How NeetCode Turned Two Months of Hell at Amazon Into a Million-Dollar Business

You had an absolutely terrible experience at Amazon. Do you blame the situation on Amazon? Do you think that lead code questions stop being useful after you get a job? AI can't one-shot your website, nicode.io. Do you think that that day is inevitable?

If it does, is it all over for you? If you've ever prepared for a coding interview, you've probably watched a NeatCode video. Navdeep Singh, or Navi as I call him, started making LeatCode explainer videos during the pandemic when his videos were just getting 50 views.

Today, he's got over a million subscribers, and Nikko.io generates so much that it doesn't even make sense for him to continue to work at Google as a software engineer. But his journey to YouTube success began in the darkest period of his life. After lasting only two months of torture at Amazon, he found himself unemployed, struggling with his mental health, and unsure if he'd ever work in tech again. What follows is my conversation with NeetCode.

Thank you, Navi, for stopping by the studio today to have this conversation. Yeah, glad to be here. So let's just get started with your origin story. So you had an absolutely terrible experience at Amazon. I think one that led to some mental health issues.

Do you blame the situation on Amazon, its culture, the team you were on, or the people that you worked with? For me personally, I don't try to look back on that and try to like blame anybody because... that's i don't feel like that's a constructive like mindset to have but i think it was just a difficult situation and we were a software engineering team trying to solve a problem

It happened to be like a stressful environment because like lack of resources and just kind of, I feel like the Amazon culture in general, things just happened the way that they did. And I don't want to over index too much on that experience because.

then I'm going to start thinking things about myself. Like, hey, I wasn't good enough or I can't be this or I can't be that because of these people. I don't really want to put myself in that situation. I want to move on and focus on, you know, things that I can do. And so when I look at that situation, I don't.

I don't take anything personally with what happened. You hold a grudge against them at all? I did for a bit. And I actively tried to let it go. And for that reason, I think I've been able to. I think a lot of people would think that...

once you got into Google, which was about a year after that, or once you got popular or made a lot of money or whatever, after that, all those problems went away. And for me, at least, they didn't. I remember... taking showers even after i got into google and just like the thought would just kind of randomly flash into my head and that

moment it wasn't so much of me like looking back on that and regretting it was more just like well crap what if that happens again at google right because that's how the brain works like your experiences stay with you and then they influence the things that you're going to do in the future and so that stayed with me for a while

and then eventually like i got over it because i did decently while i google and things worked out i feel terrible about this situation because i remember when those videos came out it was sort of like right after you would

shown your face on your channel. I think you went faceless for a number of years or a little while before you revealed yourself. And it's hard for me as a longtime Amazonian to... to like I just wanted to apologize to you I wanted to like figure out like how to contact you and just feel like I'm so sorry because it it sounded absolutely horrible and the way that

Apologists like me typically will frame it. It'll be like, oh, Amazon's so big, we're decentralized. And so you're going to have pockets of good and pockets of bad. And it sounds like, you know, you got the pocket of bad. That allows me to wash my hands of it. It allows Amazon to wash their hands of it. I think there's some truly terrible things that are going on within that company. And, you know, I don't think the loop gets closed. I think probably the team that you were on.

is still largely like that. Yeah, I could talk about the team environment all day, and I could remember the people, I could remember a lot of things about it. But zooming out, I guess, on the whole software engineering landscape, what it taught me... like in hindsight was that things aren't like a well-oiled machine like even at a company like amazon and i think like my expectations were just so different i had heard that amazon was toxic and i expected that but i thought

well, if I just work really hard, if I'm really smart, like I did well in school, like that will translate to Amazon. And it really doesn't. When Amazon... talks about their leadership principles and a lot of companies have principles and you could probably speak to this better than i can but amazon really means it they're not joking and it's true that

I don't think you have to be born a certain way to do well at Amazon. I honestly don't. But I think you have to like adapt and you have to go in with those expectations. And I think I probably didn't do that. I think, you know, I'd focus my whole life on academic stuff. I hadn't really focused on like all the other interpersonal skills that you have to have to perform well at a job. So I think...

