“Countries are coming online tomorrow, whole countries”
en
January 28, 2025
TLDR: RJ discusses ClickUp, a work and chat platform to streamline workflows and increase productivity. Stack Overflow user 'g_mauk' helped explain it.

In this episode of the Stack Overflow podcast, hosts Ryan Donovan and Ben Popper welcome RJ Tuit, the head of front-end UI and client architecture at ClickUp. The discussion revolves around software engineering, productivity tools, and the future of technology as nations increasingly move online.
Origin of RJ Tuit in Technology
- Early Beginnings: RJ’s journey into technology began at 13 with a VIC-20. His fascination led to learning programming and working with various technologies, including Pascal and PHP.
- Startups and Failures: At a young age, RJ launched startups, including a social network and an e-commerce site for vinyl records, which taught him valuable lessons in failure and the importance of understanding business at scale.
- Career at Microsoft: After a challenging path, RJ joined Microsoft, where he played a crucial role in developing MSN and Bing over 17 years, witnessing significant growth and innovation, particularly during the pandemic.
Transition to ClickUp
- Joining ClickUp: RJ transitioned to ClickUp, drawn by the company's vision of creating a unified productivity platform that integrates various tools needed for collaboration and workflow management.
- The Challenge of Scale: Both ClickUp and Microsoft deal with millions of users and must navigate issues like technical debt, performance, and product development amidst rapid growth.
The Philosophy Behind ClickUp
- Unified Workspace: ClickUp aims to consolidate numerous productivity tools (like chat, task management, and documentation) into a single platform, solving the notification overload that plagues many companies.
- Agentic AI Integration: As technology evolves, ClickUp incorporates AI to enhance productivity, automate administrative tasks, and streamline workflows, presenting an innovative take on team collaboration tools.
Challenges in Software Development
- Technical Debt: The podcast discusses how managing technical debt is crucial while ensuring that new features do not hinder performance. RJ shares insights on the balance between growth and technical stability.
- Scalability Issues: RJ recalls stories from Microsoft about scaling applications seamlessly, particularly when rapid user increases stress existing resources, as seen during the pandemic.
- Performance Measurement: RJ emphasizes the importance of developing performance measurement tools in complex applications, ensuring that engineers can quickly identify and resolve bottlenecks in productivity.
The Role of AI in Productivity
- AI Hype vs. Reality: RJ expresses skepticism over the AI hype, stressing that while AI can enhance productivity, it's essential to grasp the technical realities behind AI applications rather than succumb to marketing jargon.
- Genuine Use Cases: He highlights practical applications of AI within ClickUp, such as automating task creation from conversations, thus saving time for users and enhancing team dynamics.
Key Insights and Takeaways
- Productivity Tools Integration: Companies face notification sprawl and context-switching challenges. Unified platforms like ClickUp can alleviate these issues, providing cohesive user experiences.
- Navigating Growth: Growth comes with challenges such as technical debt and performance issues, which, if managed correctly, can propel companies to success without sacrificing quality or user experience.
- AI’s Practical Application: Companies can leverage AI to improve user workflows but should focus on meaningful implementations rather than following trends that may not deliver real value.
Conclusion
This podcast episode with RJ Tuit brings valuable insights into the evolving landscape of technology where streamlined productivity tools and AI integration play a crucial role. The lessons learned from scaling, integrating technology, and managing technical debt are fundamental for engineers and business leaders alike who navigate the challenges of an increasingly digital world.
Was this summary helpful?
Hello, everyone. Welcome to the Stack Overflow podcast. I'm Ryan Donovan, your host. And today I'm joined by Ben Popper. How you doing, Ben? Hello, Ryan. I am here hosting the Stack Overflow podcast in my new role as the Stack Overflow podcast host, Cole and nothing else. That's right. But I will be sticking around for a bunch of episodes every month. I've loved having this gig over the last six years. I've loved doing it with Ryan.
And they can't get rid of me. I'm not going anywhere. So good luck with that. Right. And all the hate mail you want. We'll talk more at the end of the episode about other ways you can find me online and what I'm up to. But.
