 Welcome back, everyone, to theCUBE's coverage. Here at Reinvent, we're on location for AWS's Reinvent Annual Conference, our 11th year at theCUBE. It's been great to watch the journey of Amazon go from, you know, not very well known, 13 years ago, 11 years going to theCUBE, and now it's like an industry event. They got so much traffic, so much going on, and the game's changed. The next gen cloud is here. It's Gen of AI, a big driver of a new experience, new expectations, role of data, chip wars, price performance, whole new layer of activity to be analyzed. Of course, we've got a great analyst panel here in the old CUBE Collective. I'm John Furrier, your host. We've got Shelly Kramer with theCUBE Research. Welcome aboard in Z as Carvella with Carvella Research, part of our CUBE friend group. Great to have you on, and also contribute to SiliconANGLE. Great stories this week on SiliconANGLE. Thanks for getting that. What to expect, you nailed it. Yeah, well, everyone always loves this event, and so get to put out a little preview piece. Let's get into it. The analyst perspective here, you guys got the briefings on Wednesday. I got the exclusive with Adam. Pretty much laid out. He had to lay down a good keynote because they were getting hammered in the press on reputation. They don't have AI, they're late to the game. Microsoft made a power move with open AI, chat GPT, selling like crazy, but the game was just a front end. There's a lot going on in the covers, so he had to lay down as a killer keynote. It was almost like a call to arms. Well, they're not a marketing organization. There's one thing they could do better is that, if there's an area that they're behind in AI, it's marketing AI, because I do think their capabilities are, I mean, they've been using AI for, as long as they've been around, right? Exactly. To say they're behind in AI, I think it's a little trite, but certainly their Redmond-based competitor there does a, in fact, the whole debacle with open AI, they say there's those things bad press, that raised the awareness of what Microsoft was doing better than anything else Microsoft could have done. You could say it's kind of negative press, positive press, but it will work for them. And I think Adam did deliver a great keynote, though. There was so much news in there, and so much of it was folks in AI. But it was interesting, though, that with AI being the theme of the day every day, across every event, it doesn't matter. It was really interesting to me that he didn't lead with AI in the keynote. It kind of came later. Yeah, it was interesting. I mean, I've observed Amazon, they're quirky with their culture. They love their keynotes, they get the bands, but they always start out on the same order of magnitude. Customer slide, yeah, and then they build up the drumbeat. But he started to start with storage. I think what he wanted to do is get it out of the way, because they had some upgrades on the storage, and we heard some from VMware today, yesterday, I mean, the decoupling, the disaggregation of memory and storage is happening. He's starting to see kind of the non-core AI stuff changing significantly, which then rolls into this core advantage when he talked about the connective tissue with Nitro, and what was that MV, the connection between this GPUs that- Oh, NVLink. NVLink, yeah, and that's, I mean, that's putting together some pretty good stuff around the chips. I think to me that was the hidden gem. And I think- I actually like starting with storage because the theme is reinvention, and if you think of an industry where there's no room to reinvent, storage would be one of them, but it does show if you're willing to continue to invest, you can find ways to innovate and things that, you know, I actually believe there's no real markets that are IT that are commodities, as long as you continue to drive more R and D into it. So for that reason, I actually thought it was a pretty clever way to drive the reinvent message. It's interesting you picked up on that because now that you say that, he did kind of make a comment, like, oh, even storage. Yeah. If we could- You're excited. Let's start, the hardest thing to reinvent, let's start there. So it's really how he flips again. Yeah, that makes sense. Amazon's culture is, and also he laid down like, we're not going to change because of the public opinion. We are who we are. This is what we do. We work back from the customer and we reinvent, which means that they don't get it right. They just scrap it. I mean, I remember Andy Jassy said, all workloads of moving to the cloud and then they launch outposts. Yeah, yeah. Remember that? Well, that's Andy. You said, that's gonna- Well, we listened to the customer. Well, you're remembering that Q&A with the analysts that you had a lot of questions about why they launched outposts and his answer was, because customers wanted us to, right? And that's the driving force here. Right. And I was talking with Pasquale de Mayo, the GMO Connect, right? The contact center product. And he was saying that too, that they don't, you know, while all the analyst accolades and things are nice, they really focus on what's right for the customer and then they'll let the market play out the way it does. Well, let's analyze, let's analyze what's going on at Reinvent. Let's get into it more. Obviously, you mentioned the layers of the stack. I thought that was interesting. The middleware there, you got the three layers, infrastructure, LLM layer, they call foundation models layer and then the app layer, which essentially, for them, it's a version of, I think, a shim layer connective tissue to enable apps. I mean, you see Q, which is a phenomenal demo. I thought that was a killer demo. Yeah, it's really exciting. I thought Matt Wood did a great demo on his piece, fast, tight. What's your guys' analysis of what they're doing here? Yeah, I really like the play they have with Bedrock because it does give customers choice, right? And I think this is, it's going to be interesting to see how the market plays out, but generally, if you give customers choice, that's better for them and you wind up the winner. I've talked to a couple customers here, actually, that had started using open AI and they moved to Bedrock because they didn't want to be locked into one model and the one thing that I can tell you is we have no idea what's