Like somebody with 10 years of experience could go into it and just say that this is toxic. This is like a crappy situation. I'll try to find a different one. Somebody with less experience like I had can't really make those conclusions. You kind of over index on it and think like, well, this is the world.

Like, I'm just not performing well. And, you know, you can draw a bunch of conclusions from that. Yeah. One thing I've learned about Amazon is, you know, principled thinking gets you pretty far, but you kind of have to be stubborn about it and you have to be okay with the consequences of it. I think the specific set that they chose leads you to situations where teams take like an entry level or somebody earlier in their career and then just put a bunch of shit on their plate.

and then say like, hey, sink or swim, because if you can't swim, then we'll just get rid of you and get somebody else. And I think to an extent, they're just kind of okay with that. And for people like me that... We're able to swim. We're able to be on the other side of that performance curve. I can give the lip service and say like, hey, I get it. You know, it's probably better for everybody that we split ways. But in another sense, like because I was rewarded for that system.

It makes it really easy for me to sort of like brush off these concerns. And so I do think that there are some very systematic problems with Amazon's culture.

On the other hand, I also believe that they're not as successful if they are less principled. And so it's hard for me. I feel very conflicted about both your situation and the situation of everybody that's sort of been trampled by Amazon. Because I do think that... there's some great things that are going on there and also the amazon is a sort of whipping boy for a lot of stuff like okay cool amazon's terrible but aws is kind of a great thing i think uh you know other than their console and

and, you know, the expense and all of that other stuff. But in terms of what they enabled for the internet, I thought it was great. I love it that my toothpaste comes like the day I order it, but there's a price associated with it. yeah i completely agree and i think as i've gotten older i can see like the other side and as i've had my own company as i've worked at another company and seen

like maybe the other end of the spectrum where people are a little bit too like laxadaisy and they don't take things as seriously. I can see why things are the way that they are. It's not just kind of random sometimes. And I feel like I have to mention this just in case somebody's wondering, like the concrete.

things that I like that happened like when i was at amazon i'll just kind of go through it quickly it's not like i joined and somebody just started whacking me on the head right like it's not like that and so so what exactly happened like while i was working at amazon the short version is i worked there for about two months I joined remotely. I think that contributed to a lot of like

lack of collaboration. It was during COVID. I joined a team and Amazon Alexa, which is not, you can speak to this maybe more than I can, but for my short experience, it's not the well-run, well-oiled machine part of Amazon. And so my team that I was in, and I can only go by what the other engineers, they were saying this feels like a thankless job. Many of them had like commits at 2 and 3 a.m. And so...

I can kind of understand why, like, if I had one question, you know, the next morning, if somebody had stayed up until 3am writing code and they had to get up for a meeting early in the morning, I can see why they would be frustrated. I would have done the same thing. But you can see...

The biggest thing that I took away was that that was like a team systematic problem. Like there was no one individual that could save the team, including the manager. It was just not going to work. And the main reason that I ended up leaving.

was not because it was like a stressful environment, not just because of that. Like there was a lot of things going on at the time for me personally. And for me, it was actually a very like uncharacteristic decision to just say, well, like I'm just going to back out of this. Normally I would... I would try very hard to make that work, but I ended up leaving and it ended up being a good decision for me.

Sometimes I do wish. What would have happened had I just stuck it out and had I just kept going? I think I could have eventually succeeded, or worst case, they would have fired me, and then I would have gotten some decent severance, which wouldn't have been that bad either. I'd like to share something important.

This episode marks the beginning of a major partnership. Linear is now the presenting sponsor of this podcast for the entire next year. This isn't just another ad deal. I chose Linear because they represent everything we talk about on this podcast. building exceptional things, like your career, with taste, craft, and an obsession with quality. When I left Amazon, I tried managing my content business with traditional project management software. And to be honest, it slowed me down.