This is a place to talk all things software and technology on the show. And today we are lucky to have RJ Tuit from ClickUp. He is the head of front-end UI and client architecture over there, formerly Microsoft. So we're going to learn a little bit about what ClickUp does. There's a lot to do with sort of centralizing all the activity that happens across an enterprise into a single chat platform. And then now this is 2025.
layering on some agentic AI or some, you know, AI assistance to help make that even better. So without further ado, RJ, welcome to the program. Thanks for having me, Ryan. Thanks, Ben. It's a pleasure to be here. Yeah. So at the top of the show, we'd like to kind of get a sense of our, I guess, origin story. How did you get into software and technology and how did you get to the role you're at today? I started out was one of those weird ones where it's like, I think I started out when I was 13 on a VIC-20. It's actually here. I don't think you can grab it from here.
It was actually like, this was the actual Commodore VIX-20. I have a tape deck too with old cassette tapes. I have a tape deck back in the day on my C64. Right. Good crew of ancient mariners. It has the aesthetics of a typewriter, not a computer. I like that. And it's like, have you ever seen that it's brown? Like all these devices are brown. You know why they were brown? Why? Because everyone smoked. So if you made something white, it turned brown. So if you made it brown, it stayed brown.
right. That's right. You want a nice tobacco juice brown. Exactly. I can always have a funny to listen to. So yeah, I got lucky enough that my parents were in the Netherlands were wealthy enough for me to get me a victory. They saw me play with the computer at some other friend's place and they were like, he seems to like that a lot.
So they got me that, then moved from that to breaking my dad's computer many, many, many times over. Learned, you know, my programming there. Went from that to Pascal to Borland Delphi, I think. And then did all kinds of weird stuff in software, a little bit 3D studio modeling, a little bit of Photoshop Illustrator, a bit of video stuff.
But all the while kind of stayed on this like web-based development. I think it was called Fusion. Originally, that was kind of like the big one for me called Fusion and PHP. Did a startup in my like early teens, not early teens, early twenties, which was the one was Sweet Lake Nation, which is a social network.
The other was a online e-commerce website for vinyl was interesting was together with one of the biggest online EDM producers in the Netherlands back then. That was like my first foray into like e-commerce and like like a large scale web. And then did that for a while, failed at it miserably. Let me say that upfront. We all we should have our failure stories, right? Like I was no idea what I was doing, honestly.
felt that it miserably kind of went into consulting, realized that I needed to learn so much more about the business world and like just how you built things at scale before I could ever go back to doing something like it. Did consultancy for I think five or six years started getting into one of the, you guys probably noticed as being podcast hosts. One of the best ways to get jobs was to go to conferences and speak.
I was almost the absolute best way to do that. So I slowly started getting into that, made it into a bunch of conferences, made it to Mix09, Mix06. Mix was in Vegas, I think, and ended up at a table with a Microsoft employee who was like, do you want to work at Microsoft and me being an ignorant Dutch guy? I was like, oh, it's just an American being nice to me. Like, you know, there are various people in the world think Americans are there just like, oh, yeah, friend.
But yeah, no, an hour later, I had an actual interview request that I got flown into like, I think a month or two later, which I bombed.
terribly, which they gave me another try, which I kind of bombed, but did good enough that they let me in. And from that moment on, I worked at Microsoft, moved to North America. It's about 17 years ago. Right. Did the interview questions you bombed have any relation to the work that you actually did once you got in there? No, of course not. No.
No, I've I ran interviews for 15 years at Microsoft and like up until the last moment I was there, it was still like running these interview questions that I know I would be excluding a whole bunch of people that would be doing perfectly fine. But we don't have anything better. It's bad. But we don't really unless we like people want to come work for us for like two weeks before they like take the job. I don't still don't really know what the solution to that problem is.
Yeah, yeah. It's an interesting question. Maybe apprenticeships or something like that are part of the answer. And so now you are at ClickUp. Set the stage a little bit there. How did you make the transition from Microsoft to ClickUp? And for folks who are listening and don't know, how would you describe it succinctly? So starting at Microsoft, going to meet my way through worked on like MSN, worked on Bing. In my time as MSN, we got to rewrite.
MSN was kind of my first foray into like large scale. It was like half a billion users a day across the world. And you're trying to rewrite something for that. First time Azure, first time cloud. We're kind of the first ones trying to go in that area. Built that, redid that. And then the team in control of MSN, they were the leader there. And I'll bring him up probably a lot because he's been very influential in my career. His name is Brian McDonald. And Brian McDonald was
to give a history, he built projects, sold it to Microsoft, built Outlook, retired, came back to do Azure, and then did MSN, and then he was like, I want to build something new, like Outlook, I want to go build Microsoft Teams.
