 Hey everyone, welcome to this special CUBE Conversation. I'm John Furrier, host of theCUBE here in Palo Alto, California. We've got two great guests, a remote interview here, talking about generative AI and using AI for legacy migrations and taking advantage of the modernization opportunities that AI is driving, codeless opportunities to move those legacy migrations and modernize them for the future of the new application environment that's developing on top of cloud. We have Chris Casey, director and general manager of industry and technology partnerships with AWS Amazon Web Services and Gary Hoberman, founder and CEO of Uncork to talk about how to move those migrations into the cloud but also modernize and up-level them big time. Gentlemen, thanks for coming on theCUBE Conversation. Thanks for having us. Thanks for having us. Chris and Gary, I wanted to have this conversation to really get into some of the advances that generative AI is happening. Obviously, we see the hype of chatGPT that shows the ubiquity of the whole west of the world seeing the magic of the future of AI, but really it's been around for a long time with AWS and others, but it's generating opportunities to step-function areas. One of them is this migration of, say, mainframes or COBOL or old legacy, I still power a lot of the IT, but need to be upgraded but modernized but not necessarily replaced. Chris, this is something that you guys got going on with Uncork, your partner. Talk about this phenomenon, why this opportunity for IT managers, CSOS and CIOs to take advantage of this usable up-gen AI opportunities today to set this migration modernization. Yeah, firstly, John, thanks very much for the time. And listen, AI and ML is not something that's new to Amazon or AWS. It's been in the foundation of what we do best in terms of applying it to real-world scenarios. And we've got more than 100,000 customers who've using AWS's ML services. One of the things that we've recently launched is Amazon Bedrock. And really what we're aiming to do there is to democratize access to some of this technology, to any builder and any developer that's trying to build on behalf of customers on the AWS cloud. Obviously that might mean that some of those developers are building brand new solutions, some of which we probably haven't even imagined yet. But a lot of the time, those developers are actually innovating and either migrating or trying to enhance legacy systems. And really part of what we're excited about with our partner network is really deeply partnering with those entities and working backwards from our customer needs to identify industry-specific solutions that really help solve some of those challenges and move some of those workloads from on-prem databases and data centers into the cloud, while also improving the user experience. You know, one of the things that Amazon and cloud has proven during the pandemic, people looking back, and this is a common theme we're hearing at every event we go to from the leaders. The folks that were in the cloud before the pandemic took advantage of a tailwind and not a headwind, and that was a huge advantage for those companies. With AI, we're seeing a similar dynamic where people are recognizing if they don't get into the AI infrastructure aspect of it, they might be flat footed if the scale and speed of change continues to be the pace. In other words, the sooner you get in, the more you can take advantage of the opportunities. This is clearly something that's happening. A lot of this conversation is going on. Gary, you founded a company that's in the, I would say the big time tailwind with Uncork. You guys have a unique opportunity with AI to change the game. Take a minute to explain Uncork, why you exist, where this idea came from, and what your current status is. Yeah, you know, thanks John, and it's a pleasure being on here and with Chris and Amazon, the partnership's been incredible, agreed completely in the pandemic, showed how we could actually use technology, digitize, move faster. But what we did in Uncork was very unique and different. So I was an engineer on Wall Street back in 94, creating and coding systems. And I did that for many years until I became a global Fortune 50 CIO. And suddenly I started to live with the results of that code, meaning you're an engineer, you're coding, you really don't care who's gonna pick it up when you leave. But the reality is the company does. And what I kept seeing was, as more and more code was produced, it was almost like this immense mountain that was created. And that mountain got bigger and bigger till the point, it was 80% of my budget. And the budget was over a billion a year. I was responsible for an influencing nine billion a year of spend. And I left that world, the corporate world, to see there's gotta be a better way. And we have to not only stop the bleeding of the new technologies, like you're saying the digital transformations, but how do you address that mountain that was built up, the legacy? And when you think about it, programming itself, a language, it's a language. And that's what Generative AI is talking about. That's what it does. There's been 9,000 different programming languages created since the existence to talk to a machine. And machines don't understand language, they understand data and binary. So, codeless to us is a simple concept. We removed all aspects of grammar, syntax, all the things that make a language look like ancient Greek to someone. We made it look like a democratized, simple data definition, which is timeless. And that's what the beauty of what on Quark is running on Amazon. And this idea of Generative AI being able to now take and tackle all of this 9,000 different