 Welcome back everyone. Day two of SuperCloud 4. It's our second day of wall-to-wall coverage. In studio here in Palo Alto is the Cube studio. I'm John Furrier, Dave Vellante. Rob Streche with Cube Research. Guys, day two, we've got another packed schedule. What a payload of SuperCloud 4 content. The theme is generative AI. As you guys know, we do this every quarter. Folks watching, every quarter we have an episode. We unpack and drop a payload of content from experts, startups, founders, industry experts, people from the community, all weighing in on the hottest trends, the most important stories. Dave, I mean, great day two. And the earnings season for the hyperscalers is kicking in. You saw Alphabet, we saw Google Cloud. We've got Amazon coming up. You saw Microsoft Numbers. Kind of a interesting time. We obviously got the war in Israel, which is kind of putting a really kind of a wet blanket on some of the tech scene, certainly on the startups and Israeli component of those of the tech scene here. So weird market. Are we in a recession? What's the cloud outlook look like? Generative AI has got all the hype. Is the bloom off the rose or is it still hyping? I mean, this is the big conversation we've been having. And it seems to me that from yesterday, it seems like it is like a web moment where people are like so enthusiastic. Confidence levels, I'd say 50-50 around where we are and how we shoot kind of pointed that out. What's the lineup look like for today and what's your analysis of the current market? Well, first of all, I'd say it's kind of playing out how we thought it would play out. Microsoft is winning its relationship with open AI has clearly put it in the front. We say it cut the line and now it's in the lead with AI. Jassy's talking about how it's early innings. It's just the first three steps in a marathon. But the interesting thing is when has Amazon not led? Amazon's always been a front runner here. So that's kind of an interesting dynamic. And Google, who's got the best AI you would think with BigQuery and Vertex AI, it would have a little bit of a tailwind from that, but its quarter suggested not the case with GCP. They were talking about cloud optimization. Azure, unbelievable. Azure's actual constant currency growth was 28%. I had it at 27% and I'm usually a little higher on Azure. And then GCP came in, the overall cloud came in at 22%. My guess on GCP, I had GCP come in at 25%. I'm just squinting through the data and the transcripts. It maybe grew a little bit faster than their overall cloud, but not much. They talked about optimization. So clearly there hasn't been a huge tailwind for these cloud guys yet because of AI with the possible exception a little bit for Microsoft. Rob, what's your take on this cloud game as well? Generally AI obviously is hot, it's hyped. Where's the rubber meeting the road? Where's the meat on the bone, as we always say? How do you see this? I think to Dave's point, I think OpenAI has been able to really help Microsoft jump the line and get people in because they can use that as a way to quickly get to chatbots and things of that nature and build out first prototypes or proof of concept type AI and generative AI products. And I think that has helped them. And I think when people looked at Vertex, it's newer from that perspective. And I think it's not as hyped. And I think the models behind it. I think we're still to use Andy's terms early innings on that because I think a lot of the open models, I think you have WatsonX coming in. You have a number of different walled gardens that are still yet to be built for people to secure their data, protect their intellectual property. And I think that's going to have a big impact on clouds. WatsonX might be the diamond in the rough here. We're actually seeing momentum for WatsonX after all the years of pain that we had to endure listening about Watson and not performing in the market. It actually looks like WatsonX has some really great potential and some momentum behind it. I mean, IBM could get a mulligan here with Watson that kind of failed to meet up to his expectations over the years with NLP. But now with the AI coming in, they're kind of maybe in a pole position. We got to keep an eye on that. Certainly, Dave, love to get your perspective on that. And the other thing I did, OpenAI is actually starting to see some competition as high as they are. They entered the market, like really came in fast. But they've come down a little bit because they've got competition, right? There's alternatives out there. There's third parties that are open source. You say OpenAI AI isn't closed. You said that a lot yesterday. I didn't really understand why you say that. Explain that. Well, okay. Well, I don't get to the Amazon. We'll come back to Amazon. But on your question, when the model started coming out, they were called proprietary models because they're proprietary to OpenAI. But they're actually crawling the web and getting data that's open. So, and as you go down to the power law, if the slide looked at the power law slide, we have that still. Love going to the slides that's there. So the power law, as we've been promoting through our CUBE research initiative, you see as these specialty models come in, all the talk at SuperCloud 4 in General AI is the data is your intellectual property. That's proprietary to the company. So words are, the semantics are, the words are getting flipped upside down. Memory, is it memory? Is it memory for the rag or the retrieval? Is it proprietary? Is it open, walled garden? So what you see in the trend line is on the specialty power law, those sets are going to be smaller, but high fidelity data. And that's going to be valuable. So it's interesting, the proprietary at the top of the head of the tail is OpenAI. So technically it's open. And then they have open in the name. But they were called proprietary because they didn't really share the data other than querying through the API. So it's like, it's just, the industry is in a weird tide moment, Dave. The tide's shifting. It's kind of mid tide right now. And I think that the AI wave coming in is going to completely change the semantics of the definitions of how data is organized. And I said that yesterday. And so, and this is why I think Amazon's flat footed because Amazon's playbook is cloud based. And you brought up the Azure success and how OpenAI kind of cut the line or as we say in the NASCAR, you know, slingshots of the front. And I think Amazon's mistake was their, their, I think they got a little bit cocky. They are