So he went from Emerson to start. It was originally TeamSpace. I have some like logos here of like, we even have a version where it was called Microsoft Office Chat Teams or something. Something terrible Microsoft, you know, those massive names. Yeah. Microsoft Office Chat. There's Yammer in their teams. Yeah. Yeah. Yeah. I was at the version of that. I was Yammer. Absolutely. Skype too, by the way.
We had all of those. Yeah, Brian started that, brought us along, and that was mostly back then Slack Compete. There were three or four groups within Microsoft that were building a chat Slack Compete. It was seen as a danger to the company, a danger to the office, like bundling strategy, like if you get someone else in there, someone else can start bundling. So doing that for about one to two years, and that's where the crazy skill kind of came in. I've seen that that's from like, we were 10 people
serving no one to like a million and then 10 million and 50 million. And then in the end, like when I left two years ago, it was like 250 million users. And the organization was like two, 3,000 people. So that crazy scale up was like probably the most interesting that I'll ever see in my life with COVID thrown in. That was crazy times. COVID during being a team's
We were just about at that plateau where we were going to start doing quality and be done with development. And then COVID hit and they were like, whatever, just do whatever, turn off features, make it work. Countries are coming online to more whole countries. Go figure out how to make it scale. We're like, but we're running out of data center space. That's fine. Disable stuff.
whatever you need to do. Wow, that's an interesting hyperscaler story. I like that. Yeah. They weren't also like, and we're building more data centers because isn't that always the case? Well, they were, but they take like a year to build. Yeah. Yeah. So it's like, we had no room. It's like they had planned for them. They had, they planned them like what, two to three to four years ahead and because there's all these things that need to happen. They went way faster back then, but it wasn't the like next week type of solution. Like we were literally running out of space all across Microsoft was crazy. Yeah. So you just see really shift onto somebody else's cloud.
They were running out of space too. Yeah, yeah. You couldn't multi-cloud, you couldn't multi-cloud it. One thing Ryan and I have talked about so many times, and it's interesting to hear you say this, is how do you pay down technical debt? How do you keep yourself from getting the sort of software bloat and sprawl? Force deprecation due to data center concerns because of growth is a really interesting moment to be like, well, what are people using the least? We got 50 features in here, cut the bottom five, just because.
Well, that normally would be such a difficult decision to make. Somebody who made that feature would stand up and defend it. But did you feel like in the end maybe it was actually good for the product? I think it was good for the product. Honestly, it's the only time in my career in enterprise software that I've seen more than a single feature being killed. Both at Microsoft and at ClickUp, I don't think I've ever seen anyone go, oh, no, that's too low usage. Let's just get rid of it.
It doesn't seem to be a thing. This is the only time in my career where it's like, we made a list and leadership was like, check, go do it. And we're like, OK, delete code. You know how much software engineers love deleting code? It's like such a fun thing to do. And now let's click up. You're also dealing with a lot of scale. Is that right?
given with millions of users. It is millions of users. And in the same way, we kind of need to go back to that. If we're talking about that, what you're talking about this, like scaling, when do you pay down debt? When do you not pay down debt? To me, that's actually the most interesting thing we do in our industry. Because perfect architectures are easy. I can define your perfect architecture tomorrow. But we're never going to build that because we don't have unlimited amount of people unlimited time. And generally, when you're in a competitive landscape,
Like you have even less time because someone else is going to move faster than you, right? Like depending on where you are. It's probably the biggest lesson I learned at Microsoft and are now applying at ClickUp where
you walk that fine line between, I do enough to not make it fall over, but I don't do too much to slow down the business too much. And there's going to be a moment where it flips, where you go like, now we need to stabilize, and we need to have no longer quality issues. But in SAS specifically, and in a space where Teams was, and now ClickUp is, we're kind of in the same space. It's even bigger, more competitors.
Like you don't really have that luxury to go like, oh no, we're going to go pay down that for six months. Like you do that, you've lost a whole swath of new customers and you're not going to be able to like get the business to the place it needs to be. And that's probably, I'd say, the hardest that we do as software engineers and software engineering leaders is like doing that, letting things fall by the wayside, letting it happen and go and like, yep, I know. And engineers go like, oh, we're doing this bad thing. And I'm like, yep, I know. We're watching it. We're watching it every day.
