 Welcome back, everyone. Cube's live coverage here in San Francisco, California for Google Next 23. I'm John Furrier here with Dustin Kirkland, our Cube analyst, our next guest. Cube with me, Gabe Monroy. VP of developer experience at Google Cloud. Great to see you. Thanks for coming back on theCUBE. Great to see you. It's great to be here. Normally we're at KubeCon, getting in the weeds and Kubernetes, but a lot's going on since we've last chatted with you and Google Cloud specifically. This event here really showcasing not just so much the cool demos and features and news, but really the roadmap to the future. And I think the AI generation's here. We're seeing that in the frothy and intoxicated open-source developer market. There's 90-something startups in San Francisco, booming, there's more meetups that I've seen since really the web 2.0 days. A lot of creative energy going into the developer market right now on AI. All under the age of pretty much 30. Except for us old guys who are feeling like the fountain of youth. 25 again. So this is really the developer market. It's certainly the boardroom to the dorm room, we say in theCUBE, but boardroom is we need AI to fuse and everything. Go make it happen, okay? Roadblocks, compliance, all in legal. We're hearing stories, but the real action is the bottoms up too. Top down, give the directive, put AI in everything. Bottoms up is just organic developer action. You're in the middle of it. Duet has been impressive here. What's your current view right now? Of the market, and then talk about the news with Duet. Sure, well I think as you're pointing out, those of us who've been around the industry for a while, we've seen these big trends come and go, right? And if you've been in this market for a minute, you can tell this is something real. Like this innovation is real. The customer benefits are real. When you look at the keynote demos that we showed on stage, and already the customer proof points that we have, this is really going to impact not just customers, but I don't need the economy overall, right? It's going to be very impacted by this. And so it's a very exciting time. I have to say, from the inside at Google, we feel very confident about our ability to leverage this technology and deliver a pace of value for customers that is going to be a little different than we've seen in the past. The speed is going to pick up the intensity, is going to pick up quite a bit. In terms of the news itself, we did announce an expanded preview for Duet AI and Google Cloud, which is super exciting. A lot of new capabilities, a lot of which were shown on the keynote stage today, but these are really things around standard developer-oriented experiences around code generation and code explanation. But new things like source citation, right? If you're getting a response from an LLM in your IDE, if you're quoting at length from a particular source, you want to know that that came from Stack Overflow or that is licensed under a particular open source license, permissive or not, right? You want to know that stuff. Lots of cool features like that that we're delivering now. The thing I want to ask is real expect than in the news. So availability on workspaces and just give quick highlights on the specifics. Yeah, so Duet AI is the Google Cloud brand for AI assistance, right? And Duet is showing up in workspace, right? Which is really around productivity. That is now generally available. And the pricing's available, et cetera. For Google Cloud, we're announcing an expanded preview of this. So we've had a lot of folks have been kicking the tires. You'll hear about Wayfair in my app dev spotlight session. Lots of customers who have a lot of great examples of how they're using this today. So that's been kind of the last few months since IO. But now we're going to open up the doors much more widely with a lot of new capabilities as well. That's where we're at on the Cloud side. Got it, okay. So we've been speculating, Dustin and I have been talking off camera and also on the preview, on the keynote analysis, that obviously Google is well positioned for machine learning. We've been covering Google and I know everyone there pretty much. Since the early days, there's a ton of machine learning in AI and a lot of other parts of Google. Yes, Google now cloudifying that. This seems to be the moment here with the AI positioning. As developers adopt AI, which is clearly on the LLM side foundation model, seeing that new ways to do say embeddings, for instance with vector database, which opens up the door to things like what you announced with extensions, right? Which folks aren't familiar with. It's pretty cool. And also you got connectors, all kinds of ways to connect data. The developer experience is going to move fast and be different. How would you characterize the developer change or new things that they'll see from Google Cloud in the development world that's different now and where it might be something that surprises them or a benefit to them? In other words, what's different? What's going to be different from their world? There's a lot of things, I would say. Just one example that you saw in the keynote demo was going from an incident that is being reported on a service issue all the way through resolving that. That usually crosses lots of different products historically. It's not always a friction-free process. And what you can see with Duet and with Generative AI, we're making that end-to-end journey seamless. And that's something a lot of people are doing day in, day out, particularly developers have to operate the software that they're running in the cloud. But when you think about the IDE experiences, the way that folks are going to author code, author, for example, unit tests, right? Unit test authoring is something that LLMs are actually extremely good at. And that's a lot of toil and a lot of burden that folks don't want to have to engage in and they'd rather spend time writing business logic. And actually Dustin and I were talking about this before the show and I think one of the things we're likely to see here is the pace of innovation for developers using technology like Duet is going to increase so much the cost of building a new prototype or testing a new idea is going to really shrink down. I'm really excited to see what that's going to do for the overall rate of innovation in this market. I mean, you're talking about developers and that developer experience is pretty incredible. But I think you're actually democratizing this to some extent to non-full-time developers to be able to take a concept and bring that to market. How does Duet fit into that world? It's a really great point and one of the things that's especially exciting and you know, frankly still early is this idea of, hey, can folks who are not an expert in one domain you know, be good enough in that domain, right? And so, I'll give you an example. In our app