# KUNAL_V1 transcript

Duration: 01:15:55

Automated transcript. Section timestamps mark 10-minute audio sections. Uncertain wording is marked [unclear].

## 00:00:00

Hello and welcome everyone to another episode of The Merge, our podcast brought to you here live from the CodeRabbit headquarters in San Francisco. Today I'm super excited because I'm here with Kunal. Hi [unclear]. So excited? Yeah, good to have you here. Kunal, you've had an amazing career. You've been making big waves in terms of the developer community. You've been a CNCF ambassador, a GitHub star, amongst a bunch of other things. So I'm very grateful that you took time today to continue our conversation, actually, from scale all the way now here to San Francisco. Always a pleasure. And this is such an incredible office, by the way. I've not been to the CodeRabbit office before. Yeah. Yeah, absolutely incredible. Thanks for having me. Well, I am honored to be here with you, and I think there's a lot of interesting things we can talk about, so.

Yeah. So excited. The weather's been a little interesting this week. It is San Francisco August. We call it Fogust, actually. So what you see outside here is actually Cal. Cal is the fog or the name of the fog, I should say. And every time in August, you know, all the fog comes in because there's so much heat being built up on the central continental mass of California creates an uplift and then it pulls in the evaporated air from the ocean and it just collects over San Francisco. Kind of reminds me of home back in London. Yeah, but no, I like California. Yeah, it's a good place to be, right? Yeah, for us coders, it's not too bad because we have one more excuse not to leave the house.

Yeah, I mean, I'm in San Francisco every month, so it's like second home. I guess third home in this case, right? Third home, no. I go, I mean, I'm in San Francisco more than I'm in India. Oh, really? Um, yeah. So this would be Second Home. Yeah, that's actually an interesting point. I think we can dive right in there, actually, because the last time we talked, which I mentioned earlier, was at ScaleCon Pasadena, and we had a very, very quick chat. We had our set up there, so you kind of swooped in and we touched on a couple of interesting points. But one thing we didn't get to talk about it was a bit of your upbringing, I guess you could call it. So I digged a little bit deeper into after you told me that you were gonna go in town.

Yeah, so I have some spicy takes, but I just I just thought like a beauty opportunity to tell the story yourself before I do. Can you tell us a little bit about how you ended up here? Well, where do I start? Yeah. So I wanted to be a doctor when I was a kid. When I was five years old, I was like in an accident and. Everything worked out fine. Since then I was like, okay, Doctors are cool, you know. I wanna be a doctor. By the time I was like 15, you have to pick a career path in high school, except you either pick like science, be that tech, or in India. be that tech or medical, or if you wanna do accounts, so commerce, or if you wanna learn like history or whatever, there is arts for that.

And I realized I am terrible at memorizing things, can't memorize anything. I had my chemistry exam because in chemistry there's a lot of rules, and for every rule there's like so many exceptions. Like, okay, this rule works fine, but it's not going to apply for this compound because it's an exception. I could never memorize things like that, you know, like bookish jargon knowledge. because I can just look it up. Why do I need to memorize, you know? And I was like, okay, I can't be a doctor because, you know, it's just so much memorizing and reading and whatnot. I was always into tech, wasn't like a hardcore programmer. Started when I was in high school with C++, but nothing major. In India, when you're in high school, you're preparing for your university entrance exams.

It's a huge country, 1.4 billion people. So we have millions and millions of students who graduate every year. It's a lot different than a much smaller country, let's say somewhere in Europe or even the US compared to India, because these millions of engineers graduate every year. And now obviously they're all looking for jobs, so it's very competitive. So from the very young age, we're put into That competitive mindset. There are various ways to get a role, I think in tech at least. throughout the, you know, last few decades, There are some really prestigious colleges in India, but because there are just so many students fighting for seats, getting into those universities is not easy. And it's not always about your interest, it's about what the university wants.

So that's maths, physics and chemistry. It's like I think the second toughest exam in the world. I think China has the toughest one. I can be wrong, I don't know, but it's like definitely up there. So yeah, that didn't work out because I didn't, I wasn't good at chemistry or as good as physics at physics, um, and then um. Then yeah, what happened, then I took a year gap. And that's sort of like looked down upon in India. I know year gaps in Europe are great. I'm gonna take an year gap. I wanna travel. Yeah, so it's looked down upon in India just like, okay, why are you taking an year gap because, you know, the race moves so fast. Took an year gap, tried for the exam again to get into the top IIT, Indian Institute of Technology.

Didn't work out because... I mean, a million students take that exam every year. I think they have around 5000 seats. Look at the odds. No. Even if you're great, you won't get computer science. But there's one option for me, like I could get a really nice college, but then I wouldn't be studying computer science. I would be studying something else like metallurgy or, well, God forbid, chemistry. Yeah. I don't hate chemistry. If there's any chemistry lovers, I love chemistry, but... It's just I can't memorise the compounds. But anyway, then I was like, okay, let's just get whatever I'm getting. So I did that and I was like, OK. In life, I'm always like, whatever happened, whatever has happened has happened. I mean, sure, that is just life.

If sometimes things don't work out according to your plan. You can't time travel. Didn't get a college, didn't get a college, but what would I do, you know? So I can't time travel back. Whatever you have in your resources right now, Just use that to make the best out of it and that's what I do. So I was like, okay, I'm in a decent college now and there are examples that I've seen. The CEO of Microsoft didn't go to IIT. Satya Nadella, he went to MIT, it's a private college in India. I got into MIT. So I'm like, not MIT US, so Manipal Institute of Technology in India. I'm like, okay, these people are doing great, you know, things in life. The IIT tag is not the only thing.

So I was like if they can do it, I can do it. Yeah. And then I was like, Okay, I'm just gonna start learning and building and just. Do a bunch of cool stuff. That's when I found open source, changed my life. Um. And yeah, I mean, then that's, that's how it started. Yeah. How do you say open source changed your life? I mean, it's there so much competition in the market, you know. To think from a student's point of view, they're looking for internships first, then they look for full-time roles, standard thing. But when you go for an internship, you have to send a resume in. And in the resume they mentioned What have you done in the past? And the student's like, I haven't done anything because I didn't have an internship.

But you need to put something in your resume to get the internship. Just like Deadlock, open source is great because... You don't need to interview at a company? You can just pick a nice cool project, contribute to it. Put that on your resume? Career takes off. So if you're a contributor, if you're a hardcore contributor in Kubernetes, like you wouldn't have issues at least getting an interview from... some top tech companies, you know, like I got from Red Hat, some other companies. And it just opens a lot of doors for you. And you get to meet a lot of cool people, and I think that was the main main thing. Yeah, I contributed to Kubernetes when I was in sophomore year. And then yes, then my whole career was in cloud native.

Still am, now obviously AI. And yeah, which is a bit interesting because... The complaint I see from people is like, open source is so complex, like these projects are so big to contribute to. But it doesn't mean that it's not. Open to beginners. I mean, Kubernetes is, I think, second largest project in the world after Linux. It also is one of the most Welcoming That's the word I'm looking for. So they are always looking for contributors. There are a number of things you can do, code contributions, non-code contributions. The key is just getting started. The one last thing is like, if you don't put in efforts yourself, I mean, the community people are there to help you, but if someone sees this person is not putting efforts themselves, no one's going to help you.

