

10 Nov 2026 · 47m · 18 chapters

Stack Overflow’s CEO Prashanth Chandrasekar discusses how ChatGPT (Nov 2022) disrupted the internet and forced Stack Overflow to evolve from ad-driven Q&A to an AI-era “trust layer,” including enterprise knowledge products and agent workflows.
Prashanth is CEO of Stack Overflow (first CEO role starting 2019). He previously worked in operations/consulting and finance (banking), then moved to Texas in 2012 to work at Rackspace in cloud growth (AWS, Azure, Google Cloud). He’s an immigrant who came to the University of Maine on a scholarship (1998).
Stack Overflow banned ChatGPT-generated Q&A to prevent “AI slop” contamination. Trust is the biggest blocker to enterprise AI agents (only ~6–7% highly trust AI answers internally). Stack Overflow’s moat is curated, validated knowledge; licensing deals with OpenAI/Google improved answer quality ~40–60% and reduce token costs.
spikes from real events (e.g., PHP security issue); “Stack Overflow for Agents” launched by two engineers; “Stack Internal” launched at OpenAI Dev Day; his “chief repeat officer” communication practice (weekly emails).
Written by AI. May contain mistakes.
Jessica introduces Prashanth, CEO of Stack Overflow, and their upcoming discussion.
“Welcome to TruthWorks, where we dig into the nitty gritty of leadership and work.”
Prashanth shares his journey from being an immigrant to leading Stack Overflow.
“Welcome to another episode of TruthWorks.”
Prashanth discusses his diverse career path leading to his role at Stack Overflow.
“Wonderful, wonderful to be here with you and looking forward to the conversation and sharing it with your audience.”
Prashanth explains the challenges Stack Overflow faced during the pandemic.
“And some of my earliest experiences in college gave me opportunities to work for very large companies.”
An overview of Stack Overflow's history and its role in the tech community.
“And then finally, after doing that for a few years of my early career, some of my mentors said, look, what are you interested in doing?”
Discussion on the significant changes brought by AI technologies like ChatGPT.
“And you joined like shortly after the pandemic or during the pandemic?”
Prashanth describes how the Stack Overflow community adapted to AI-generated questions and answers.
“And then the enterprise side is effectively a private version of that in that we wanted to effectively help establish a knowledge foundation for companies.”
Insights from Prashanth on leading a company during transformative periods.
“It's about what information do you trust now through your favorite AI tools.”
Learn how effective communication and leadership are crucial during company transformations.
“It's kind of like almost like going to, you know, the gym a little bit, you know, with a trainer, because if you, if it was too, it's not an easy process to transform your body.”
Explore the challenges of trust in AI and how Stack Overflow is addressing them.
“So I think we do a survey every year in our community.”
Learn how companies can effectively shift their organizations to embrace AI technology.
“Well, I guess, you know, there's so many companies, you know, I mean, every company, there's not one company out there right now that's not trying to navigate this change.”
Discover how empowering small teams can lead to successful AI product launches.
“You know, and and how do I how do I help my people?”
Understand the importance of creating a company culture that embraces AI.
“And it was basically, hey, some people were already leaning into it.”
Explore the difficulties larger companies face when integrating AI into their workflows.
“And they're doing all sorts of innovative things, including the product that we just launched a couple of weeks ago with Stack Internal.”
Examine the impact of AI on job creation and the evolving workforce landscape.
“it is murky at best by the way still because like you know i know because all these companies are like we're spending all this money you know um we haven't seen revenue go up yes you know we haven't seen sales increase.”
Discuss the importance of a well-rounded education in the age of AI.
“Like, I don't need to keep doing my job because AI does that very quickly.”
Learn how to align personal interests with the opportunities AI presents.
“So everybody's got their own passions and their pursuits.”
Discover a perspective on personal journeys and collective experiences in life.
“And, you know, right now he wants to be a ninja.”

Get the gist, explore an idea or ask about something said.
On TruthWorks, what are the key points of this episode?
Automatic transcript. May contain errors.
0:00November of 2022. Yeah, I very, very specifically remember it. Our community very prudently decided to ban using ChatCPT to ask and answer questions on the site. How scared were you? Quite terrifying, to be honest. You know, what you've really been building for like a decade has all of a sudden just changed with OpenAI and Google and like all the big cloud companies and so on. And so they pay us for ongoing access to the data, close to a 40 to 60 percent improvement in the quality of the answers when people used our data. I don't have a chief of staff, but my agent is my chief of staff. I should be a much better CEO post-AI than I was pre-AI.
0:42What would happen if we just told the truth? Welcome to TruthWorks, where we dig into the nitty gritty of leadership and work. And what needs to change? I'm Jessica Neal. And I'm Patti McCord. Our journey together started in HR, but trust us, it's evolved into something wild, honest, and well, a bit rebellious. So throw out the handbook. We're here to redefine rules to work for us, not against us. Let's dive into another episode.