Spent way too much time setting things up and then telling my team what the new process was. Linear is different because they said no to endless customization. They optimized for the use cases that people who need to ship actually need.

not on things like generating reports about why they're not shipping. Whether you're building software or building content, the fundamentals are the same. You need tools that help you focus on the work, not manage the work. This is why OpenAI Cursor and Perplexity Basically, every AI company reshaping our industry runs on linear. When you're moving that fast, you can't afford tools that slow you down. While I'm definitely not open AI, I have moved my entire video production workflow to linear.

Every episode, every guest, every deadline, my entire team runs on it now. And here's the nail in the coffin that convinced me to make them our presenting sponsor. It took me 15 minutes to set everything up. Not days or weeks, 15 minutes.

Whether you want to try it on that side gig or use it at work, visit linear.app.ale and get six months of linear business for free. Thanks again to Linear for sponsoring this video and the podcast. You've been really public with your mental health struggles during that time period. Do you think more software developers need therapy? And then also on top of that, like what ended up working for you? Yeah, that's a really good question.

And I've been waiting for the day somebody asked me this question because I honestly feel I've helped people when it comes to algorithms and data structures. If I could go back in time and help myself, the biggest help I could give myself would be in the mental health. space just because I feel like the way my brain is wired maybe there's like a hint of autism thrown in there maybe not I don't know a little autism yeah but but but like

That's how I would describe it. I think maybe a lot of other people fall into that bucket. Like, I feel like if in programming, there's like a rule book, right? Like you have your loops, your conditionals, you have all this stuff. And those are like the basics. Before you can move any further, you just have to know what the basics are. And then once you can do that, there's like infinite complexity. You can program all sorts of things.

The way my brain is wired, I have to learn things like in a very deliberate way. Like sometimes I just don't pick up on like the things I don't read between the lines. And so I learned a lot about...

something called cognitive behavioral therapy for me it was life-changing it's the type of thing that i knew about for a long time but just kind of put it off like throughout teenage years and in college and stuff because i thought like i didn't need this this is stupid like what can you learn by reading a book or whatever

But it really changed my life. It didn't just change my life when it came to dealing with people, but also when it came to work. This is something that's really, really obvious to me now, but it wasn't obvious in the past. If you show up to work... And you want to be maybe closer to your colleagues or you just want to like have some kind of relationship with people. Right.

If you see somebody that you know, just smiling at them or just saying hi, it's such a small thing. The logical part of your brain doesn't understand that that's like the doorway to all sorts of things, but...

like it really is like once people get familiar with you then it just opens up a lot of doors and the logical part of your brain just doesn't pick that up and at least for me it didn't and so I had to like learn these things and then once you kind of like start doing them then you just like learn all sorts of stuff so

That's kind of my big take on it. I think a lot of people in programming, there's a meme of how programmers don't have social skills and things like that. I was probably one of them. Maybe I still am. But I think anything can be learned. Just like programming, you can learn social skills as well.

society really encourages you to shove your emotions down below the surface. A lot of people will be like, okay, well, you should, you know, oh, I love people that are really in touch with their emotions. But if you cry, you know, in a workplace setting or in a social setting, people will judge you. Society is really rewarding you to have mastery of your emotions.

