 Hi, this is your host and welcome to here for let's talk. And today we have with us Prashanth Chansekar, CEO of Stack Overflow. Prashanth is great to have your show. Thank you, Slapnel. Wonderful to be here. Yes, I'm looking forward to this discussion. Of course, primarily we are going to talk about overflow AI. But since you are here, and if you look at Stack Overflow, first of all, you folks predate a lot of these modern technologies that we talk about all the cool boys, you know, Kubernetes, cloud natives, everything else. And also, if you look at the whole developer movement, you folks played a very big role there as well to kind of build a community there. We started talking about a lot of culture movement we talked about. DevOps movement, DevSecOps, and SRS platform engineering. But now we, I mean, I record a lot of shows and we are talking about bringing that developer experience back again. So what I want to hear from you is how have you seen the whole evolution of this developer journey from the day when Stack Overflow was created and when we are today? Great question. We've obviously got a tremendous amount of historical knowledge and data on a platform for about 15 years of operating since we were founded in 2008. So we've seen a lot of evolution in software programming when we think about concepts like abstraction where it's gone easier and easier to write code over time. And I look back to the time when I was writing code where, you know, we were like most people that pre Stack Overflow, you know, we didn't really have a lot of resources to leverage other than, you know, really figuring it out on your own or, you know, working with other people in person to do it. But over time, and Stack Overflow has been part of this journey where we've actually accelerated in many ways and enabled people to build more and more elegant software, you know, really sort of improving on the prior versions of certain capabilities that have been invented. And it's been fascinating to see that notion of getting it sort of making it easier and easier to grow, so to sort of build software. And along with that, what has happened is that the number of software developers have actually increased fairly dramatically. So it's that's what we've really sort of observed. Because as it's gotten easier, more people have actually gotten into the field of writing code, which makes a lot of sense because as it's gotten easier, more companies want to leverage technology to be more, to be closer to their customers to deliver great experiences from a digital standpoint. And now we're entering this, this new phase around the AI era, which is also going to be a tremendous platform shift, which is going to affect all things in software development, because again, companies are going to leverage this sort of compounding effect of collecting data, leveraging the fact that, you know, they've now got access to very, very sort of affordable compute storage and networking through cloud computing. They've got transformer technology. Now that is, you know, obviously with neural networks have been around for a long time, but all this is compounded to a place where now all companies can leverage all these pieces and these building blocks to make really, really powerful, intuitive customer experiences happen. And then of course, don't forget the startup ecosystem, you know, of course, in India and the rest of the world, this is an absolute explosion of startups building great things in this new and enabled by this new space. And so it's really, really exciting time. And one of the last point I mentioned Swapnil is that, you know, we believe the number of developers are just going to go up very, very meaningfully. We're probably talking about a multi, you know, three to five times type of, you know, number because so many more people are going to feel comfortable writing code going into this next phase. This is kind of a loaded question, but a lot of folks say that he's going to take away jobs. I don't believe so invention of build did not take away jobs actually led to the creation of company like Tesla and drivers and a lot of things. So generally AI is also going to create more job than take away. But but how do you see the emergence of Genetic AI because you folks have been leveraging AI for a long time. But Genetic AI, sometimes a lot of technology, they come like a blip and FT is a good example, but then there are a lot of technologies like Kubernetes or Linux kernel, which is actually concrete. How do you see it because you have to, of course, embrace latest technologies. But in reality, how do you see Genetic AI? We believe Genetic AI is a tremendous platform shift that's going to define this next era of computing. So I think we are we don't believe it's it's a fad and we don't believe it's, you know, sort of a blip on the radar. It is going to be transformative in many, many ways. Now, that doesn't mean that there isn't a lot of hype associated with this. That's just the world we live in with, you know, information flowing very freely. But it is transformative. There is something, as we like to say, there's a there there in the context of the technology. And so the way we have thought about it, Swapnil, is that Genetic AI specifically has some tremendous benefits. You know, it's got summarizations, discoveries, things through, you know, semantic search is a really key capability. And many other elements that make it really, really powerful to leverage very, very large, you know, data sets in a way that can be really sort of insightful and really provide produce new insights and new capabilities for companies and developers to use. And of course, generate code and all these other use cases that we've seen. At the same time, we believe that it has a set of deficiencies, which we believe we should be filling with community. So, so community brings things like, you