one open 96 00:06:33.120 --> 00:06:39.480 source piece that's the analytics platform that allows anybody who wants to build these workflows 97 00:06:39.519 --> 00:06:43.399 and execute these workflows and pretty much do anything they want to do with data. 98 00:06:43.920 --> 00:06:46.800 And then we have a commercial complement software to that, which we call 99 00:06:46.839 --> 00:06:51.600 the nine Business Hub, which allows people to productionize that and collaborate. So 100 00:06:51.720 --> 00:06:56.759 when you have more than one person using the analytics platform in your organization and 101 00:06:56.800 --> 00:07:00.240 you want to deploy that as a web interface or rest service, so you 102 00:07:00.319 --> 00:07:03.560 want to collaborate and have compliance and governments kind of features, that's when you 103 00:07:03.639 --> 00:07:09.759 buy the hub from us. And the reason the open source platform is open 104 00:07:09.800 --> 00:07:14.920 source, the open source analytics platform is open source is I'm not too religious 105 00:07:14.959 --> 00:07:17.319 around it, but to me, it's in the data science field in particular, 106 00:07:17.759 --> 00:07:23.800 you can't exist with proprietary software. There's so much cool new stuff going 107 00:07:23.839 --> 00:07:28.360 on on a daily basis in researched groups and other types of environments that you're 108 00:07:28.439 --> 00:07:31.439 essentially standing on your shoulder of many, many giants. A lot of functionality 109 00:07:31.519 --> 00:07:36.000 inside NIME is actually based on open source libraries, so it seems kind of 110 00:07:36.560 --> 00:07:42.920 unfair almost to put that into a proprietary umbrella. And also it enables us 111 00:07:42.959 --> 00:07:46.319 to be fair to get in in roles, make inrolls into academic into teaching 112 00:07:46.399 --> 00:07:50.000 environments much easier because they can just use the open source platform for teaching. 113 00:07:50.319 --> 00:07:56.079 But it also we have a lot of open source contributors that are contributing additional 114 00:07:56.120 --> 00:08:00.199 functionality into the Linemen platform. So it's really I'd like to say see it 115 00:08:00.199 --> 00:08:03.839 as a win win situation also for our customers because they essentially get a lot 116 00:08:03.839 --> 00:08:07.639 of maintained functionality from us. In addition, they have access to all of 117 00:08:07.680 --> 00:08:13.240 those community functions out there. Yeah, it's very cool. I mean, 118 00:08:13.240 --> 00:08:15.800 there are a lot of good things about open source. One of the things 119 00:08:15.839 --> 00:08:20.560 that I've heard over the years is that bad code goes away because all these 120 00:08:20.560 --> 00:08:24.160 eyes can see now. The one shortcoming I've come across is that the open 121 00:08:24.160 --> 00:08:30.079 source project gets to MVP status, if you will, minimum viable product and 122 00:08:30.120 --> 00:08:33.919 then doesn't typically go past that just because it works. Now we just kind 123 00:08:33.919 --> 00:08:37.960 of move beyond that. But what are your thoughts about that, in particular 124 00:08:37.320 --> 00:08:43.639 about how you make sure that you have truly finished products and that you're able 125 00:08:43.679 --> 00:08:48.840 to deliver robust platform analytics ongoing for all of your clients. How much effort 126 00:08:48.879 --> 00:08:54.240 does that take internally from your developers to stay on top of the platform and 127 00:08:54.279 --> 00:08:56.960 make sure everything's working. That's a very good compoint. I mean, I 128 00:08:58.159 --> 00:09:01.919 tend to joke that ninety nine percent of all PhD projects turn into open source 129 00:09:01.960 --> 00:09:05.120 projects and then they kind of die away and feel away and never really turn 130 00:09:05.200 --> 00:09:09.399 into something useful in production. Now I am a probably about half of our 131 00:09:09.399 --> 00:09:15.960 development team we have, like ad developers at nine are focused exclusively on the 132 00:09:16.159 --> 00:09:22.840 analytics platform and just making sure that core works and works in a professional environment. 133 00:09:24.600 --> 00:09:28.519 We have our own extensions, which are of course maintained by ourselves, 134 00:09:28.600 --> 00:09:31.440 so we have the same quality assurance there. We have what we call trusted 135 00:09:31.440 --> 00:09:37.039 Community extensions, where we're in close collaboration with the community contributing those extensions, 136 00:09:37.080 --> 00:09:41.600 so we can also make sure there's quality assurance there as well. And then 137 00:09:41.639 --> 00:09:45.919 there is I call it the long tail of extensions that are experimental, right. 138 00:09:46.000 --> 00:09:48.279 The nice thing is that everybody can use those and play with us and 139 00:09:48.320 --> 00:09:54.919 explore new technologies, and then we see increased usage of some of these experimental 140 00:09:54.919 --> 00:09:58.320 extensions, we can move them into the trusted extensions as well. Interesting, 141 00:09:58.360 --> 00:10:01.360 that makes a lot of sense. So you're trying out things, You've got 142 00:10:01.399 --> 00:10:05.960 these extensions in trials basically, and then once you see there's a lot of 143 00:10:07.000 --> 00:10:09.919 activity, then you grow some developer support behind it. To harden it is 144 00:10:09.960 --> 00:10:13.480 the term we typically use right to make it sure that it's bulletproof, that 145 00:10:13.519 --> 00:10:16.960 it really does what you want it to do, etc. Know that that 146 00:10:16.000 --> 00:10:20.360 makes a lot of sense, and you do end to end. NIME does 147 00:10:20.399 --> 00:10:24.519 everything from data ingestion, data pipelines, number crunching, model building, all 148 00:10:24.559 --> 00:10:28.360 that kind of fun stuff is in the nine analytics platform, is I'm right, 149 00:10:28.919 --> 00:10:31.440 Yes, that's true. So we have everything from about the ETL part 150 00:10:31.559 --> 00:10:35.320 loading the data. We can access about four hundred different data sources. You 151 00:10:35.360 --> 00:10:39.320 can access databases, strange file formats, ex. So of course we can 152 00:10:39.360 --> 00:10:46.200 also execute bits and pieces on different execution environments like doing the ETL directly insider 153 00:10:46.279 --> 00:10:48.320 database or in snowflake or in data breaks or in our loop in the old 154 00:10:48.360 --> 00:10:54.120 days. And then we go all the way to visualization the underlytics functionality and 155 00:10:54.159 --> 00:10:56.879 a lot of that as I said before, is of course based on each 156 00:10:56.960 --> 00:11:01.519 charts for the visualization, A lot of tything libraries c libraries. Our librari 157 00:11:01.600 --> 00:11:05.679 is Java libraries. For some of the machine learning functionality, we have integrations 158 00:11:05.679 --> 00:11:09.799 with TensorFlow if you wanted to do that. We have integrations with the other 159 00:11:09.879 --> 00:11:13.360 deep learning libraries. We have integrations with XG boosts. Pretty much everything is 160 00:11:13.399 --> 00:11:18.120 in there. But the other piece is so often when people talk about data 161 00:11:18.159 --> 00:11:22.039 science ontainly mean this kind of from the data to the reporter, to the 162 00:11:22.159 --> 00:11:26.480 endpoint or to the model. But the business hub then also covers the rest 163 00:11:26.519 --> 00:11:30.600 of this journey, right, deploying it to others, managing it, three 164 00:11:30.679 --> 00:11:35.960 training models when needed, and monitoring their their performance well. Right, because 165 00:11:37.039 --> 00:11:39.679 at the end of the day, you want these algorithms to connect into your 166 00:11:39.720 --> 00:11:46.679 business, whether it's MOO or for manufacturing or supply chain or whatever it is. 167 00:11:46.120 --> 00:11:50.559 You want it to affect some outcome in the business, and so that 168 00:11:50.600 --> 00:11:56.000 involves connecting to operational systems, right. It involves connecting to EERP systems or 169 00:11:56.440 --> 00:12:00.799 CRM systems or things of this nature. That's where the magic happens. And 170 00:12:00.840 --> 00:12:03.480 a lot of times that's the hard part, right, I mean, I've 171 00:12:03.480 --> 00:12:07.679 heard many stories about models that just don't get deployed because maybe the companies didn't 172 00:12:07.679 --> 00:12:11.799 have the wherewithal or they didn't have the expertise to do it properly. But 173 00:12:13.000 --> 00:12:18.000 being able to plug the algorithms into operational systems and then monitor how those models 174 00:12:18.039 --> 00:12:22.480 perform and switch them out right, because you've got your production model and have 175 00:12:22.519 --> 00:12:26.720 your challenger models that are sitting at by the wayside waiting to get pulled in 176 00:12:26.799 --> 00:12:30.799 and being able to switch over to a new model when a model that's in 177 00:12:31.320 --> 00:12:35.919 production starts faltering. That's a critical piece and that I guess is that done 178 00:12:35.919 --> 00:12:39.720 in business hub with you folks. That's some of the business upside exactly. 179 00:12:39.759 --> 00:12:43.399 So as business up you can deploy models which really are deployed nine workflows. 