 There are folks which are now teaching AI and generative AI pretty much in their lower school, right? So, right, in this case, right? And you see like basically like a skid with very little skills being able to build basic AI applications, right? I think that is trend which is going to continue to develop. Hi, this is your host, Abdul Bahatian. Today we have with us Peter Zaitsev, founder of Percona. Peter is here to have you on the show. Thank you, glad to be here. It's my pleasure to host you today. We have covered Percona in past, so our audience, they do know about the company, but I would love to just quickly remind them what is Percona all about before we dive into today's discussion. At Percona, we help people to run the most popular open source databases, MySQL, MongoDB and Postgres successfully. With that, we have our own branded approved version of software as well as solution and services, right? Which you will need, especially running those databases in high load environment and mission critical applications. And today's theme is around AI to understand, when we look at database administrators, what role AI can play there, or if you look at AI as a workload there. So let's quickly talk about from your perspective, let's first talk about the whole state of database administrators, DBA. How mature is the market? What are the challenges, especially when it comes to, because mostly about skills, data engineers, data scientists. So let's look at it from that perspective. Adoption challenges is skill gap. I think when it comes to their databases, right? And we speak about AI, I think we often are talking about three very different things, right? Because one is if you think about the DBA perspective, rather somebody who is running a database, right? In this case, there is a desire to use AI to run the database most successfully. For example, can AI tune the database? Maybe these parameters, maybe queries, maybe in terms of generated AI help us to create better queries, right? And all this kind of stuff, right? And I think that is, there is both a lot of interest in development and also a lot of hope for a future improvement in the last recent years, right? The other one, which in a lot of case, comes for analytical databases saying, hey, well, if you are the data scientists, right? In many cases, you are looking to make some projections about the future, not just a report about the past and you are looking for database to support integration of AI models you would want to run, right? And that is a different angle, I think in a lot of, there has been a lot of work has been done in the recent years. Now, probably the hottest thing, the hottest thing which is the newest thing, right? I think which came just after the chat GPT was released, is saying, okay, now we have a lot of developers building those generative AI applications. And they need to kind of to integrative database and create with a very special kind of workloads. This is where we see a lot of vector databases, for example, coming up, right? I mean, if you follow that like startup ecosystem, you can say, wow, in a database space where have been now many databases coming up with those focus and they're raising a lot of capital. There on the other side, you see the majority of those like old timers databases, right, which have been in this market, they also provide some extensions to supplement their database with their workloads. And I think this is kind of where we will see an interest in development over the next few years to see like what approach is actually going to win, right? Are we going to see more of their traditional databases, are going to grab this market of those applications, or is it going to be the new generation of databases, which will come and create a huge market, right? Or maybe it's going to evolve. Now, can we also talk about, but not the question, also talent and skill gap? Of course, we talked about early in 2023, a lot of layoffs were happening, which actually meant that a lot of big companies were freeing a lot of resources for startups and not that big companies. But if you look at the whole skill set availability, are you seeing that they're still a shortage or you feel that now they are enough, engineers, scientists, who can help these companies with their database challenges? I think, look, right? If a market, which is blowing up as it is right now, there is of course, skill shortage, right? Especially for the highly skilled engineers, right? I know like a lot of folks those days, right? Expert in the generative AI, right? And like a deep AI space, right? They can have, you know, absolutely tremendous, tremendous sellers available for them. But I think you also have, should look at a little bit different things right here, because there is AI skill, which corresponds to building AI platforms, right? And that is where we see a lot of, you know, competition. And hey, if you, for example, want to throw a challenge to open AI and build something better than chat GPT, well, you know, you better be damn good at it, right? Like in this case. Now, if you look at the other market though, right? There are, you know, AI is kind of, especially generative AI is kind of like an internet, right? Or like, you know, mobile applications. Now, every company need to think about how they're going to integrate AI in their business, or they're going to be left behind. And you know what? And in that case, we actually have a lot of innovation happening in terms of having the people with less skills being able to do that, right? Like for example, I recently attended, you know, one, like webinar and somebody showed how they use Zapier and as a no code framework, so they could use AI for their recruiting process, right? And it was like, wow, that is amazing, right? What we have that interest in integration