Well, how do you feel like that now changed? Because I mean, it seems to me, I have my open claw running. I can just, you know, pick up the repo and tell it to make a contribution, like how... If this happened now, if my story happened now, I was in a uni that wasn't as good. And now I'm like, okay, I wanna look for a role. Okay, that's a good question. What would I do? Um... I feel like the way of contributing changed, right? Yeah, the way of contributing has changed, but I think the whole... The narrative is the same in terms of building because Programming languages and how we interact with the systems is just, it's a mechanism, you know. 20-30 years down the line, like sorry, not down the line, 20-30 years before people were like coding in very low level languages.

Let's say even before that there were punch cards and whatnot. So the problem solving mindset was still there. So to answer your question, I think gone are the days when companies are going to be like, I'm going to hire this person because they write better Java than this other person. I don't think that's gonna be the case anymore. It's more about how much of a generalist are you, how can you...

## 00:10:00

I mean... how well you communicate with others, for example, you know, other skills like empathy, problem solving. I love the AI generation because I don't have to worry about syntax and programming languages and things like that. I can focus on engineering. I can focus on building and, you know, problem solving and stuff like that. So AI takes a lot of that off my plate. So yeah, yeah, I can write code, as you know, you can review code, do a bunch of stuff, write docs, testing and whatnot. So it's just giving people time. You know, to build stuff. In the very short period of time, I mean, my mother can build a website now. You know, it's very simple, easy for her. There are so many, so many tools out there.

So it's just giving people more control. to, you know, Bring their ideas to reality. And this is a big thing for small businesses. I read it somewhere, like when something becomes easier to do, more people do it. And, um... Yeah, it's just, how do you give access to... You know the right tools to people at scale. Right, and that is a beautiful arc, I think, to the next chapter of your life, which then focused on developer education, right? Yeah. So can you tell me a bit more about how that, how you got started, how, what got you into that, and how do you scale it today? Ja, um... Sophomore year I was in Kubernetes, I was contributing to Kubernetes, and then... I was I got it in this program called Google Summer of Code.

With the Red Hat, it's like an open source program, you contribute to. projects, and then Google pays you. And then I was a mentor in that same program. And then I realized when I was learning how to code that I love going into the details of how things are working and whatnot, and I love teaching. So I started volunteering at this local boot camp in Delhi and I realized, like, I love teaching. People, people told me, Gunal, you're so good at teaching. I know, I watched one of your videos. You know what I really, it was, Crazy, I think one of your very early videos four years ago, you have this scratchpad, digital scratchpad on like a Windows or something and you just draw like little figures on it, but it was so well explained.

And here we, we're all sitting there. I could use a lot of videos as well with crazy animations and so on, and I felt like, no, it can be so simple. Yeah, it's a People don't watch what you do, they watch how you do it. Even though there's a lot of tutorials on Kubernetes, if someone can explain Kubernetes in a in their own way, I think they'll get a lot of views and they'll get their own audience. So this is what the question people have like, hey, why would I make a video on React or something like Codex or whatever, when there's already creators making videos on it. But the topic doesn't matter. What matters is how you do it. Yeah, but I realized I'm very good at, like people say I'm very good at teaching.

And then COVID came. COVID came, I was in San Diego 2019 for KubeCon and then 2020-ish, early, COVID came. So all these bootcamps I was teaching at, they had their own online portal. So all the kids now were learning online. Um... And now they're like, Kunal, you are not on the online portal. I'm like, yeah, I didn't record videos for them. I just teach online in the building, in person. They're like, no, no, we're gonna learn from you. And I'm like, oh, okay. And I started my YouTube channel. And I'm like, okay, if you wanna learn from me, it's COVID, I'll just teach on YouTube. I remember I uploaded five videos on Python. Yeah. And each video was like four hours long. It was like full Python course, in the depth.

With some like 10 projects or whatnot. Like if you don't know anything about programming, you could be like... Interview ready, that kind of a... I think because of that I got zero to 100k subscribers organically in like two and a half months. And I realized, okay, this is, this has been good. People have been very kind. Yeah, you found your niche. This is what should be doing. Yeah, people have been very kind to support and then I got into DevOps and then just started pushing more content, teaching more people, making education more accessible. Um... And then I was like, OK. Now I have a few and it just kept growing. In one year, in like a year, I had like half a million or more than that.

And then I was like, okay, all these people are now asking me questions. Yeah. So, what do I do? Because... If somebody is asking me a debugging uh wants me to debug, help them debug something or whatever, the YouTube comments are not the right place for that. Like I'm like, okay, I'll start a Discord server. Started a Discord server. I don't remember what it's called, it was called something. And uh... That grew to like seventy, eighty thousand people. And then that converted to vMix. So like we have this whole community. I started giving them challenges. I was like, okay, let's do some hackathon or something, you know, because we have this big community and... Learning online is great, but What I would recommend is to everyone, even in today's age, because you asked me earlier, like what I would do in the AI age if I was a student now.

I don't think much would change than what I did earlier, honestly, Hendrik. It's just the medium would change. I would be using a lot more AI tools. But apart from that, problem-solving, contributing to open source, going to hackathons, creating content, writing blog posts. speaking at conferences as a student, just get started. Yeah, that resonated with me was like, I think, learn in public. Is that something? Yeah, build in public, learn in public because you get, learning in public doesn't mean like, this is a mistake people make. They're watching my videos, they're like, today I learned support vector machines from Kunal Kashaha. No one who is watching that post cares. That's what the post would be. Today I learned about two pointer from Kunal Ghosh. That's not learning in public.

That's just you sharing about your progress, which is great. Yeah. But. Think about it like someone who is watching that post. someone who's a really good engineer, senior person, it's not giving them enough value, right? It's just helping my promotions, which I appreciate. You know, I appreciate that. Please keep it coming. But no, jokes aside. When you talk about learning in public, it's more about... sharing your knowledge out there. So I learned this computer course from Kunal or whatever. If you find something interesting, share about that. Write that in a blog post, speak about it at a conference. credit the creator if you saw the video from them if you want to, no pressure. But that's what learning in public is all about. And one other thing is because you get feedback from people in the industry.

And quite often, I mean, Twitter can be X or LinkedIn can be a little... What's the word I'm looking for? Cringe. Not cringe, but like interesting place. Because sometimes if you ask for feedback, people will have no filter. They lack empathy, is what I'm saying. Oh yeah, especially the YouTube comments. I don't know what's wrong with people on YouTube, but you need to stop. Like, guys, this is too much. No, the other good thing is I get so many comments, I just don't see a lot. Criticism is fine, always welcomed. Just make sure I'm just saying like you will find people. Yeah, you will find Folks saying like, hey, this is not good, but if you're a student, if you're just getting started, I mean, learning club is not just for students, like anyone, even if you're launching your own startup, getting feedback online.

I mean, this is actually, learning public is what people do actually with this go to market sort of like activities because they do launch on product hunt. And they get, sorry, building in public, which is sort of like you're learning from your customers or whatever. But you're building in public, you're doing soft launches on product hunt, get a lot of feedback. You're doing hackathons and stuff. You're doing go-to-market events. You're doing uh some webinars. You're doing uh conference sponsorships. So you're learning from the community in the public, right? So it's just a different scale at what beginners do it at. But yeah, make sure the posts that you're making are valuable. Don't spam people, don't use too many hashtags, please stop the AI slop. It's just not gonna work.