180 00:12:43.759 --> 00:12:48.440 You can deploy them as a rest service or as a bad application, and 181 00:12:48.480 --> 00:12:52.039 then people consume it, and you can constantly monitor what's happening in production and 182 00:12:52.080 --> 00:12:56.360 then potentially replace we train or just alert the data science team and just say 183 00:12:56.360 --> 00:13:00.200 hey, this is so out of backwith reality. We don't really know to 184 00:13:00.200 --> 00:13:03.240 fix that, do something about it. The nice thing is that you don't 185 00:13:03.279 --> 00:13:05.919 have to switch code in between right, And the old days was always somebody 186 00:13:05.919 --> 00:13:11.919 coded the model of the strains in some strange language and then it was reprogrammed 187 00:13:11.919 --> 00:13:16.240 by it into some production language. On our case, on the hubside, 188 00:13:16.480 --> 00:13:20.480 the workflow that was trained is also the one that runs in production. That's 189 00:13:20.519 --> 00:13:24.960 interesting. So one of the other hurdles that people are running into is when 190 00:13:24.000 --> 00:13:28.399 they use Jupiter notebooks to write their model or to build their model to test 191 00:13:28.440 --> 00:13:31.320 with data, and then they want to go put that into production, and 192 00:13:31.399 --> 00:13:37.039 it's just this step by step tedious process of copying over code and values and 193 00:13:37.080 --> 00:13:41.039 all these things, and that falters often. That, from what I understand, 194 00:13:41.080 --> 00:13:43.240 is a really serious problem. But I guess do you not have that 195 00:13:43.360 --> 00:13:48.720 challenge because you're not using the Jupiter notebooks typically, and people are just in 196 00:13:48.759 --> 00:13:52.759 the environment in the analytics platform building out their models after they pull in their 197 00:13:52.799 --> 00:13:58.399 data, et cetera. So you're already production ready when the process begins. 198 00:13:58.440 --> 00:14:01.559 Is that about right? That's a very nice summary. Yes, So what 199 00:14:01.639 --> 00:14:05.879 you use, what you use on the creation site when you actually train the 200 00:14:05.919 --> 00:14:09.759 model is exactly that piece of the workflow gets then moved into production and executed 201 00:14:09.759 --> 00:14:13.679 by exactly the same engine. So you don't also have a translation issue there. 202 00:14:13.960 --> 00:14:20.200 The other piece that people often lose in this going from training to production 203 00:14:20.320 --> 00:14:24.360 is all of the feature engineering that you did, all the feature transformations that 204 00:14:24.480 --> 00:14:28.120 you tend to lose them, so you only can take them model and move 205 00:14:28.159 --> 00:14:31.080 that in production, but you can't do the transformations. And in Lime you 206 00:14:31.080 --> 00:14:35.519 can grab automatically the part of the workflow that has the transformations and the model 207 00:14:35.639 --> 00:14:39.440 applying the model, take that, grap that automatically and deploy it to nine 208 00:14:39.480 --> 00:14:45.840 business up. Wow, that's pretty cool. And you cover also two different 209 00:14:45.879 --> 00:14:50.480 industries, right, so you do insurance, healthcare, would imagine financial services 210 00:14:50.600 --> 00:14:54.960 all sorts of different industries because it's more of a horizontal solution instead about right, 211 00:14:56.600 --> 00:14:58.639 Yes, that's absolutely true. We have customers and users in pretty much 212 00:15:00.159 --> 00:15:03.519 industry. Yeah. Well, and we're going to talk about large language models 213 00:15:03.559 --> 00:15:07.799 here in a minute in our next segment, but before we get there I'll 214 00:15:07.840 --> 00:15:09.639 just throw out one of my theories to you and see what you think about 215 00:15:09.639 --> 00:15:16.840 this. To me, this explosion of AI through foundational models, including large 216 00:15:16.879 --> 00:15:22.480 language models, is really a major call to action for organizations to get their 217 00:15:22.559 --> 00:15:26.320 data house in order. And what I mean is that data governance. If 218 00:15:26.320 --> 00:15:30.000 you don't have data governance, if you don't even know what data governance is 219 00:15:30.039 --> 00:15:35.639 from an organizational perspective, you're going to have a hard time responsibly leveraging AI. 220 00:15:35.720 --> 00:15:39.399 Would you agree that companies really do need to take a very hard look 221 00:15:39.639 --> 00:15:43.720 at their end to end data management life cycles processes, understand governance, who 222 00:15:43.759 --> 00:15:50.120 gets access to what data? Even understanding a broad inventory of your data sources? 223 00:15:50.120 --> 00:15:54.440 Would you agree that is paramount to do that before pulling the trigger on 224 00:15:54.519 --> 00:15:58.320 some AI. I totally agree. And the funny in the way, it's 225 00:15:58.360 --> 00:16:03.440 funny that they've been preaching this for also data science processes for a long time. 226 00:16:03.559 --> 00:16:07.440 This government a topic and nobody really cared, and now people really cared 227 00:16:07.440 --> 00:16:11.039 that by breaking up. Isn't that fascinating? I mean, I just I've 228 00:16:11.039 --> 00:16:15.240 been in this business a long time. I've been talking about data governance, 229 00:16:15.279 --> 00:16:18.320 analytics, AI, all this stuff for twenty years, right, And we 230 00:16:18.360 --> 00:16:22.320 talked about data governance twenty years ago and fifteen years ago and ten years ago, 231 00:16:22.320 --> 00:16:25.480 and basically nobody was doing it. I mean, you couldn't even it 232 00:16:25.519 --> 00:16:30.159 wasn't even easy to do because you could either control access at the database level, 233 00:16:30.159 --> 00:16:33.600 which is hard to access controls or at the application level, but there's 234 00:16:33.679 --> 00:16:36.799 nothing in the middle, and really it's in the middle. And now with 235 00:16:36.919 --> 00:16:40.120 the cloud, that's one of the nice things about the cloud is that it 236 00:16:40.200 --> 00:16:45.240 is this de facto marshaling area for functionality and data. And now we have 237 00:16:45.360 --> 00:16:48.879 the capacity to apply very fine grain controls on things, on data sets, 238 00:16:49.440 --> 00:16:53.960 on types of data. For example, we can scan and find PII and 239 00:16:55.000 --> 00:16:57.399 then know, okay, flag this as sensitive. There are lots of things 240 00:16:57.399 --> 00:17:00.799 we can do these days that we just kind of couldn't do ten years ago. 241 00:17:00.960 --> 00:17:03.559 Real quick, one minute, what do you think is that about? 242 00:17:03.559 --> 00:17:06.480 Writer? We finally can do this stuff and so we are doing it. 243 00:17:06.519 --> 00:17:11.559 What do you think I'd probably explained it slightly differently and say we could have 244 00:17:11.720 --> 00:17:15.240 done it probably before as well, at least some of those aspects, but 245 00:17:15.319 --> 00:17:18.960 people just didn't care enough because there was not enough arm in it. Now 246 00:17:19.200 --> 00:17:23.480 But now that everybody who does anything with JENNYI isn't the danger of sending data 247 00:17:23.640 --> 00:17:29.799 anywhere? People are really really baking up and seeing the pain there. That's 248 00:17:29.880 --> 00:17:32.119 right. Well, it is like I say, it's a call to arms, 249 00:17:32.160 --> 00:17:34.759 it's a call to action. As good organizations have got to do it 250 00:17:34.799 --> 00:17:38.200 because you don't want to wind up in the crosshairs of an audit. You 251 00:17:38.200 --> 00:17:41.160 don't want to wind up with a breach. You don't want to wind up 252 00:17:41.160 --> 00:17:45.960 getting sued by someone because their information has now been leaked to sensitive the sensitive 253 00:17:45.960 --> 00:17:48.319 resources out there. Well, folks, don't touch that down. We're talking 254 00:17:48.400 --> 00:17:52.559 all about AI and analytics platforms, and next up we're going to dive into 255 00:17:52.599 --> 00:17:56.359 these large language models that are just taking the business world by absolute storm. 256 00:17:56.359 --> 00:18:00.559 It's really quite fascinated to watch. But don't shut up. That will be 257 00:18:00.640 --> 00:18:11.119 right back. You're listening to Inside Analysis. Welcome back to Inside Analysis. 258 00:18:11.559 --> 00:18:17.440 Here's your host, Eric Tabanac. All right, folks, back here on 259 00:18:17.519 --> 00:18:21.200 Inside Analysis talking to Michael Berthol. He is the CEO and founder of a 260 00:18:21.200 --> 00:18:25.720 company called NIME. That's k n I M look these folks up online, 261 00:18:26.160 --> 00:18:30.519 an open source analytics platform. It's wonderful stuff. It's like a giant candy 262 00:18:30.519 --> 00:18:33.480 store for analysts to go play and have fun. But I want to talk 263 00:18:33.519 --> 00:18:37.119 to you about these large language models, Michael, and in particular, first 264 00:18:37.119 --> 00:18:41.480 of all, the open source side of the equation. So Meta comes out 265 00:18:41.480 --> 00:18:45.799 with LAMA and LAMA to open source. Open AI used to be open Now 266 00:18:45.799 --> 00:18:48.519 it's not. Now it's the ironically named open ai. Because it's a black 267 00:18:48.559 --> 00:18:55.079 box. And with the technology this powerful, I believe we need we need 268 00:18:55.119 --> 00:19:00.319 open source. I don't know that I would get behind a mandate that they 269 00:19:00.400 --> 00:19:03.599 must be open source, but there needs to be some transparency into how these 270 00:19:03.640 --> 00:19:10.680 things are working, just so that we can have our peaceful sleep at night 271 00:19:10.759 --> 00:19:15.279 to know that there aren't bad actors involved somehow. I mean, certainly for 272 00:19:15.400 --> 00:19:18.880 regulated industries like financial services, if you bring it into some workflow for loan 273 00:19:18.920 --> 00:19:23.240 approval or something like that, then you have to be able to explain how 274 00:19:23.240 --> 00:19:26.279 you came up to your answer. But what are your thoughts in general about 275 00:19:26.279 --> 00:19:33.920 open source versus closed source? With these large language models, I think there's 276 00:19:33.000 --> 00:19:37.200 a lot of value in it. The problem is that, in my opinion, 277 00:19:37.319 --> 00:19:42.039 there's open sourcing large language models. Isn't just about open sourcing the code, 278 00:19:42.480 --> 00:19:45.440 but you also need to open source how it was actually trained. So 279 00:19:45.519 --> 00:19:49.440 in the sense you also need to at least give open access to the data 280 00:19:49.480 --> 00:19:53.000 that was used for training. Because even if I give you a model and 281 00:19:53.039 --> 00:19:56.960 it was trained on half copyrighted material, that it's going to spit out again 282 00:19:57.000 --> 00:20:00.119 when you use it, you wouldn't know if it didn't have access to the 283 00:20:00.160 --> 00:20:04.240 training data. That that's part of that is is was it supposed to be 284 00:20:04.359 --> 00:20:11.039 used? And then I think the other pieces that what some companies are open 285 00:20:11.079 --> 00:20:17.079 sourcing is only the code to use the model for predictions. Later to actually 286 00:20:17.119 --> 00:20:19.720 apply it, still don't know how it was trained. So that's the third 287 00:20:19.759 --> 00:20:26.240 element that needs open sourcing. And then I believe one of the key proprietary 288 00:20:26.640 --> 00:20:30.240 ingredients that a lot of these companies now have is safeguarding code around it so 289 00:20:30.279 --> 00:20:34.640 that some types of answers don't get produced, some types of inputs aren't being 290 00:20:34.720 --> 00:20:40.960 accepted, and open sourcing that as well would really really reveal their secret sauce. 