coming and so the ability to go pick and choose the model you want for the specific use case is actually better for the customer. That kind of flexibility, I think, you know, is going to serve them well long-term. Well, and as are the guardrails for Bedrock that they announced. And I think that's a huge deal and we talked about this earlier, sort of the role, the importance of security and having security baked in foundationally and I think that's what we see. You know, safe user experiences, that's what everybody wants and needs and so I think that was a really important thing. And you pointed out, Shelley, when we were on our earlier interview that the race to the top vibe was very strong. I love, I mean, no one wants the race to the bottom but you can say chips are going there but not really they're being differentiated. So it's a whole other ball game and I think another interesting point I want to get your guys thoughts on, I don't know if you caught this, but they kind of pulled a Microsoft on Microsoft. Project, the big, the announcement was Q, okay? Well, Project Q is the big thing in OpenAI that's around the corner that there has been talking about. So I wonder if they wanted to get the name out there. Well, they kind of like, is it the same project? Well, Q's been used with QuickSight now for a while and I was just talking to QuickSight and they said, Quick, Q grew up, right? So instead of it being used for a tool just for database analysis, now it's being used for everything across Amazon. So in this case, Q was named after the James Bond Q, right? Who's the assistant in the gadgets. But it is, I do think having a single, generative AI client there that works across all your products. Yes, all the layers of the stack. It makes them, and it furthers, we talked about this last year, right? One of the things Amazon's doing better now is they're having better integration across all their building blocks and I think Q's a good example of that. Well, and I think too, when cost is such an issue and really maximizing not only effectiveness but minimizing cost as you can, I think this is an important thing with that. I mean, I'm trying to figure out what's real in the demos because there was some stuff there that I was squinting through, I'm like, okay, that business thing looked too good. I mean, three steps and you're in. I mean, where do you think, how real do you think that is, in your opinion? I think it's very real. I think that's the power of natural language, right? There was a lot of things that you just couldn't have done a couple of years ago but now with just a couple of steps, if you want to query your sales tool and ask it what my most profitable region is, a line of business person can do that instead of having to have an analytics person in the middle, right? That's amazing. From a generative perspective, it's the rise in natural language and its ability to bring content together that actually takes something that was a whole bunch of steps and makes it three steps. So Dave and I were debating on our keynote analysis this morning, how much he left on the table and how much meat on the bone he left for Swami tomorrow. So Swami's keynote's tomorrow. So I'm thinking the zero ETL is kind of a telegraph to what might become behind the scenes because to make all that work, you got to have a data management strategy. So what's that look like in the new era of generative AI? Yeah, so the zero ETL vision actually is you leave your data where it is, right? But then you got to create the middleware, if you want to call it that, to manage it, right? And so Clean Room does that with multiple companies, but they do need something, I think, to tile those different data. Because that's the one thing that I think plagued companies today when I was talking on one of the big hotel chains about it and they were saying that we have data for this, data for this, data for this. And so they're trying to normalize it, aggregate it. They put it all up in the cloud and I said, is that scary? He goes, yeah, it scares the crap out of me because now I'm at a single point of failure, right? But I think if companies are going to actually try and use data as their competitive advantage, they got to find a way to bring those data sets together. Because that is, to me, that's the killer of what companies are dealing with right now is their data. Yeah. Use your experience. You're starting to see that shift with the generative interfaces, streaming answers, the demos I thought with the code whisper were phenomenal. Everything is of use, yeah. Everything's simple, ease of use. You know, really kind of democratizing. I kind of hate to use that word because I feel like it's really overused a lot. That's the right word, though. But I feel like I was really excited about Amazon Connect, Amazon Q Connect, and really, you know, the opportunities that exist for Amazon in Contact Center, which I know you and I both talk about a lot. I mean, there is so much opportunity here for Amazon to really win in this space for a variety of reasons. And you know, one is it's ubiquity, right? How many companies are running on AWS? And so with a solution like this, and then with the queue baked into that and making all of this AI-powered, it's pretty exciting. Yeah, I mean, I think we're going to see some stuff tomorrow. I hope to see more custom models like healthcare, some of these industry-specific, they kind of tease some industry stats out. I thought the Pfizer component was great there shown there. And then, to me, what jumped out of me at the end there was when they did the demo of JavaScript, 1,000 applications over two days. Yeah. Okay, I'm like, okay, that's a game changer, number one. And then they weaved in dot net to Linux. Yes. Okay, that's going to be a huge reduction in licensing fees. If you're an enterprise, that is like a silver, that is a silver bullet. It's so exciting. I mean, that with the impact of the cost side, you got that kind of code migration efficiency, productivity, the chips now with energy, just energy saves, you want to stack a bunch of GPUs with Nvidia, you have two choices. Put it together on your own or go in with the new MV link cluster. I mean, the cost savings on that alone and performance. So we get the performance game going on here, speeds and feeds are back. One of the interesting things about