know, creation of new knowledge, it creates talks about things like, you know, how do we connect the dots between various things and come up with with conclusions and make decisions of things. So the intersection between Genetic AI and community is really where Stack-O-Flow plays. And why is that? It goes down to, you know, every year we we conduct a developer survey, and this year 90,000 people from our community responded. And they said that 70% of them were interested in using AI tools or were already using AI tools. And, but only 40% of them trusted these AI tools, because they're very much sort of black boxes. They have no idea, you know, the amount of risk and let's say the code being generated or information being provided. Is it hallucinating? Is it not? Is it right? Is it wrong? There's a lot of that, you know, in the responses. And it makes sense, because, you know, we as technologists and developers are on the hook to build production level applications, things that are, you know, are very, very mission critical for companies. And so you want to feel confident that what you're actually putting into your code is actually going to be what is going to work. Number one, and number two, you can always reference back and be able to figure out if something does go wrong. How do you sort of really sort of debug it and also engage with your colleagues to that do that successfully. So, so this intersection of community and AI with Stack-O-Flow in the middle is really rooted around building highly trusted solutions as part of that, but has attribution and citations. It is highly accurate and high quality. It has a great feedback mechanism, just like we've been doing on Stack-O-Flow for 15 years. It has personalization and ultimately has recognition, recognition for the contributions of people putting content, creating new content, you know, innovating new languages. And all that existing on Stack-O-Flow is the foundation of how we've built what we built with O-Flow AI. So O-Flow AI is this, you know, we have announced five different capabilities under O-Flow AI, all with these principles that I just described to you. And I'm happy to talk about exactly the five different announcements around O-Flow AI, but specifically that's how we have built this with this intersection of community and AI with humans in the loop, as we like to say, to provide that reinforcement learning and feedback and attribution along the way. Yeah, I would love to talk about O-Flow AI and how it will be used. But before we go, one more question about the whole developer news articles we keep reading about, hey, is Genetic AI going to augment or is it going to replace the jobs? There's always a fear and something new comes. There's a whole strike going on right now with the, you know, actors and, you know, writers' association, their fear. How do you see that is Genetic going to take away a lot of developers' job or is it actually going to empower just like Photoshop? It gives us, actually, it enables more people to be photographers now because they can easily do those things that you could not do earlier. I think that it is absolutely going to, especially in software development. I can speak to what I'm close to. We believe it's going to increase the number of software developer jobs dramatically, right? And this is because it's going to make it more accessible, more inclusive, more open to people with various skill levels to be able to just get on the journey of being a software developer because the bar is going to be a lot lower. But at the same time, it's lower and also higher in that everybody's going to be able to do, you know, a minimum amount of coding in a way that's quite powerful as a result of some of these tools. So it's actually quite a very interesting time to be in in technology. And overall, you know, we believe that same, that virtual cycle I mentioned earlier because more people get in. What they're actually doing in coding could be different, by the way, Swapna, it may not be what we were all doing, let's say five, six years ago, because with the rise of prompt engineering, which is also an area we've now created on Stackflow flow called the Genai Stack Exchange, and the national language processing collective or sub community where people can have discussions and learn about the subjects. So with prompt engineering, the types of things that people will do will be different over time as they sort of really leverage to harness the power of these LLMs to produce tremendous outcomes. And then companies using this, the startups using this will create this kind of virtual cycle where companies will be able to unlock a lot of value as they build great high value experiences for their customers. And so we are very bullish on what this means for job growth for in specifically in technology. You know, as it relates to other industries, I think it's, you know, your point in that people will have to adapt to leveraging these tools. I think people who don't use these tools will probably get left behind. And those when you think about job loss, I think it's basically people who are reticent to leveraging these tools to be more effective and productive. And people who do use these tools will be orders of magnitude, you know, more effective in their organizations. So I think it's like anything in technology, I think you have to adapt and leverage these tools for your benefit to go faster and do more with what you have. So I think it's going to generate a tremendous productivity boom for people. When I was listening to you, it just reminded me the whole, you know, old times of, you know, digital transformation journey or cloud native adoption journey where a lot of companies who were not, you know, and suddenly they're especially during COVID times, only they almost ran out of, you know, business because they did not have any digital online presence. So I think