Yeah, doing strategically bad things to make sure the business is interesting. And maybe we can segue here into a little bit of what the business does. I was doing some research and this is certainly an experience I've been familiar with at many companies. I've been at the overload of different services and notifications that you get from different departments or even within your own department that force you to be constantly context switching.
in Slack and get a notification about a Google Doc, but I have to update the Monday sheet, which will go to the air table, which pings the, you know, Jira ticket. And it's just like, man, you know, there's got to be a better way. So what is sort of ClickUp's philosophy and how does this play out in, you know, sort of maybe like the MVP version of the product, if you could describe it? You just named all of our competitors, which is kind of crazy to think about like every single one of those. Sorry, that was not intentional. I was just those are no.
No, but it's good because you're actually describing, like, to me, when I try and explain, click up to someone, like it's sometimes hard to explain how large the space we're in, because we're in all those spaces. We are chat. We are task management. We are docs. We are meeting. We are calendar. We are like all of those things combined, which like solves the, I'm going to go back a little bit because the story here to me is fun.
In teams, the reason it was started is because Brian McDonald came up with this idea of Metas. He had seen the Windows thing and then he'd seen the iOS phrase and it's like, it's this thing where you go like, if you have this operating system, it doesn't need to be a physical one. It can be a virtual one that has everything in it. That's where you have market attached, not just a great user experience, but it's like where you truly get like market and people attach.
And Teams originally was set up to be this, like this app of everything. We started out with chat, and then we added in meetings, and then we added in SharePoint for files. And then we were trying to do calendar and other things, and that ran into both organizational and other issues, because everyone has their own, I don't know if you've talked to any Google or Microsoft or other big company people, but like you run into these, the VP of X and the VP of Xbone, or both own a problem, and integrating is not in their favor.
And then COVID happened too, so like, and then COVID made everything meeting in chat. And we kind of lost that like vision of this like one productivity hub, like the one place where you do 90% of your work. Not 100% because you want depth places, right? Because trying to do everything 100% is going to not have you succeed. Like you want maybe even 80%, like 80% of your work and all the tools in one.
So the reason i left microsoft was it became too big and too large and it was like bureaucracy but the reason i joined click up is because once i start talking to zeb i realized that his vision was that. Even though they started out as a jira compete probably just as team started out as a slack compete division was always there is this like one app that is going to do all of them and back then when click up started doing that which was almost six years ago.
It was a radical idea. None of the VCs liked it. They were all like, you know, how SaaS has been for the longest time. Like you need to go deep singular problem solving. And that's like the thing that the VCs want to fund because they understand that and like doing to do too much things allows you not to focus. But the click up focus has always been try and solve the holistic like product problem is like, what is your day to day workflow? And not end up in these places where Microsoft and a Google and even a Salesforce end up, which is like, honestly, they're iframing stuff.
That's how they scale. Like they have hundreds of engineers here and hundreds of engineers there. And the way they scale with each other isn't by having them all work in one code base. It's by basically giving each other iframes, which works mechanically, but it never gives you this like cohesive experience where you can go like, no, I want a chat in which I want to insert a doc and a card. And then I want to be able to close that ticket right there that triggers like a calendar event. And that'll visually just kind of like works.
That has always been that vision of ClickUp. And for me, when I try and describe ClickUp, I take a lot of words to do it because I don't think a product exists today. That, that is what the ClickUp like vision is. And we're getting very close. Like we've, we've, we're, we've added chat, which is rolling out, like we've added calendar, we've added syncups. And we only have a couple of more small things to do. And then like the broader surface area of ClickUp is there. And you can kind of see that, like,
Like when that happens, I had the same thing happen in Teams. Like at some point you realize that you're spending 90% of your workday, me as a person building Teams or building ClickUp, you're starting to spend 90% of your day in the product and you go like, it just happens at some point. It's very hard to explain. I recognize it now because it's been, it's happened twice, but it's like there's this moment where you go like, Oh, I spent 80 to 90% of my time in the product while developing the product.