languages into a unified data definition and ended going forward means reduction in costs, speed to market, all the things everyone's been dying to get to with technology, the promise of it. Yeah, and my mental models, I think about COBOL, I mentioned that, we were talking about that before we came on, it's still powering a lot of applications, banks or whatever. Give an example of this AI powering approach, because it makes so much sense. I'm trying to visualize it. Give me an example of how it works and how it's playing out for you guys. Yeah, and when you think about COBOL, so Assembler was an abstract, COBOL was an abstraction of Assembler and you could think about Java and all this abstraction layers we've been going on, the best abstraction we all know is the cloud. It's what Amazon did, it's what, and the reality is what cloud did was said, hey, there's a software-defined infrastructure, that's what cloud is, but it stops at the infrastructure, it never addressed the applications. COBOL, the actual apps that are being built, the green screens that are there. And when you take it up to the next level, as we did with on Quark and with this utility that sits above the cloud, what you start to see is you could actually understand the meaning of these applications and describe them in terms of readable data, democratized data. And an example would be as you could take an onboarding application at a bank or maybe in an insurance application or underwriting application or you walk in a hospital, federal government's a new customer. John, my favorite example was we refactored into on Quark, the entire COBOL system for New York City for marriage licenses. And we became the books and records marrying 38,000 people last year and the year before during COVID. So it's digitizing and it's a great example. I was one of those people, let's just say, going through this, and it worked brilliantly, but when you think about what it can do, that's the beauty of it. And John, what people don't understand is when you're creating software today, it's about 20% of your budget and 80% is running it after. So everyone thinks the finish line is we're live, like we're live, great, finish line, the reality is it's when the hell starts. It's like that's the pain, the suffering, the end, we're looking to end that with Amazon. That's where we're saying there's a better future coming. I can imagine the sales calls are pretty easy. Hey, I'm going to come in and help you guys out. Come on in, no one's brought a solution like this before. This is a game changer, and I can appreciate what you're doing because I can imagine the impact that you have in these meetings. But the question comes up with these legacy apps, do you maintain it or do you innovate it? And it used to be mutually exclusive, right? Do we maintain it or do we completely rewrite it? You're kind of coming at it and saying, no, no, we can do both. You can maintain it into the cloud while innovating it in real time and then giving them the choice to use the higher level services say on AWS to add stuff to it. Am I getting that right? Is that the core issue? You're absolutely right. Could you imagine being able to drag in all the bedrock features and services? We have an amazing demo with an image talk getting generated from Dolly, but then giving it to AWS reconcile and getting back or recognize and getting back all the information on the image. And we're basically scoring Dolly to see how well it did. But in reality, John, when you think about bringing in the best of technologies, it's like, that's where this relationship comes. And most importantly, what we're embracing and our success today is what we call enabling business led IT transformation. And this is the taboo, gotcha, let's just call it, John, which is the controversy, who leads IT? And the reality is business leads IT because they control the budget, the requirements, the ways what we do with Uncork and Amazon is we enable the business to participate in software development. Actually take an active role. Maybe they own the rules, maybe they own, and no longer this bureaucracy built up of walls between the groups. We bring them together for the first time. Yeah, and it really is an innovation story as well as a business value. It's really incredible. Chris, you know, I kind of get tingly because Dave Vellante and I have been talking in years, going back to say 2015, 2016 timeframe. And we were kind of, people stared at us like, what are you talking about? We were saying, you know, the cloud's horizontally scalable, you get the scale with the cloud, but the data is where the vertical specialization is in industries, you have domain expertise. You're starting to see that now with Uncork and these examples of Amazon where all of a sudden it's been like this enlightenment moment or this explosion of, wow, I get it now. It's like it's clicking. AI is now the app that sits on top of the cloud so that the domain experts can apply their data in whatever way they want to protect themselves but also innovate. This is like brand new. I mean, it wasn't as strong a couple of years ago. It's now clicking into high gear. Do you agree? What's your reaction to that? And what's your view of, you must have the same feeling of, you know, it's finally happening, it's happening big time. I do. And it's really exciting, John. I mean, I don't think we've even scratched the surface on the use cases that eventually can be unlocked here. And I think really what we're seeing is part of the recent launches you've seen from AWS is democratizing access to some of this technology. And certainly there are a lot of foundational models available. There's a lot of different large language