It's not working. Use AI to write content. I'm totally for that, okay? Use that to refine your content. I know English is not the first language for a lot of people. So you can use AI to sort of like help you with that. But some people just completely copy-paste. We can tell. I mean, I think then also the bot traffic and the bot commenters, especially LinkedIn, their banning algorithm is not so good yet. Yeah, it's not that good, but it's actually doing you more harm than good. So you say I am very pro-AI. but don't list copy-paste comments, but yeah, for learning in public, getting feedback about your projects. One best thing about learning in public is you make so many friends. I think that's the best thing.

You make a lot of friends and you build your profile. And the world is so small. Like ei. Every month in San Francisco I run into someone that I knew from someone else. I mean, San Francisco is a surprisingly small city. Well, yeah, I think from that arc, moving on, you talked about, you know, how you built this passion for learning and then how that translated into your community that you're building now. Can you tell me a bit more about that community? How big is it? What are you guys doing exactly? Yeah, so when I was at conferences and stuff... When I was doing booth duty. It was great. It was just a lot of salespeople in the room, which is fine. Salespeople make it happen.

They are the ones who bring money in. But like if I wanna get developer feedback, If I wanna get like... What's wrong with the tool? What can I change? How easy was it to contribute? How easy was it to onboard?

## 00:20:00

and things like that. I think that's where you need to obviously get your tool in the hands of builders. So I think these sort of like go to market or engagement events that companies do, let's say do like a workshop thingy or a webinar. Always talk. Always think about how engaged do you think people are. Some people they just do a webinar live. And that doesn't really... Convert to anything, you know, the argument they give is we got hundred email addresses. Yeah, how many of them will reply to you? You know, try getting a reply from everyone how many of them will get on a meeting with you. So, the coefficient factor for like how much engaged a person is. I think that's something that you need to think about for your marketing efforts for developers, not like you in general, like people in general.

Um... Hackathons are great because We we did one. I think it was 2021. We got like um 2,000 something people and I hadn't even done any choice. Was it remote or like? It was remote from my YouTube. I think we have like 2 to 3,000ish or something, and it was with Elastic, I remember. um Actually, it was via Arvind. Oh really? Arvind, for context, is the previous VP of growth at Grammarly. Great person. Great person, and it was my first, it was a very good friend. He was like supportive and it was very small, like event we did, the budget was a thousand bucks. We got so many people share feedback and stuff. And then we just started doing more and more events, I think last year.

Because... WeMakeDevs community, we renamed it like last year, but it was so big that we were trying to figure out, okay, what do I want to do, okay? I want to do workshops, I want to do in-person meetups and stuff. We did one or two more hackathons and I realized, okay, This is like the thing that people want. The community wants this because when we do hackathons, that's the most engagement we get. So, around 10,000 people. Yeah, they try out the product, you can be sure that everybody who attends will actually have physical touch with the tools of the sponsors. Yeah, so for the attendees, and then I was like, okay, let's just do hackathon. That's what's working, let's just do that. There are two strengths to it.

One is why would the audience care? Why do the companies care? So for the audiences, I think they get to, because there's the trend is shifting so fast, There's a new tool come out, the new inferencing tool will come out, a new monitoring tool will come out. The way we look at it differently at BMIC Devs is we are riding with that trend. Something cool comes out, we give people the platform to build with it and share their feedback. That's what I care about the most, because we have this incredible community behind us... The feedback that they give us is worth so much. to, well, essentially all the companies. In the world, the tech companies. Because you know, when a VC firm wants to invest in a company, they will go to like the research firms and some independent people.

They pay them like $200-300K for doing that background check about the industry analysis and whatnot. Big, big companies like, yeah, like Gartner and stuff, they publish their reports and they work with enterprises and share those findings and stuff with them. What we have found working with companies is Imagine this, you are a company. We're doing a hackathon together. Get ten thousand people test out your platform. signups, GitHub stars and whatnot, I genuinely don't believe in vanity metrics. I don't think it matters as much. May look good on the paper. But the kind of numbers if you want to expect from a hackathon or like any conference you're sponsoring, it doesn't matter if it's not, I mean in the end it doesn't matter if it's not Directly relating to some sort of business outcomes.

You know, because in the end that's what people care about. That's what's gonna pay the salaries and whatnot. Well, I mean these events often they're there for generating pipeline for companies, right? Exactly. Yeah, but it's sometimes, you know, most of the time, it's like indirect as well. So let's say if you're doing a hackathon with students. we do like Professionals, senior engineers and students and everyone included, but that ROI you will get when the students graduate. And now they're working at a company, and they would know, like, okay... If I wanna do... Take an example, I wanted to do code reviews. I did code reviews when I was in my uni. And I used CodeRabbit. So they, you know, kids, they become like a community for...

der Career. It transitions into the career because when they are new in the career, they have to answer all these questions. I've given this task and what do I do? Oh, I used this tool like a few years ago. Why do you think all these big companies like You know, all the hyperscalers, they all have their student programs. Oh yeah, I know. I, for me, I studied mechatronics in university and I remember still working with Autodesk for my CID projects. Yeah. So the translator just carries on and the time moves by so fast. These people will graduate, they'll become seniors, and now you have your own sort of like community members. So even though if you like, let's say you are at an event, you're just scanning people, let's say you're at the booth duty.

Someone comes in. You're like, hey, can I scan you? I'll give you a T-shirt. And the person is like, yeah, sure, scan them, give them a t-shirt, and then you go to your VP, like, hey, we got a thousand lead scans. And then you reach out to the thousand lead scans and there's just no reply, nothing. It just looks good on paper that, hey, we got thousand leads from the event. Looks good for the CMO or whatever, but... Is that translating into anything? So how do you promote ROI from your events? So for us, it's about the depth rather than just the numbers. There's this thing called value hours. The idea is that you did a hackathon or you did a webinar, a hundred people came, one hour was the average view time, and there's a coefficient to it, which is how engaged do you think the people were?

Webinar people are not as engaged as in-person workshop. So let's put a coefficient of 0.5. hundred into one hour into point five fifty hours of value hours. So that's like measuring the kind of depth, how engaged the person is. In your event. For our hackathon specifically, it helps because it's hands-on. The most important thing when we work with like enterprises, open source projects, dev tools, we've got a lot of professionals and senior engineers. So we need to give incentive for them to come. So one of the incentives is you're riding the AI wave. You're basically any new cool tool comes out. We do a hackathon around it. Any new industries or like term comes out, we do a hackathon around it. We get experts on the stream and we share resources.

That's the second part. Making it easy for people to contribute. making it easy for people to build with these tools. No one's judging you. I think that's why people come to hackathons. They wanna learn, they wanna see what's out there. Surprisingly, I thought it was gonna be just students, but we get like... So many engineers and professionals. The one we're doing in SF, the one we did at GitHub. Thirty-five percent executives came. thirty or something other person, senior engineers. And a lot of developers and Yeah, over here, you even see a lot of VC funds. They sent some of their scouts to these Tabi events too. Very interesting trajectory. I wanna pull us a little bit further because I think one of the most interesting things that you experience every day through running all of these events, you're kind of so close to the developers of today and the developers of tomorrow, really.

So can you give me an impression, like maybe you can start sharing of one of the most impressive projects that you've seen in the latest weeks or two weeks or month? And then I'll be interested also to learn a little bit more how you see, like the development workflow shifting on these events as well. Yeah, so what to what I was mentioning earlier, like measuring ROI, it's around... For a dev tool, if I take as an example, care about we want feedback. We want quality people like uh from all backgrounds, but then I mean quality people is people who are actively engaging with the platform. It's can be students, can be engineers. We want feedback from them. We want to see use cases of our tool from the community.