But they have not given you the tools to do so. And mastery of your emotions and really pushing them down so that they don't surface and repressing them, those both look like the same thing from the outside. I... Think more people should spend money on therapy than they should spend it on textbooks on how to program better

Yeah, I agree. I've had it like in the past, like in a relationship where people will kind of give you that advice where it's like, be more emotional, be more in touch with your feelings. And it's not that like... me or many other people can't do that but it's it's about like trying to reconcile what you said where it's like no you're supposed to be big strong and macho and like in programming

i don't know what it is about programming but for some reason it's like a very masculine thing where it looks like okay you sit there you spend hours by yourself alone just programming it's very logical and like emotions generally don't really help you with programming and then it's like okay now you need to be more emotional now you need to like let go of the logic and it's it's like

It's like you have to give up the thing that got you there in the first place, and it's hard to reconcile those things. I have two young children. They're one and three years old. And when you're a toddler, their emotions are about to overrun. You want them to like name their emotions. And so I'm like, okay, cool. And I realized I didn't have the vocabulary to like name these finer grained emotions. You know, the difference between being happy or elated or frustrated.

or angry or like i didn't have literally have the words and so somebody like me who's like essentially emotionally stunted gets thrown into the world, you know, and to navigate all this, I just feel like society has let us down in terms of just, you know, learning how to like go about the world in a way where.

you at least know what the terms are so that you can start to improve yourself. I think it's hard to know, at least for my age group, and people even younger than me, it's even harder. But when I was a kid, at least, there was like... an idea of like what what a man is supposed to be right like

whatever like stereotypes they were and whatever you saw on TV, but it was pretty consistent. It was like people mostly agreed like, okay, a good man is supposed to do this and that. I feel like nowadays it's very complicated. Everybody has their own definition, which is great, but I feel like it's hard to be a kid nowadays, especially with social media and whatnot. So let's talk about your content. You do, obviously, leak code and coding problem content.

What makes your content different than, say, some other person that might be doing an explainer on a leak code problem? Yeah, I think at this point... The biggest difference is just that I think I did it before everybody else. I remember it was the pandemic. I can remember it like it was yesterday.

i'm seeing a bunch of videos i don't know if you had started youtube just yet but i remember i was seeing a bunch of viral videos everywhere people doing day in the life videos getting five million views and i was just sitting there making my like leak code video getting 50 views a video or something and i just really liked it and i just saw that it was helping people and like

Because somebody watching a day in the life video, they might enjoy it, but they don't really get anything from it. It was entertainment. Whereas my video, I remember seeing people like, hey, I could not solve this problem. I was stuck on it forever and then I saw your video and it saved me so much time and I really liked it and I'm going to subscribe and I'm going to keep watching future videos. And I could feel...

I could empathize with them so well because I was doing the exact same thing. I remember going on LeakCode. I'm seeing these solutions. Nobody's explaining it. You have to read it line by line to finally understand what it's doing. and i was crazy enough to to put my time to do that and so i found like a pain point that was so strong and i was solving it so well i would have never expected that it would have like gotten as popular as it did and i think that's why

like nobody else really did it because nobody expected those videos to get popular and nobody was really doing it the way I was. And when I say the way I was, it was just about going really, really in depth and explaining things in like an intuitive way where it's not just.

you memorized the algorithm. Now you kind of actually understood why it worked. You saw the picture, you saw like be drawing all the lines and things like that. So basically, you're the first person to explain the solution rather than just give the solution. Yeah, I think so. Sounds so easy when you say it like that. Do you think lead code questions stop being useful after you get a job? I know personally, when I had a job, I never really had to interview at other places.

I loved picking up lead code questions because it was like a Sudoku. Like, why do people do Sudoku and drink their coffee? It's like, do they just want extra stuff in there? There's something fun about a puzzle.

But if I'm like brutally honest, like I didn't need any of the leak code stuff once I had a job. Yeah, I agree. And I think honestly, for a lot of the algorithms.

Because like you said, it's like a puzzle and people will call it with leak code pattern matching, right? So if you see a problem, okay, well, which pattern can apply to it? And sometimes you do have to do creative things to get the problem to work. But a lot of the time... it's just well do you know the pattern because if you don't know the pattern well then you're not going to do well at it and so it's okay so then you have to go through and learn all the patterns

So it's like, just before you can even play the game, before you can even play Sudoku, you gotta learn the entire English language and the alphabet and stuff like that. And I don't know if it's the most useful way to interview people. Obviously I'm biased, so I'll mention my bias.