models but a lot of them are not a one size fits all strategy. So really what we've been working with our partners like Uncork is to make sure they're following sort of a mental model of four key pillars, one of which is specializing. So identifying very specific industry use cases and workloads that they see that generative AI can actually help those customers innovate and change the status quo of maybe how a business process has been done before. Then familiarizing with the right foundational models. As I said, they're not all one size fits all and sometimes they need to be tailored for a specific use case to drive the right business outcomes, then get to building. And I think like, you know, a lot of our customers are very eager to start to run proof of concepts to actually try this out and figure out where it fits into their business processes. And really the third part of that is that customer engagement. You know, we don't know all of the potential capabilities that this technology can bring. And you know, I think with generative AI it's actually potentially changing the definition of a developer to Gary's point. You know, the business can now get much more involved in iterating on different potential user experiences which is really, really exciting. But that means the different permutations of where a solution might end up in terms of its state and the user experience is excitingly variable in that vein. And certainly companies that are even looking further beyond that as to, okay, we've got a different user experience that we're putting into production, how we then thinking about maintaining that solution to Gary's point long-term is going to pay dividends into the future if they're thinking that far in advance. Yeah, and the democratization angle is interesting. The data science went through this same thing. Oh, you've got to be a unique person to be a data scientist. Now that got democratized here, machine learning, you have to be really unique. Now you can democratize it with the cloud and you can see examples here. Gary, I want to get back to the codeless idea. I like this idea of codeless because what you're implying is is that voice could be doing, hey, computer. You know, it's like a very Star Trek-like. You know, give me more power, Scotty kind of thing. So enterprises have debt, technical debt. You mentioned that earlier. Talk about the fact that you might have to run these legacy apps and then all of a sudden the regulations changes. Seeing regulations all over the world around locating in certain regions or applications might have privacy issues, whatever they are, there's new regulations. How do you bolt that on? And this is something that I see as an opportunity. How do you guys look at getting rid of that tech stack to modernize for the regulations and or changes that got to be coded into the apps? You know, John, myself and my team worked in highly regulated industries our whole life and it's where we started on Quark. It's really like when we talk about on Quark, it's like, yes, after your live in production, it's a 65% reduction in your cost and 100 times less bugs. We have eight times less defects than Linux operating system when we go live. That's the results. But what you're describing is why we created on Quark, which is imagine the world where in a codeless world, there's no more code generated. There's no more code allowed. And this was the premise while we built on Quark, which was if we could eliminate code, what it means is the upgrade path. So getting you on to the next version of on Quark becomes something of a click of a button with no risk. So for example, we've upgraded since 2019 on Quark 320 times, 320 version upgrades without a single customer ever knowing, but the beauty of that means all customers from Goldman Sachs and the finance side and CVS and then healthcare through the federal government HHS through New York City, they're all running on the same version. They're all running on the same code base, which means we're almost like a bubble, a protective container. So at the bubble level, what we're doing is we're the ones facing security requirements, ethical hacking, penetration testing, same as cloud, same as Chris does with Amazon, but it means at the bubble level, we're stuck to type two privacy shield, GDPR compliant, privacy 2020, Fed ramp certified and announced two weeks goes authorized operating federal government, HIPAA compliant, Wyrm compliant, ISO compliant, which means we launched an HR startup launched entirely on Quark and Amazon about two weeks ago. It went through the ethical hack with the flying colors, not a single issue. And first time I've never seen it done. That's the beauty is you could actually take out all of the concerns, the security testing, all from the customer and move it to us. Chris, how about the dynamics relationship and the dynamics with AWS? Because I mean, think about gen one of cloud, Dropbox, Airbnb, there's SaaS applications, ISVs and you are too, I guess, by Adam Sileski's definition, but you're really kind of building on top of Amazon's hyperscale infrastructure investment. I mean, you're taking advantage of all that work that they did, that's kind of like a super application. It's a super, because you're essentially a legacy software app that's putting a wrapper around these capabilities and handling for everything for the customer. I mean, you're leveraging AWS. What's the Amazon connection there? Because I mean, you're essentially an extension of AWS by default because you're sitting on top of them. We are, the way we view ourselves, just you said, John, we're a boring infrastructure layer on top of cloud. And most of our clients are AWS clients and that's been a