right now with the no code frameworks are racing to integrate the AI, right? And I think that's both going to continue happen, right? So on one extent we have a continuing market growing for extremely skilled AI professionals, but on the other hand, we are looking to enabling people, right, with limited skills to build the applications in terms of AI. And that is a progress happening very, very fast. You know, maybe another small interesting thing I wanted to highlight it was interesting for me to see is there are folks which are now teaching AI and generative AI pretty much in their lower school, right? So, right, in this case, right? And you see like basically like a skids, a very little skills being able to build basic AI applications, right? And I think that is trend which is going to continue to develop. When we look at generative AI because AI and ML is now traditional, right? Generative AI is the new buzzword. When we look at generative, I want to talk a bit about the impact on DBAs. Now, I want to talk about this impact in two different ways. One is generative AI as a workload. A second is generative AI as a tool to help DBAs. I think this is interesting, right? Now, if you look at database industry as a whole, right? I think we had this interesting trend going on for years, right? On, if you look at, I don't know, let's say 80s, right? 90s, right? Often you would have DBA who has, you know, couple of Oracle database, right? Or something like that. And they just really keep a very good care of them. Right now, as amount of data, amount of workloads explode, you are relying on a lot of automation in this case. And you also see what in many cases, the DBAs do not even exist. In many companies, for example, you see folks running database as a service, right? And have all that database management outsourced either to cloud providers, right? Or maybe to their partner providing you managed services, right? This kind, right? And if you look at from this standpoint, right? They're still continuing going down this trend, right? Even with all those changes, we still have a lot of opportunity to better quality of our data environments, right? Improved security, availability, right? And so on and so forth, right? Because you can probably see, there is still a lot of the hacker attacks, data leaks happening, right? Or you will find that things are not working because of some sort of related to the database failure, right? So I think in this regard, AI helping database is very helpful. Now another thing, and I think this is kind of also goes to the generative AI is what we often see is what it is not AI on itself, but it is a human plus AI, which is a best combination, right? Because AI can generate a lot of stuff, at least at this point, well, but it doesn't really have a lot of, I would think about that's like a critical thinking, right? In this case, so, you know, if it's something which is kind of not very sensitive, you may be able to, you know, like generate AI coverage in let's say, I know, let's say football games, right? Well, it didn't cover it completely, who cares, nobody dies, right? Now, if you're thinking about AI when it comes to wherever like a medical stuff, right? Or I would say even kind of to a databases, you want to make sure there is a human in the loop for critical decisions. So you are avoiding those problems, right? And I think that the same approach if you're coding, you would say, hey, you know what, okay, you may be able to have a AI to generate the code from you, right, but you want also for human to take a look, right? Because you may have other unsaid requirements, right? You know, like security and performance and so on and so forth, right? One thing which I think is very interesting, right? In general about the humans and machines is, I think it is this, right? Because as a human, when they operate in the work environment, wherever, right? There are a lot of unsaid, let's say assumptions. You think about that, you know, like culture, right? How we do things in this kind of, in this company, right? And then you give your prompt to AI to generate things, right? You often do not mention those because you don't even maybe honestly, you know, recognize that consciously, right? And that is one of the reasons where you provide there, you get the outputs from generative AI, even if it's perfect, right? It doesn't recognize that, right? And I think that is for me is the big reason why generative AI will have to have a human in the loop for a long time. No, that's very well said. And I was actually about to ask you that when we look at DBA, as you also said, you know, some company, they don't even have that kind of roles there. Do you think that Genetic AI will fill that role? Or it's not going to be a replacement, but Genetic AI is going to be a tool, but you still need a hand to deal with that tool, to use that tool. So can you talk about that, that it's going to be a versus story? Hey, you know what, Genetic AI versus DBA? Or you think it's an and a story that Genetic AI will help DBAs? How do you look at it? Yes, well, what I think in this case, right? And I think that is also like a good to look at there, like a history in this case, right? Because of course, what we are looking as a consumers, right? In many cases is, you know, simplicity and the self-service which comes with that, right? Like for example, if you think about something like, you know, cars 100 plus years ago, right? Well, every driver would have to be kind of mechanic, right? Or using computers, right? You think about early days in the sixties, right? Well, everybody will have to have a skills and debugging, right? Because bugs will get into the system, right? And well, now with many of those things, we are on in the kind of the most common use cases. We don't, we don't need