soon we'll start talking about, hey, you should be on the gen AI transformation journey, you should have those strategies in place. I would love to, as I said, overflow AI. But when we talk about these teams, you know, or as you said, organizations, they should or not necessarily should, but organizations, you know, they do want to embrace a lot of these technologies. What kind of challenges you see are there, which kind of becomes kind of hurdle in the adoption of these generative AI technologies among developers. We've got a Swapno, not only the two parts of our company, it's important to explain this because we, you know, we have the context in that we have a public platform that everybody knows about, which is, you know, tens of millions of developers come to that. And we've got 58 million questions and answers on the public platforms, a lot of like the historical context there over the past 15 years. And then we also have a very large software as a service product that's called stack overflow for teams. And that's leveraged by 15,000 organizations to effectively share knowledge and collaborate with our platform internally within organizations. So effectively it's a private stack overflow for companies to use. And with these both these vantage points what we have discovered through our research and our work as we launched overflow AI, is that especially in companies, there is absolutely an interest to leverage generative tools like that 70% statistic I mentioned as part of our survey very consistent. But there is a lot of reticence to fully trusting, you know, what's coming out of these tools and can they actually run mission critical critical applications is so those are so trust is a broad theme. And then you look underneath that, there is this notion of privacy, security of your data. There is this, you know, real explosion of the amount of code that's going to be generated using code gen tools, without really any control to be able to figure out how do you debug this when you need to are you going to get pulled into a room and something breaks in production. And where's all the context that's missing, you know, which is why what we've discovered that the need for something like stack overflow for teams as the knowledge base for all things technology information is so high. And we've seen a tremendous interest since we announced and we've been working on that overflow AI and even a few weeks before that as we were working with customers to preview some of this. People are super excited that they now can leverage a capability that is the source of truth within the organizations that they can leverage that data and leverage our AI capabilities are generally capabilities on top of that trusted data, which which really gives them a lot of comfort as they leverage these leverage and harness these tools. So generally speaking, privacy, security, trust and attribution, and really feeling that they can have control over the explosion of amount of code that's going to get created here with high context knowledge base. And since you're talking about teams, I also want to quickly talk about how, how does it integrate with teams with the organization because we talk about a lot of things go wrong. We talk about things like chaos engineering where they big break things on purpose, they collect a lot of data, we talk about the whole observatory where they do see what's going on wrong in the company. But when they do that, they also look at answers what went wrong, how to fix it. So how does stack overflow teams, does it still remains like, okay, it's private, but does it somehow integrate with their internal processes or it remains a totally different entity. It completely integrates with their developer workflow. And with what we have announced with overflow AI even more so and we can talk about that next. But with the currents that just you know, pre overflow AI stack overflow teams integrates with your code repository with GitHub, it integrates with you know, JIRA for your ticket thing systems it integrates with octa integrates with Slack integrates with Microsoft teams. So the idea is to really sort of make sure that you are in the flow of your work and you're not context switching. And you just have access to accurate information the right place right time irrespective of where you are in the developer workflow. Sometimes what happens is that when organization like stack overflow you embrace our technology, it also sends a message to the to the to the rest of the world. Hey, the people who are using these tools, it breeds confidence between those tools, those technologies as well. So let's now talk about overflow AI. Yeah, so you know, we were super thrilled to announce overflow AI. Last week at the we are developers conference in front of thousands of our developers in the community. And this has been by the way months in the making. We had carved out 10% of our company to focus on building these amazing solutions. And now the response has been absolutely tremendous over the past, you know, a few days since we've actually made the announcement publicly. So overflow AI is us leveraging genitive AI capabilities in both our public stack overflow as well as stack overflow for teams. And it's really bringing together those two things I mentioned earlier, which is community and AI the power of both those things, the intersection of both those things to our users and customers. So the five things we announced. Number one is really providing fantastic search capabilities on the public platform. And what does this mean overflow AI search leverages semantic search where they're able to have a natural language conversation with stack overflow. You know how do you do this thing in Python, and you're going to get a generative AI answer that is going to summarize content from our 58 million questions and answers from our public platform, provide