focus. We have others on other providers, but smaller numbers, but it means with AWS, we have the ability to bring in every single one of their tools instantly. So private, we could do private links so that we're actually behind the scenes, secure and connected to firewalls. You have the, of course, the whole bedrock concept of AI tools coming in. We have one of the biggest mortgage companies is working with us in Amazon using Appflow, which is the entire engine of actually how do you integrate and orchestrate? And it is beautiful when you see it working and tying together. So the relationship there is we're able to bring their tools and developer tools into a codeless world for others to just drag and drop and use for the first time. You're a cloud without having the infrastructure. Amazon's the cloud, you're the ISV. I mean, the lines are blurring, Chris, when you have this kind of next gen capability, they're got so much power. This is the benefit of this kind of next level of AWS because you have all kinds of data protection now with the AI. One of the biggest concerns right now is privacy and intellectual property. Bedrock has the capability to manage data on VPCs. All that's under the covers. So the speed game is okay. This is a huge unique nuance point. Can you guys talk about this speed game because Amazon set up with Bedrock to do first party, third party and long tail open source. So like all the models are running there. Yeah, I mean, speed does disproportionately matter in general, John, and especially in this space. And I think you've just hit the nail on the head there. The lines are a little bit gray in terms of AWS services and partner solutions. And ultimately that's actually great for end customers. One of the key aspects of what we're trying to bring with our partner network and even the AWS marketplace, which I know you're very familiar with, John, is the end customers are looking for complete solutions to help solve their business problems here. They're not really looking for a piecemeal where they have to put it together themselves. So the more that we can really improve that experience for an end customer, where yes, they might be leveraging the preview of Amazon Bedrock in terms of accessing those underlying foundational models. But maybe the customer doesn't even know that that's powering some of the user experience that they're seeing as they're iterating on an application with Uncork. Ultimately that is bettering the end customer outcome there. Chris, talk about the option with Bedrock and Amazon with from the data perspective, if you're going to be managing the legacy modernization path, you got to have that option. I mean, that's the biggest fear right now with each of the AI techniques is, where's my IP? Where's the data? Yeah, and certainly we take that very seriously and certainly the way we've designed Bedrock is allowing customers to access these foundational models through an API. They can point those models to certain data that's sitting in their AWS environments and get the results from those models without having to worry about that IP extending beyond their AWS environment, which is really critical when you're talking about some of these enterprise applications for Generative AI and certainly John, that is certainly one of the questions that I get asked most from customers when they're thinking about the utility of this new technology and they sort of understand it from sort of a social you and I are playing around with something like chat GPT, but the next frontier from an enterprise application, that's certainly very much front of mind for customers and certainly from our side, it's not only just the Amazon story and sort of how we secure our services and the cloud itself, but working with partners and Gary mentioned a whole bunch of certifications that uncork have to make sure that the security of the data in the cloud is still first and foremost at the front of mind for our partners and the end customers too. Gary, you know your peers, they think of the same thing, you got which foundation model to choose, what's your angle on this whole IP data? Where's that store? What's your reaction? Yeah, I mean, we always get that question because we have, we're OEM and a lot of other software you would never know existed today where they're using us behind the scenes or even in front to face the customer. So we designed it and when you think about leveraging Amazon and the work that Chris mentioned, we designed uncork to be single tenant because cloud has basically taken away the reason to be multi-tenant, which is a risk. Multi-tenant is this, it's a potential security risk depending on how you actually do it. To us, we're single tenant which means the customer gets to own the encryption key to their IP and their data that's unique to them, which means there's no keys to the kingdom behind the scenes. That is the most secure model. And we like to, you know, when you think about what that means is we can now store data for customers, PII data we're storing for banks and socials and data bursts, we're able to store that very securely with all the regulation and compliance needed in the Amazon bedrock, in the actual Amazon services here. If someone wants to upload a W2 document, we're storing it in S3 buckets behind the scenes and then we're making sure it's worm compliant, right once read many. And it's that relationship between the app and the cloud, which was missing all these years. When you think about cloud migration, when I was there at the earliest days when I was a city group MD seeing cloud come in and I watched a server move from under the desk to a data room to a data center to the cloud. And whole concept of that