dev spotlight session, we have a video that we cut of folks who are using Duet AI in the trusted tester program. And the last stands of the video is an operator, right? It's like a system administrator type. And they say, you know, I'm not a developer by trade, but with Duet AI, I was able to dabble enough in the code, able to debug an issue, that it made me feel almost like I was a trained software engineer, right? And so there's these questions of, you know, what is that bleed through going to look like and how much more productive can we make folks not just in their discipline, but how much can we democratize the adjacent disciplines, right? Cyber security is another great example, right? We want security to be core to every app. We believe that fundamentally in Google Cloud that security is core to our design principles. I want developers to be able to understand how they can use security tools in a way that isn't going to be disrupted to their workflow. On the app modernization wave, obviously that's been discussed, I would say pre AI, that was cloud. As NextGen cloud becomes AI cloud, what does the app modernization mean to you guys? I mean, how do you frame that now? Obviously the demo you referred to was pretty compelling. You showed some automation in there. We were talking about it before we came on. That crosses, that's hard to do because you got compliance and legal issues and you let people work together with data across applications and platforms. What's modernization? What is the modern app from an AI perspective? How would you classify that? You know, it's a great question and one of the things that we're seeing now with our customer engagements is every app is going to become a generative AI app in some way or it's going to have generative AI and some AI capabilities as part of the broader stack. And that's creating this incentive in enterprises to take a fresh look at applications that might not have been on the table for getting modernized. Right now every app is on the table for getting modernized and one of the cool things is Google Cloud is now a part of these conversations in a very real way because these companies, they want to make sure that they're partnering with someone who's going to have their back over the next three, four, five years as the rest of this journey plays out. I think it puts Google Cloud in a very good position. But more generally, you're seeing in the keynote demos a lot of tooling specifically applied towards making app modernization easier, right? We saw in the demo migrating a date time stamp from Oracle format into a format that supports AlloyDB, right? That's a small thing on some level but when you're doing that at scale as part of a broad enterprise digital transformation program those little capabilities add up to meaningful cost savings. Can you connect the dots and go all the way from that front end experience all the way back to the data, the BigQuery, the looker, everything that Google Cloud and data is built upon? Yeah, I mean look, the way I look at this is the data and things like BigQuery, transactional databases, operational databases, these are synonymous with the business criticality of workloads, right? And so for us, we want to make sure that we're not just getting the compute parts of workloads, because that's interesting to customers but the data is where the real value is. It's really synonymous with value. So we want to make sure that we're actually stitching together products experiences that span from that inner loop of coding, how I build an application, how I debug, how I operate that application, but also how I store data. What's my schema layout? How do I think about indexing and the query optimization? And there's a set of seamless experiences that we can deliver around optimization. And can help optimize that. Absolutely, these are capabilities we already have in the Duet preview today. It's not just around query building, we can also do things like optimization of queries as well. Where's the innovation going to come from in your mind from the developer community? Is there white space out there for entrepreneurs to come in and participate? How would you share, because everyone wants to know, I said Google's making big power moves here. You got the data, you got the developer. Where's the white space for opportunities in the ecosystem? That's such a fantastic question and honestly, this is a really old startup, I've got a guy here that's so sorry. This is my background a little bit, but what I'll say is it's such an exciting time because the answer is we don't really know everything that's going to happen in this market space. Vector databases and embeddings, they've kind of popped on and seen relatively new. I mean, generative AI itself, in terms of the level of popularity, also relatively new. Who knows what the next six to 12 months are going to hold? And I think that creates a lot of opportunity for startups. The thing that's most exciting to me is as you saw on the keynote stage, 70% of these generative AI unicorns are building on Google Cloud because they look at what we're doing in terms of the hardware, in terms of services like Vertex, and they want to be betting on a partner who's going to accelerate their ability to innovate and create value in this market space. Gabe, how are you guys looking at the developer market from a data perspective, if, and we've been saying on theCUBE prior to Google Nets all year, since KubeCon literally last year, we introduced the idea of data developer. DevOps changed the game. What's kind of going on with AI is very DevOps-y. It's like agile, fast, highly accelerated. So if data is going to be now programmable, because what you're basically talking about is programmable data. Okay, that's a DevOps concept. So data as code kind of seems like infrastructure as code. So okay, go to the next level. If you believe that's happening, does the developer shift left? Is there a shift up or down? If the developer had to shift left for security, we're seeing a lot of parallels between security market and data, more guardrails, we hear words like guardrails, automation, policy, compliance, legal, security kind of brought that to that. Now data is now programmable. I got a code with data in mind, which was not the mindset of a developer when you're dealing with silo data warehouses. You know, it's a fascinating point, and I talk about this a lot in my Aptev spotlight session tomorrow, but this idea of shifting left, it manifests as developers getting a lot of added responsibilities, right? We want to model a bunch of things in source code, CD, security, governance, it's like all shifting left. And look, it makes sense, right? You want a lot of this stuff in source control, you want it version, but it's also adding a lot of cognitive load to developers. And I think your point is this is creating the same risk in terms of the data science and data analysts