And that's super strong. To me, that strong product feeling and just getting to that place where you go, oh, this is going to work. When this happened in Teams, I realized we had something. And the same thing has been happening in ClickUp, where we've been a great product so far. But that part is what makes it
amazing. Yeah, I mean, the messaging, the chat apps are such a key part of it. And you start attaching things to that chat app. I remember at my last job, you know, trying to do internal documentation, I was like, where do you go for answers? And it was slack for like two thirds of the people.
because that's where everything happens. Yeah. I guess, you know, one thing I'm curious about as you're, you know, bringing all this stuff together, you mentioned spending 80% of time and then sort of feeling like, yeah, you've gotten into a flow. Like I enjoy editing video and audio sometimes because I end up just doing a two, three hours inside of one thing. I'm not constantly pinging back and forth and checking email and, you know, shooting off a slack and then writing on paragraph and Google document and then running that loop over and over again.
Do you have ways that you try to measure flow state and attention and productivity and then feed that back to your customer so they can understand the impact the tool has? We do and it's it's a very contentious subject as you guys probably know because there's also a lot of people that think it's being used as work tracking.
which is why we've had a lot of discussions internally on how to actually do it because it's difficult, right? Is the boss going to start using it as a way to see if I'm doing enough work or is the person going to use it to make me more productive? Those are two very different views on that same problem. They solve the same problem. They use the same data, but it's a really important
that we fall into one category and at the other. Because if you get too many users hating your product, it doesn't matter what the bosses and the companies want, you're probably going to not get there either. So you have to walk that fine line, which we spend a lot of time internally discussing. Like, what is that right? Like, like measure that we can do. But yeah, so we track where we have another piece of the product that isn't out yet, which is teams, which actually tracks teams, like usage of the product.
And we even import external data. So we have GitHub data because we have like search across. So like one of our products is that you can actually import data from other companies and then search through them. We have a large search index and we also use that data to aggregate like statistics from that as well.
So we're going to start giving you, like if you're thinking software engineers, like we can get at some point very close to like what the X can do, which is like give you a number of PRs, but also how much did you chat about those PRs? What is the documentation you've written about them? Which tasks were associated with the PR, which for us is already linked. Like the way that my workflow is is like, I am in click up, I create a task, which like gives me a button to create an actual GitHub PR and that PR then automatically gets linked. Anything I do in both,
gets updated on both of them. And then when it closes, same thing happens. That ticket closes. It properly gets the status set through all the rings. So you can see how we have all that information in ClickUp. We've been playing with it internally a lot to see what the valid data points are. I don't feel like we fully hit it yet, but we're close.
It's very difficult to define productivity as a single number across everyone. Yeah. And like you said, it's hard to keep that as a measurement and not as a target, right? I want to go back. There's something in your, in your, the initial pitch about the scale of things you're working on where you start finding bugs in other people's software, right? There were two that I mentioned. You found one, uh, working on teams in kind of bug and chromium.
Those there for a while. Permium was here the other way around, but yeah. Okay. And then you found one in Akamai that was 20 years in the hiding. Can you tell us about those bugs? So like you said, like when you start doing things at scale where it's like both your code base starts getting really large in a space that was never really intended to have that much code. Combine that with like hundreds of millions of users as you get into these like weird problems that like no one's just ever run into the scale yet.
Or it's been there, but no one's been large enough to observe it. The Akamai one was kind of one of the latter where
We were probably the first large enterprise app that was completely web-based. And then we were electron-based. And so we were experimenting a lot with what that looks like. Because the world we used to come from was being an MSN, was like single bundles, don't download too much, try and make it very efficient to start up. When you go to web apps, that trade-off becomes very different. And we were kind of exploring what that looked like. And one of our explorations was we can make everything faster, especially when
I don't know when HTTP 2.0 came out. But when HTTP 2.0 came out, it no longer mattered how many connections you made. They were all going to go over the same connection. And therefore, we could now, instead of giving you one large bundle and dealing with that, we could just give you the bundles as tiny as they are. So we started scaling that down. What that meant is that every user started downloading hundreds and sometimes thousands and sometimes 10 thousands of files every time
They connected to teams and we shipped every week. So this started happening at like a really high cadence and we started to see in these like bundle downloads, random failures.
Like across the board, like it was probably like 0.1%, but it was enough at our scale that users were complaining. And very often it was solved by a refresh. It was really hard. So we spent all this time trying to dig in, trying to add logging, trying to add other things in.