And if they end up paying in the future, become like a sales pipeline or whatever, that's a different story. The last hackathon we did, to answer your question. So interesting story. We did it on Saturday at the Bright Data Loft with... Port and we had BrightData obviously, we had people from Anthropic and stuff. It was a really nice big hackathon. We had a lot of people registered, I think over 800. Then you can already hold a... So you had to cut people off at the door? No, no, no. We had a wait list beforehand. So we had a... Yeah, just beforehand and this one kid came. He was I think 12 years old. Oh wow. And he was on the waitlist, so he didn't get in, okay?

And he's at the door and he's like his dad dropped him off and he's like, hey, I registered for the event and I'm looking on Luma. This guy is like 12 years old, on the Luma, he's not accepted. You know, and then I couldn't say no to him. I'm like, dude, just, yeah, just come, I'm coming, I'm just, come on in. Yeah. So I came, he came in and then, He was sitting at his desk with some of his friends as well, like I just sit there and build. In my mind, I was like, okay, he's a kid, I mean, 12 years old. But what was he gonna do? When I was 12, what was I gonna do? You know, his project was top three projects that I have received.

In the last ten hackathons we did, his project was insane. This project was insane. He blew me out, he gave me imposter syndrome.

## 00:30:00

And I don't get impulses. All right, tell me all about it. I want to hear what's it about. It was incredible. It was not mainly just about project. It's how, first of all, his demo video was so cool and cute. It was like a role play, and he was role-playing, and he was acting, and it was cool. It built like a go-to-market sort of like app where you put an idea in his app. It's gonna build the whole website and stuff. It's gonna do research. It's gonna pinpoint people in like San Francisco and stuff that can, you know, that you can engage with. The UI was incredible, how he used the whole pipeline for both Bright Data for scraping the data from the internet and the UI was so good.

And the documentation, you know, small things like that, because you have your project on GitHub, right? Small things like code quality and just how well you use the PRs, the documentation, readme. And this was done in eight hours. Wow. And I'll leave, you'll see for yourself. I mean just the guts to go there and heck with like people that are way, way older than you building entire products. Yeah, so AI won't take your job, these kids will. Well, yeah, interesting that you say that because I think that's my next point that I want to talk to you about it. There's a lot of anxiety in the developer ecosystem right now about AI is potentially taking their job, like what? What does it mean for me as a developer, I spend so much time learning...

C++, I've spent so much time understanding pointers and, you know, the complex relationship between hardware and software in these lower-level languages, and now all that time seems like... It doesn't account to that much anymore for a lot of people and that triggers anxiety. How do you feel about that? If there is a very good point, I would say The number one benefit of going into these details for the basics, for example, Is you develop a superpower. You know what that superpower is? So, the superpower is you learn how to escape the hype cycle. Because you know the fundamentals, you cannot be persuaded with the hype. And that's very important So I always recommend focus on fundamentals. You don't have to be an expert in every single thing.

Obviously, there's Pi-shaped learning, know something about. know everything about something, something about everything, so pick a niche. But I think with AI now is... I'm not gonna say that it's not like a time to worry, but it depends where you are coming from. There are a lot of people in India who are like... working at very lower wages. Because I mean, there's a big divide in India amongst engineers. Some of the people are running companies. some people are you know CEOs obviously at the big tech companies, but also there are a lot of outsourcing agencies. that pay very little amount compared to their home country to developers in India. They get a lot of stuff done. They don't have a choice, because they know that if I quit this job, it's paying me three thousand dollars a year.

I'm not gonna find another one. Right? The question you need to ask yourself is why is it that someone from India came from a village? And they went to US. They went to London. They, they obviously did something. You know. I didn't come from a rich background. We had debt and everything and Grew up in a small apartment. No, no, in no way whatsoever I came from a rich background. Didn't even go to uni in any university outside India. No master is nothing. Didn't have, you know, it's expensive studying in the US. So... With AI, I think all those people are the ones who are going to be impacted quite a bit. So... It is time to worry, but it depends where you are coming from.

Coming back to my earlier point, like you are not going to be hired if you know one programming language better than the other. I think it's more about... Being a generalist, in all honesty. So... If I'm giving an if I'm giving like a Task to a student right now. Student is like, hey... I have an interview and whatnot, and how do I prepare or what task can you give me, or like just in general when I'm approaching founders and recruiters and stuff, give them a website and ask them, break down this website for me. What would you change in this website? I don't wanna talk about code. Don't talk about code with me. Talk to me about this website. What would you break down? How would that change the market?

How would that change the industry? And what changes would you make if you had full creative control and things like that? Yeah. So I think that's one. And obviously the second, the obvious answer is that how well people work with these AI tools. So if you're looking to get a job in law. What kind of AI tools are being used in the law industry? If you are a Kubernetes, how do you use AI to make Kubernetes smarter? Right, if you're in testing, how do you do that? If you're a... So a data science person, so there is a lot of stuff that junior data scientists used to do that now AI can do for you. So, what apart from AI? Do you bring to the table is the question you need to ask yourself.

Here's the other thing, you know why this is very important, the issue with people and students right now, you see them, hey, a new model came out, and my tools and my pipeline and everything is 30% more faster, 40% more accurate. The model that just came out that I just used in cloud is now 40 times more smarter, 40% more smarter. this new inferencing layer that came out, or whatever, this inferencing tool that came out, is 20% more smarter or faster. This rooting mechanism that came out that does everything for me, I just write one prompt, is now 10% more smarter. What about you? Are you smarter than what you You know, then what then what you were yesterday? What are you doing? to improve your own personal self.

So using these AI tools as a helping hand is I think the key. Don't worry about Don't worry about having n number of things to learn. The industry will always change. There will always going to be something that you don't know. And that's fine. So it makes us human. You know, do you think I know everything about Kubernetes? I don't. And I was field CTO at Cevo building data centers. I don't know everything about Kubernetes. I don't think many people do. It's such a big project. I don't know everything about Linux. I don't know everything about networking. But I know A lot from my hands-on experiences. Right? So if you're looking for, if you're scared about losing your job to AI, If AI can replace Any programming task that you're doing right now, 100 percent you will get replaced.

No question about it. If you're just good at C++ or Java... And AI can do that for you. You don't bring any sort of other problem-solving skills to the table, any sort of like, you know, teamwork skills. Empathy is obviously a crucial one. How do you interact well with developers, your users. UI, UX and whatnot, just the whole sort of action nannigan. ya if If AI can replace you, it will. And If you use AI as a helping hand, to sort of like, empower stuff that you're already good at. I think it's great. So taking back to the example of that 12-year-old kid, he used Clojure or Codex or something pretty heavily. but the thought process behind it, the infrastructure behind it, the architecture behind it.

A lot of questions, a lot of conversations recently around tokenomics. People spending, because you know, when Kubernetes came around 2014 and then the cloud world, people are money, sending money on the cloud bills and whatnot. And now even today, there are so many companies active. There's a whole FinOps Foundation. The whole idea of a FinOps foundation is reducing your cloud bills. There are so many companies... You worked in that space before, right? Yeah, I was at Cast.ai. I was leading developer relations and business doing some amazing. So many companies doing similar stuff. So now we're seeing that with AI as well. They have big budgets. Did you see that news? 500 billion. Five hundred million or five hundred billion on cloud bill. Half a billion, something like that.