space. What we want to do at Google Cloud is we want to shift down. We want to shift these capabilities down into the platform. Let the platform do the work for our customers so they don't have to take on this extra cognitive load. That's certainly true in the developer space, and I'd argue it's true in the data space as well. So a line of code could activate a duet call or some AI feature, or even a co-programmer. I mean, basically that's what we're getting to, right? I think so. Some of the futuristic stuff in this space that we're working on in more than the research realm is just incredibly exciting. Some mind-blowing innovation that's going to be coming to the fore here. Can you measure it yet? Can you measure what does duet do to an average, level five, level six engineer, level three, level four? We actually can, and honestly, I'm less interested in what Google's measuring, although we do measure internally because we have a lot of these tools. This is classic Google innovating in part on the basis of what we do inside of Google. But we also have customers who are, you'll hear from Mark Quigley, Director of Engineering and Enablement at Wayfair, who we've been working very closely with as part of the duet AI program. They've been measuring the productivity enhancements that they're seeing from duet AI. Some really fascinating numbers on this. And I got to say, the early indications are, the productivity gains are real. They're real, they're durable, they're sustainable, and we've only just gotten started. So imagine where we're going to be in six to nine months. Everyone we talk to is unequivocally admitting and driving the point home that the productivity is smashingly great. I mean, the code training code, going through code, but there's different approaches. One person said to us last week, train your good code, not your bad code. So you got to have the right data set you're working with. How important will that be? Because that's going to require some cognitive thinking on the developer. Am I actually in the right data set? Or the policy IT department? What are the data IT department looks like? Then that's another question. What does the data IT department look like? But assume there is one. Here's a set of stuff you could use. This model, that model, there's going to be choices. And you got to have good data. Data is now at the center. I mean, I'd argue it sort of always was, but now it's obviously at the center of a lot of this new innovation that's coming out. There's going to have to be a lot of evolution in terms of how folks are giving policy and governance around that data. But one of the biggest indicators that we found of the quality and the accuracy of what's coming out of these LLMs is in that quality of data. And that's why we announced with Duet AI the fact that we fine-tuned Duet on GCP specific documentation, best practices, API reference architectures. It makes for more accurate results. So if you want to access a storage bucket in Python, let's say, inside your IDE, we are going to print out the right code for that because we fine-tuned the model to do it. It's like the data is no longer passive. The data is now actively part of the solution. And one of the other elements of this is techniques like Retrieval Augmented Generation where you can actually reach out into a vector database, find some embeddings, and customize these LLM responses. That's going to become an important part, we're already seeing it become an important part of how customers are going to have to tweak some of these LLMs to their enterprise-specific use cases. Gabe, it's great to have you on the queue because you're like an analyst, right? We're like an analyst session here. Tim Hawkins over there is waving to us, hey Tim, you guys are both analysts in my mind. You're doing a great job here sharing the data. We appreciate that. I got to ask you the final question for both of you guys. I'm curious, how does Duet and the Duet AI aspect revolutionize Google Cloud for your customers? Where do you see it having the most impact? Obviously the code generation thing is great. I see some awesome obvious things, but if someone says, okay, this is a game changer, what about it will revolutionize? Not incrementally improve, revolutionize Google Cloud. You want me to take a stand with that for a second? You said something earlier I wanted to come back to which is training models on good code. You want to train models on good code. Actually with annotations, you can train models on bad code too. Just note it as like this is what problematic code, this is what insecure code, this is what leaky and efficient code looks like, this is what good code looks like, and the marriage of those two things is how you double down on the positive patterns and avoid the anti-patterns. For my money, I'm loving what Google is doing around AI and generative AI around code and design, and then as a product manager, the idea of being able to take an inkling of an idea and yield a like workable demo before I even write the memo, I love that. You know for me, I think to answer the question around what is going to get radically changed here, if you look at the cloud market as it exists today, it's not just Google, all the different cloud providers, you've got 300 different services typically that developers and engineers have to wade through to get something done. Developer comes in mind with a very specific orientation, I want to build a new app, I want to modernize an app, I want to migrate, I want to troubleshoot an issue on my app. And today, the set of gestures that you have to go through to do that, it's very ad hoc, and it's not necessarily consistent. With generative AI experiences like what we're doing with Duet AI, we have the opportunity to completely reimagine these experiences from first principles, oriented around what is the job to be done that our customer is doing and what's the most friction-free experience that we can deliver to get that job done. My belief here is Google Cloud is going to accelerate out through our ability to make just the overall experience of using cloud and order of magnitude better using generative AI. We've been saying on the opening, we'll wrap with you guys, nail the developer experience, get the solutions for customers, not compete on services, and nail the ecosystem that's showing robust action. That's a trifecta. That's that you hit that, then the game changes and the world's looking different for Google Cloud. Gabe, thanks for coming on, I appreciate it. Thanks for having me. Gabe Munro, VP of developer experience. Changing the game, AI cloud is what's happening, AI generated applications will change the world and it's on different demographic, different cultural shifts, so innovations here, we'll be right back with more cube. Getting the data, we're like AI bots now, Dustin. They're gonna replace us. We'll be right back after this short break.