I don't even know. We spent months, I think, tracking this one now, because we thought it was something in our code. We were like, what did we do? What is happening? We replaced Webpack with some custom downloader. We added in all kinds of plugins to see if we could detect it. We even added custom C++ in Electron to see if we could maybe find it there.
In the end, what it ended up coming down to, and I don't remember, I was thinking before this podcast, who was the one that found it? I think it was Sophia, but I'm not 100% sure. They ended up
going, just digging, like, you know, how some engineers just can actually dig into full stack. Like, yeah, he's like amazing people that just they can go top to bottom. And he ended up looking at a networking library and just running at scale of those networking libraries and see what happened. And we came to a place where I think one of the Linux network stack libraries has a bug in there that occurs like one in a million times or something.
I tried to look up what the exact bug was, but it was something that existed for 20 years, something that we've all experienced, but it generally is like one request out of a million, not specific to one user. So it's a different user every time. So each of us here sees it.
every six months. And when you do, you probably go like, whatever, it's my Wi-Fi hit refresh. But yeah, we were at that scale where it was happening often enough, where we basically found a problem in an Akamai library that then got fixed within like this, the funny part, like most performance in software, which you guys probably know, like the hard part isn't fixing it. Sixing is easy. It's like, it's finding it. And that's like,
one of the absolute hardest things to do that requires real talent like that like finding of things like people who can do that well are they are special. I treat them as very special. So RJ one of the things I wanted to ask is that I think anybody who's worked at a company you know that
Avails itself of enterprise software products pretty quickly becomes familiar with the sprawl and the notification overload that you talked about and the appeal of click up is kind of obvious. It's all in the execution, right? But one of the things that I saw when I visited the website was that a lot of emphasis, at least in the marketing on, you know, the homepage was on bringing AI into the mix and how that would add value. So I think our users are a little bit sick of hearing AI.
But I also think that Ryan and I increasingly are seeing it productized and even used inside of Stack Overflow in a way that's not baloney, where it's like, no, that actually helped us do something we couldn't do before, made it faster, cheaper, whatever. Talk to me a little bit about how ClickUp is using AI and where you think they're substance instead of hype. I like that. I think problem one is that we call it AI. I think most software engineers, that's where it grinds us the most, where we go like,
But that's not the AI I was reading about in sci-fi books has nothing to do with it. This is like an LLM that knows how to regurgitate content very well. And it does it in a way that is really useful. Like I use it every day, but all those AI things you're saying it does, it doesn't really do. Like I think I don't know if that's what you guys are feeling a little bit, but like that's what I feel talking to other software engineers. That's kind of what we get tired of. Yeah. I think I know what you mean, like sort of anthropomorphizing it and, you know, we're on the road to the singularity and AI agents are going to take over.
It's just the hype of it, right? Yeah. Yeah. Yeah. And it's how for engineers don't generally like hype, like, especially if we understand what's going on and we understand, like, if you understand how an LLM works, like it's not magic, like we don't necessarily fully understand what it creates, but like what it does and what it's good at. And what it's not good at is relatively well documented. Unless we have another like massive, like revolution, which we had with the original LLM, like this is about where things are going to stay and they're going to incrementally get better.
To me, bringing that to ClickUp is the thing that I think makes ClickUp specifically good at what today we call AI, is we have a lot of data. We have a lot of user content, and that's the problem that most companies are dealing with that are trying to do something AI. If you're just piping chat GPT with their content into your product, then that's part of that hype cycle that I go like, interesting.
But if you look at, click up, like you look at how much data we have of you. We have all the stuff that you just talked about, like about usage and how people work. Your, all of your content, all of your information is in there. We even like take in like a whole bunch of other data from other places. So if you, if you put an LLM on top of that, like it can do really, it can, it can have the same fidelity that chat GPT has, but on your content. Right. And that's where in my mind, AI is useful. Like I use AI.
like every day. I use it at cursor and via code and co-pilot and like some of the others because it's useful as a code regurgitator that is better than me. But in ClickUp, I use it every day to create tasks. Like I no longer create their own tasks. What happens is we have a chat, we talk about things for like five minutes and then we go like, okay, this is the work we need to go do. I hit right mouse click, I say AI create task. And honestly, 95% of the time, based on what it knows, it knows how to create it, who to assign it to,
which space to put it in, which status to assign. And I go like, that to me is useful AI. It's just saved me 10 minutes of work, sort of like an admin for the meeting that's giving action items to the correct people. Great. Yeah. You still have to review it though, right? Yeah, you can review them quickly if they're not right, tweak it. I personally love AI, but it's the hype around it that is difficult to deal with sometimes.