Like, the longer that coding interviews, like, stay the way that they are, like, it's good for me. But if there were a better way, somebody would have proposed it by now and it would have been implemented. And maybe in some companies it's already being implemented, but... I feel like it's a harder problem to solve than people think because it's definitely like LeetCode questions are not useful on the job to answer your question. No, they're not. So, yeah.

It just runs out of steam, right? It's just like, my big saying is that the clubhouse does not resemble the gatehouse. The gatehouse is how you get into the club, but you stop having to prove yourself in that way. At work, you have to do something else, right? And how many times has somebody asked a question that they would never be able to answer themselves? It's like the classic, the interviewer is going to do that. I've been guilty of it. I try to correct myself. I've done nearly a...

thousand interviews and every once in a while i'll you know i'll ask a question that i don't know the optimal answer to but yeah i think it would be you'd be really great if there's a registry of questions and you had to like prove that you could solve the question before you asked it, right, to make it absolutely fair. So when you boil down necode.io, the website itself, there's not a ton of engineering that goes into it. I think there's some interesting problems. But at the end of the day,

You know, it's a website, but it's not a really challenging engineering problem. Correct me if I'm wrong. I think it's, you know, it's not because of your skill. It's just that's just the domain. Right. And it sounds like on this. Google team, there wasn't a ton of engineering that was going on. It was a little bit more figuring out what was already there and configuring it and trying to get it to do the thing that you needed to do.

maybe a different formulation of my question is are is there some part of you that misses the a place where you can do some deep engineering problem because i feel like at the at the end of the day like leet code problems coding problems they give you that sense of satisfaction but at the end of the day leet code is

Nothing changes. It's stateless. You've solved the problem or you don't solve the problem. Nothing about the world changes. So like, are you craving a place where you can make an impact through these deep engineering skills that you've cultivated over the years?

yeah i think that's interesting because i've never actually like intrinsically i've never like people are drawn to programming because they can create an impact but i've actually never intrinsically craved creating an impact or cared about that specifically i care more about just like the joy of the craft or just like for leak code for example

I don't even like programming, how I would describe it, actually. Like, I like problems. It comes out. Yeah. Like, if I could go back in time, and if money wasn't ever an issue, I'd definitely study math or physics or something. But with elite code, it's kind of like...

You feel close to computer science theory, and that's what I really enjoy. Now, you also said the NeetCode website that I'm running. It's true, it's pretty cookie-cutter, basic web app stuff. And I don't enjoy that as much as I enjoy leak code. But when I'm creating an app, there's very low overhead and bureaucracy. It's just like, just go like as fast as you can get this deployed, get it working. If there's things are rough around the edges.

clean that up afterwards it's not that way at google it's the exact opposite just clean like polish everything and okay it's not even gonna ever get deployed it's like okay i just spent like weeks doing that and it's not gonna go live and some executive just changed the feature actually know we're doing this and it's like well okay like it's just not fun and yeah because everything just goes really really slowly at google like from the little that i remember at amazon

It was definitely more enjoyable at Amazon because my team was using Lambda functions and other AWS services. And the culture at Amazon felt like, you know, just go, just do it. There was some designing involved as well. And we did unit tests and stuff. I don't do unit tests for my own project because why would I do that? Because it's not, like you said, it's not even that complicated. And if I'm writing all the code, usually I can catch the bugs. I think...

That happens at every large organization. Essentially special specialization comes in and then you're not working on the mobile app or even the Android app. or even any of the functionality, you're working on app startup and trying to cut that down by another hundred milliseconds. And it takes an incredible amount of work.

But at the end of the day, it's a tiny piece of a much broader thing that you're giving out to folks. And so by the time you slice it all down and then you want to talk about your particular impact, it gets diluted. But it sounds like you're totally okay with not really having to flex your engineering muscle for your day-to-day. It's more about the understanding. Yeah, I think...