those, right? But at the same time, on a high end, we still need those, right? If you have like your Formula One or NASCAR racing, right? There are actually mechanics involved, right? In this case, right? When things get serious, like an, you know, airplanes, right? Well, are constantly working with mechanics, right? Because risks are high, right? And I think we have like a two trends, right? If you, on one extent, we have increased this race of simplicity. If you are building simple applications, it becomes easier and easier and easier, right? You don't need to be DBA. Hey, you don't even need to be a coder those days. You can build, you know, like a fantastic applications with, you know, no code frameworks, not knowing anything about the coding later on the database, right? But at the same time, then we are thinking about a mission critical complex system where is still those skills needed and they're going to be continued to be needed, right? Why? Because if you look at the amount kind of of a software are we having this world, right? And also like a potential we can do in this case, right? And potential we can use the data is continuing to explode, right? There is still a lot of industries where we can innovate with using the data and software, software better, right? So I'm not kind of concerned so much, right? I think in this case, as a society, we will be able to really absorb that kind of increased productivity which we have, right? For, you know, for good of society, good of a planet and everything. How does Percona kind of help organizations to have that stability that they need for business operations? At the same time, allow them to dip their toes in latest technologies, you know, such as Genitive AI where their teams don't get overwhelmed and lose focus on what they're supposed to do. If you look at their corner, right? In this case is we don't have the most sexist role in many cases, right? In a lot of cases, databases are not healthy, not sexy, right? I want to just say, right? And I sometimes compare a database to like a plumbing, right? You do not, you know, spend a lot of time when you're thinking about your planning but you know what? If you have your toilet overflow, right? That is going to be a big mess, right? Like in this case, and you wish you would be taking a better care about that, right? And that is something what you can think comparably what we do at Percona. And through years, we support a lot of different kinds of applications, right? And you type of applications being built, you know, moving to, you know, mobile or raise of social, that's all happened on the Perconas, you know, time horizon. And then the same thing is here, right? As people are using the major open source database in this regard, you know, Postgres, MySQL, MongoDB, we are looking at how people use those with generated AI applications. And we are there to support them in those joining. Can you talk about what are some of the major trends that you are seeing in the space customers or the folks who are building these technologies in the realm of Genetic AI and DBAs? Yeah, so what I think in this case is interesting, right? As with many technology, when it kind of blows up, right? It's often goes as well. It's, there is not a lot of, you know, guardrails around that, right? You think about things about security, right? Or you think about the things about the IP, right? Intellectual property, right? And I think that is, if you look at from my standpoint, right? We often have this kind of stuff. We have a tremendous breakthrough in the AI, especially in a generative AI, right? And some of those, if you kind of regulations, or I would say, you know, just like some boring compliance stuff about that is still keeping up. And that is where I would see a lot of development, right? Because a lot of things need to be cash up, right? For example, we have seen the generative AI coming up with the idea as well. How do we ensure, right? What when you share the corporate data, customer data with AI, it's not intermingled, right? If something else, right? Maybe in a way we don't, you know, we don't totally understand, right? You think about like security, right? Or, you know, thinking about where, you know, the content AI generates, right? Where that can be, you know, or inappropriate, offensive, you know, biased, right? And I think all of those works that people are continuing to be working on, right? Or, hey, who owns, right? Or what even the content we can train on, right? Now, you know, for us, we look about in the open source space, right? And I think that open source and AI, that's also when we are kind of raising a lot of, I think, getting questions, right? Because you can think about the code on one extent, but then there is like a training data, but then there are also weights, right? Which have been computed, right? After training has been performed, right? The, you know, concept of an open source that not completely, you know, connect to all of that, right? And I think we are still yet thinking about, you know, how to approach that in this space and what models are going to be done, right? Because if you think about that, well, while we have something called like open AI, right? As a company, right? Well, it's not really kind of quite open, right? It's open also in a sense like, well, you know what? You can pay us. And if you choose to, right, we'll go into give you that access to our technology, right? Not open as an open source, right? Unrestricted innovation, right? And, you know, avodian login and so on and so forth. Peter, thank you so much for taking time out today. And of course, give us a great overview, great insights into genitive AI and DBAs. Thanks for all those insights. And I would love to chat with you again. Thank you. Yeah, welcome. Thank you. It was a pleasure.