citations and attribution to where the answer is coming from or how did it actually come up with the answer that it did. And here's all the links to those. It's embedded in the answer, the generative answer, including the experts, you know, the world's best JavaScript expert actually contributed to that generative answer as an example, in an implied way. And it gives you the ability to provide feedback based on that answer, whether that helped you or not. You can if it didn't completely solve your problem, you can have a natural language conversation with overflow AI straight on stack overflow, where you can extend that conversation for deeper provide more context even add in your code. And it will provide even better search results and more refined generative AI and kind of summarize search results and answers. And all of that is sort of a single player type of experience where until you can just engage with the generative AI overflow AI capability. And if none of that work, then you can still post it in our community and and get the opportunity to somebody an expert to respond. So this is number one, which is search capabilities on public stack overflow. And what this does is it opens up the opportunity for people of all levels of capability to engage on stack overflow. Historically, the level had to be pretty high for you to engage on stack overflow you'd be an expert and so on. But now this is like anybody can have this experience and still get the best of what the platform offers again based on the 58 million questions and answers. So that's number one. The, the number two piece on the public platform is that we've also launched an entire area of AI community discussions where people can really sort of engage on as I mentioned earlier, generative AI topics like prompt engineering, natural language processing. We now have the ability for people to have real discussions on the public platform with their subject matter experts. This is this is related to the AI announcement. But we wanted to really make sure we promote learning for people because the space is moving so rapidly that we want to provide that capability. So that's that sort of focused on the public platform. Now as we move to stack overflow for teams, we have the first piece is something called overflow AI enterprise knowledge ingestion. So what we're able to do with this is really use generative AI to point to existing knowledge bases in the company could be confluence or SharePoint or GitHub or, or Google Drive and really ingest all that information into stack overflow for teams, instantaneously create Q&A pairs, automatically create tag suggestions, subject matter experts, and then give the option for the for the customer to choose whether to ingest these are not based on the quality and also dial up and down, which are, you know, what are qualified to actually be ingested. And this does this happens in minutes. And so what this allows you to do is to bootstrap a stack overflow community within your company in very, very rapidly, right? So there's no cold start problem that many software products have. And this is really tremendous way to leverage this to get your community flywheel going within your company. So that's the second large announcement. The third one is around enhanced search capabilities, like on the public platform that I explained. Now what we're also able to do in stack overflow for teams is provide ask a natural language question. How do I do this on, you know, could be Google Cloud X, Y, Z service. And the answer is going to come back from the public stack over 58 million questions answers but also from your stack overflow for teams, you know, repository within your own company, and it's going to summarize it in a generative AI answer with again links and attribution to all the source, you know, the source content, but also, so importantly, also bring back a links to other repositories within your company, even outside of stack overflow. So you could provide a confluence link, a GitHub link, it could provide all these other elements and search in our opinion and companies has been broken for a long time. So this we really wanted to address that by building this amazing enhanced search capability, which really sort of, you know, is in this kind of this phase in the in the sphere of enterprise search, etc, which is really what's been in that category has been around for a long time. So we're very excited about this capability. And then the next two announcements, our capabilities are all about keeping developers in their flow of work. So it's about what we did what we did announces this overflow AI visual studio code extension, where now for the first time you have the ability to surface stack overflow content, both public stack overflow from 58 million questions and answers as well as your private stack overflow content in your IDE, through a genitive AI capability and conversational interface where you're able to ask again a natural language question, it's going to provide you again with all the answers in the flow as as your writing code or generating code on the left side of the screen in the IDE and visual studio code, it's going to provide links to all the relevant content from stack overflow public as well as stack overflow private, and you're able to really then have a lot of context as you encounter issues and you want to start things and so on. And then the final one is providing is providing the same capabilities genitive AI capabilities in your slack in your slack interface. And this is really again you're able to have natural language conversations in your slack interface, people spend a lot of time on in their chat ops, and again surface public stack overflow as well as private stack overflow content, and have a conversational interface to be able to do that so those are the five that be that we announced so really sort of a robust, you