1:20Hi, everyone. Welcome to another episode of TruthWorks. I'm your host, Jessica Neal, and I'm going solo today, but that's okay because my guest, I get him all to myself. So today I have Prashanth, who is the CEO of Stock Overflow on the show. I've really been looking forward to this conversation because Prashanth is leading a company that, as we described it in the pre-record where we were just chit-chatting, has kind of had a little bit of a hands up on all the things that we're experiencing in the world today, but especially with AI. And so I want to unpack all of it with him because, you know, he, I think he's learned some lessons.
2:03He's maybe got some scabs on his knees. I don't know. We're going to find out today. And the other thing I found out is we should have done this in person because he's in Texas right down the road from where I am. So next time, Prashant, We're doing it in person. Okay, perfect. Well, welcome to the show. How are you? I'm well, thank you. Thank you. Wonderful, wonderful to be here with you and looking forward to the conversation and sharing it with your audience. It's been quite the journey for Stack, so happy to be here. Yeah. Well, first of all, let's go back a little bit. I always like to kind of explain how you got to where you are.
2:42And, you know, I'm sure most of the listeners know what Stack Overflow is, but you're also can tell us a little bit about the company. But go back in time and take us to how you got to where you are today. Okay. Even before Stack Overflow? Well, yeah. Whatever you think is important for us to know. Yeah. No, I think, you know, perhaps I'll pick a couple of pieces that may be relevant. I think along the way, I grew up, you know, I'm sort of an immigrant to this country. So I came here on a college scholarship to the University of Maine back, you know, in 1998. And with the American dream, you know, very much, you know, as the focus and all things that are possible in this wonderful country.
3:25And so I, through that journey, have always thought about exploring new uncharted waters, people that, you know, my family never came to the States. I was the first one to sort of move here on my own. and it was this kind of sense of exploration and really thinking about new things to go do, new things to build. So newness and novelty and building a future has always been sort of part of my DNA and my excitement. So I think perhaps that's important in the context of my, let's call it, somewhat entrepreneurial journey over time, especially with Stack Oilflow. It's probably the closest I can call that, even though I wasn't the founder of the company.
4:05But I had a fairly, basically my education is me starting out in engineering and computer engineering, but realizing I didn't want to be a computer engineer or software developer full time for the rest of my life. So I really was fascinated by how companies had built themselves over a very long period of time. And some of my earliest experiences in college gave me opportunities to work for very large companies. And I was just fascinated by the fact that these companies were created and then they became these massive economic machines within their communities. And they really created so much value for the U.S.
4:46and the world. And so that really, I think, put me on a business journey after my underpinnings of technology background. And I was just curious about operations and I went to consulting for that. I was curious about finance and I went into banking for that after business school. And it was all, again, it always had technology as a through line throughout that because either it was implementing technology after an operations consulting operations improvement project or it would have been advising technology companies on M &A and capital raising in my banking experience. And then finally, after doing that for a few years of my early career, some of my mentors said, look, what are you interested in doing?
5:27And I said, someday I want to run one of these companies and do something like that. And they said, if that's the case, you've got to go get some operating experience. And that's kind of what brought me from New York City to Texas back in 2012. And I moved to work at a company that's based here in the cloud services space. And it was a public company, Rackspace. And that brought me to this part of the world. And seven years after doing various entrepreneurial things at Rackspace, helping them build their very fast-growing businesses around the cloud, around AWS and Microsoft Azure and Google Cloud, especially, it really excited me to be part of these big growth efforts and transformation efforts.
6:11And so that's what, when I got the call about Stack Overflow, I thought it was an amazing opportunity. Great, great website, publicly well-known, big impact around the world. It positively impacts so many people. and the big challenge was can you take this public website which is ad supported and you know very traditional content site and make it more into an enterprise oriented company given some of the work that I'd done in Rackspace and so that felt like a great challenge and you know my first CEO job in 2019 and and I went for it and it was it was it's been an amazing journey seven years since I joined Yeah.
6:52And you joined like shortly after the pandemic or during the pandemic? Right. Right before. It was three months before the pandemic. And so, you know, welcome to your first CEO job. Yeah. You're thrown straight into it. And, you know, I just remember, you know, that two business lines were advertising and talent job listings. Right. And that was what we had when I first joined the company. And, you know, we had an early experiment with the enterprise product that the company had launched. And both those first two businesses immediately were affected in the pandemic. You know, people stopped marketing and they stopped hiring, at least initially.
7:28And so we really went, we made a huge move to effectively use that crisis to our advantage and really lean into the enterprise story. And so that's it. Yeah. Explain to people what Stack Overflow does. Explain to them what it is. Yeah, I think the roots of the company go back to 2008. It was effectively brought together by a couple of amazing founders who decided that when people were writing software, they shouldn't be pulling out their hair in a lab all by themselves or having to ask people just next to them. But we should really work as a community to help each other solve these problems when people are stuck.
8:07And so they created this wonderful website, StackWolfo.com, which has this amazing corpus of questions and answers. And it began with a few people that began to contribute to this site. And then it grew over the next decade to something like 100 million users with now close to 80 million questions and answers on every possible technology topic. We have some of the oldest questions and answers like COBOL to the most recent things on LLMs with Codex on OpenAI and Cloud, et cetera. So it's the entire library, if you will. And it is highly, highly curated and trustworthy knowledge that has been very painstakingly put together by a community of users who are all validated and they are trustworthy sources.