That would be like half of their... No, no, no, $500 million. $500 million, obviously. I think like because they only have like somewhat billion of revenue right now, which is crazy already, but... Yeah, it can't be half a trillion, obviously. Half a half a billion dollars on cloud bill and stuff. And even Sam Altman, all these people are now like, People need to talk about, you know, reducing their token costs. a lot of uh agent hardnesses having the narrative around that. So AI won't replace you, personally, a person using AI will. I don't know who said it, but it's a very good term. But also, something becomes easier to do, a lot more people do it. You can spend more time in problem solving. There are so many problems in the world.

Solve those problems, use AI as a helping hand. I want to dig a little bit deeper into that because I think the topic is really, really important and you made some great points, but I want to be a bit more practical. What can I do if I've been, you know, you are an experienced engineer, you're not a junior. For a junior, it seems somewhat almost easier because you are not constrained to the same ways of thinking that that 12-year-old has no restrain in his head of like, oh, this is the way I need to build this fast API server first and then I can think about how I scale all of this. No, he's just like, I wanna, he almost has a natural product sense, which I think is what is needed also for what you said in terms of problem solving and building something people want.

## 00:40:00

But if I'm in my 40s and I build my career on being very knowledgeable in Kubernetes, for example, what do I do to remain relevant? Do you have a practical advice? First of all, I'm 27, so... If I was in my forties, again, it's about... How you communicate with the tools. So, code is just communication. You have to write code, Java, C, C++, and it just interacts with the systems. Now that's transitioning into natural languages. Everything else behind that, the architecture, how do you want to system your, you know, architecture system, how do you want to maintain it, and things like that. That's all still there. There's always going to be a need for a human to do that. So My answer is just... in two parts.

First one being, don't let go of the basics. Yeah, that's first. Critical thinking, you get to escape the hype cycle, you get to know the basic foundations and fundamentals. And the second thing is, like use AI to automate the mundane tasks that you don't like. Um, and Double down on problem solving is what I would recommend. So if you're 40 years old, good news for you because You're so into your career. Now you have more time. to do stuff that was on your backlog. The last thing I would mention is, I don't know at which point we are now where you can just, I think we are there, that you can just approve the code that AI wrote. Oh, OK, that's an interesting take, yeah.

I mean it depends for, I mean with code review tools, I think they're quite smart enough to figure out there are any vulnerabilities and whatnot. But in the end, there's always going to be a human who's approving that PR being merged or whatever. Right, so when you are using these AI tools to write code, Just make sure it's checked. You know, and apply your intuition. Apply your own intuition. Because you're the one who still owns the responsibility, right, for whatever is shipped, not the agent. You cannot blame the agent, but you can also not fire the agent, but you can fire yourself. That's the other thing with AI, like agent AI. How do you give access to AI? How do you give... Identity access to AI agents, you know, if agents mess something up, there needs an identity attached to it.

You can call someone. Yeah, I mean, that's a whole different discussion. I did have the pleasure of recording an episode with WorkOS, the founder of WorkOS, they do authentication for agents. I recommend people checking that out. I'll be happy to send it your way as well. He's an amazing founder. The company is used by companies like Anthropic, OpenAI, and so on. They use his product. But I wanna briefly stick around because I think the topic is very interesting. Let's talk about what you just said regarding code review and code review agents being good enough to review the code so you don't need to actually, like you can just approve. I've been hearing a lot of talk about that not being the case. Recently there was the AI engineering code summit here in San Francisco, welfare, I think it was called, yeah.

I was there. You were there? Yeah. Remember the talk from, shout out to Dex Herfee. he made an amazing talk about why agent software factories fail. And the reason he brought was very interesting because he said that basically... the reward system, how we do reinforcement learning today, it does not take into account the long feedback loop that you have when Agents write unmaintainable code. Because the fact that you have to do a refactor at some point because things like performance slows down, or maybe you had some security vulnerabilities that your tests didn't catch and it deployed fine. All of that is not built into the loop because those issues arise three, four, five, six, seven months later and that hence they are not being downgraded in the reinforcement loop.

And that's why today you still have to look at the code and you cannot just um deploy, you know, lights off, software factories, because as you, and I can talk to that too, uh, we have the saying here at CodeReby, the entropy in your code base gets bigger and bigger as you deploy more and more background coding agents into your code base. I have a project that I've been working on with coding agents for about a year now and I can speak to that. Like at the start when I first build it out, like I felt like I was really using agenting engineering and uh like putting the build it out very fast, build out the backend very well, build out the ETL pipeline that it needed to process some things very well.

And I really understood it. Now my agents are in there all the time while I sleep or while I go to my daytime job. And I just feel like the entropy has increased so much. So I don't, I'm not, I'm sometimes worried when I click on the, even though all the flags are green, I don't know if do you feel the same for some. Take an example, like cloud is running, let's say cloud is running your Slack and connected to your email. It's essentially the same thing. You could argue that, hey, what if cloud send something messed up? But if, you know, and like, yeah, so that's the thing. It does something very, but people are still like, there's one entity. There's like, I'm not gonna give access to my services to AI.

There's a local AI is getting so popular, you know, because of that. And then the others are like, okay, I don't care. With code reviews in itself, I think it depends on the What is at stake? So for example, I think we will go into a era, into an era where humans are not reviewing code at all. I don't know when that will be. Ten, twenty, thirty years down the line, we will definitely reach there. That will happen because of how fast that things are moving. There's a lot of, you know, you mentioned the context increasing and increasing and increasing, the whole narrative around like company brains, obviously a lot of talk now around agent memory so that you don't have to give the whole context again and again in every conversation.

But... Human, I could be wrong, but human beings not reviewing code and just agent handling everything. I don't know. You're saying that won't happen in the next 50 years. Oh no, I'm pretty sure. Next 100 years. No, no, not even. I'd say like, I don't think we're gonna ship like that. Actually in... I work at a company where we build our foundation on the fact that that will not be the case anymore, right? So we're building a lot of tools. We just recently launched a tool called ChangeStack, which is a completely new way of how to review a PR. So we don't, we want to strip away from you taking the looking at the code line by line and we are abstracting multiple layers. And you have to, you have to because these tools are so awesome.

Why would you not use it if a... If an AI tool has that much potential and power to scan the vulnerabilities and check what happened in the industry and make sure that you are not making the same mistakes and stuff, no one's asking you not to review it, but my point is around... How much of the, what is at stake? So I'll give you an example. Black Friday, okay, you are a billion dollar company, a twenty billion dollar company. You have a big sale on Black Friday. Okay, or something like that, and you get paid at 2 a.m. At page like 2 a.m. that something's messed up. You have to fix it, otherwise customers will get affected. Let's say guaranteed, let's just say guaranteed you will lose $500 million in sales, guaranteed, if you don't fix this.

Now you have two options. You can just ask AI, hey, please fix this, you know. Make no mistakes, make no mistakes, please fix it. You are awesome, very smart. You are now working as a senior debugging on-call engineer and whatnot. I don't think when a billion dollar sale is on the line, even though the AI can do everything perfectly, that human will double-check. There's so much on the line. you know, so it's more about the emotional aspect of it as well. Let's say 10 years down the line, you get into the same issue and... Oh yeah, I think there's there's there with CEO of Vercel. He made this post, Guillermo, uh there's like, if you're not looking at code anymore, one of these five things is true.