conservative with the trends. They'd like to see things develop before they make their move. But OpenAI was such a strong move. They were so flat footed. And then their response from a PR perspective and their analyst relations perspective was anemic. And then if you look at what they're doing now, I see Jassy posting about the Today Show. Their entire PR strategy is focusing on mainstream press. I don't get why they're doing that when all the action's in the industry. So it's like, they're, it's almost as if they got the wrong playbook on comms. What are they trying to convince on the Today Show that Amazon's got good AI? Like, what, like, what? I was going to say, what if somebody's going to go out from the Today Show and go and, you know, my mother is not going to go use something from Amazon. Your developer mom? Yeah, my developer mom is not going to go out. If anything, it's worse to use chatGPT. She can't use, you know, anything from Amazon for this. And I think that's, that's, that is, and we talked about this when we were at Google Next, is when we look at it, Google really has pervasiveness with Vertex. And I think that really has helped it, how it goes through a lot of their services and helps across many different services. Amazon, just based on how the stack is built of the services and how independent all 300 services are, so they can do the two pizza teams and they can go fast, really is going to be a tough thing for them to go and build into that. Even just co-pilots is going to be tough for them to build into these and integrate into these services just based on how they're built. But do you think- But Microsoft's running the table on integrating across the product line, Office 365 and the co-pile. I mean, Microsoft's got the formula down. And the other thing about Microsoft, and one of the guests yesterday, I was talking to him offline, I forget his name, VJ. VJ from, from Howie Shoes Panel. He was saying they made a bet years ago not to scrape email, to fairly double down on privacy. And he said, that hurt us for a while because we had all this data that we could have scraped but we didn't. We said, no, we're going to be true to privacy. And I mean, it sounds very self-serving now. You know, kind of Apple same thing, but I think it's true. Whereas Google, you're typing your email and it's, it's finishing the email for you, suggesting replies. So Google's scanning your emails. Now, I don't think Google's, I kind of trust Google, but maybe I shouldn't. But- What do you think Amazon should do then? Do you think Amazon's picking the wrong approach and comms, you think that they're messaging to the wrong audience? I mean, I think, you know, the today's show and these kinds of outlets are more like, could do more harm to Amazon because it's going to scare people. Like, Amazon's too powerful. What Amazon's always done well, and I, I got to believe it's going to do this at re-invent, is it shows proof points. It shows products and it shows customers that are using them, it trots them up on stage and it convinces people that this stuff actually works and it's the best out there and we're the fastest, we're trusted. We got to see that. And, you know, I've talked to people inside of Amazon. They're still not using their own bedrock. They're like, yeah, we're still testing it out and, you know, the components in bedrock and Titan, it's, we're still out there yet. Yeah, we're using SageMaker, but, so it feels like it's just not ready. I think to that, I think you're dead on with where re-invent's going to land in, you know, about a little over a month from now, right, is that they're going to have proof points but they're going to double down on their community, which is the builder community. That's where they've gotten that real momentum always is from people building. And I think what has happened is people aren't building the complex, huge language model, chat GPT style stuff. They're doing these SLMs, the segmented language models and those small ones. And I think, you know, on the longer tail. So I think they're playing for the long tail. They're not playing for the front end. And I think that's their strategy. I think it's going to hurt them in the, you know, in the run up to getting to the tail. But again, if they play it right, they still have a really good play in there as a platform. I think Amazon, I mean, we can get into the research in a second, but I wanted to say that Amazon should go to their true north star and that is the IaaS. They have to address the cost issue and performance. They nail that right now. That's what I would be saying to Jassy and Dean and Adam, nail the silicon, nail the performance, let the data sets, whether it's in Bedrock or SageMaker, sit in there as an app enabler. So the question that I'm looking at for reinvent and love to get your perspective guys is, what's the apps look like? Okay, because there's a lot of talk about them having solutions. Yeah, the call center, they trot that out all the time. It's contact center, whatever the depth got going on there. But what are the apps going to look like? What are those AI apps going to look like? What, in Gen one cloud, it was SaaS. That was well understood. Post on EC2, that's three. I mean, you do a SaaS app. What's the AI app look like on Amazon? Again, Microsoft has the strategy down. I mean, they've nailed it. They're saying, hey, we have the apps. We're going to layer AI on top of that. Boom, they've got a captive market and they're going to do really, really well. I think it's, they got to go back to their roots. It's the developers because Amazon doesn't have that up the stack, the software portfolio. They rely on developers to go build that stuff and compete as SaaS companies. And so they've got to get back to the core. By the way, when you look at the data and you talk to customers about their intent to use them, they definitely want to use Amazon. Amazon's well positioned. They've got to execute, they've got to deliver and they've got to show it, reinvent that people actually are using this stuff. And then I think they'll do great. Rob, I'm going to see what Swami's going to say, but we heard Bratton say yesterday on the keynote from Amazon, it's not about the models. It's the end to end reaction to that. What's your thoughts? Yeah, I think, you know, again, they're going to look at how do they bring it to SageMaker, how do they bring, you know, more advancements because SageMaker, even with Canvas, was still fairly clunky and convoluted. I think the pricing thing is definitely an issue. They're always