8:54They're not just anonymous people. These folks are people who have been vetted by the community saying, OK, this person knows what they're talking about because it actually worked when I tried it. And so that has been the foundation of the company. And then with that comes ads and all that on that side of the house. And then the enterprise side is effectively a private version of that in that we wanted to effectively help establish a knowledge foundation for companies. This is even pre-AI era. We thought that was important inside companies. But now, especially as AI is central to everything, it is very, very important to have this kind of foundational knowledge intelligence layer, as I'll explain, in our new enterprise product.
9:35Yeah. And so when we were talking before we started recording and you said you kind of had a little bit. I said you might have had a leg up. You said a heads up on the whole I thing. But my guess is that you kind of saw a lot of this coming just because you are in this community and you're hearing the questions. And maybe you could start to see some of these organizational or even competitive kind of impacts that this whole, you know, evolution, whatever you want to call it, is going to have. So was there like a moment when you were like, oh, wow, maybe we need to pay attention to this? Oh, absolutely.
10:21November of 2022. Yeah, very, very specifically remember it. Okay. Yeah, it was the moment when ChatGPT came out. It was very obvious that this was going to have a huge impact, not only on us, but just in general across the world. Because you could very quickly see the, you know, at a broad level, the, you know, the agreement on the web, if you will, is that people have content websites. And then you have search engines and you have, you know, people go to search engines and then you send traffic to your website. Your traffic then, you know, is able to monetize the ads and all those sort of things.
10:56Thankfully, we're not just a content website. We have this big enterprise business. Otherwise, it would be even bigger of a change for us. But generally speaking, I think the internet basically changed in early 2022. I mean, over the last decade or so, it's been changing, but not as dramatically as it did when ChatGPT came out. Because effectively, the agreement of that traffic coming to your website has sort of changed. It's not all about traffic anymore. It's about what information do you trust now through your favorite AI tools. But very specifically in November, we noticed actually a very dramatic increase of traffic on our site, which is very odd and counterintuitive.
11:37Because the whole game on Stack Goal Flow is you are answering other people's questions or you're asking questions. And you get recognized for that work. And then you get reputation points. And then that keeps building. and we saw this huge flow of chat GPT generated content on the site immediately, right? People were answering questions with chat GPT. People were asking questions on chat GPT. And what is now termed as AI slop, and people talk about that a lot nowadays, you know, people are like, oh, there's so much AI slop on the internet. And so our community very prudently decided to ban using chat GPT to ask and answer questions on the site back then in 2022.
12:17to and we've maintained because we we thought it's almost like uh one of my actually founder we were talking to a few years ago about this and he said it's kind of like radioactive metal you know like pre a certain date uh you know that metal has not been tainted you know and then post a date you know everything's been tainted after that and so like the question is how do you actually preserve this really really high quality corpus that's human curated and validated and not just polluted, if you will, by a generated slop. So that was the moment, Jessica, when we realized that, okay, this is going to be a big deal.
12:54And of course, we started seeing the adoption of AI inside our customer base, especially as we were talking to them about our enterprise products. And it was very early pilots and trials and experiments and so on. And so we realized, okay, having a very foundational knowledge layer is going to be very important for sure as well. So, you know, we've had to do a whole bunch of things on both sides, which I can talk about, to effectively, you know, completely evolve the business. Yeah, because, I mean, this is a massive shift. Yes, yeah. How scared were you? Oh, I mean, it was, I think, quite terrifying, to be honest.
13:34You can call it the O-S-H, you know, asterisk T, or maybe the Code Red moment, you know, whatever you want to call it but it was like absolutely it's been that way for a while actually you know it has been a sustained period of it was an initial shock because like you know what you have really been building for like a decade has all of a sudden just changed you know over a notice you know that's just life I suppose sometimes but um and so absolutely shocked for sure scared man I'm not sure about scared it's probably a little bit scared you know you know what the unknown but you know you very quickly after you realize that okay a lot of people are depending on you figuring out the answer to this list let's go get get into solving the problem you know so i think you right you get through that yeah and so how did you um how did you lead through that right yeah yeah because that's a lot to lead the company through and that's a transformation yeah yeah yeah no i think it's uh you know i it is it has been easily one of the hardest things I've done as well as one of the most gratifying things that I've done at the same time.
14:45Because I think it really challenges you to sort of really put yourself in the shoes of other human beings and think about how do you really persist through this kind of very challenging period while keeping hope, keeping, you know, North Stars in mind, keeping, you know, key goals that you can navigate towards. It's kind of like almost like going to, you know, the gym a little bit, you know, with a trainer, because if you, if it was too, it's not an easy process to transform your body. So it's going to be very painful, in fact, to do that if you really want to hit like certain goals. And it's very easy to say, look, like, I just can't do this.
15:30It's just difficult or whatever. But the reality is that if you want, there is always a way to get to the end zone. It's possible if you put your mind to it, right? And I think inspiring people with that level of optimism, while being very realistic about where we are and telling the truth about, look, this is the reality of the situation, and doing what you have to do to sort of keep the engine going. And that includes all sorts of things, you know, that includes making changes to the business. It may be including, it may include taking expenses out. It may be all sorts of other things that you've had to sort of navigate.