Even though the problems have been relatively simple, I've honestly, people won't believe this, but I've honestly learned more just building neat code, doing my own stuff, reading my own things, than working at Google. People will say, well, if you work at Google for X many years, like...

then you'll learn all these things. But it's not working at, at least at Google, it's not working at the company that's going to teach you all those things. You're still going to have to learn them yourself. And like, for example, with distributed systems, nobody's going to go into Google.

90% of the time and just start learning distributed systems from their team because everything is solved at google every like all those problems are solved you use borg to do this and use spanner and you use like all these tools that are already created that already manage everything

that like the microservices problems are solved at google they have bock for that and you start doing arcane things you don't know why you're doing them but it's a solved problem and so when you step outside of that then you can actually like learn things from like first principles rather than, okay, this is the way Google is doing it. So right now, AI cannot one-shot your website, nicode.io. Do you think that that day is inevitable? If it is going to come...

Is it all over for you? Yeah, that's interesting. So if it was going to one-shot my website, a lot of the easy stuff, like just creating... of front-end and having a database, having Stripe payments added to that. I think a lot of that can be built right now. The hardest part probably to clone would be the code execution functionality. And I'm using mostly just tools for that. Like I don't have my own execution engine. I'm actually using an API for that.

But I'm pretty confident that these AI tools are not going to actually be able to do that. If it can, it's going to take a smart person to plan it out, to basically tell it what to do, and then do that.

Yeah, I think even if we were a couple years from AI cloning my website, nowadays it seems like that's not really where the value of a business comes from anyway. It seems like nowadays... uh there's a lot more involved uh with the business um so i'm not really worried about that now if it does clone my website

We're probably at a point where coding interviews probably aren't going to be so important anymore. So at that point, my website is dead anyway, which is fine with me. I'm not so married to the idea of coding interviews. I've enjoyed coding interviews, but I am happy to do something else with my life. Yeah, I think that my take on all of this stuff is like, I 100% believe that AI is going to do 80 to 90% of the work that's necessary. No problem.

and that day is coming if you want to be more efficient with your stuff it just it doesn't make sense not to have ai do the stuff that you don't want to do like all of the annoying stuff and i worry the most that it's going to remove like tinkering from people that are just getting into the industry. I think back to high school, I made a Tetris clone in high school in Visual Basic, Visual Basic 6.

It was a very fine vintage of visual basic thing. I didn't know any better. Anyways, so I made this Tetris clone. It was a little dodgy when you try to rotate when it's getting close to other bricks, but for the most part it worked. And I learned a ton. about it and now it's almost like hey go build a tetris clone and one shot it is a way to evaluate all of these different models and so i think back to like steve that was like 16 years old

Trying to tinker with this stuff. Would I actually try to build that thing, or would I try to go and one-shot it with Claude and then show my friends instead? And that kind of scares me a little bit, that we're losing that. Yeah, that's a good point. I have a brother who's doing college. He's doing computer science. And the big thing I've seen is that there was always, I think, a culture of cheating in school. But I mean, it's gone like way up now. And it's kind of like if you're not using GPT.

to basically do your assignments for you. You're the odd one out. You're the dumb one. And I mean, I completely agree where it really just creates a culture of people not being able to do like basic things. I mean, the question is like, will those basic things go away completely and people won't have to know them anymore? Maybe they will. But if they don't, it just kind of creates a generation of people that are like mentally handicapped to...

to do anything beyond, like, basic stuff. And it seems like a waste of college. It seems like a waste of money to go through that and not have any skills afterwards. Well, what's going to happen is, you know, I'm old enough that... when you took a test in college you had to go buy these scantrons and you had to like fill in the bubbles and then hey you have to go and get an essay book and write the essay in class and that was the thing that was graded so i think

You know, there's all this news about AI and GPT helping you cheat on programming interview stuff. They're just going to go back and do what they used to do, which is fly people in for the interview and then give them a whiteboard. The other option is you say, use AI. Go ahead. I dare you. Use AI. But now the bar is going to be really, really high. They're going to be looking at that 5% or 10% that AI can't do as the way to sort of measure you. Yeah.