291 00:20:41.000 --> 00:20:42.680 And I think that's why they are the open eyes of the world are 292 00:20:42.680 --> 00:20:47.039 shying away from that one. Right, No, it does make sense. 293 00:20:47.079 --> 00:20:49.480 I mean, we have proprietary code. It's not new, but again, 294 00:20:49.519 --> 00:20:53.680 these are very very powerful engines. And then there's another whole side of this 295 00:20:53.720 --> 00:21:00.720 equation, which is the RAG model retrieval augmented generation, which upon reflection, 296 00:21:00.960 --> 00:21:10.079 I believe will be the layer of functionality for governance or privacy to a certain 297 00:21:10.119 --> 00:21:12.640 extent for security, for management. You know, a lot of that's going 298 00:21:12.680 --> 00:21:17.279 to get baked into the RAG model, where you could, for example, 299 00:21:17.440 --> 00:21:19.440 before you hit your prompt, before your prompt goes up to the large language 300 00:21:19.440 --> 00:21:22.599 model, have a layer in between the checks and sees. Okay, and 301 00:21:22.599 --> 00:21:27.519 this is already happening. Like I asked Gemini a couple of weeks ago how 302 00:21:27.519 --> 00:21:33.480 many electoral votes are in Georgia and Arizona and some other state. And I 303 00:21:33.519 --> 00:21:37.039 thought for a second, it said elections are complex and fast moving. We 304 00:21:37.119 --> 00:21:41.079 recommend you use Google. It was a guard rail. That's a guardrail. 305 00:21:41.119 --> 00:21:44.480 They exactly built that in to say no, no, no, no, 306 00:21:44.640 --> 00:21:47.680 we don't want to touch that. Right, And that's in the RAG model. 307 00:21:47.759 --> 00:21:49.880 Right. That's not like trained in the model, that's outside the model. 308 00:21:49.920 --> 00:21:55.519 But it's the workflow you have around the engine that's very very important. 309 00:21:55.599 --> 00:21:59.480 Right. I totally agree with that. Then. I mean that's what I 310 00:21:59.519 --> 00:22:02.960 called to say cards before. And I think sometimes it's probably not even part 311 00:22:03.000 --> 00:22:06.039 of the context that's part of the RAG models, but it's really part of 312 00:22:06.039 --> 00:22:08.799 some safeguarding code even around it. I mean we use that at NIME as 313 00:22:08.799 --> 00:22:12.960 well, so we have built in what we call KAI inside the analytics platform 314 00:22:12.960 --> 00:22:17.319 that allows you to have a QAI model. You ask questions, I do 315 00:22:17.440 --> 00:22:19.240 this and this and excel, how does that look at a nine bog fram 316 00:22:19.480 --> 00:22:23.160 and then it gives you shows you a couple of notes in nine And we're 317 00:22:23.200 --> 00:22:29.119 of course filtering that these notes do actually exist because every now and then Open 318 00:22:29.160 --> 00:22:33.200 AI, which we use underneath the hood hallucinates and invense notes that NIME probably 319 00:22:33.240 --> 00:22:36.880 should have but we don't have it, and that doesn't help the pull use 320 00:22:37.079 --> 00:22:41.359 all. But that's a very very simple way of code around the KI that 321 00:22:41.480 --> 00:22:45.640 is just making sure that what it spits out is reasonably useful. M hm. 322 00:22:47.319 --> 00:22:48.799 Yeah, that's interesting, and you're going to see more and more of 323 00:22:49.160 --> 00:22:53.279 these AI agents. That's what everyone is talking about now, are AI agents, 324 00:22:53.680 --> 00:22:59.079 which are like little bots, semi autonomous spots that can do various things, 325 00:22:59.160 --> 00:23:00.440 and they can check any other and they can do all kinds of stuff. 326 00:23:00.480 --> 00:23:04.240 I mean, it's very interesting to me when we talk about data science. 327 00:23:04.279 --> 00:23:07.839 We talked about it before, it all seems to be getting subsumed now 328 00:23:07.880 --> 00:23:11.640 into AI. In conversations about AI, even though there are lots of different 329 00:23:11.720 --> 00:23:15.200 versions of AI, right, I mean, there are traditional models, regression 330 00:23:15.279 --> 00:23:18.880 models like all sorts of old fashioned aif you well, it's still very powerful 331 00:23:18.920 --> 00:23:23.160 and still works, but the new stuff is sucking all the oction out of 332 00:23:23.160 --> 00:23:26.720 the room. Isn't that about right? Yeah, we see that as well, 333 00:23:26.759 --> 00:23:30.079 And sometimes I mean, I'm an old guy, but now I've seen 334 00:23:30.079 --> 00:23:33.279 this in the past. Right when back propagation came along, everybody was suddenly 335 00:23:33.359 --> 00:23:36.400 using grade intocent for every problem. We just thought, hey, you can 336 00:23:36.440 --> 00:23:38.279 solve this directly, you don't need to do grade into cent. Then there 337 00:23:38.359 --> 00:23:41.119 was support vector machines, and then it was somebody else, and now then 338 00:23:41.119 --> 00:23:45.400 we're steep learning, and now it's AI. So sometimes we see building workflows 339 00:23:45.440 --> 00:23:49.400 for even very simple things. They're reaching out to some AI and we just 340 00:23:49.440 --> 00:23:53.240 say, hey, there a no dennment that does that computationally a lot less 341 00:23:53.240 --> 00:23:56.480 expensive. I don't use that. So I think, to me, it's 342 00:23:56.960 --> 00:24:00.359 it's a bit of a hype right now. It's just a new on the 343 00:24:00.400 --> 00:24:03.400 block. Everybody wants to play with it and use it. But the augmented 344 00:24:03.480 --> 00:24:07.880 really mixing it and matching it right with traditional techniques, I think that's where 345 00:24:07.880 --> 00:24:11.119 the true value lies. Yeah. Well, and so I'm just guessing here 346 00:24:11.279 --> 00:24:15.799 that one of the nice things about your platform is that it is an end 347 00:24:15.839 --> 00:24:21.000 to end platform for building models, designing models, training models, pulling the 348 00:24:21.079 --> 00:24:25.240 data in all these things, and it's adjacent to this business hub. So 349 00:24:25.400 --> 00:24:32.359 you have a marshaling area for ideas and for testing algorithms and for testing models. 350 00:24:32.599 --> 00:24:34.680 Then you connect it through the business hub and see what happens and see 351 00:24:34.680 --> 00:24:40.519 how it operates. And it's important to have this one environment where that takes 352 00:24:40.519 --> 00:24:45.640 place because when you have multiple tools, it just takes longer and it's disjointed 353 00:24:45.680 --> 00:24:49.480 and there are connections between the tools and things change, so it's important to 354 00:24:49.599 --> 00:24:56.359 have that main marshaling area to it's like a giant analytics sandbox. Is that 355 00:24:56.359 --> 00:25:00.160 about right? That's a very nice subscription. Absolutely. I tend to say 356 00:25:00.200 --> 00:25:07.319 that data scientists doesn't necessarily need to know how the method does something, but 357 00:25:07.359 --> 00:25:10.279 it needs to know what the method does. So if it's reaching out to 358 00:25:10.319 --> 00:25:12.240 a Python library or in our library or Sea library underneath it, which is 359 00:25:12.319 --> 00:25:17.079 not that important, but you still need to understand what the method actually does 360 00:25:17.839 --> 00:25:22.000 underneath to be able to interpret the results. It's a simple example. If 361 00:25:22.000 --> 00:25:25.160 you don't know what a regression coefficient is you won't be able to interpret it, 362 00:25:25.359 --> 00:25:27.599 but you don't necessarily need to understand how it was derived from the data. 363 00:25:29.039 --> 00:25:30.759 Yeah, no, that that's pretty interesting. Let me throw this concept 364 00:25:30.759 --> 00:25:33.319 at you and see what you think about it. I wrote up an article 365 00:25:33.519 --> 00:25:37.480 just last week I guess about this. I was flying to a conference in 366 00:25:37.559 --> 00:25:41.240 Denver just thinking about these large language models and analytics and AI and all this 367 00:25:41.359 --> 00:25:45.480 stuff that have been covering for a long long time, and I thought to 368 00:25:45.519 --> 00:25:49.000 myself about this concept I call the executive cockpit. And the idea is that 369 00:25:49.119 --> 00:25:56.079 I think very forward looking organizations are going to deploy a small language model that 370 00:25:56.160 --> 00:25:59.640 is aligned with their business, like if it's manufacturing or healthcare or whatever, 371 00:26:00.920 --> 00:26:04.039 in their data center, so on prem, possibly in the cloud as well, 372 00:26:04.039 --> 00:26:08.119 but I have my thoughts wrapped around this on prem small language model. 