16:06But I think it's just staying, it's like an entrepreneurial journey. You have to just keep getting to, okay, why are we here? What is the mission and purpose of why we're all still together? Why does this company matter in this new era? I think being very clear about that with folks and constantly, and by the way, that answer is like easier said than done in terms of being able to answer that because things are changing so rapidly. So you have to sort of maybe keep making adjustments. Thankfully, I consider myself a chief repeat officer along with my title of CEO. So for me, a lot of it has been very consistent communications across the company, them knowing what's on our minds.
16:44I sent an email to the company every Friday. I've been doing that for seven years straight. And so people generally, I think we do a relatively good job of communicating being transparently, and that's one of our core values here, be transparent with folks on, like, exactly where we are and just so that people then can, you know, come into the boat with us and, like, let's roll together, you know. Right. Which is, you know, I mean, AI or not, like, clarity is how you should, you know, show you care for your organization. And it's, you know, it's something I think that we talk about it a lot on the show, but it sounds like a simple thing to do, but it's incredibly hard to do.
17:28You have to really build a muscle around being clear and communicating. And I think a lot of companies, maybe they do it a little bit, but then they let their foot off the gas. And like you said, you kind of have to be a chief repeat officer because, you know, and I learned this lesson at Netflix, And it was like, you know, you're saying the same thing over and over and over again. But the and you're bored by it. Right. You were like, God, I have to say this again. But but you have to keep saying it because people forget or they weren't in the room when you did say it. Or, you know, this person was just hired.
18:13They never heard it or, you know, but but you want people to have that clarity so that one, they want to jump in the boat. And then when they're in the boat, they know what to do. They know what role to play. Right? Right. Absolutely. And I would say, you know, one thing around it being boring, I actually have found it to be quite useful, actually, to communicate often because and be the chief repeat officer, because, you know, there's this, some folks mentioned how if you want to really be crystal clear about your own thought process, you should try about any concept, try and teach it to somebody else, you know, and I take that and I really believe in that, because I think the more you explain what's in your mind, because our minds are complex things, a lot floating around, especially when so much information is now overwhelming, in fact, to keep up with the amount of developments that are going on, or the number of developments going on, that for your mind to be able to distill that, and then to be able to be as clear as possible in your communications over and over again, in fact, it allows me to further refine my own thinking as I am communicating to folks.
19:19So, you know, the strategy becomes clearer. Why do we exist becomes clearer. And maybe it even flags what's not that clear, you know, to folks. You get feedback back saying, look, I didn't really understand what you meant by that. And so, okay, okay, I need to work on that then. So like, I think it's just like a, it's a loop. Yeah. I love it. Chief Repeat Officer, I'm going to steal it. Today's show is sponsored by Manifest Law. From startups to scale-ups to enterprise, Manifest Law is the AI native immigration law firm. Just think about that. AI native immigration law firm? Come on. This is great.
20:01There is no such thing, but there is Manifest Law. And they have all of your immigration needs handled. It's real-time visibility. It's fast filings. And did I mention the AI? It's got an AI-enabled workflows, a tech board experience, and straightforward pricing. So if you are in HR and you are doing a lot of immigration, Manifest Law is the solution for you. If I was running HR, guys, I'd be getting it. All right? Go out there and check it out. So I guess, what is the competitive advantage now in this new era for Stack Overflow? Yeah. So I think we do a survey every year in our community. And there's a very well-known survey in the industry.
20:51It's called the Stack Overflow Developer Survey. It goes out to tens of thousands of our community members who respond to it every year. And every year, for the past few years, we've been asking a series of questions about, of course, AI. And so I would encourage your audience to go check it out if they haven't. It's a very, very good, detailed piece of work that our team does. And it's kind of unbiased. It's vendor agnostic. It's just by us. We are the Switzerland, so to speak. And one statistic that stood out over the past few years to us that we have glommed onto, which is around trust. And so when you think about the age of AI and what it means, it means tremendous potential and possibility when we harness the power of this amazing capability.
21:37And really, it's like quite, I mean, it's just like absolutely mind bending to think about the possibilities of, you know, what you could do with this in the long run. However, the road to the nirvana state, I think, has got a lot of potholes and issues. And one of the big issues is around trust. And so when you think about the statistic that we, that we around trust that we asked this year, I mean, it was quite fascinating that only, I think it was about, you know, six to 7 % of folks who actually said they truly would highly trust the answers of AI within their organization. And it was just, it's such a shockingly low number where people just don't want to give it.
22:21Especially from your community. Right, no doubt. And this is a very skeptical community. the technology community, but the usage of AI is very high. Just don't get me wrong. Right. Yeah, they all use it. But, you know, and, you know, because they probably use it more than the average person. Right. You would think that they would have higher trust around it. Exactly. But they seemingly, there is a concern that for it to be just truly for you to give it full autonomy to go, agents especially, to give them full autonomy to go do whatever they need to do within your organization. it is a little bit dicey for you to be able to do that.