And he like listed out either you don't have customers, or you don't have revenue, or you're still experimenting or something. So even if 10 years down the line, we reach a point where... You know, you don't have to look into the AI code review, which we probably will. So that was my earlier point. Like right now, we are heading in a direction where you don't have to review the code. But the thing is that if something breaks, The human is gonna be in charge. But don't you think that we have to understand the change, still? You do, but it again depends on the context. If I'm changing a button's color, I'm not reviewing much. Okay. So we're building a lot of tools here around helping you understand the blast radius of your changes, like the abstraction layer of what you touch, maybe like you're doing a backend change and trying to surface that maybe your agent changed the front-end file, so bring that to your attention instead of you having to dig through it or find out in production later.

These type of things, that's the way I see personally the world going on that pace. And then next to that is also then, of course, the security part, which, yeah, a lot of... The number of security vulnerabilities that have increased in an age of AI is just so much, so many secrets crawls and whatnot. So Yeah, it depends on the blast radius, but obviously in the end, whoever is approving the PR is a human. So that person will be responsible when something bad. You know, the light LLM attack that happened as well. So that was kind of, that was kind of crazy. Yeah. So I'm very pro-AI, but am I comfortable right now just for... It to run my whole systems without me being or my engineers being like, okay, hey, we are the ones who are doing the final approval.

No, but it works. Hand in hand. The review that the code has given, read by human, checks everything is fine, merge the PR. And if you're not using like AI code reviews to review code, I mean, why not? Because they're super fast and stuff, you know. Why not use it?

## 00:50:00

But in the end, it's a human. Taking the check and to answer the point. Ten years down the line. This is a Black Friday sale. Agent can solve the bug, but the person will go on the laptop and see themselves, you know. There's so much on the line. So if you don't have paying customers, you don't care. Ja. No, that's all well put. Moving on the lines, I think another point that I had in my mental notes of things that I really wanted to get your talk on is, you're running so many events all over the world. Europe, London, Paris, Berlin, San Francisco all the time, India. Global events. So you're very close to a lot of incredible builders that you just were telling us about earlier as well.

How do you see them building and how have you seen the shift of them using the tools, like, you know, using cloud was very different today than it is from a year ago and the same goes with like some tools that we don't, we stopped using maybe. I remember, for example, Windsurf having huge popularity last summer and then now it kind of died down and I guess Devin has taken over. But then as well, like the approach of how we use tools, it used to be. Taboo to kind of open up massive PRs and people be like, oh, you should break that down and da da da da da and distribute it so people can understand. Now it's like you see on Twitter, you see CTOs posting 400,000 lines of PRs or whatever.

So tell me, how do you see the shift? ja, i mean in Europe. I don't think in the US many people care about like data sovereignty as such. The US has a big argument around that, even India with their data protection laws now. At our hackathons, we mainly see people have now enough time because they have just one day to build in the event space, you know. So... The idea is how do you really how well do you integrate the sponsor tools in your project. And I don't think many people are spending too much time on UI. As much because they can just wipe code UI and UX, every single UI we see is like incredible. Um... And something that would take like coding from scratch, like I know you didn't code it from scratch, I know you used some sort of like AI tools for that.

One of the cool thing that people do is like if they're working in a team, they would divide the teams, they would like you use, you work on the testing part, you work on the review part, I'll work on the infrastructure. But one other aspect to it is how- Are you saying they're not doing that anymore or they're doing that now? They are doing that anymore. No, they are doing that. Yeah, at our hackathons. I mean, that's how I used to go to hackathons as a student all the time as well. Like I would build out the back end and then this other guy would build out the front end. Yep, because usually in the big in-person events, there are many sponsors. Yeah. Right, so someone's doing this one thing, someone's doing monitoring, someone's setting up something else.

Yeah. And then there's infrastructure, so where do I host it? Yeah. And if somebody doesn't come through with their part, you're screwed. Yeah, someone's just doing, we've seen this many times, someone's just the marketing person. They're the ones- You need them too, though. They're very important, very important. Yeah, because then if nobody can speak, you don't win. They're telling you all the content about how the direction should be and about the... Video, when they do the recording for the demo. So we see also, we also see cracked engineers. One person will come alone. And they will win. So that's, I think that's the main point, for that person to be working as a designer, for that person to be working as an infrastructure person, as a developer, I don't think the right test during the hackathon, really.

But that one person is now doing all these things, and that's made possible via AI. In eight hours, being able to delegate all these tasks so you can focus on what my theme is, what my problem solving is, what my narrative is. So you see more solo developers now, more solo participants. 50% are now applicants. That's crazy. It used to be, like, you, people would get anxiety because they go to an event or a hackathon, they knew they had to find a teammate, otherwise they would have no chance. In VWing Devs, 50-something percent come by themselves. Yeah, so that's a huge shift. Yeah, and that alone person can... You can do all these things in eight hours because of AI. Their main question is, I need to develop a sleeve in, I need to make cool projects, I need to integrate all the sponsor tools.

I need to work on my idea. and presentation and whatnot and And obviously how these tools integrate with each other. So it's not like they're delegating the entire thing to cloud. I don't think that's gonna work. It's just that the way they communicate has changed. They were writing code before, they're still writing code, reviewing code, but Coding tools are doing most of the heavy lifting. Interesting. What is, maybe give me some idea, like what type of tools do you see most popularly used on the market? Is everybody on cloud or? Yeah, a lot of people on cloud, Codex is up there. Are there any Cursor users you see on your events? Cursor? Yeah. Not as much, no, claw definitely and codex in our events. And these are also not like sponsors, like they just bring their own cloud and codec setup.

um Ja, interesting. Any review tools? Any review tools. We did a hackathon with CodeRabbit. I think during hackathons, people don't care much about... Code quality. No, you just got to ship it. That's fine. You just got to ship it in eight hours, make sure it works. Make sure it works for the demo that something breaks afterwards, that's fine. Then you onboard the users and everything crashes, but yeah, it's all right. They don't really care that much about code quality. They just want to ship, ship fast. They care about UI a lot because that's, you know, it looks cool so the judges are happy. They care about deploying so people can use it easily. Vercel is popular. Vercel is quite popular, yeah. No, that's cool. How do you see, I guess, in the trend of all of these people that are participating, Are there any patterns that you see in folks that come to your events where you feel like, okay, these people, they might be coming quite successful in what they're gonna do?

You can definitely see seasoned hackathon people at events. One trait that I have seen is when you have let's say three or four sponsors in a hackathon, like about online or in person, you have a limited amount of time. The number one mistake I see people make who don't win. They would have a lot of features in line. Yeah. And they would be like, I'm gonna implement all these and whatnot. The person who wins. build a minimum viable product first, very simple UI, simple project that narrates the idea they had initially, has a working feature, the main feature, simple UI, integrates all the tools, save that as version one for backup, and then they build on top of that. This helps a lot with stress, because if you're...

Building something for eight hours. You will get stressed a little bit because the time is ticking. This helps with stress quite a bit. And the other thing is make it easy for people to judge your project. some people will have like a sign-up page that they would ask the people to like pull up a docker image. The judge is not pulling a docker image to run your project. Keep the demo open-ended. And the last thing would be... The idea does not matter that much. Oh, really? Yeah, I have, because I have seen people with the simplest ideas and perfect execution win hackathons. Okay, so it's not like it just works and it's seamless and it's a good user interface and UI and stuff. There was one hackathon where it was like There was just a lot of Chrome.