progression and abstraction was the beauty, let us do the work for you as a service. Like that's, so we've taken it with Chris just up to the next level is let us store your data. We have a new feature coming out, John. It's auto indexing. Let us auto index your data based on usage as opposed to where you think and get better. That's the benefits working with us in Amazon. I mean, soon AI was generally I was going to have all certifications for Amazon in any way Chris, it's like, you know, I mean, it really is a wonderful thing. I have done right and it's a data problem and opportunity. You manage the data, accurate data in, you get outcomes. So it's input, inputs are huge in this, Gary, right? You got to have the right inputs, but the data because the outputs will take care of itself if you handle it properly. This is why I think I like the bedrock approach because you can handle the inputs and engineered and architect that. Absolutely. And you know, think about how much paper you as a consumer are still seeing in your life as much as we read about generative AI, you walk to a doctor's office, you're handed a pen and paper. And so like the beauty is, you know, with the solution with Amazon on Quark we could tackle those transformations that are still paper. And now you could use generative AI to ask those questions and get the responses as you're saying, the accuracy back. There's a great book called The Score Will Take Care of Itself about Bill Walsh. He focused on the inputs, not the output. He didn't care about the score where he cared about what was the inputs. I think we always love sports and allergies. So that's kind of the data AI game right now. It's not so much trying to replicate the large language models. It's get the inputs done, take advantage of the different foundation models. And that seems to be a great approach. You guys are doing great. I guess my final question to both of you guys, Gary in particular, what's next for on Quark? Where are you now? What's going to happen next? Because if you continue to do this, you're going to be the legacy cloud. You can just be like everyone come to you and just everything. And you're going to modernize all the legacy apps. John, like I built on Quark to fix the mess I left I saw for 25 years. I mean, I was like, that was the reason I went there was that. And what would be left is to actually make cloud migration successful because like the idea of a data center still being needed. It's a company that's confused in many ways to say why is that their strategic advantage? So that's where like to me, the excitement is actually changing in industry. We know codeless is the future of all software. It is clear as day to all of us and we see that that's the future coming. And to me, the future really becomes when as a consumer, you know, my son passes his driver's task he just did. I should have got a text automatically from DMV saying, congrats and by the way, we see you have Geico. Would you like us to add him to your pause? That's the world I'm from. That's where I wanted him to. I'd get many speeding tickets for sure. My insurance would go up for sure. John, we're watching your every moves rolling through those stop signs in California. Chris, your take real quick. I know you got a lot in your mind and a lot on your plate with a lot of these industries. Is that the same pattern in all the different industry partnerships you have that's scale everywhere kicking in? Is that Gen AI wake up call here? Same pattern, different use cases, starting to get very specific too. And I think each of the customers in all of the industries, one of the things they all have in common, they're very excited to reap the benefits of this and they're very eager to get started. And really from our perspective, we think partners are very much critical for AWS to mutually bring that value proposition for those customers. And we're ready to support our partners and customers through proof of concepts and funding and our working backwards approach to make sure what we're building is going to deliver the outcomes they expect. But yeah, certainly the commonality across the industry is they're eager to get started and eager to reap the benefits. The devil's in the detail though, in terms of which foundational models are right, how you customize them and tweak them for the specific use case that you have in your specific industry. That's the exciting challenge ahead of us, but it's going to be a fun ride over the next months. It feels like the early days of cloud, but a whole nother level of scale with AI, using AI for legacy. I love this conversation, AI and codeless. Gary, I'll give you the last word, last minute. Give the pitch for Uncourt, for the folks watching out there that want to be customers are swimming and sauteing in the frog and boiling water, going, oh my God, I got this legacy. What do I do? What's your pitch for Uncourt? Yeah, we democratize technology without compromising enterprise. And it's through codeless, which is another way of saying we convert programming language to data or data-driven software. And that's the future, is data-driven software, data is the best asset in the world? We make it your biggest asset in your company. It's all about the data. Chris, Gary, I really appreciate you coming on this special CUBE Conversation, Genitive AI that continues the conversation. We'll be talking about this for a long, long time. We're going to continue to talk about it. Thank you for spending the time and coming on the CUBE Conversation. Really appreciate it. Thank you for having us. Thank you, John. I'm John Furrier, host of theCUBE. We're here in Palo Alto, California for CUBE Conversation. Thanks for watching.