22:56So then what are the issues around trust? I mean, it's around privacy, it's around security, and of course it's around the knowledge that you have. Are these agents working off the most trustworthy knowledge within a large organization? Are they working off of outdated information, people that just said something somewhere? So there's a lot of scattered knowledge across companies. There's a lot of information that is out of date, hard to find. And it's not as simple as let's go just plug in your AI tool into a bunch of places and it's just going to work. You have to really curate, validate, etc.
23:32And for us, that is what Stack Overflow has been doing for over a decade, curation and validation and producing this highly trustworthy knowledge. And I'll give you a very specific example around this. and so that's what we have decided our moat is our magic sauce so to speak is that to really do that really well to be that trust layer to be that layer that produces decision grade knowledge if you will and that's where we've doubled down with our new product called Stack Internal which we launched actually at OpenAI Dev Day this past I saw that on your LinkedIn congratulations yeah thank you it's kind of ironic that we launched it there given what I mentioned to you about November 2022.
24:15Right. Right. So, but that's how it works. So one data point around this real quick, Jessica, is the, you know, our public corpus over the past few years, we have done, our public site, we have done a lot of different things. One of the big things that we've done there is partnered with all the big AI companies. So we noticed a lot of our data had been used by the LLMs for pre-training and so on. So we struck these big licensing agreements with OpenAI and Google and like all the big cloud companies and so on. And so they pay us for ongoing access to the data on a yearly basis. And so that generates revenue for us.
24:51That was effectively to make up for the advertising issues that we had, because if you think about the traffic obviously went down. And so then we had a drop in advertising revenue. So the data licensing revenue stream basically became kind of a replacement for that in some ways. But we still have an advertising revenue. We still have tens of millions of people come to the site. But I'm just giving you a relative from relative to the heights of this platform. And have you seen like, you know, because obviously accessing, they're buying your data and accessing this data. You can access it on, you know, OpenAI or Clutter or whatever.
25:28But are you still seeing the number of the community kind of grow and sustain, even though that's the case? Yeah, the questions on the site are a lot more advanced than they used to be. For example, a lot of the question flow, there's a viral negative chart about Stack Overflow, which is like this question count drops very precipitously. which is true because like the simple questions that people used to ask previously don't get asked on the site as much, you know, or any actually. So basically people ask them on ChatGPT or Claude and so on. But, and again, we're powering those platforms so then they are getting our answers, you know, through a different mechanism.
26:13Right, but they're still your answers. They're still our answers. Yes. Along with, of course, I'm sure a couple other sources that people tap into. But the thing around that is that the simple questions that people were asking went away, but the more complex questions are growing quite rapidly on our site. So we see, for example, AI-related questions around the foundation models, around your harnesses, around all the latest tools that you know in the news. All those questions have gone up quite dramatically. We also see major spikes in certain events. For example, yesterday, we had this whole PHP security issue that was going to work away.
26:48Yeah, I saw that on the news. And we had a massive spike in traffic on our site because, you know, people are all and I think our team did a really good job of giving them guidance on how to navigate things. So it is like it is the canonical reference for these technology problems that happen. And I think, you know, you're going to need that combination of human and, you know, AI related knowledge. So going back to the reason why I came on this track, I want to mention to you that when we did those licensing deals, what we noticed around the impact was that there was a close to a, and these are white papers that people have written, close to a 40 to 60 % improvement in the quality of the answers when people used our data, right?
27:30Wow. Things like there were certain evaluation scores called Human Eval, etc. So the LLMs got extremely better, or got materially better, I should say, through our data when it was used. So that same concept applies inside organizations with our product called Stack and Tunnel. And so just going back to your question of what is our moat? Our moat is to produce this trustworthy knowledge, which actually matters because it produces this highly accurate outcome. And then, by the way, that highly adequate outcome then reduces your token cost because you're not having to hit like all your context every time that you're hitting like this really high quality context to do your work.
28:08So that's the full experience. That's great. Well, I guess, you know, there's so many companies, you know, I mean, every company, there's not one company out there right now that's not trying to navigate this change. What is your advice to them when they are trying to shift their organization, when they're trying to figure out, how do I design these agents to do what I want them to do? And can I trust them? You know, and and how do I how do I help my people? Because, you know, there's some people in the organization who are really excited about this. And then there's some people in the organization who are kind of not that interested or, you know, just, you know, kind of keeping it at bay for now.
29:02Yes, yes. I think this is such an excellent question because I think pretty much all of what we're going to go through over the next few years is very much a people problem. It's probably overstated now, but I would say that it's absolutely true because how you use this technology to your benefit, I think determines the success of organizations. I think most individuals, I think, use AI very extensively. You and I probably use ChatGPT and whatever else, Gemini and other things in our spare time. But I think companies need to accrue the benefits of this as well, right? Because that's like the next organizing layer.