Amazon and the connect product to a good example of this is there's a lot of people out there that think connect is a builder led product. And when you say Amazon to people, everything is builder led, but they've been doing more and more work to actually, obviously there's always going to be a builder led component, right? But more, you know, with more and more of their products, line of business can use it. And I think connects a great example of that where you can use a turnkey out of the box. Business intelligence at your fingertips, right there. They've got a low code, no code, so your software power user can use it. And I think that the addressable market of who can use Amazon is growing from just developer, right, to your average software power user, now to line of business. And that actually puts them in a much different position competitively right now, you know, you can see them competing with like the big application vendors that tend to have a lot more CEO focus than an infrastructure provider would. You guys have been in the industry for many waves, you've seen the major inflection points. What's different here because we're talking about next gen that's legit next level here. We're talking about what is going to happen next? What do you think is going to happen? Obviously, Nathan, I compared to the web, you had online search provides like AOL, they were proprietary, the worldwide web with standards. There's all kinds of ways things to discuss, open versus closed. Is it similar, you know, maybe not, but there's a big debate. So the hugging face founder on Twitter is saying, here's what's going on in open. It's not just open source, it's open models, right? The power law, Silevsky basically validated our power law of specialty models. Okay, I agree with that. I think choice, multiple models integrating together, that middle layer. What happens next? What do you guys see? Because we're still embryonic in this whole shift. Well, over time, any market you've been in, when it goes open, that's when the hockey stick adoption curve, right? So I think open is the right approach. I do think your ability to be kind of a combination of open and turnkey, and Vinny is a great example of that, right? Everything they build, they put a reference architecture together, and that's why arm was so important to them, and but that reference architecture, then they take to their OEMs and say, you can build it as well. And so the customers get the benefit of a turnkey solution in an open model, so they still have choice, but they get the performance benefits of turnkey. And that's not, we've never really seen that in IT before, right? The best of both worlds are. And what about developer enablement? Obviously there's some code whisperer, you should mention call center. Not a lot of app conversations, so I think Silevsky's going to lean in towards letting the ecosystem do its thing. What do you guys see developing at the top of the stack? Well, I like what they've been doing. I like what they've been doing in apps, because for a lot of their apps, like you remember they launched supply chain last year, they're just, they're apps they use internally, right? And then customers go, wait, we like that? Yeah, I kind of like to what you do that, and they flip it around, and it's already fully based. Yeah, and they sell it as a service, and because it's consumption based, their bear entry is really low, right? And then you just start to build it from there. And I think there's, I'm trying to figure out what the roadmap is here, because they could get into logistics, they could get into HR, I mean, they build a lot of their own applications, and I think this is an area I was disappointed we didn't see any atoms. That was one of the points I brought up in my re-invent preview. I think more and more we start seeing AWS roll out Sass apps that are targeted at that line of business person, as well as continuing down the path of developer. Developer role's going to change though, and this is something I'm curious to see how well their developers long onto that, because if indeed you can do 1,000 Java app, you don't have migration to that in two days, that would have been, if you think of the team of people that would have been working on that for years before, so that whole role is going to change. Yeah, I mean, it's a whole nother ball. You can't fight it though, right? Yeah, it's good for the company. No, it doesn't make sense to fight it. And I think too, one of the things that Soliski said is really, their focus is all about reinventing the future of work. That's a big bucket, okay? But everything across the whole tech stack, all the infrastructures, so much of this. And I think that that kind of telegraphed a little bit of what's ahead. And I think that's right on the money, because that's the key app is the user experience. And Amazon's DNA is remove all the heavy lifting to get into business. And here, to your point, it's not just techies doing a startup, and it's the choice between data center or put my credit card down and build the next Dropbox or Airbnb. Now it's like, well, I could actually revolutionize my department or company with the creative idea, test it out, but it's more scale. I mean, the scale's a huge advantage here. I think we're really close to living in a world where in natural language, you see, I need an app to do this, this, and this. Here's your options. And it goes and builds it for you. Yeah, yeah. I mean, I remember the 3D printer, first time I saw a 3D printer, I really fell out of my chair and said, this is incredible. I think that's the kind of moment we're kind of seeing now where it just happens. Deploy me a web app that does some, or, you know. Well, and again, built on a foundation of security and respect for privacy and data privacy and all of that. And I think that is what certainly CISOs are looking for. But that's really such an important consideration that I think that having that be your foundational strategy. Well, I think we're going to have a lot of conversation around the data strategy, around to unpack what this is all going to mean, latency, the physics, the brave new world. It's the battle for AI supremacy. Thanks for coming on and packing this with the analyst angle. We're back with more coverage here in Las Vegas, back to the studio. Thanks for watching, stay with us.