You know how there are a lot of Chrome tabs? In fact, the one project person built like some extension to manage all those external Chrome. Oh, I have a friend of mine. He needs that package manager. Sounds like that. Simple, simple tool, simple execution won the project. I don't know the hackathon was this. So idea matters. It's not like a crazy thing that's gonna make you win. I mean, in my idea, like in my world and perception that I've done a fair amount of hackathons to say that, but... I felt like the idea was getting more and more important, because I feel like the problem-solving skill, the idea of... Focusing your attention on something that... Brings value to people and is kind of somewhat novel or breaks the pattern even, um, is something that is more and more important as the execution gets easier and easier because of the AI coding agent, don't you agree?

Yeah, but if you wanna win... The judges These, if you go to like a sponsored event and you want to win a sponsored track prize, They wanna see how well you integrated the sponsor tools. And these seasoned hackathon people know that. So the idea does matter. It's just that if it's purely just the idea and the execution is not there, then... You're not gonna win. But the main thing is about if you're learning, I'm gonna say something cringe. If you're learning something new and you're having fun, you're a winner. I mean, that's true. Yeah. But these people are seasoned hackathon people you'll see at big prizes. If a hackathon has like 100K in prizes, you'll see a lot of people who are seasoned hackathon professionals.

They have their own setup, whole templates and everything ready. The hackathon sharks, those are these guys that come with their own keyboards to the event and have a little stand I'll bring a second monitor. They will bring a second monitor and Their ideas are always simple, nothing too crazy. And we don't have themes in our hackathons as such. We keep it open-ended. You can do anything. Okay. But obviously if it's a themed hackathon, like hey, build the best project for city...

## 01:00:00

Infrastructure. Yeah. Obviously then how unique your idea is matters, but idea alone won't do anything if you are not executing it well. Someone with a boring idea and perfect execution will win over your great idea and lack of execution. Well, I guess there's a tangent that... You can apply it to the professional world as well if your boss gives you a task and you do like the craziest idea of how to solve it and the craziest angle that has super much potential, but you don't execute it well. Ship fast. This is what Zuckerberg also did with the meta, the meta glasses thing, the whole AI universe he was doing. For the whole AI universe he was doing and people were like, oh no, his demo failed and whatnot.

You know, Zuckerberg's demo failed and should have waited and this and that. No. It's fine. Build, shape. iterate on it. That's good. Well, that's true. I mean, it's also a fair statement to make when you have the entire distribution engine of the world on your product, but... Exactly. But yeah, actually, that's true. If it's like a smaller company, that would backfire a little bit. Yeah. Which is why minimum viable product that you know will work, and any other fancy other features you wanna add, do it later. Yeah. But yeah, key is moving fast. Ship fast, move fast, break things, learn from your users. Great. Yeah, maybe starting to wrap up a little bit. I'd love to finish this wonderful conversation on a couple of encouraging notes and thoughts.

You just mentioned a few of them, ship fast, learn fast, you know, this entire, we hear that a lot. Do you have some practical advice for people who might be listening, thinking like, Hey, You know, some of the things you just said really resonated with me. I'm in that spot where I'm somewhat in my career, but not in truly AI native yet, and I don't feel like I need to get somewhere. Where do people pick up? What do they do? Yeah. If you're looking for your next role. I would say list down what kind of role do you want, what your ideal company look like, what does that role look like, like what does the person need to do, and then see what kind of AI tools are being used in that role.

If you are just starting out, the advice for you is very simple, it does not matter. Go get started. Get a job at the Apple store. Learn everything about the Apple ecosystem. Build Apple apps. Then go to your next job, right? Does not matter. The question around which tool I get started with language do I get started. Does not matter. Just literally doesn't. Just get started with building. That's my recommendation for juniors, like students. Just get started with building. And as you move forward, just experiment with different tools. Expand your horizon. No one expects you to be an expert when you're a student, 17, 18 years old. But what do they do expect you to do is have you know, drive and explore, like they want to see what has this person done based on the resources that they already have available.

So I go to hackathons, meetups. Speak at conferences, learning in public, which we mentioned earlier, it can be a little bit awkward speaking at conferences, intimidating, but it's going to help you, you know. And these conferences have special tracks for first-time speakers, so that's working in your favor. And if you're a senior engineer, mid to level engineer, I mean, you're already there, you're already using AI in your sort of like workflows. But... Now you have more time for problem solving, and now you can ship faster. So it's really about focusing on this product mindset, I think is what I'm hearing you right. Yes, product mindset, ship fast, use AI to achieve all the tasks and goals, but make sure you are... You're not completely relying on it, using it as a helping hand, and when you're using AI to review code tools and whatever, yeah, do a final check because if something messes up, you're the one who is responsible.

And one thing I mentioned earlier in the three points that I was mentioning, when AI is writing code, this is also true for junior engineers. AI is writing code, especially if you're a student, and AI can build the whole website for you. That is great. I think you should do it, but... Whatever the WebAI has written, make sure you read through that and understand and learn from that so that you are not just... A code monkey. Yeah, that's the word. So if you're a student... Use cloud to code 100%. But what it's generating after it's done doing it, you can ask it follow-up questions like, hey, explain to me how this works, what's the function you wrote. That's an interesting segway into also what you talked about the next frontier of engineering not being in the code writing anymore, but in the architecture level and truly understanding as a software engineer, like what is my next big thing that I should learn, study, where do I bring the most value today?

Yeah. How to... If there's one thing I would say is... Learn how to run LLMs locally. Oh wow, that's a good one. That's so interesting. So that's one thing. Learn how to learn LLMs locally. Then how to fine-tune models locally. Um... And how do you transition that pipeline? At scale, what can I do? Because that's where the people are headed. You know, a lot of talk around running AI locally in secure environments, and there are n number of things you can do to run that. to run LLMs on like infrastructure. So for example, when Kubernetes came out, there are so many ways by which you can do it. You can have your own virtual machines and Kubernetes on top of that. You can use managed Kubernetes platforms like EKS or whatever.

You can run it on that. You can use all these neo clouds and smaller providers, you can do that. So when you go to a job and the boss asks you, like, you're just starting out, and hey, we wanna run LLMs, we wanna use inferencing tools. We have hundred options, which one is the best? So now if you have the whole industry landscape idea, like hey, you can be like, we can just use an API. No, no, no, we won't use an API. We'll run it on our own hardware like this. No, we won't use that. We're gonna use Ceribrares. We're not gonna use Ceribrares, which is gonna pay Codex a monthly bill or whatnot. And umm Yeah, I mean, you know about it, like running Kubernetes, there are so many ways to do it.

The same thing with AI. I mean, just run, learn how to run AI models locally. The other thing is because this is the most basic thing, the fundamental. It's like running a model on a machine is a fundamental. So it brings us back to your earlier point. The basics. So it will teach you a lot. um Which in itself is A lot of, uh, is it, you know the book, not the book, the repository Kubernetes The Hard Way by Kelsey Hightower? So. If you're running Kubernetes obviously in production, you're Let's say a billion dollar company is most probably using AWS. Like Netflix uses AWS, but those people know how to run Kubernetes like on minikube or locally or in their own little cluster, right? So that learning journey that you get, you get to see things inside.

