 Welcome back, everyone, to theCUBE's live coverage here at San Francisco, Google Next 23. This is theCUBE's team coverage. I'm your host, John Furrier, with my two analysts co-hosts here, Rob Streche, heading up theCUBE Collective Research and Dustin Kirkland, CUBE analyst contributor. Lisa Martin is here as well, and we've got the team, Rob Hove, Mark Albuson, the entire SiliconANGLE team on the floor, getting the stories. No stories too small for theCUBE. We're going to break it down here as the keynote analysis, day two segment. And also, we're also going to analyze, zoom out and look at the big picture relative to the cloud industry and some of the technologies that are folding into place for each of the hyperscalers, AWS, Azure, and Google. Google's doing great, we're day two. A lot of enthusiasm, Rob and Dustin. I got to say, I'm enthused. I'm converted over to Google Cloud in terms of my appreciation for their momentum. You know, when I see winning hands, I like to look at it and stare at the cards and go, you know, they're doing a good job, and I like how they're playing their cards with the cloud. I think they're smart. They got a lot more work to do. They're clearly number three in the marketplace, but they are hitting all the right notes, in my opinion. What's your take on the keynote analysis? This is about developers. It was an interesting keynote. They had to, in Google Flare, you never see Amazon do this, by the way. They do walkout songs, but you never see Amazon do a musical, Rob. Live performance, on stage, three live singers. There was a piano, there was a sousaphone. It was the real deal. There was a tuba, chief tuba player. I've never seen a chief tuba officer, chief tuba player, or whatever it was, but yeah, I thought it was interesting that they were leaning into the legacy, you know, legacy is bad kind of thing. I mean, I don't know how that's gonna play over with the people in the audience. With the developers, it definitely plays well, but when you start to look at it and go, having been in the insurance industry, and I had something that was running on NTFORM, we had no idea who had actually built the app, and you had to kind of still, we called those heritage apps, not legacy apps, so it'll be interesting to see how it plays. Let's explain what happened first. First, they opened the keynote with a musical live performance of artists, Googlers, I think, they were Googlers, I believe. Yeah, all of them, yeah, our former Googlers. Classic Google, a lot of color, and I love that in there, a lot of commentary, a lot of flair, a musical called Living in Legacy Land, where there was a singing and dancing around, legacy infrastructure, clever script, and then they weaved in the demos, but they also weaved in the lower stories, the lower of Google, kind of showing that they have chops. So legacy that flexing their muscle, Living in Legacy Land, again, that was the focus of the entire keynote, how to manage legacy environments, and the pivot was build your legacy. So, interesting theme, I kind of liked the positioning, I thought the musical was kind of cringe worthy, but okay, that's me, that's me, but I did like the approach. Well, if legacy is very relatable, and I think that was what's pretty incredible. Yesterday was all about here's what's next, here's what's in beta, here's what's just hit GA. Look, if you're a developer and have access to write something new from scratch, it's quite a blessing, it's quite freeing to start something from scratch. The truth is though, that in the enterprise, most developers aren't tasked with, here's a blank slate, product manager development team go build something new. Most developers are tasked with, there's something out there running, go make it better, and you're going to inherit something, you're going to maintain something, you're maybe going to migrate something. So there was something very relatable about this today's keynote theme. Yeah, I think it was relevant, and what they didn't really tie, at least in my mind effectively too, was the DevOps movement. They did bring SREs up on stage with the champions, they were SREs, they called them champion innovators. It wasn't clear if they were Google employees, or part of the community, like VMware has Vexperts, so it clearly had the stakeholders. But I liked how the DevOps angles tied in there, because if you think about DevOps, IT, remember the old stat rob, 70% of the dollars go to running the business, versus investing in. So operating IT used to be kind of a sunk cost. DevOps are essentially running things. So yeah. I think they got into it when they started doing the demo of Cloud Run and Cloud Deploy, and they got into the Dora metrics a little bit, and to your point, I think they could have gone deeper on that stuff, but I thought, again, to Dustin's point, you're not going to really have those opportunities to build net new apps that often. Yeah. When you can, it's great, but. It's refactoring, where maybe a piece of the app is being rebuilt. So I think that the demos, although they talked about legacy and converting legacy, didn't really show converting, they showed building new. And I think that was a miss, but I thought it was, I mean, the demos were good. The app sheet resonated with me, of all the things there, because I think there's instances where you need to put a front end on something that you have plumbing for. So if you have APIs, and it made me start to think about like, did I architect that right? Am I set up for that? So it kind of, there's an underlying question that you're asked, like, can I even do that? Like, okay, where's my, it's in reports, okay, they have that access to data. So do I have that API pipe? Yeah. What webhugs are they in place? How many times I went and built macros in Excel to go and do something, and this would have been so much easier. Well, this is a place where I think Google can really dominate the low code, no code, you know, minimal code. One of the cloud champions that was on stage, she