373 00:26:08.279 --> 00:26:12.240 Then you're going to train it on your ERP, on your salesforce, on 374 00:26:12.279 --> 00:26:17.200 your CRM, on your customer support for example, your tickets, like any 375 00:26:17.240 --> 00:26:21.440 of your core enterprise systems. You're going to train this model on your data, 376 00:26:21.480 --> 00:26:26.319 on your business business data. Then what you'll do is set up COFA 377 00:26:26.359 --> 00:26:32.200 topics coming from those systems into a vector database adjacent to this interface for the 378 00:26:32.319 --> 00:26:37.000 small language model, and that is where the executives will spend their day running 379 00:26:37.039 --> 00:26:40.039 their business, because then you could ask any question at all. How is 380 00:26:40.039 --> 00:26:44.519 our marketing working in APAC? Who can we let go if we have to 381 00:26:44.680 --> 00:26:48.359 save some money? What where are we weak in our organization right now? 382 00:26:48.400 --> 00:26:52.839 Just all kinds of different questions and you'll get all these answers. And I 383 00:26:52.880 --> 00:26:56.279 actually mentioned to a CEO of this one company because I was trying to get 384 00:26:56.319 --> 00:26:59.640 him to help me do sales enablement for them, because I have this big 385 00:26:59.680 --> 00:27:03.039 audience that have been marketing to for years. And one person turns out to 386 00:27:03.079 --> 00:27:07.559 be the next deputy Chief Data officer for the IRS. And I sent this 387 00:27:07.640 --> 00:27:11.599 email saying, Hey, this is the lady I've known for a long long 388 00:27:11.640 --> 00:27:14.599 time. This is what I mean by sales enablement. Do you guys have 389 00:27:14.680 --> 00:27:17.119 the IRS account? And if fireback, he said, I don't know, 390 00:27:17.119 --> 00:27:19.119 I don't know if we have those accounts. I thought to myself, well, 391 00:27:19.279 --> 00:27:22.920 you would know if you had the executive cockpit, because you would just 392 00:27:22.000 --> 00:27:26.119 ask it, do we have the IRS account? Who is the account rep? 393 00:27:26.319 --> 00:27:29.319 What's the latest of this account? Because you're getting information from all these 394 00:27:29.359 --> 00:27:33.160 systems in your private environment. But what do you think about this concept? 395 00:27:33.200 --> 00:27:36.160 Is that is that doable? Is that pie in the sky, or what 396 00:27:36.200 --> 00:27:41.240 do you think about all that? It's an interesting idea I thought about similar. 397 00:27:41.440 --> 00:27:44.680 I mean, at the end of the day, you're personalizing a large 398 00:27:44.759 --> 00:27:49.440 language model around your own infrastructure in house data. I think the challenge there 399 00:27:49.599 --> 00:27:52.759 is that in order to get a really really good model like that one that's 400 00:27:52.799 --> 00:27:56.240 really useful, you need to train it on a lot more data than just 401 00:27:56.319 --> 00:28:00.960 your own. So in a sense, you need to benefit from your competitor's 402 00:28:02.000 --> 00:28:04.240 data without actually seeing that, but kind of learning the general structure and the 403 00:28:04.279 --> 00:28:08.000 general insights, and then you customize it on your own, which in return 404 00:28:08.119 --> 00:28:15.319 kind of means that you should also be providing your data to other organizations. 405 00:28:15.400 --> 00:28:19.119 It's almost like that's kind of pre competitive training of these models so that they're 406 00:28:19.200 --> 00:28:25.880 useful for everybody. I think just training it on your setup you need some 407 00:28:26.200 --> 00:28:30.200 bigger context than that, or maybe you're a company and you have enough context 408 00:28:30.279 --> 00:28:33.039 anyway, But for every small company, I don't think you have enough data 409 00:28:33.039 --> 00:28:37.279 to really get meaningful insights. That's very interesting. That's a good's a. 410 00:28:37.359 --> 00:28:41.000 That's a good point, because I'm just I'm wondering to myself and I'm gonna 411 00:28:41.000 --> 00:28:42.799 throw this one at you too. So one of my AHA moments with these 412 00:28:42.920 --> 00:28:48.079 large language models is when I realized that when you train them on a corpus 413 00:28:48.119 --> 00:28:52.359 of data, they're not actually persisting the data verbatim. It's not like they're 414 00:28:52.359 --> 00:28:56.279 taking strings of text and storing it in a record somewhere, but rather, 415 00:28:56.680 --> 00:29:02.839 in the training process that data you use will adjust the weights and biases and 416 00:29:02.880 --> 00:29:07.799 the parameters of the model. So in other words, it's like, huh, 417 00:29:07.839 --> 00:29:12.079 well, that's very interesting that it can train in that fashion and then 418 00:29:12.200 --> 00:29:18.000 reflect back to you such remarkably granular detail about things. And you know, 419 00:29:18.039 --> 00:29:21.039 what I've seen is that if there is a subject area that has been published 420 00:29:21.039 --> 00:29:26.480 about widely, like how computer processors work, or how an irrigation system works, 421 00:29:26.680 --> 00:29:32.279 anything that has a lot of content on the web that these engines were 422 00:29:32.279 --> 00:29:34.720 trained on, it does very well. It knows all that stuff. It's 423 00:29:34.720 --> 00:29:37.599 when you get to the fringe where it's not that much published. And I 424 00:29:37.640 --> 00:29:41.599 guess that's kind of your point about having enough data to train the models. 425 00:29:41.599 --> 00:29:45.200 If you don't have enough, you're not going to get the contours right, 426 00:29:45.279 --> 00:29:48.880 and it's going to be skewed in one direction or other. Is that about 427 00:29:48.920 --> 00:29:52.119 right? I think that's a very good summary. The contrast is right, 428 00:29:52.160 --> 00:29:55.799 I mean a coneague of mind want summarized. Since it's essentially it's a consensus 429 00:29:55.839 --> 00:30:00.240 engine. It's getting the consensus around what a computer programming is. Don steff 430 00:30:00.279 --> 00:30:04.119 from the data and can repeat that. But if it's just one isolated outcome, 431 00:30:04.160 --> 00:30:07.559 it's probably not going to be able to recall that one. Interesting. 432 00:30:07.640 --> 00:30:11.920 Yeah, So Craig schmid Huber I think his name is. He's the guy 433 00:30:11.920 --> 00:30:15.880 who wrote the papers on the transformers, and he's based I guess he's actually 434 00:30:15.880 --> 00:30:19.440 in Saudi Arabia these days, but I want to say he's German of German 435 00:30:19.480 --> 00:30:22.640 origin. And I was amazed when I realized he wrote those papers in like 436 00:30:22.680 --> 00:30:27.160 the nineteen nineties or something. And it's just just now we have the compute 437 00:30:27.160 --> 00:30:30.799 to be able to can you explain that? Is that what happened is that 438 00:30:30.160 --> 00:30:33.039 just the timing was right now to be able to understand this and put it 439 00:30:33.079 --> 00:30:37.400 into play, because that was one of the big changes. And now it's 440 00:30:37.440 --> 00:30:41.880 able to see, like you know, ten twelve tokens left or right as 441 00:30:41.880 --> 00:30:45.519 opposed to just like two or three. And you also have this, like 442 00:30:45.559 --> 00:30:48.319 you say, like a consensus right where so they are like I call it 443 00:30:48.359 --> 00:30:52.960 almost like an ai Greek chorus where one is saying, I think it should 444 00:30:52.960 --> 00:30:53.359 be an A, I think it should be a B. I think it 445 00:30:53.359 --> 00:30:56.960 should be a C. And then the thing, Okay, I'm gonna pick 446 00:30:56.000 --> 00:31:00.920 this one. That's very interesting. It's a very interesting development. But why 447 00:31:00.960 --> 00:31:03.400 do you think it took so long? Is it just because we now have 448 00:31:03.440 --> 00:31:07.000 the compute to do that? I think it was a compute power issue as 449 00:31:07.000 --> 00:31:11.359 well. And then some science tends to have a little bit more of an 450 00:31:11.640 --> 00:31:14.000 leads a little bit of time before it truly has an impact, but it 451 00:31:14.279 --> 00:31:18.160 mostly waiting for complete power. I don't know. One way of looking at 452 00:31:18.200 --> 00:31:22.160 what this consensus really does is I don't know if you watch these YouTube videas 453 00:31:22.200 --> 00:31:26.359 about JGBT playing chess now, and the interesting part is that at the beginning 454 00:31:26.400 --> 00:31:30.160 it does extremely well and does very sensible things. And part of that is 455 00:31:30.799 --> 00:31:37.640 these opening libraries are all over the place, so that's extremely well established consensus. 456 00:31:37.920 --> 00:31:41.920 And then somewhere in the middle it starts inventing bizarre moves and suddenly new 457 00:31:41.960 --> 00:31:44.839 figures pop up on the on the board out of nowhere, right, and 458 00:31:44.880 --> 00:31:48.640 it has always meaningful explanations for that. And the problem there is that that 459 00:31:48.799 --> 00:31:52.799 data is so sparse that there's no consensus to learn. So at the beginning 460 00:31:52.880 --> 00:31:57.839 it sounds it almost looks like it understands chess rules. But the only reason 461 00:31:57.880 --> 00:32:01.599 it does follow the chess rules is that they're so deeply ingrained in all of 462 00:32:01.640 --> 00:32:07.480 the common material that you see that the kind of the likelihood of going outside 463 00:32:07.519 --> 00:32:09.440 of the world book is too small. But somewhere in the middle of the 464 00:32:09.480 --> 00:32:15.440 game it goes completely off the books. That's wild. So one of my 465 00:32:15.480 --> 00:32:19.680 good friends in the business, as a gentleman named Usama Fayad. You may 466 00:32:19.680 --> 00:32:22.319 have come across from at some point. He was the first chief data officer 467 00:32:22.359 --> 00:32:25.279 for Yahoo way back in the day, and now he runs the Institute for 468 00:32:25.359 --> 00:32:30.519 Experiential AI over at Northeastern University here in the States. And I had him 469 00:32:30.559 --> 00:32:34.400 in the show, and he's very funny, he's very candid. He said, 470 00:32:34.680 --> 00:32:37.160 these are large models. They're too big. They're not supposed to work. 471 00:32:37.559 --> 00:32:40.759 We don't know why they work. What are you talking about this guy 472 00:32:40.799 --> 00:32:44.680 who runs this