Yeah. And then you just use the managed service to run that at scale. OK. So learn how to run LLMs locally. It's a lot of fun and you'll learn a lot. You're going to see how it how the networking works, how it interacts with other agents, how you run multiple agents, and then you just sort of like see, okay, how do I scale this now? And yeah, but it's always comes down to learn the basics, how things are working. And when you work with more complex tools, whatever the complex tool gives you, ask it, look at it. Study from it, especially if you're a student, study from it. and then implement that. Yeah, an interesting quote comes to mind here. I had Simon, the founder of Assistant UI here not too long ago, and he said, Cloud can get you 80% there today, but all the alpha is in the remaining 20%.

And that 20% determines who gets hired. Yeah. Wonderful. Well, Kunal, it's been an amazing conversation. Thank you so much for taking the time. This was a lot of fun. Yeah, I have a little bit of a ritual that I do with every guest, so I have a couple of rapid fire questions that I'm going to ask you now if you're okay with it. Yeah, rapid fire. Yeah, and you can answer them either short or if you want to expand, that's also fine, okay? Yep, sure. All right, favorite programming language? Golang. Golang? Yeah, I just love Golang. Let's just keep it there. I don't want to start a debate. OK. Actually, shout out to Rust. Yeah. Shout out to Java also because my career started with Java.

So it's emotional. OK. All right, good. We'll keep it there. Favorite AI model? Fable five. So good. It's very expensive. Yeah, but it's good. Cool. Favourite coding harness? Claude? Cloud, yeah, okay, cloud code, cloud code, yeah, I like that. Okay. All right, no open code, no nothing like that. I like Cloud code. Okay, that's fair. I guess I feel like also now that they, like, obviously they train the RL the model on the harness, so it just gets better and better over time. I like the cloud community. I think that's another thing. Like the community is quite. There's a lot less noise. These people are just building and shipping. Where do you connect with the cloud community? Reddit, and LinkedIn, and there's a lot of people.

We have seen cloud very popular in VeeMake Devs. So a lot of our users when they Get stuck somewhere, they have a question. Probably I'm like, what are you using, using cloud? Codex is up there as well, and those are the top two. Okay, what do you do when you don't code or run your business? I'm a foodie, love trying out food. I'm full-time on [unclear] now, so I'm not, I'm working for myself. Yeah. Left my job on June. June was my last. So I'm like, okay, we are scaling. Close to two hundred fifty thousand people now, so I'm like I'm just gonna do my.

## 01:10:00

Yeah, so I thought... If for someone who thinks like you're doing your own thing, you have all the time in the world. Now it's like a full-time. I'm now doing, I think I'm working more now than I was before. Yeah, I am a foodie, so I love trying new restaurants. Love shopping, yeah, for silly stuff. Um, just like toys and whatnot. TV shows, love TV shows, movies. I don't go to cinema, I watch it at home. Spending time with friends and family is something I love quite a lot. And because I'm traveling a lot, anytime I get, I just relax. I live in London, so there's a lot of things to do, like musicals, and I don't drink, so not really pub kind of a person.

I know you're from Germany, I don't know if you drink, but I have a German friend, and I drink the whole pitcher and he would not get drunk. I'm like, how are you not, dude, how are you not drunk? He's like, it's just beer. Beer doesn't make me drunk. It's water for us at this point. And I'm like, so I'm like, if you wanted to get drunk, you have to drink hard liquor. He's like, yeah, beer I can drink as much as I want, I won't get drunk. I'm like, wow, superpower. Yeah, I don't know about that. I don't know about that. Yeah, but he's like, he drank this whole pitcher of beer. Yeah. And he was fine. Shout out to Daniel. Uh, yeah, no, I'm a foodie and I love, uh...

I think my life revolves around my family and friends. So just spending time with them no matter what we do. is what I love and yeah, foodie, movies, video games, Call of Duty. Um... Running badminton, shopping. Nice, that's a lot, yeah. That's it. Keeps me busy. Thanks for sharing that. I think this is kind of the last point where I wanted to give you the opportunity if you want to say anything to the audience here that you got coming up. Anything people, where can people follow you? Obviously, you got your YouTube channels and so on. And yeah, that's basically it. Yeah, I'll just say two things. The first thing I would say is go to remake.dev.org and check out all our events, hackathons. You can sign up and create an account.

We have built a cool platform. We're going to do more cool stuff. And if you're a company who wants to get your tool in the hands of like 200,000 engineers. We have an incredible community and they would be more than happy to... You know, give you feedback, build with it. The third last point I would say is don't worry. It's gonna be fine. That's what I would say, because... Indian life just works out and Don't worry about things that are not in your control. AI is coming, AI is replacing jobs, AI will replace some fields. It's not in your control, then why are you worrying? You know? I mean, it sounds silly. Someone is losing their job and I'm like, don't worry. You know, can't you can't say that.

But the thing is, if something is not in your control... What can you do? But wouldn't you say that... Like, let's say, I'm, we just talked about this, and I feel like we wanted to wrap up, but now I need to go back into it. Go for it. Yeah, you... There's always something you can do. Exactly. So if I'm about to, like, if I'm on the verge and I feel like, oh, my relationship at work with my manager is not going so well, like my performance is apparently not looking up so much, like... There should be something I can do about this, like, you know, go watch your tutorials, go come to one of your hackathons, go like, get CodeRabbit to review your code so you don't get any like all these things, right?

So I think that that that there's almost for anyone who's trying to, who's not completely strangled in something or maybe works in a huge tech firm where they just get cut off. Even then there's something they can do, which is train yourself up and go look for the next opportunity. In fact, if your company, if you feel like that right now, I think I should argue you should be proactive and try to get something somewhere else already, right? Yeah, it's, it's, it's like, life is not fair to everyone. There's always hardships. You get laid off. And sometimes, you know, things are not in your control. This is a very personal advice. I understand it's very hard to follow for a lot of people. Human beings are wired differently, but this is how I am wired.

If something is not in my control, it does not make me sad. Worse things could happen to me in my life, but if it's not in my control, there's no point of worrying, you know. But that's, it has taken me some time to think that's how I like myself. But there's always something you can do. There's always, people are always looking for cool people, so people are always looking for good people. You can always do freelancing. You can always upskill yourself. There are people getting hired out there. There are people getting jobs out there. The question is just why are they getting hired? Why am I not getting hired? So obviously there's one good thing is because of all these AI resources. The barrier to entry has shrinked quite a bit.

So for you to study, for you to learn, for you to research, for you to upskill yourself, it's very easy now. I mean, you have so many options. So, silver lining I would say. That's the silver lining. But if you want my personal advice, things that are not in your control, don't worry about it. Make sure you give your best. I'll give you one last example. Let's say John Wick has an interview and Spider-Man has an interview for the same job. John Wick is very focused, so John Wick is spending 100% of his time preparing for the interview. Spider-Man is spending 50% of his time preparing for the interview, 50% worrying about, oh, what would I do if I don't clear it. And let's imagine they both have the same intellect.

Who do you think has a higher chance of doing? Yeah, I mean, you can also overprepare. That also happens. But yeah. Yeah, things that are not in your control, don't worry. And then, yeah, that's it. OK, I think we'll leave it at that. Thank you so much for coming on the show tonight. Thanks everyone for tuning in and yeah, it won't be the last time that you see us, I think. I will be back. Great.