29:38So I think individuals use it in their spare time. But how do companies benefit from this, where your company can actually make the leap to the next level versus getting crushed? I think it comes down to a few things. One is you mentioned, of course, the trust. And we've talked a little bit about that, which is how do you ensure that you have enough of a balance between your own mind and your abilities and using these powers, you know, in your benefit within your organization. So I'll give you an example within our own company. You know, we really had, we were like, I would say, when we started a couple of years ago around this topic of AI adoption, very much, I would say, you know, probably behind average, in fact, within the AI native companies of Silicon Valley.
30:28And it was basically, hey, some people were already leaning into it. And we had a small crew that was experimenting with things. And a lot of people were skeptical for a long time and saying, look, I don't know about AI and XYZ because it's a lot to navigate. But then we started really empowering small teams to go and build great products with AI tools. So I'll give you a very specific example. One of the things that we launched with what I like to call inside our company a small and mighty team is something called Stack Overflow for agents. So basically Stack Overflow, the public website, we created an agent equivalent rather than for humans, for agents to share information and validate each other's learnings on that public website.
31:12And that has been one of the fastest growing things that we've launched. And it was launched by literally two engineers on our team. And it was launched with a small supporting team of other designers and a couple other folks. But it was all written with AI and specifically, you know, opening a chat GPT codex, which they love so much that they did a case study on us and how well we did that. You know, that's posted on the web right now. And so that small and mighty team was a way in which we could communicate to the broader organization saying, look, this team shipped a product that is now reaching millions of people in literally a record amount of time with a very small number of people.
31:56And that's the power of AI. Right. Right. And and you're using these are very smart engineers. We had some of the best people on this. But so how do you use your brilliance of a human, your human ingenuity, plus the heavy lifting of the AI tools to remove all the craft and like, you know, the work that you perhaps don't want to do as much to produce great outcomes for users and customers? And I think doing it by example, giving them people, giving inspiring people on, you know, with sharing those wins. And it comes down to leaders who talk in a repetitive way every Friday about that. is how we ultimately got now, I would say a very big shift in the organization.
32:37Pretty much everybody uses AI in our company two years later. And they're doing all sorts of innovative things, including the product that we just launched a couple of weeks ago with Stack Internal. Which is so exciting. And I tend to agree. And when I'm not doing it like you, I get to just advise people now. But I have been advising people to kind of start small and, you know, kind of go one team at a time. But I think what's a little daunting about it, and maybe you can help shed some light on how companies might want to think about this, is that if you're a bigger company and you've been working a certain way for a very long time it's really hard to like shift and work a different way and um and i mean that's everything from not only the engineering team but to you know the finance team and you know the marketing team and and what have you so how do you think companies should think about redesigning how they work Right.
33:52Yeah, I think this is such an excellent question. This comes up all the time as we engage with our large enterprise customer base. So we serve in our SAC internal business, our enterprise business, all the world's largest banks, tech companies, healthcare companies, you know, they're customers of ours, you know, they have anywhere between 50 ,000 to 100 ,000 employees each on that enterprise platform, right? Which is a lot. Which is a lot of people. And so when we engage with them, we definitely, when I talk to CTOs and CIOs, it's all about this conversation of all of them really leaning in in a big way.
34:27The technologists are definitely leaning into these places to establish the foundation that they're trying to effectively bercolate across the organization. A couple of years ago, it was about experiments. A year ago, it was about pilots. This year, it's all about like in production everywhere. and it's like it's a go big or go home situation because you know that the clock is ticking on their investment also right yeah to spend a bunch of money so the pressure in the system is very very high to say look you just spent a bunch of money and what am i getting out of it exactly it's it's it is murky at best by the way still because like you know i know because all these companies are like we're spending all this money you know um we haven't seen revenue go up yes you know we haven't seen sales increase.
35:12They're the same as last year. And, you know, product innovation's the same. So what are you telling them? Yeah, I think it's the, you know, it's the other phrase that is constantly used, which is like, you know, you underestimate things in the short term, and you overestimate it in the long run, right? Yes, effectively, or maybe it's the opposite. You overestimate it in the short run, and you underestimate it in the long run. And so in that context, it's very much like people are expecting a lot in a short period of time. But to your point, it is a human problem. But once you do change the way in which you actually re-engineer how you actually work from ground up, how do you use these agents as like really as teammates to do a lot of the work that you're planning on doing?
35:56Like I don't have a chief of staff, but my agent is my chief of staff, right? I don't have my own product manager on things that I care about, but I have an agent that does that for me. It goes and pulls all my calls that I did with customers. It validates what, you know, it was relevant from a product standpoint. It documents it for me. And then I'm able to share that with my product team to say, look, here are all the calls I did. And here's exactly what I think we should be building. So those sort of things give individuals tremendous leverage. And over time, if you think about how that compounds, it should mean that as an individual, you should be able to do a lot more.
36:32I should be a much better CEO post-AI than I was pre-AI because I have a lot more things at my fingertips than I did before, right? Or at least faster and better and so on. Yeah, well, are you trending? Yeah, I guess somebody else should be the judge of that. But I would say, no, I believe, I feel a lot more empowered with these tools to be able to get to the truth a lot faster. And so that CEOs sometimes, bad news gets them slower. you know, you've got to go several layers down to, you know, a pretty flat organization, but larger companies, like you're saying, I think it reduces the, you know, like the kind of the layers within the organization, because you're able to get the ground truth a lot faster.