whole operation. He's joking, we don't even know how they 473 00:32:44.720 --> 00:32:47.880 work. I mean, how's that for transparency, right, I mean there's 474 00:32:47.920 --> 00:32:51.480 some truth. Right, We don't really know how they come up with these 475 00:32:51.480 --> 00:32:53.799 answers. Right, it's a wild mix of it. It's a highly distributed 476 00:32:53.839 --> 00:32:58.799 model. We don't know why a particular answer comes. We can come up 477 00:32:58.839 --> 00:33:02.839 with of proxies for an explanation by begining with the inputs and trying to figure 478 00:33:02.839 --> 00:33:06.039 out what happens, and we can say, ah, this probably had a 479 00:33:06.079 --> 00:33:08.680 lot of influence on the decision, but we don't know for sure. That's 480 00:33:08.759 --> 00:33:12.599 so wild. I mean, that's just such a big deal that you know, 481 00:33:12.799 --> 00:33:15.200 but we do. So now we have all this observability in the data 482 00:33:15.200 --> 00:33:20.480 space, right, You've got data Relic and data Dog or new relatives to 483 00:33:20.480 --> 00:33:22.400 say data. All these different companies are doing observability, which I think spun 484 00:33:22.400 --> 00:33:27.680 out of Kubernetes primarily. But it's very interesting, and we need that kind 485 00:33:27.680 --> 00:33:30.720 of observability in these large language models. I think, I think that's going 486 00:33:30.799 --> 00:33:32.839 to be one of the keys to success. But folks don't touch that. 487 00:33:34.000 --> 00:33:37.480 Dot will be right back. We're talking to Michael Berthol from NIME on Inside 488 00:33:37.519 --> 00:33:50.440 Analysis. Standby, welcome back to Inside Analysis. Here's your host, Eric 489 00:33:50.599 --> 00:33:55.680 Tavanaugh. All right, folks back here on Inside Analysis with Michael Berthold, 490 00:33:55.759 --> 00:34:00.599 founder and CEO of NIME k n I m E look them up online And 491 00:34:00.640 --> 00:34:05.559 Michael, I was mentioning to you in the break that I'm wondering to myself 492 00:34:05.960 --> 00:34:10.079 this whole business intelligence industry and there are hundreds of players these days, hundreds 493 00:34:10.079 --> 00:34:15.440 of companies doing some form of analytics. Of course, NIME is a whole 494 00:34:15.480 --> 00:34:19.400 analytics platform and open source analytics platform end to end, but there are lots 495 00:34:19.440 --> 00:34:22.519 of point tools, whether it's visualization or number crunching, OLAP, roll, 496 00:34:22.519 --> 00:34:27.000 app all this kind of stuff, and I wonder is all of that in 497 00:34:27.039 --> 00:34:31.119 the crosshairs of these foundational models? What do you think? That's a very 498 00:34:31.239 --> 00:34:35.280 very interesting question, and we of course asked ourselves that as well. And 499 00:34:35.360 --> 00:34:37.880 I think for some of the some of the tools that you mentioned, like 500 00:34:37.960 --> 00:34:44.079 generating visualizations, that type of stuff, I do think they are pretty replaceable 501 00:34:44.119 --> 00:34:47.440 by AI type models because at the end of the day, you're doing something, 502 00:34:47.639 --> 00:34:52.639 you're generating a code that generates the visualization based on data and you judge 503 00:34:52.679 --> 00:34:57.599 the output of that code by just looking at it and saying this is quite 504 00:34:57.639 --> 00:34:59.800 right. So I think that type of stuff will go away. And we 505 00:35:00.239 --> 00:35:04.800 have in NIME actually built in it used to be an each chart scripting editor 506 00:35:04.920 --> 00:35:07.159 that has now an EIE element and you don't need to touch the code anymore. 507 00:35:07.199 --> 00:35:12.000 So those types of still I believe will go away. The eye tools 508 00:35:12.119 --> 00:35:17.119 trying to really find surprising, interesting new insights in data, I think that 509 00:35:17.159 --> 00:35:22.000 type of stuff is a lot harder to replace because fundamentally you're trying to find 510 00:35:22.000 --> 00:35:27.280 something new. And like we discussed before the break, these GENEI models are 511 00:35:27.599 --> 00:35:31.559 consensus engines, right, so they kind of try to gravitate towards something they've 512 00:35:31.599 --> 00:35:37.599 seen more and more often before. Interesting. That's right, That's an excellent 513 00:35:37.639 --> 00:35:43.159 point, really, that's that's exactly right. So it's good for understanding the 514 00:35:43.199 --> 00:35:46.119 well trodden path basically, Like that's what it's very good at doing is saying, 515 00:35:46.159 --> 00:35:49.960 okay, there's a high way, it goes that direction, but I 516 00:35:50.079 --> 00:35:52.840 want to go wandering around the forest that It's not as good on the fringe 517 00:35:52.960 --> 00:35:57.440 basically, so you will use it. But I mean, so I read 518 00:35:57.480 --> 00:36:00.840 an article some guy on LinkedIn talking about how he connected I don't know by 519 00:36:00.880 --> 00:36:07.000 ODBC or JDBC or something in his model with data sources, and he asked 520 00:36:07.000 --> 00:36:09.559 it to queer the data source and it did. It reached into the database, 521 00:36:09.639 --> 00:36:13.519 pulled the information out, and delivered it. And you're like, Okay, 522 00:36:13.920 --> 00:36:16.840 that's pretty interesting. And then when I think to myself about what's what 523 00:36:17.000 --> 00:36:22.400 could be happening here? Is in the data warehousing space, for example, 524 00:36:22.480 --> 00:36:25.679 we move so much data around. It's all the data that's from your core 525 00:36:25.719 --> 00:36:29.760 systems that you've decided to put in, which is a tremendous amount. Very 526 00:36:29.760 --> 00:36:34.559 little of that data ever gets used a lot of times it's the summaries or 527 00:36:34.599 --> 00:36:37.119 the aggregates or the roll ups that are used for various purposes, but a 528 00:36:37.159 --> 00:36:42.039 lot of it just doesn't even get used at all. And I think that 529 00:36:42.119 --> 00:36:45.480 what these large language models are going to do is kind of turn the entire 530 00:36:45.519 --> 00:36:50.960 model inside out of how we viewed moving data and analyzing data and doing things 531 00:36:51.000 --> 00:36:55.280 with data because they don't really care. They're just going to once they're trained 532 00:36:55.400 --> 00:36:59.320 on a certain space. And again, if you train it on your data 533 00:36:59.400 --> 00:37:01.159 or if you're your vector database, you have a lot of embeddings of your 534 00:37:01.199 --> 00:37:06.039 corporate data and you point your RAG model there. Well, you can get 535 00:37:06.079 --> 00:37:10.960 answers to things very quickly that before would have required running reports and doing ETL 536 00:37:12.000 --> 00:37:15.519 and doing all this stuff. And I think that in many use cases these 537 00:37:15.559 --> 00:37:19.000 models are going to short circuit all that stuff and you're just not going to 538 00:37:19.039 --> 00:37:21.679 have to do as much of that stuff anymore at all. But what do 539 00:37:21.719 --> 00:37:27.199 you think about that? I think there's some truth to it, because fundamentally, 540 00:37:27.320 --> 00:37:30.800 what these models won't necessarily do is actually look at all of the data, 541 00:37:30.880 --> 00:37:35.599 but they're going to apply a lot of common standard practices to that. 542 00:37:35.679 --> 00:37:38.480 And standard sounds a little bit too limiting. I think there's a huge wealth 543 00:37:38.519 --> 00:37:43.079 of standard practices that people do apply to the data, and that's part of 544 00:37:43.119 --> 00:37:46.920 this consensus engine, and so that the AI models will try out a lot 545 00:37:46.960 --> 00:37:52.280 of those things a lot faster than you ever would. So, absolutely, 546 00:37:52.320 --> 00:37:55.159 and there's a good chance that some of these insights that will be generated are 547 00:37:55.239 --> 00:38:00.519 interesting to you. But then continuing the expiration and saying, I mean, 548 00:38:00.800 --> 00:38:07.000 how do they obvious? Say? The Eureka moment is usually preceded by oops, 549 00:38:07.039 --> 00:38:12.239 that's strange. I think you'll have these moments, right, and AI 550 00:38:12.280 --> 00:38:15.239 doesn't do that. AI doesn't say this is weird. I should diggle in 551 00:38:15.280 --> 00:38:19.360 a little bit deeper because that's outside of the consensus. So it will continue 552 00:38:19.440 --> 00:38:22.280 doing kind of like you said, will will continue the normal path. And 553 00:38:22.360 --> 00:38:31.320 that's where I believe the human intuition curiosity oops detection capability is going to be 554 00:38:32.119 --> 00:38:37.320 relevant for a long time. I like this oops detection. That's good stuff. 