What do you think people focus on the hype as opposed to the technical details of it? Because it gets you funding. Yeah. People love the hype. They don't want to be bored by the implementation details. Well, you kind of have to. Like to be able to do the work, which is there's a lot of work that needs to be done. You need to get paid. You need to sell these amazing things. And therefore, you need massive hype and the massive hype will get you VC dollars. And the VC dollars allows you to iterate and see if you can find a solution, right? This is the core of Silicon Valley.
like in like a three word, like into a three sentence, like cosmos. And even beyond that, now you go to sell to an enterprise and you're like, well, we have budget, but it's mostly budget for gen AI tools. Do you have that in your product? And you're like, yes, we, yes, we do. Does everything we said before? But now it fits into your job. So there you go. Yeah. Exactly. Gen AI and federated search, like those are like the key terms. When you fit those, you do federated search.
And look out for that one. RJ, I guess, you know, is there something that your team has done over the last year or that you're looking forward to building on this year that you'd like to share with our audience? Just something you think, you know, for developers, we're listening team managers, CIO, CTOs. One thing we were like, man, that was a really cool, you know, product we put together or I'm really excited, you know, because this year we're going to try to roll out.
The funny thing, the first thing to come to mind is something we already did, but that was a short one. It's the one you asked about we didn't get to, which is like being able to find like the Chrome memory leak and being able to help hot fix that. That was an engineering perspective, interesting. But from a this year perspective, I think what we're going to be spending a lot of our time on is like performance tooling. Performance is always this like,
interesting, like super important, but kind of side aspect of software engineering that the tooling is very rudimentary. Yes, you have MenLab to run your like memory leak tests, but it's kind of limited. You don't have that for a lot of other things. You don't have that for your style recalculations. You don't have that for your other things that you would get out of a Chrome performance profile. And with the complexity that we have as ClickUp,
That's where all of our performance issues are. This is always the discussion I have. People go like, Oh, memory leaks. Don't just go unsubscribe to all your observables. I'm like, that's not the fact those memory leaks are fixed. We fixed those a long time ago. That's not what we're stuck with. It is a component in a component owned by a different team with a component and another component that back reference is something that back reference to component that like if someone changes something somewhere in that chain, this whole tree plus window gets leaked and we're screwed.
Just like that. Just find that. Just find that. Just find that. Just find that. Just find that. Just find that. Just find that. Just find that. Just find that. Just find that. Just find that. Just find that. Just find that. Just find that. Just find that. Just find that. Just find that. Just find that. Just find that. Just find that. Just find that. Just find that. Just find that. Just find that. Just find that. Just find that. Just find that. Just find that. Just find that. Just find that. Just find that. Just find that. Just find that. Just find that. Just find that. Just find that. Just find that. Just find that. Just find that. Just find that. Just find that. Just find that. Just find that. Just find that. Just find that. Just find that. Just find that. Just find that. Just find that. Just find that. Just find that. Just find that. Just find that. Just find that. Just find that.
We're going to be able to actually write some tooling at the beginning to write for about a decade. That is going to allow us to catch some of these performance issues early on and directly point to here was the problem. Here's engineer go fix it or, you know, who knows AI go fix it. I'm joking kind of, but like that is to part and probably most excited about. It's interesting that companies, you know, don't always care about performance until suddenly it matters. You get to a sort of scale where suddenly it's, oh, this is a problem.
You know, I saw a video that was like, here's all these big companies talking about rewriting their entire application stack for like 10, 50% performance gain. And at that point, it matters, right? It does. It's what I did at the end of teams. We rewrote it from AngularJS to react to get that 15% performance gain. Exactly what you said. But we already owned the market. The competitors were already like no longer there. And then you can take your time to spend like two years rewriting. You don't always have that much worry.