know, set of announcements. How much of these, you know, kind of capabilities that you build were also kind of based on you did touch upon some of those pain point that are actually facing today so if this is coming through from your users and less from you. We are here to serve our developers ecosystem of community, our ecosystem of developers in our community as well as customers who we've been working with on stack overflow for teams. So we've had, you know, a great feedback loop with them for, you know, obviously years. And so in that context, we have been observing what challenges that they've been facing. We have a developer survey like I mentioned to you that was responded to by 90,000 people this year. So we look at all of that. And we have been engaged with both developers in our public community, as well as our customers and stack overflow teams to build these overflow AI capabilities over the past several months. So they've been very much part of giving us the right input to make sure that we we ultimately build something that's highly valuable to them. Of course, there are certain things that you cannot talk about this point there, but the things are in the pipeline, the roadmap. But in general, if I ask you what are the next thing we can expect from you, okay, you're like, these are the things that we are working on to further improve, you know, developer experience or to help organizations in that journey. And what we've discovered is that the value of the, we've got a long list of ideas, firstly, that we've got and that we are always looking at and exploring the goal for us to make sure that we firstly launch these things here. And then in August to make sure that we open this up to as many developers and customers as we can. And then go from that process all the way to general availability over the next several months, and really sort of iterate based on even more customer feedback to your earlier point, which is we want to make sure that customers and users drive in many ways, you know, how we how we continuously build this. But overall, you know, we just believe that the opportunity set is significant here, because the value of Stack O flows historical data that we have from 15 years, we've been running several experiments on that topic behind the scenes as in the power of layering in certain data sets on top of let's say open source models. And what does that actually produce and how efficiently can it produce because one of the issues with with general AI is that it's extremely expensive proposition at the moment right specifically for companies trying to build some of these capabilities. And we've realized that our data is fantastically powerful. You can build very lightweight models which produce tremendous outputs and high quality outputs that you know right you know that are commensurate with what you see in the news with some of the very expensive models that people have built. And that gives us a lot of flexibility to do things for our users and customers, especially from a from a unit economic standpoint. So that's just to give you a sense what notes around how to be provide more and more insightful, right place right time context for our for our users and customers and do it sort of at very, very large scale with with cost in mind. And as you said, no, we will be leveraging a lot of genetic AI, but you also kind of stressed on the importance of actual people, actual developers, subject matter experts. And, of course, adoption of detail. What role do you see is tag or flow or these technologies that you folks are building will build in within organizations as they will be leveraging those technologies because they will need a lot of help. And with cloud cloud what they had done is it has enabled a lot of organizations who would otherwise not been able to get on journey to get started very quickly. So talk about the role that you folks will be playing in future. We believe that we've got this large data set firstly on the public platform, which is the place where people are able to really build new technologies answer each other's questions as we develop new innovation happens very rapidly on our platform, much like it does within companies as people try to, you know, go through technology transformations and they want to learn from new things and their colleagues, etc. What we notice is that what's shifting is this notion this nature of you know self service right where people are at lowest the barrier to entry as we talked about with people sort of being able to access information in a very reasonable way with Genitive AI. So yes it's going to open things up it's going to open things up where people the friction of actually learning and engaging is going to be so much lower because you're engaging with something like overflow AI, and you're getting sort of this collective wisdom of, you know, tens of millions of developers who have who have contributed to that, and that knowledge over the past 15 years. So developers will be able to learn very rapidly as a result of this be more productive, efficient. I think all those things are powerful both within companies and just generally in the ecosystem. And that is going to result in, as we talked about earlier, a just a huge productivity boom, where people are just going to be able to do a lot more with, you know, very, a lot faster than historic than history. And that's all very exciting for us as we as we keep moving into this phase. Prashant, thank you so much for taking time out today. And of course, I really appreciate the whole discussion on not just the developer journey, but also how they can leverage Genitive AI and how you folks actually could help them. So thanks for all those insights. And I would love to chat with you folks again, because I can see there are a lot of things in the pipeline. Thank you. Indeed. Thank you Swapnil. I appreciate the questions.