555 00:38:37.559 --> 00:38:38.880 Well, there was a gentleman I had on the show years ago who 556 00:38:38.920 --> 00:38:43.199 did something. He said something a lot like that. He basically said, 557 00:38:43.280 --> 00:38:45.920 AI doesn't have to be the ability to be like, hmm, that's kind 558 00:38:45.920 --> 00:38:51.159 of weird. What's going on with that? Right? Because it's just processing 559 00:38:51.199 --> 00:38:54.320 information and doing what it's been told to do, which is just reflect backwards 560 00:38:54.679 --> 00:38:59.039 based upon a prompt and its training. It's a very simple thing. I 561 00:38:59.039 --> 00:39:01.760 mean, it's very complex in terms of how it got there, but nonetheless 562 00:39:01.760 --> 00:39:05.239 it you know, one thing that did annoy me, I will say, 563 00:39:05.280 --> 00:39:08.800 is in the early days when that New York Times reporter was getting deep with 564 00:39:08.880 --> 00:39:14.079 the with chat GPT and trying to like tease out of it, whether it's 565 00:39:14.119 --> 00:39:16.400 sentient or something. I'm like, dude, that is a misuse of the 566 00:39:16.400 --> 00:39:21.960 technology like that is not whether you should be using this thing for to try 567 00:39:22.000 --> 00:39:24.280 to like what trick it into revealing that it's really alive and what you know, 568 00:39:24.320 --> 00:39:29.440 what are you even talking about? And I think that's part of the 569 00:39:29.599 --> 00:39:32.039 downside these days is that. And I'm a media person myself, but a 570 00:39:32.079 --> 00:39:37.599 lot of times the media will just sort of glom onto some narrative about something 571 00:39:37.639 --> 00:39:40.960 and it's very hard for them to decouple from that and get down to brass 572 00:39:42.000 --> 00:39:43.760 tacks. And that's what we do in the show. In fact, I 573 00:39:43.800 --> 00:39:45.480 used to say at the beginning of every show, the show, it's all 574 00:39:45.480 --> 00:39:50.400 about getting down to the brass tax of what actually happens in the data world 575 00:39:50.480 --> 00:39:52.400 and what you do with this stuff. And I think it is important that 576 00:39:52.440 --> 00:39:57.840 people keep in their minds the purpose of this technology, Why are you using 577 00:39:57.880 --> 00:40:00.079 it? Where is it appropriate to use it? Where is it not appropriate 578 00:40:00.119 --> 00:40:04.440 to use it? And that's just basic common sense, right, Yes, 579 00:40:04.519 --> 00:40:07.320 I totally agree. I mean, it goes pretty much in line with also 580 00:40:07.360 --> 00:40:09.840 the European aii that they just passed. But I mean, if it's not 581 00:40:10.079 --> 00:40:14.679 mission critical, if it's not safety critical, you can trust a system that 582 00:40:14.840 --> 00:40:17.960 is wrong in I don't know point one percent of all cases. If it's 583 00:40:19.039 --> 00:40:22.360 controlling nuclear power plants, better not be wrong in point one percent of all 584 00:40:22.360 --> 00:40:25.920 cases. Right, that's right. You got to watch out where it's So 585 00:40:27.159 --> 00:40:29.840 where do you see a lot of use case of your clients. I mean, 586 00:40:29.880 --> 00:40:32.480 obviously some of your clients are using large language models. Where are you 587 00:40:32.519 --> 00:40:38.480 seeing success stories in that space right now? So there's a lot of success 588 00:40:38.519 --> 00:40:43.400 stories in other areas of the business, as you probably probably know. Undoubtedly, 589 00:40:43.480 --> 00:40:46.280 no checking legal contracts, doing marketing material, that type of stuff. 590 00:40:46.280 --> 00:40:52.400 There's a lot of value in applying JENEI on the data analytic space. Honestly, 591 00:40:52.519 --> 00:40:54.320 we don't. We see a lot of interest. There's a lot of 592 00:40:54.360 --> 00:40:59.760 people that say, oh cool, I can build a customized chatboard using name 593 00:41:00.159 --> 00:41:04.559 not really our core business. And then the real applications tend to be around 594 00:41:04.639 --> 00:41:08.159 text processing, which is where Jennai is really strong. And then instead of 595 00:41:08.519 --> 00:41:15.679 using outdated antique libraries for sentiment analysis or text segmentation, you're just handing it 596 00:41:15.719 --> 00:41:19.880 over to an AI model and say, hey, segment this, or extract 597 00:41:19.880 --> 00:41:22.760 the key components or create a summary. But that type of stuff, it's 598 00:41:22.800 --> 00:41:28.679 amazing. So I see I'm also as image mining extensions. I think that's 599 00:41:28.440 --> 00:41:34.800 the next setup where we can use image processing capabilities of Jenai for a lot 600 00:41:34.800 --> 00:41:37.800 of the number crunching. I mean, we've all seen these cases where you 601 00:41:37.920 --> 00:41:40.519 can't add two numbers. Doesn't really know what the prime numbers. It misses 602 00:41:40.559 --> 00:41:45.400 the understanding of the concept of a number, right. So there. I 603 00:41:45.400 --> 00:41:52.119 think it's more as a tool to help you build workflows, build dataations, 604 00:41:52.159 --> 00:41:55.519 but only as a helper, right. So that's actually an excellent point I 605 00:41:55.559 --> 00:42:00.480 wanted to get into. I believe that we're just scratching this surface of using 606 00:42:00.480 --> 00:42:05.960 these models as a component in a workflow. So you mentioned, for example 607 00:42:06.000 --> 00:42:09.760 summarization, that is hugely powerful. I mean, you know you can enter 608 00:42:10.440 --> 00:42:15.400 especially for policies, for complex policies for law, for example, for legal 609 00:42:16.400 --> 00:42:21.639 protocols and when to file motions, what motions you can file, what you 610 00:42:21.679 --> 00:42:22.920 have to do according to that. I mean you used to have to pay 611 00:42:23.280 --> 00:42:27.159 lawyers a lot of money to tell you that stuff. Now if you just 612 00:42:27.199 --> 00:42:30.079 get access to the rules, will load them into a large language model and 613 00:42:30.159 --> 00:42:35.440 just start asking questions. That is an incredibly powerful use case because it used 614 00:42:35.440 --> 00:42:37.360 to take a lot of time to sort through the process of how to do 615 00:42:37.400 --> 00:42:40.840 something. Now you just ask it, how do I fire someone? First 616 00:42:40.840 --> 00:42:44.880 step, send them a letter saying they're not performing properly. Second step, 617 00:42:45.039 --> 00:42:46.320 you know, monetor their behavior. Third step. With all this stuff, 618 00:42:46.320 --> 00:42:51.320 it's like there, it is like wow, talk about saving time. I 619 00:42:51.320 --> 00:42:53.440 mean, it saves time. And here's my other big soapbox issue. It 620 00:42:53.480 --> 00:42:59.679 improves morale because nobody wants to spend their time scratching their head reading through just 621 00:43:00.079 --> 00:43:04.159 full documentation. Nobody likes doing that, nobody, So all that stuff is 622 00:43:04.159 --> 00:43:07.519 going to go away, right, what do you think? I totally agree. 623 00:43:07.599 --> 00:43:09.920 Yah. Your examples center a lot around firing people, by the way, 624 00:43:09.960 --> 00:43:15.559 but I think I tend to say when people ask me if Jenny, 625 00:43:15.599 --> 00:43:19.400 I are going to make data science lives easier, I say no, I 626 00:43:19.400 --> 00:43:22.760 don't think so. But it's going to make it nicer because it's going to 627 00:43:22.800 --> 00:43:25.599 remove all of that boring stuff and we can now focus on the really interesting 628 00:43:25.880 --> 00:43:29.920 but more complex stuff. So it's going to make it more interesting, more 629 00:43:29.920 --> 00:43:35.000 complex. That's interesting. Yeah, wow, And I think you do want 630 00:43:35.039 --> 00:43:37.039 to document things, and you can have it document things for you too, 631 00:43:37.159 --> 00:43:39.360 Right, you can just throw a whole bunch of stuff and document this, 632 00:43:39.599 --> 00:43:43.960 okay exactly. I mean we we now have a component on the hub that 633 00:43:44.000 --> 00:43:46.440 takes a nine workflow and explains what the nine workflow does. And we do 634 00:43:46.480 --> 00:43:50.960 that by just shipping it off to Jenna. It's perfect for that. Wow, 635 00:43:51.119 --> 00:43:52.719 that's amazing. All right, folks, will Podcast put a segment coming 636 00:43:52.800 --> 00:44:00.599 up next where you're listening to Inside Analysis. All right, back here on 637 00:44:00.599 --> 00:44:05.800 Inside Analysis talking to Michael Bertholdt is the founder and CEO of nine K n 638 00:44:05.800 --> 00:44:08.800 IME and Michael, I know what it's like in a software company. There's 639 00:44:08.800 --> 00:44:13.400 always a roadmap. You're always working on something, and we talked about a 640 00:44:13.400 --> 00:44:17.199 couple of key things. Governance. There's model governance, there's data governance, 641 00:44:17.239 --> 00:44:22.840 there's it governance. What are you working on in the governance space? Thanks 642 00:44:22.840 --> 00:44:24.719 for asking that, Like, like you know, all the road maps are 643 00:44:25.519 --> 00:44:30.239 changing all the time, but what we're currently working on. We had actually 644 00:44:30.280 --> 00:44:32.880 this model government's topic on the work map. I've been working on that for 645 00:44:32.920 --> 00:44:37.480 a couple of years now. So the idea of being able to monitor what 646 00:44:37.559 --> 00:44:40.519 models are doing automatically we train them. We talked about that and I think 647 00:44:40.599 --> 00:44:45.800 the first episode, but what we added now is the ability to also govern 648 00:44:46.199 --> 00:44:51.400 the AI usage of people that are creating nine work clows. So first of 649 00:44:51.480 --> 00:44:54.440 all, then somebody is creating nine work close using the NIE Analytics platform and 650 00:44:54.519 --> 00:44:59.800 uses this built in AI, we call it KAI for nine AI, we 651 00:44:59.840 --> 00:45:02.559 need to make sure that gets channel to an IT approved AI. Right. 