37:16Yeah, I mean, you know, a lot of companies are, you know, sort of saying, and I believe this to be true, I think companies will be flatter over time. And guess what, everybody's a manager now, because even if you're not a manager of a person, you're a manager of an agent. But there's a lot of stuff in the system and things in the headlines about you should be able to not have to hire and you can probably lay off 10 % of your workforce or more. I haven't seen that to be true yet. But, you know, there's still a lot of human need today. 100%. And I've been pretty vocal about this is that I think for certain parts of the world, I think this is quite a devastating thing for jobs, by the way.
38:13So going back to places where, you know, things like call centers and places like that, I think it's going to be absolutely brutal for those folks because you're going to be able to automate a lot of that away with tools. And, you know, it's going to be much even even those were originally sought after as low cost regions to go do that work back in the BPO era. It is going to I think it is going to be a negative thing. But especially for the, let's say, for the U.S. or the Western economy and so on, or even the globe, I would say, generally speaking, it should generally democratizes knowledge and intelligence in such a large way that there should be a lot more people that should be able to do a lot more things.
38:53And so the whole, you know, the notion that there's going to be an infinite number of problems that you can go solve, because that's just humanity, right? Like, it's like, if you just project out, I love, I'm a big Star Trek person. So for me, you know, if you just project out, you know, 100 years out or 200 years out, you know, there's so much to go keep doing, right? It's not like, oh, we solve all the problems, you know, we're all out of jobs. I mean, no, there's going to be plenty of jobs that we don't even realize what they are at the moment. So I firmly believe that AI actually creates a lot more jobs in the aggregate and over a long period of time.
39:26And I don't think it's going to be the opposite. I also don't think it's going to be a straightforward path because there's going to be obviously a lot of jobs reskilling and people who are going to say, oh, I guess this is not a thing anymore. Like, I don't need to keep doing my job because AI does that very quickly. Okay, so then what do I do with my time? Okay, you got to then do this, which is a higher value added thing in the same chain of events. And so it just, I think, requires people to be comfortable with change to then embrace this in a way that then they can take advantage of it. Like, there's so much more that can happen and so much more that you can go build that, you know, you should just go fully into it and 100 % immerse yourself.
40:08One other point I mentioned here is it may be that companies will try to do more with less, like larger companies as they try to prove out ROI and so on. But as you're seeing, there's a complete explosion of new company creation. In fact, it's only going to increase at this rate. Whether that is individuals. Another interesting stat is that we saw a lot of freelance developers in our survey this year, and we saw an increase. So meaning people feel a lot more empowered to go and work on many new projects. They're able to do it on their time with AI enabling them to do that. And so I think it creates this new job creation with new startups, new companies, I think will be the job engine.
40:52And that's kind of, you know, I think it puts what the U.S. has really been amazing about in the past on steroids, in fact. And I think new job creation accelerates. So I think if you do it correctly, you should be able to put folks who are, let's say, not working at larger companies or asked to not work at them anymore into the new, back into the economy and make, you know, get them to be productive again by working on these newer things that are going to come up. So I'm an optimist in general on this subject. Well, so am I. So I have a seven-year-old as of yesterday. He just turned seven. I don't know if you agree, but I think he needs to study the liberal arts.
41:36I mean, you know, because I think like, you know, judgment, creativity, critical thinking, like all those things become more important in this next era. But I don't know if you agree. Yeah, no, I'm a general believer that being well-rounded is very important. My whole education I mentioned to you previously came from a technical background and then went to a business background. But that's me. Everybody doesn't need to be me, right? So everybody's got their own passions and their pursuits. I think people should pursue what's interesting to them, and generally speaking. The great news about AI is it's fairly universal across.
42:18You can imagine it having an impact across pretty much any fields that you can think of in a hopefully positive way. And I think following what actually is interesting to you, I think at a minimum, I think is good because then you can do some amazing things with AI as like this accelerant for your work. So that's that's one foundational point. With regards to liberal arts or science and math and those sort of distinctions, I actually think it is, I think it's, I would advise people, again, beyond their own passions to think of these as like equal opportunity things. They're all going to be very useful because science and math and computer science, economics, all these topics are, I think, very useful because as you think about using these tools, I don't think you should use them blindly either.
43:05So, you know, a very simple example that I've used previously is, you know, when I was in high school for the first time, well, actually in college was the first time I was able to use a calculator for, you know, an exam, right? Right. Then it was just like, you know, you handwrite stuff in high school and so on, which was, you know, I went to India, I went to high school there. and it's not like, you know, if I had fat finger two plus two equals, you know, if I made a mistake when I was tapping it into my calculator and I got a wrong answer and I just like blindly just followed that, that'd be foolish of me.
43:38So thankfully I learned arithmetic, you know, through schooling. And so I know when it's wrong. So like, you know, I catch myself. So I think it's important to, you know, really learn the fundamentals in whatever you're doing. So like if you are going to be writing code through agents, I mean, you should go learn computer science. I don't think it's a bad thing. I think it's a very good thing for you to know exactly how these things are the building blocks that have been built up over time. And back then it used to be assembly code and more and more abstractions. And now we've got the highest level abstraction with, you know, AI agents writing code.