introduced herself as a teacher. I got the impression she's not a Googler. She's, you know, sort of volunteering a community, a power user, but she brought a hell of a lot of expertise about a particular problem and then was able to draft an application with a handful of, you know, narrative voice prompts. I mean, it was scripted, of course, but, you know, it's got the wheels turning that like, this low code, no code environment, Google has built a ton of brilliant infrastructure around. Yeah, I mean, I thought the first day two keynote, I didn't really give it high marks, at least from a content perspective. I did like how they put that theme out there, legacy. I did like that. I think that's important. I think you had a good point that there's a lot of things that need to get fixed that people are working on. And that begs the question to our question yesterday. Was that a developer conversation? Or was that more of a DevOps conversation? I mean, there was some SRE in there for sure. John mentioned the lore. Of course, Google, like any company, has, you know, legends and lore and horror stories. And we were treated to a couple of those from Google, one in particular, one of the SREs responsible for helping identify and patch the log4j vulnerability back from December of 2021. She gave a pretty detailed play-by-play. You raised your hands that you remember where you were at that moment. I remember the very moment and exactly where I was. Log4j day, for sure. No, I think also in the lore where they talked about the ARP storm and how that really caused them to go out and build their own load balancing. I think it became a very interesting, I thought there was a dichotomy there about, hey, we couldn't get the vendor to be named later, to go and fix their load balancers. So we went out and built one in software. I think they're, you know, hitting on. That's a very Google response. By the way, I mean, there are a few companies in the world can say, you know, this cornerstone of technology is broken, so let's just rebuild it from scratch. But what they went into, and I kind of wrote it down into kind of how to develop faster, right? And they kind of had that theme for developers around plugability, open source and APIs, and reuse of extensibility layers. I thought to me, that was them going back to their roots. And you would know from being here, I mean, that to me was like motherhood and apple pie for Google, reaching out to developers and saying, and also to other people, bringing them in and saying, hey, the AI is going to help you do these three layers and keep up with that. I mean, again, they got into a lot of the security stuff, which I thought was interesting. To your point, I thought, you know, it was the same. She has ex-Google, what did you think of the keynote? So, I mean, I've been on that stage, or stages like that. I know what goes into the preparation. A lot of efforts put into selecting good speakers, internal, external. The rehearsals go on for weeks. There are people here that haven't slept, you know, for the last few nights. I saw one former colleague nodding off because I'm sure she's been working around the clock on this. So a lot of work goes in. There's a lot of prep. It's polished, it's live. You know, that's also, you know, kind of a key differentiator. Stuff goes wrong. In general, you know, I thought the topic was relatable. I, for, you know, it was a little cheesy, but I enjoyed the effort that went into the musical performance. I wasn't, I'm not too hardcore on it, but, you know, I want to see more meat in the bone. I was hoping that I'd see much more stronger developer posture. Like, look it, here's some bang, bang up, new code, here's some AI, go. You know, it was a stretch with app, well, the app thing to say that's AI. I mean, I guess it's AI, but I think it's like chatbot. They went and tied it into Vertex and how they were able to do the image recognition and build the app that could, you know, go and say, hey, what, you know, describe what's here. I think that's a small chatbot like use of it, but it was interesting how they showed you that there was different things under the hood happening, even though it looked like it was all in the same interface. I thought to me, that's how you build a solution, and I thought that's one of two good strengths. I think my big takeaway was that that made me think, are you set up to do those things? Because I think they're showing a use case of dynamic building stuff, right? And I think the question is, are your data sets available? What's your architecture look like? And then this comes back down to the whole cloud versus, you know, selection. I mean, we were talking with John Truro from Montreal yesterday, and we had this discussion around, do you pick your platform? Like, are we getting at the point where it's like, Google does this, AWS does that, Microsoft does that, because at some point, when did they cross over and being an island? I mean, there's no multi-cloud talk here, you know why? Because it's Google Cloud, they don't want you to talk about multi-cloud. Because they don't, I mean, we talked about Red Hat, but I mean, for the most part, Amazon's not going to talk about multi-cloud and reinvent, and certainly Azure, I mean, they're like all in on open AI, so it's like, it just seems it's a weird time in the cloud game. Yeah, you know, I think part of the AI approach, and we talked about this yesterday, it's a great opportunity for Google to take advantage of. Google has so much of the data necessary to train the models, to initiate the models, now adding, you know, this garden of models right there, and then the user and developer productivity. One theme has been trying to do more with less, and we hear about that a lot, like hey, let the AI help you do more with less when a time we're downsizing and cutting. I think I'm actually looking forward to doing more with more, but those things, that work well, why do more with less? If that's working well, and you can multiply