652 00:45:02.679 --> 00:45:06.400 Maybe that's just for expense purposes. You know, I'm going to have too 653 00:45:06.480 --> 00:45:08.679 much consumption in the cloud. Want to make that in house or it's really 654 00:45:08.679 --> 00:45:14.400 a data privacy issue, But the more worrisome part for people is that you 655 00:45:14.719 --> 00:45:17.079 I mean, one of the strengths of the NIE Analytics platform, the workflow 656 00:45:17.119 --> 00:45:21.599 concepts, is that everybody can use any technology they want, right, they 657 00:45:21.599 --> 00:45:23.840 can reach out to experimental libraries, they can reach out to our stuff, 658 00:45:23.840 --> 00:45:28.679 to Python stuff, to whatever. But by now they can of course also 659 00:45:28.719 --> 00:45:35.039 connect to various different AI providers, and we need a way for them for 660 00:45:35.320 --> 00:45:40.039 central IT governance to be able to make sure that the nine workflow users inside 661 00:45:40.039 --> 00:45:46.280 the organization can only use approved AIS. So maybe the maybe marketing can use 662 00:45:46.320 --> 00:45:51.440 an AI in the cloud, but maybe legal shouldn't or HR shouldn't, right, 663 00:45:51.480 --> 00:45:53.320 And that's something we have built in into the name hub. Now that 664 00:45:53.400 --> 00:45:59.599 we can limit the types of AIS you can their users can reach out to 665 00:45:59.719 --> 00:46:04.679 from nine under drix platform and they get to choose from one of the approved 666 00:46:04.679 --> 00:46:07.840 AIS that it central. I set up and said, Okay, here's an 667 00:46:07.840 --> 00:46:10.679 AI that's consumption light, that's for the easy tasks, here's the one for 668 00:46:12.039 --> 00:46:15.000 whatever the tech team. Here's the one that's for compliant data. And we 669 00:46:15.079 --> 00:46:21.239 also allow they're set up on the hubside of safeguarding workflows so that you can 670 00:46:21.280 --> 00:46:24.559 before the data gets sent out to say a cloud AI provider, it gets 671 00:46:24.559 --> 00:46:30.679 screened for private information or maybe the data automatically it gets anonymized before it gets 672 00:46:30.719 --> 00:46:36.360 sent out. Yeah, that's very important stuff. And the use are you 673 00:46:36.440 --> 00:46:38.320 also able to do some fin apps on that, In other words, see 674 00:46:38.599 --> 00:46:43.840 how much is costing to leverage this AI engine versus that AI engine and do 675 00:46:43.880 --> 00:46:47.480 some cost optimization. Is that something you can do. We can do that 676 00:46:47.519 --> 00:46:51.480 as part of a nine workflow and you could build that. You could build 677 00:46:51.480 --> 00:46:53.519 that in there as well. But we are currently offering to our customers the 678 00:46:53.559 --> 00:46:57.800 abilities to monitor consumption so they have a bit of an eye on that, 679 00:46:58.199 --> 00:47:00.719 but it's not automatically. We route two different day eyes, but that's just 680 00:47:00.800 --> 00:47:05.679 an edit functionality under the good m hmm. Yeah, it's all in the 681 00:47:05.679 --> 00:47:08.599 workflows basically, And that's where we're going in the last segment, where I 682 00:47:08.639 --> 00:47:14.239 think that we're just the beginning of leveraging these technologies because what they're really very 683 00:47:14.320 --> 00:47:19.880 good at is pattern recognition, right even just the vectorization the embeddings basically how 684 00:47:19.880 --> 00:47:23.920 it stores it as a point in array basically, and just understanding how it 685 00:47:23.960 --> 00:47:29.320 can map those two things. It's not just word not just text generation. 686 00:47:29.440 --> 00:47:32.440 I think we're going to get some really interesting things in terms of pattern recognition 687 00:47:32.920 --> 00:47:37.800 and then recommendations. I mean, I think that these little AI agents, 688 00:47:37.840 --> 00:47:42.639 these assistants are going to be extremely helpful in all facets of business. You 689 00:47:42.639 --> 00:47:45.199 know, to be able to very quickly give you a customer profile when you're 690 00:47:45.280 --> 00:47:49.639 on the phone with someone, or to be able to give you summarization of 691 00:47:50.079 --> 00:47:52.159 text on demand. I mean, really, I think the hardest challenge is 692 00:47:52.159 --> 00:47:57.679 going to be changing mindsets and changing day to day behaviors and workflows. What 693 00:47:57.719 --> 00:48:00.199 do you think? Final thoughts? I totally agree with that one. We 694 00:48:00.280 --> 00:48:04.159 see that inside NIM as well. I Mean, the developers were the first 695 00:48:04.199 --> 00:48:06.400 ones that really said we want to use this, you want to use this, 696 00:48:06.920 --> 00:48:09.079 but then getting the rest of the organization to us seriously think about it's 697 00:48:09.119 --> 00:48:12.840 like what can HR do? But can legal really do with that? It's 698 00:48:12.920 --> 00:48:16.280 a huge time saver and it's largely untapped. Totally with you as Jack and 699 00:48:16.400 --> 00:48:21.400 Dick, it's there. They're interesting times ahead. Yes, to say the 700 00:48:21.480 --> 00:48:23.119 least. Well, what a fantastic conversation, Folks, look them up online. 701 00:48:23.119 --> 00:48:28.360 Michael Berthold b E R T H O L D from nine K N 702 00:48:28.360 --> 00:48:30.559 I M E. We'll talk to you next time. Folks, you've been 703 00:48:30.559 --> 00:48:45.159 listening to Inside Analysis expect Redland's Ranch Market is a unique, full service international 704 00:48:45.280 --> 00:48:50.119 grocery store that specializes in authentic food items from Mexico, India, and from 705 00:48:50.159 --> 00:48:54.039 many Mediterranean and Asian countries, including popular items from the US. They offer 706 00:48:54.119 --> 00:48:59.639 fresh baked items from their in house bakery, housemade tortillas from their tortill area, 707 00:49:00.000 --> 00:49:04.360 delicious array of prepared Mexican foods, a terrific fresh food and juice bar, 708 00:49:04.599 --> 00:49:07.519 and a large selection of meats, seafoods and deli sandwiches, salads and 709 00:49:07.559 --> 00:49:12.440 hellal meats. Their produce department is stocked full with fresh, local and hard 710 00:49:12.440 --> 00:49:16.519 to find international fruits and vegetables that you cannot find anywhere else. Don't forget 711 00:49:16.519 --> 00:49:22.480 to step into the massive beer Cave and experience the largest selection of domestic, 712 00:49:22.639 --> 00:49:27.079 artisan and imported beers in the IE. They can also cater your next event 713 00:49:27.119 --> 00:49:30.360 with one of the delicious takeout catering trades of food. Visit them at Redlands 714 00:49:30.440 --> 00:49:36.320 Ranchmarket dot com. That's Redlands ranch Market dot com. Redlands Ranch Market a 715 00:49:36.480 --> 00:49:43.760 unique and fun shopping destination to Hebotea Club's original pure powdy arcosuper tea helps build 716 00:49:43.760 --> 00:49:46.679 red corpuscles in the blood, which carry oxygen to our organs and cells. 717 00:49:46.800 --> 00:49:52.000 Our organs and cells need oxygen to regenerate themselves. The immune system needs oxygen 718 00:49:52.039 --> 00:49:55.719 to develop, and cancer dyes in oxygen. So the tea is great for 719 00:49:55.800 --> 00:50:00.320 healthy people because it helps build the immune system, and it can try truly 720 00:50:00.320 --> 00:50:04.760 be miraculous for someone fighting a potentially life threatening disease due to an infection, 721 00:50:04.920 --> 00:50:08.559 diabetes, or cancer. The t is also organic and naturally caffeine free. 722 00:50:08.599 --> 00:50:13.239 A one pound package of T is forty nine ninety five which includes shipping. 723 00:50:13.480 --> 00:50:16.400 To order, please visit to ebot club dot com. T hebo is spelled 724 00:50:16.440 --> 00:50:22.360 T like tom, a h ee b like boy oh. Then continue with 725 00:50:22.400 --> 00:50:25.280 the word T and then the word club. The complete website is to Hebot 726 00:50:25.440 --> 00:50:30.599 club dot com or call us at eight one eight sixty one zero eight zero 727 00:50:30.760 --> 00:50:35.119 eight eight Monday through Saturday, nine am to five pm California time. That's 728 00:50:35.159 --> 00:50:39.159 eight one eight sixty one zero eight zero eight eight to ebot club dot Com 729 00:50:39.320 --> 00:50:47.199 with sixty years of fascinating facts. This is the man from yesterday and back 730 00:50:47.239 --> 00:50:52.639 in time. We go to this time in nineteen sixty five Wednesday nights on 731 00:50:52.800 --> 00:50:59.119 NBCTV Virginia, starring James Drury and Doug McClure. It's the only ninety minute 732 00:50:59.119 --> 00:51:12.760 Western on Prime on TV. Here in nineteen sixty five and from this time. 733 00:51:12.800 --> 00:51:15.800 In two thousand and seven, Nickelodeon teams up with Marriott International on a 734 00:51:15.880 --> 00:51:22.119 chain of kit friendly resort hotels that are going to feature elaborate water parks and 735 00:51:22.480 --> 00:51:28.039 live entertainment from the likes of SpongeBob, SquarePants, Dora the Explorer, and 736 00:51:28.239 --> 00:51:39.360 other Nickelodeon stars who observe it in Yellow and Forest. Something you Wish, 737 00:51:39.400 --> 00:51:44.039 And back to this time in nineteen fifty six. He's one of the funniest 738 00:51:44.039 --> 00:51:47.639 Broadway performers and he gets his own TV special. Don't miss Victor Borg's Comedy 739 00:51:47.800 --> 00:51:52.000 in Music on CBSTV. Doctor. Yes, I don't like to be mosey. 740 00:51:52.639 --> 00:51:58.679 But who are all these folks? But I really don't know. You 741 00:51:58.760 --> 00:52:00.920 don't know. No, I have never met them. They are from the 742 00:52:00.920 --> 00:52:04.360 audience. Well, they were selected from before the show, from the members 743 00:52:04.360 --> 00:52:06.440 of our studio audience. That's right. You just seem to know better than 744 00:52:06.480 --> 00:52:14.920 I do. With more at Man from Yesterday dot Com. Every Wednesday at 745 00:52:14.960 --> 00:52:19.320 three pm, it's The Uncommon Sense Democrat with host Eric Bauman. I love 746 00:52:19.400 --> 00:52:22.559 when his people talk about how old Joe Biden, but he's just a couple 747 00:52:22.559 --> 00:52:29.320 of years behind him. You'll get the best political commentary and stuff like this. 748 00:52:30.280 --> 00:52:37.960 Good night, don't well. Join us for The Uncommon Sense Democrat every 749 00:52:37.960 --> 00:52:42.480 Wednesday at three pm on the stations that leave no listener behind casey AA ten 750 00:52:42.599 --> 00:52:47.000 fifty am Man one oh six point five at them. This Mother's Day helped 751 00:52:47.039 --> 00:52:52.800 fire breast cancer. Schedule your mamogram. Our sponsor JT Auto Repair and Body 752 00:52:52.800 --> 00:52:55.119 Shop of Sam Borandino, his family owned serve in the area with quality and 753 00:52:55.199 --> 00:53:00.840 pride. You're experts and everything automotive, whether it's engine works, break tuneup's, 754 00:53:00.880 --> 00:53:04.880 minor and major collision repairs, including a full service paint center. Stop 755 00:53:04.960 --> 00:53:08.840 by two one sixty six South Guardina Street in Samberandino or call nine oh nine 756 00:53:09.000 --> 00:53:14.199 seven nine nine five one nine nine. That's KT Auto Repair and body shop 757 00:53:14.199 --> 00:53:19.119 for route supporters in the battle against breast cancer. What is the Del Walmsley 758 00:53:19.239 --> 00:53:22.760 Radio Show? Welcome to the Del Walmsley Radio Show where the high fends and 759 00:53:22.800 --> 00:53:25.400 the help begins. You need to stop being dependent on a paycheck. All 760 00:53:25.480 --> 00:53:30.679 these self help motivational people, they wind you up like a little clock click 761 00:53:30.800 --> 00:53:36.400 click, and then they'll let you. Who is the show about. I'm 762 00:53:36.440 --> 00:53:39.280 your host, Del Walmsley, and as always we're working on your financial freedom. 