44:10Right. But, you know, I studied computer engineering. So I studied assembly. I studied basic and C++ and Java and whatever else. So I think that's important as a foundation always. So I'm, again, equal opportunity on those. Okay. All right. Well, I am happy you said that because now I'm thinking differently. And, you know, right now he wants to be a ninja. So, you know, I don't know. We'll see what happens. Follow his passion. We'll see what happens. So we're running out of time, which is sad because I've really enjoyed this. I'm going to ask you, and I always put a little pressure on my guest to leave us with something inspirational.
44:53And so as people are going off in their week this week, what is it that you want to think about? Yeah, I think, you know, maybe taking a step back, you know, humanity in general, I think that we're all. I think this is like a big human experiment on this in this world. We're all kind of together going through our individual stories. We've got our context. We've got kind of our past. We've got our experiences. And I think know that everybody around the world is going through some sort of a test of some kind. And it's not just all about you right at the end of the day. We're all here because we've got, I firmly believe this is sort of like a school of sorts where we're all going through it together.
45:37And so in that context, know that you're not alone. Know that, you know, if you put your mind to it, you can do it, whatever it is. You have the ability to. It's just a matter of trying to figure out how you get there. There's always a path. Just don't give up and just, you know, just do whatever it takes to get to whatever goal you set your mind to. And it doesn't matter what that goal is, as long as it's the one that's for you. And then at the end of the day, we're all going to the same place, you know, when we're 100 years old. So don't fret. Don't take it too seriously. don't take yourself too seriously but enjoy the journey along the way it's going to be a lot of you know change along the way but embrace it and it's going to be a lot of fun I love that do whatever it takes people well Prashant thank you so much for coming on the show I really enjoyed this and then next time please we have to do it in person for sure and we have to get together I mean this is we live too close to each other not to see each other so thank you so much thank you Thanks for listening to TruthWorks.
In November 2022, ChatGPT came out and traffic on Stack Overflow went up, not down. A huge flow of ChatGPT-generated questions and answers hit the site, and its community decided to ban ChatGPT. What the company had been building for about a decade changed, in his words, "over a moment's notice."
Prashanth Chandrasekar is the CEO of Stack Overflow. He came to the US on a college scholarship to the University of Maine in 1998, worked in consulting and banking, spent seven years at Rackspace, and took his first CEO job at Stack Overflow in 2019, three months before the pandemic. Host Jessica Neal talks with him about leading through the ChatGPT shock, the trust gap in AI, and why the shift ahead is a people problem.
WHAT YOU'LL LEARN
CHAPTERS
0:00 Preview
0:32 Truth Works intro
1:20 Welcome to Truth Works
3:00 An immigrant on a scholarship to the University of Maine
5:31 "Go get some operating experience": Rackspace
6:52 First CEO job, three months before the pandemic
7:40 What Stack Overflow is
9:35 November 2022: the ChatGPT moment
11:26 Traffic went up, so the community banned ChatGPT
12:17 Radioactive metal: protecting a human-curated corpus
13:24 "Quite terrifying": how scared was he?
14:16 Leading through the hardest transformation
16:32 Chief repeat officer: an email every Friday for seven years
18:27 Teach it to learn it
19:42 Sponsor: Manifest Law
20:32 The trust gap: only 6 to 7% highly trust AI
22:40 Privacy, security and scattered knowledge
23:41 Stack Internal: decision-grade knowledge
24:09 Licensing deals with OpenAI and Google
25:15 The viral chart: simple questions gone, hard ones growing
27:10 40 to 60% better answers with Stack Overflow data
28:11 Advice for companies: it's a people problem
30:10 From behind average to everyone using AI
30:41 Stack Overflow for Agents: built by two engineers
32:46 How big companies should redesign how they work
33:50 50,000 to 100,000 employees: go big or go home
35:01 Where is the ROI?
35:50 "My agent is my chief of staff"
36:41 A better CEO post-AI?
37:15 Will AI take the jobs?
40:16 Freelancers and new companies as the job engine
41:18 Liberal arts or computer science?
43:01 The calculator lesson: learn the fundamentals
44:44 Something to take into your week
ABOUT THE GUEST
Prashanth Chandrasekar is the CEO of Stack Overflow, the developer question-and-answer site founded in 2008. He came to the US from India on a college scholarship to the University of Maine in 1998 and studied computer engineering, then worked in operations consulting and, after business school, in banking. In 2012 he moved from New York City to Texas to join Rackspace, where he spent seven years building its cloud businesses around AWS, Microsoft Azure and Google Cloud. He became CEO of Stack Overflow in 2019 and has led its move from an ad-supported public site to an enterprise company, including its enterprise product, Stack Internal.
ABOUT THE SHOW
Truth Works is hosted by Jessica Neal. Each episode is an honest conversation about why work is hard and how leaders can make it better.
This episode is sponsored by Manifest Law, Findem and True Search.
#TruthWorks #Leadership #AI #StackOverflow #FutureOfWork #Developers #AIAgents #CEO