about how do we lower the cost and build out apps. I think when you start to look at BI, and I mean, that's Adam's, you know, from his Tableau history and heritage, coming back, I would not be shocked to see some advancements with the AI they'd already announced within that product line. And I think that when you start to look at that end to end, like you're saying, they do have a little bit of a stack on that, but it is the developers. They got to make the tooling that much easier. So I think- It's the picks and shovels market, isn't it? It is, and we're going to hear, I think a lot of co-pilot-ish type stuff for building your bedrock applications with example use cases. And I think that's what they're going to focus on. But I do think Amazon has to do a better job of integrating its data estate, right? It's still got a very bespoke data estate. When you look at what BigQuery's doing with Vertex, AI, what Microsoft's doing with Fabric, what Snowflake, Databricks, they have a much stronger, in my opinion, integration story. Amazon's starting to get there, but they're kind of gluing things together, but it's still, that to me is critical because then it makes it easier to layer AI on top as opposed to have to munch your data. Well, quite literally called gluing, right? Yeah, quite literally called gluing. But I think to that point, that actually may or may not be a disadvantage for them. You think they can leverage that? I think they can leverage that, especially where you have Databricks is so huge within Amazon and you have other estates there and they have so much data. And they benefit from those other partners. I mean, Amazon has nailed the partner and the go-to market. They're blowing Google away. They got ecosystem, no doubt about it because they don't really compete with them much at all. I got to ask you guys the question we asked yesterday to all the experts and I want to get your reaction to the question. So we talked about the step up function of AI, how we brought that up in his panel. We mentioned it to the other founders. Before the cloud, you had to provision a data center, put a box in there if you were to start up. Cloud came in, start up, said, oh, I don't have to just pay for those boxes. I can use my credit card and start a company. We know all the history of the web 2.0 and beyond was cloud-based. So it was clear value between paying for machines, the tax of starting up a company versus the cloud, cloud won and was great. What is the step function value of AI in the cloud now? So take the progression, data center, cloud minimizes that for the startup. What's that benefit that the cloud with AI brings to the startup? That's going to be that step function that's going to differentiate which cloud. Now remember, Amazon was kind of the only cloud at the time, so it was obvious you go to Amazon. The only game in town. Now with the competition, you got Azure, Google, Amazon, and maybe even Oracle in the mix there, the startups now have choice. What's the step function value that makes that decision? So the cloud, it was time to value, right? I mean, it wasn't so much about TCO. Maybe it was, maybe we can argue that, but it was really about agility and time to value. It became so compelling to get stuff done quickly. That was the cloud value proposition. The value proposition of AI is cutting labor costs. I mean, that's it. That's where the step function is going to come in. I can cut headcount or I can reduce the need to hire people and that's going to drop, that's going to increase productivity and drop right down to the bottom line of the income. And specialize talent too. I don't need PhDs to run. But I'll take that and say, okay, back to the startup question, right? I think what you're seeing is expectations out of the venture community for five to six customers out of the gate to go get your seed funding and things of that nature. And what it's helping people do or not is build that first alpha beta product out of the gate without having to have the massive team. So to your point about labor costs and things of that nature, when you start to look at it, it's augmented. It's how do I get to that first, that minimally lovable product out of the gate? Awesome. I think we should probably set up day two. Well, no, well, I want to spend the last minute just taking a quick public service announcement. We still can angle the cube and Wikibon has been in the three brands. We have the new cube research that's going to replace Wikibon. I want you guys to explain the last minute we have here. What is the cube research? Rob, you're working on that, Dave. What is the cube research? What do we bring it to customers? And what are we following? Yeah, I mean, we're adding to our capabilities, which we have the cube, the cube community, the insights that we get from the cube. We have a partnership with our data partner, ETR. And we're layering value on top of that to help people. We got research, we've got advisory, we do private consulting, we help them get the word out with the cube. We're combining those in what is essentially the industry's best service to take data, data-driven insights and broadcast to the world. Rob, we've been facing this quite, we haven't really launched it, but what's your experience? What are we offering? What are you talking to customers about? Yeah, I think a lot of, it's meeting the customers and providing them value. I think typically in advisory services, you would get a subscription or an annual contract and it's, hey, we say nice things about you. What we're focused on is really ROI to those customers and helping them build their go-to markets, helping them get the word out and helping them understand how they're competing within those markets. And I think the fact is, we see so much of that signal and it's mixed up with the noise and it's extracting that as you guys would always say and how do you help them understand this is where you position yourself and doing it in an honest and way that has integrity as well. I think that's- And then we've got the technology, we've got the video technology, the video data lake, the AI coming out, combining the technology, the research and the video. I'm looking forward to hearing more on news about the research. All right, day two, setting it up. We've got a great lineup of experts. SuperCloud 4 day two, so much coverage being dropped in these two days on General AI, so much there. Share it with your friends, check it out, consume the content. It's going to be a great community site. You can update two starts right now.