Maybe that's a better world or maybe not. So let's talk about advice in this shifting. What's the advice that you give your brother that's in college now? Is he studying computer science? Yeah, that's right. So, sure, he's the odd personality because he's not using GPT to do his homework, or maybe he is. Either way, it doesn't matter if your brother's a cheater or not. But, like...

How does he buffer himself against like the quickly shifting landscape, especially with AI? Yeah, I mean, I think in terms of like my advice, it's been to mostly just like learn how to learn. Like I try to kind of give him... like the programming like mindset that it took me a long time to figure out like math for example is very different than programming when you want to like learn to program and it's easier to do nowadays with like the ai tools they can kind of

like regardless of what you're trying to learn like the AI tool can be like the first step for you almost every single time but it's more about just like learning how to learn where there's like you're learning math you kind of have to like think very hard but then there's programming

which took me a long time long to learn this but it's you you should approach a lot of programming the same way you approach like a video game where you're kind of just like mashing buttons where you're just not thinking super hard about it you're just trying things you're making mistakes like don't be so

caught up on like writing this piece of code the exact way go ahead write it the wrong way maybe duplicate your code or something you can always go back and then re like learn stuff if you're just learning and you're not making like a really important project or something So that's kind of been my advice. Focus on that and also to just focus on doing things that you enjoy because then you'll be able to do things that are truly impressive and it'll be easier for you to do that.

so nowadays with ai tools it's very easy to create like a crud full stack web app and even get it deployed i think that's great for somebody uh just trying to learn like that's the first thing you should do then you kind of know how an app works end to end but then

Try to do something interesting. Try to do something hard. Try to do something that maybe other people haven't done before. I think that's really good to develop your skills. Also, I think the social aspect of it I think is very important. I think... For any college student right now, try to socialize with people. If you have a professor that's really interesting or they teach a really interesting class, please talk to them.

Just have a conversation with them. It doesn't have to be networking. You don't have to try to get something out of it. If you get something out of it, that's great. But just build a relationship. There's nothing that could go wrong with trying to do that. Would you advise somebody to...

take computer science as a major if they were making a decision about what to study today? I think it would probably depend. If you enjoy computer science, I think I definitely would. If you're just doing it for the money... I think it's still a reasonable choice because, I mean, if your goal is just to go to college to get a reasonably good paying job, I think computer science is still a decent choice.

But I think people should know about like the job market. Like right now, people say the job market is stressful, is not great. At the same time, I'm definitely seeing plenty of people, students get jobs. So it's not like... I would say even the majority of CS grads are still getting jobs, but it is harder than it was before. Don't expect to just coast your way to a job. Don't expect your degree.

to be enough to get a job there are other things you're gonna have to do and those other things are more important than the degree um but i i think it's still a valid choice so What if you're one of those folks that want to get into tech and they're trying to land that first job and it's just not working? You've, you know, I scroll Reddit all the time, you know, CS majors.

That's subreddit, experienced devs, that's programming. And there's this sort of thing in the air where people are really discouraged and they're kind of spiraling right now. And I completely... sympathize with them i've i've applied to 1000 jobs and i have not gotten a call back like what should they do yeah i mean

It's definitely a difficult situation. I was kind of in the same situation as well after I left Amazon for almost a year. I was kind of just applying and doing leak code, which I wouldn't recommend to other people. I think it would... again like depending on the exact situation but in general i would figure out what you haven't done yet like what haven't you tried because i think it's it's easy to just keep pressing the button

the one that's in front of you. But there are other buttons available to press. There are other things that you could do. Is there some way you could put yourself out there? Is there some way you could go to some event to put yourself in front of other people that could be hiring or are also looking for jobs? Could you somehow do that?