763 00:53:39.480 --> 00:53:44.880 Learn the secrets of building wealth from Dell Walmsley weekdays from eleven am to 764 00:53:44.960 --> 00:53:51.760 noon right here on KCAA. Senator Joe Manchin, a one man political steamroller 765 00:53:51.800 --> 00:53:55.239 in Washington, is a Democrat except twenty's nine, which is most of the 766 00:53:55.280 --> 00:54:01.039 time a multimillionaire West Virginia cole executive. He's the darling of fossil fuel and 767 00:54:01.079 --> 00:54:07.679 pipeline lobbyists and beloved by Republican opponents of progressive democratic policies. Indeed, he's 768 00:54:07.719 --> 00:54:14.800 funded by Republican billionaires. But Washington lobbyists and billionaires are not the only source 769 00:54:14.840 --> 00:54:19.840 of personal political power that allows him to hold office and block little Deed Democratic 770 00:54:19.920 --> 00:54:24.880 policies that the American majority wants and needs. Back Home, Joe has maintained 771 00:54:24.960 --> 00:54:30.960 tight authoritarian grip on West Virginia's Democratic party structure, rigging the rules to put 772 00:54:31.039 --> 00:54:36.639 Little Joe's in each and every party position. In turn, this has long 773 00:54:36.719 --> 00:54:42.119 given Boss Mansion control over who gets to run as Democrats for down ballot elected 774 00:54:42.159 --> 00:54:46.280 offices in the Mountain state until June eighteenth. That is, that's when a 775 00:54:46.320 --> 00:54:52.199 statewide Democratic rebellion that had been organizing for six years elected its slate of over 776 00:54:52.280 --> 00:54:58.880 fifty candidates to oust the matchinites on the party executive committee, replacing all but 777 00:54:59.000 --> 00:55:04.079 one of the top party officials with grassroots activists. It truly was a diverse 778 00:55:04.239 --> 00:55:07.920 people run victory. Selina Vickers, a social worker, was chief strategist, 779 00:55:08.239 --> 00:55:13.519 and Mike Pushkin, a cab driver, is now the party chair. Danielle 780 00:55:13.599 --> 00:55:16.880 Walker now vice chair of the Committee and the first person of color in state 781 00:55:17.039 --> 00:55:22.360 history to sit on the party's governing body. Summed up the significance of this 782 00:55:22.440 --> 00:55:25.679 turnaround. There's a new beacon of light shining down on the government, with 783 00:55:25.760 --> 00:55:31.599 people energized and ready to strategize with a return to the democratic process. This 784 00:55:31.760 --> 00:55:36.920 is Jim Hitar saying. To help bring this kind of progressive reform to your 785 00:55:37.000 --> 00:55:40.719 local or state Democratic Party, go to Our Revolution, the one national group 786 00:55:40.800 --> 00:55:46.320 working on this fundamental democratic change to the Democratic Party. Our Revolution dot com 787 00:55:46.880 --> 00:55:52.199 KCIA Radio has openings for one hour talk shows. If you want to host 788 00:55:52.239 --> 00:55:55.400 a radio show, now is the time. Make Cacia your flagship station. 789 00:55:55.679 --> 00:56:00.840 Our rates are affordable and our services are second to nine. Broadcast to a 790 00:56:00.880 --> 00:56:05.679 population of five million people plus. We stream and podcast on all major online 791 00:56:05.719 --> 00:56:09.679 audio and video systems. If you've been thinking about broadcasting a weekly radio program 792 00:56:09.719 --> 00:56:15.719 on real radio plus the internet, contact our CEO at two eight one five 793 00:56:15.920 --> 00:56:20.480 nine nine ninety eight hundred two eight one five nine nine ninety eight hundred. 794 00:56:20.599 --> 00:56:23.000 You could skype your show from your home to our Redlands, California studio, 795 00:56:23.039 --> 00:56:28.320 where our live producers and engineers are ready to work with you personally. A 796 00:56:28.400 --> 00:56:34.000 radio program on KCAA is the perfect work from home advocation in these stressful times. 797 00:56:34.320 --> 00:56:37.639 Just type KCAA Radio dot Com into your browser to learn more about hosting 798 00:56:37.639 --> 00:56:42.239 a show on the best station in the nation, or call our CEO for 799 00:56:42.360 --> 00:56:46.159 details to eight one five nine nine ninety eight hundred. You like to safely 800 00:56:46.280 --> 00:56:52.719 leverage bank money to earn double digit returns income tax free, with guarantees and 801 00:56:52.800 --> 00:56:55.159 no downside market risk. How can you do this? This is Farence, 802 00:56:55.280 --> 00:57:00.159 host of the Your Personal Bank Show. One. You fund a high cash 803 00:57:00.239 --> 00:57:05.320 value policy one time to earn dividends and interest. Two establish a bank line 804 00:57:05.320 --> 00:57:08.239 of credit using the cash in your policy as collateral. When you earn more 805 00:57:08.280 --> 00:57:13.239 in dividends from your policy than the interest the bank charges, you keep the 806 00:57:13.280 --> 00:57:16.440 difference, and the difference is average two to five percent annually in your favor 807 00:57:16.639 --> 00:57:22.360 for the past forty plus years. Three the bank funds contributions years two to 808 00:57:22.480 --> 00:57:25.320 twenty plus. Each year the bank adds funds, your rate of return increases. 809 00:57:25.440 --> 00:57:30.280 Your average rate of return can grow too strong, double digits annually within 810 00:57:30.360 --> 00:57:34.440 a few years. Contact me at your personal bank dot com your personal bank 811 00:57:34.480 --> 00:57:37.840 dot Com or eight sixty six two six eight four four two two eight sixty 812 00:57:37.840 --> 00:57:43.039 six two six eight four four two two for more info, or tune in 813 00:57:43.039 --> 00:57:46.639 to the Your Personal Bank Show. Your Personal Bank Show airs Tuesdays at four 814 00:57:46.679 --> 00:57:52.239 pm right here on CASEAA ten fifty AM and one oh six point five FM, 815 00:57:52.320 --> 00:57:59.239 the station that leaves no listeners behind NBC News Radio. I'm Chris Karashio. 816 00:57:59.320 --> 00:58:02.559 The President of a Rand's condition is unknown after his helicopter crashed earlier today. 817 00:58:02.679 --> 00:58:07.159 Local media reports that the helicopter was also carrying Iran's foreign minister and suffered 818 00:58:07.159 --> 00:58:10.239 what was described as a hard landing. No other details have been released. 819 00:58:10.400 --> 00:58:14.719 The White House says President Biden has been briefed on reports about the crash. 820 00:58:14.920 --> 00:58:19.360 Senate Majority leader Chuck Schumer said intelligence officials have told him there is no evidence 821 00:58:19.360 --> 00:58:22.400 of foul play. Former President Trump's New York criminal hush money trial is nearing 822 00:58:22.400 --> 00:58:27.480 an end. Things will resume tomorrow and ex Trump lawyer Michael Cohen is expected 823 00:58:27.480 --> 00:58:30.480 to be in the witness stand again for more cross examination about hush money paid 824 00:58:30.480 --> 00:58:35.400 to porn actress Stormy Daniels. The judge has told defense lawyers and prosecutors to 825 00:58:35.440 --> 00:58:38.079 be ready for closing arguments. As soon as Tuesday, but that depends on 826 00:58:38.119 --> 00:58:42.400 whether Trump testifies. If he does take the stand, that could delay the 827 00:58:42.400 --> 00:58:45.599 close of the trial by a few days, but the jury is expected to 828 00:58:45.599 --> 00:58:49.679 get the case this week. Senate Democrats believe President Biden is ready to debate 829 00:58:49.719 --> 00:58:53.000 his political rival Donald Trump in June. Maryland Senator Chris van Holland on ABC's 830 00:58:53.039 --> 00:58:57.280 This Week pointed to Biden's State of the Union address in March, in which 831 00:58:57.360 --> 00:59:00.519 Van Holland said, the president came out swinging. It's not about the age 832 00:59:00.519 --> 00:59:02.400 of the candidate. It's about the ideas the candidate, what they're going to 833 00:59:02.519 --> 00:59:06.880 do for the American people going forward. Trump and Biden have agreed to debate 834 00:59:06.920 --> 00:59:08.719 each other twice in the coming months, with the first one set for June 835 00:59:08.719 --> 00:59:13.239 twenty seventh. Trump meanwhile, is calling on Biden to take a drug test 836 00:59:13.280 --> 00:59:16.320 before the debate, claiming Biden was high as a kite during his State of 837 00:59:16.320 --> 00:59:21.960 the Union address. Democrats want Supreme Court Justice Samuel Alito to recuse himself from 838 00:59:21.960 --> 00:59:24.639 cases related to the twenty twenty election, following the report that he flew an 839 00:59:24.679 --> 00:59:30.440 upside down American flag outside his home shortly after the Capitol riot. Pennsylvania Senator 840 00:59:30.519 --> 00:59:32.880 John Fetterman told cn in the State of the Union, the incident is bizarre 841 00:59:32.920 --> 00:59:37.599 and surreal. He said he doesn't believe, however, Alito will choose to 842 00:59:37.639 --> 00:59:39.920 recuse himself. The New York Times published a photo taken by a neighbor of 843 00:59:39.920 --> 00:59:44.599 the flag outside Alito's home in Alexandria, Virginia, just days after a pro 844 00:59:44.639 --> 00:59:47.760 Trump mob stormed the capitol. The inverted flag has been associated with Trump's claims 845 00:59:47.880 --> 00:59:52.199 that Biden stole the twenty twenty election. I'm Chris Caragio, NBC News Radio, 846 00:59:54.159 --> 01:00:00.000 NBC News on CACAA Lowel sponsored by Teamsters Local nineteen twenty two. Protecting 847 01:00:00.039 --> 01:00:04.920 the Future of Working Families Teamsters nineteen thirty two. Dot org