the productivity of those people with that, then put more people with more bots helping get that work done. I think, again, I felt a little differently on the demos. I thought the demos were, to your point, having been on a stage similar to that, those are really hard to do live, and the fact that they were flying without a net, I give them a lot of credit. I think there's only a couple little hiccups, but I think they were showing how to use the AI in interesting ways, and that was to me, and simply. Better than vaporware. Yeah, well, I'll tell you what, here's what I really wanna see, and it'll probably be next year before we're able to answer this question. So the idea that these code assistants, co-pilots, duet, Cody can generate a working demo is awesome. And as a former engineer, now product manager, the idea of taking some basic requirements and spitting out a prototype, a proof of concept, that I get. What I don't yet get is, is that gonna write the production code that builds the next Google search engine, that builds the next expense report SaaS. Can we take it all the way there? And I think we're gonna need six to 12 months of using it in anger. We gotta check it out, that's a great point. I mean, hey duet, solve my technical debt problem. Yeah. Boom. That's the holy grail. That was the piece that was missing for me in the demos is they talk about legacy land, and it's how do you get from legacy land to non-legacy land? If they had taken some old code and put it in there and then redone it, to your point, using it in anger, I think that to me is where this could be. Help me understand this code because there's so much legacy code that was written by people who are no longer with your company. So I played with Bard maybe six months ago with a partner. I've got some legacy code written in 1998 for dividing up the bills when you go to a restaurant. We used it for roommate splitting. This is ancient PHP in post-credit code. Venmo. No, no, no, no. This was years before that. But I thought about, you know, let me ask Bard in Vertex AI to port this forward. And the little stuff that got right, it helped me re-architect the database schema. But boy, getting into all of those nuances in corner cases that took years to get right, I have found myself needling around in new code that was working well enough. I think that's a great use case. And I think that's the one that's going to be the most exciting is that you train the code as a language model. I mean, they call it programming languages for a reason. So if it knows the code, it should be able to engineer the code. It's going to be very easy to see. And then how do you train that? What's good code? And so you got to kind of play with that. And you brought that up yesterday, but I said, get the good code or the good data. And you're like, actually the bad data is good too. Because that's what bad data is. Don't do that. That's right. I mean, we saw a bit about security vulnerabilities in the scanning in the demo. I think that's another great use case of, don't do, it's more than just trying to do more with less. Do more with more. If this solves a set of security problems, pour gasoline on that fire and let it run. Yeah, and to our discussion yesterday, the fact that when they were showing those vulnerabilities and also going into GKE enterprise and showing, hey, here's your security posture, your secure, secure software supply chain and getting into some of that and what's in the containers. I thought that was pretty interesting stuff. I didn't get enough of, okay, where is that running? Like if it's a SaaS delivered management station, what if I'm air gapped in where I'm running GKE enterprise or something like that? How does that run? There's still a lot of pieces that are missing that I didn't get out of the keynote and there was nothing on data. Once again, I was a little bit disappointed coming out of it that there's this whole gap between the infrastructure and the app, which is the data. And okay, they talked about how you can engineer, you can use the AI to go through and cycle through a Postgres database. Okay, well, that's great, but that really didn't get to, how do you better use the data that is underlying to AI? How do I connect this all together? Because it's not all going to live in Google. Well, the good news there is we've had some fantastic guests this week so far who've come and talked about some of those third-party solutions around data, data pipelines, data streaming. I can think of a couple just right off the top of my head. We've got security conversation coming on with Sunil Padi coming on and Mandian is mentioned too. They have their own event coming up. The queue will be there in DC for the Google Mandian Security Conference and we have the heavy hitters on the infrastructure site coming on, Sachin Gupta and Mark Lohmeier from Google Infrastructure, the new TPUs are coming on. So they got a lot going on, right? We're going to keep analyzing. We'll wrap up at the end of the day today. We'll look at the stack and we'll identify where the gaps are, we think they need to work on. But clearly, there's some things they got to work on. I mean, obviously they're not 100% there, but it's looking good, looking good. So day two keynote about developers. It really is about the use cases. Again, this is about the roadmap of Google. This is kind of where they're going. You can kind of see the 20 mile stair, if you will, and kind of what the direction is. You can see these things filling in and use case demos are very helpful, but they're putting it to work. They're not producing reference architectures. They're putting stuff out there and they're building. So again, Google thumbs up on keynote. Day one was smashing. I thought day two was obviously hard to beat day one. But guys, thanks for the analysis. All right, day two continues. As we roll into the afternoon of theCUBE here in Moscone for Google Cloud Next, this is theCUBE. Stay with us for more coverage after this short break.