RJ Ryan, we're going to put a timestamp on this Tuesday, January 21st, 2025. RJ, we'll have you back in the year. If you have a mid-level engineer who's AI, I will obviously have to discuss it. I know you were kidding, but also we're not kidding. We'll see. We'll see how this year goes. Have you seen the LinkedIn ad of someone trying to hire an AI agent? If you haven't looked it up, it's hilarious. It's terrible and hilarious and kind of scary.
Yeah. Yeah. I mean, this is another piece of the hype that everybody hates is every time I go to YouTube, it's like seven ways to make $10,000 in three weeks with AI agents. And then they just build some toy app that breaks. And it's like, no, everybody can do that. Like this is not going anywhere for the long term inside of a company, you know? Yeah, exactly. But my favorite ads are the ones that can bind the blockchain to it, where it's like double hype.
And then you're forgetting maybe some AR VR, like animal together. And then, that's right. Now you're about to get funded. You better stop talking. We get to the crypto, the crypto bots, the actual robots. We're in trouble. All right, everybody. This is the end of the show. Thank you for listening. As always, we'd like to shout out a user on Stack Overflow who came on adoptable knowledge and earned a badge.
Today, we'd like to shout out Joshua Fox for earning a lifeboat badge. They came in and they got an answer score of 20 or more to a question that had a negative three. They saved this question with their lifeboat. And the question is, get the current hour and minutes from instant.now. If you're curious about how to get the current hour, go check out this question. It's already helped almost 20,000 people.
Very cool. As always, I am Ben Popper. I am a host of the podcast here at Stack Overflow. This will be the first episode. We are changing things up a little. I am over at a new startup builder.io. You can check me out on Twitter. You can learn about that. But I'll still be doing lots of things with the Stack Overflow team and content, which I'm very excited about this year. I'm Ryan Donovan. I am the still editor of the blog here at Stack Overflow co-host of the podcast.
You can find the blog at sacoverflow.blog. And if you were to reach out to me with episode ideas, criticisms, comments, further questions, ask, you can find me on LinkedIn.
My name is RJ to it or my mom called me over John. There are still some people who like no actual original name. I work at ClickUp and the head of front end and the client architect. If you want to find me, you can probably find me on LinkedIn. There aren't a lot of RJ to it. I think there's literally one. So it should be easy. Come check out ClickUp. It's an amazing product that is
And not as well known among software engineers as it should be, but it will be in the next year. I think we're going to have a breakout year this year. Try it out. It's there to make you more productive. Clickup.com. Well, thank you for joining us, RJ, and thank you for listening. Talk to you next time.
Was this transcript helpful?
Recent Episodes
Why build your own vector DB? To process 25,000 images per second

The Stack Overflow Podcast
Verkada is a cloud-based video security company. Back in the innocent days of 2021, we spoke with a company that makes smart dashcams. See how far video and image processing has come. Congrats to Reg for earning a Lifeboat badge for their answer on What is the difference between JSP and Spring?
February 07, 2025
Will the web ever be the primary delivery system for 3D games?

The Stack Overflow Podcast
Tres.js is an open-source 3D engine for Vue built on Three.js. Find Jaime on LinkedIn or GitHub or explore his creative lab.Push is a browser-based identity security platform that detects and blocks identity attacks, enforces security controls, and monitors employee logins to cloud accounts.Shoutout to Stack Overflow user zwol, who earned a Lifeboat badge with an excellent answer to How would you write the equivalent of this C++ loop in Rust.
February 04, 2025
Feature flags: Theory meets reality

The Stack Overflow Podcast
Discusses LaunchDarkly co-founders Fynn Kaiser and Ben Kehoe about extending feature flags for entitlement management and monetization on Localstack.
January 31, 2025
How the internet changed in 2024

The Stack Overflow Podcast
Celebrates user for providing insights on Cloudflare's platform; suggests visiting John's posts on Cloudflare and connecting with him.
January 24, 2025

Ask this episodeAI Anything

Hi! You're chatting with The Stack Overflow Podcast AI.
I can answer your questions from this episode and play episode clips relevant to your question.
You can ask a direct question or get started with below questions -
What was the main topic of the podcast episode?
Summarise the key points discussed in the episode?
Were there any notable quotes or insights from the speakers?
Which popular books were mentioned in this episode?
Were there any points particularly controversial or thought-provoking discussed in the episode?
Were any current events or trending topics addressed in the episode?
Sign In to save message history