If you can, you should ask yourself why you haven't done it yet. Probably because it's uncomfortable. Probably because you thought you could just spam jobs and get a job the easy way. And I wish you could. I wish that was still an option. But if it's not, you've got to try the other way.

other thing try to get yourself in front of people try to show people what you have to offer and if you feel like you don't have a lot to offer then at least you have time to work on like skills try to do something truly impressive if you have a month right now

You could be able to do something original within a month or try to get in front of other people and see, hey, what problems do you want solved? Like somehow on social media, maybe see what other people are working on. Try to join them. Try to ask them, is there a problem I can solve for you?

I wouldn't necessarily recommend like unpaid internships or unpaid opportunities because that's like a slippery slope. But that is an option if you're just trying to get experience or trying to get your foot in the door. I've seen it work for other people where they... They got rejected, but they said, hey, what if I work for free? What if I do that? I think even the interview coder person, Roy Lee, did something similar.

you know if you can't get a job it's been a long time and you're desperate there are no rules anymore like feel free to throw the rule book out the window start breaking rules try doing things yeah that's that's um That's what I would say. But I guess I don't know for sure without knowing like the exact person, but I think that's like the mentality I would have. I bet if you got one of those people that claimed to have applied to a thousand jobs and you just hovered over their shoulder.

you'd probably just be like, that's how you're trying to do it? So there's probably some more instructions necessary. They aren't maybe being forthcoming with everything about their situation. But I love that framing, which is just like, they're trying to spam a button in a very specific way. So if it's not working, do something else. Yeah. Right. And so, you know, I've done nothing. You know, what's that one line? It's like, I've done nothing and I'm all out of ideas or something like that.

So, yeah, there's, you know, I also think that it's, there's never been a better time to just do a thing on your own. Yeah. for me i quit about a year ago and i've had about six ideas that i'd like six businesses that i would like to start and i don't have time i have all this other stuff that i need to do

But I wish I had a month where I didn't have to do anything and I knew a little bit about code and I didn't have any obligations and some free time. I would love that. And so I do think that... The thing that you have if you're young and unable to find a job is a bunch of time with no commitments. You can buy a bunch of lotto tickets, not literally, but figuratively. Yeah.

Awesome. So let's move into the rapid fire session. So what's one thing that's really overvalued right now that everybody's talking about that's kind of driving you insane? I guess the vibe coding is like the first thing that comes to mind. I think people say like you need to learn how to use AI tools and I've used a handful of them and all I can say is

They're the most intuitive tools to use. Like, I guess you kind of learn like edge cases with prompting and like, okay, like, hey, I was using AI wrong. AI the wrong way before I guess what you're supposed to do is tell cursor to make a plan first and then it executes the plan and that's not a learning curve if you're a programmer like that's not a learning curve like you could i feel like you could go to sleep from now until next year and if you need to ramp up on the latest ai tools

You can do it in like a day. It's like I just used Claude Coat for the first time. It came out a few months ago, but I just used it for the first time. And it's so easy. It's a terminal tool, but I feel like even people who've never used a terminal could use it because that's how easy it is. It's just like a chat box. And so I think people are focusing way too much on that. Like somebody...

AI is not going to replace you. Somebody using an AI tool is going to replace you. I don't believe in that. Are you team like vibe coding means... Not looking at the code and then just accepting all of the commits? Or are you team vibe coding is just computer-assisted coding with AI? Yeah, I guess to me it's computer-assisted coding with AI. Yeah. Amazon just came out with their product where it's basically like you just come up with the tech spec and you get into the requirements space.

And then once you're good there, then it comes up with a plan and then breaks it down into smaller pieces. And then the AI goes and implements those smaller pieces. Versus, you know, you're trying to one-shot something with these two-sentence requirements. And so it'll be really interesting to see where you go. But I completely agree with you in the sense that if I just wake up in July of 2026,

The world is not fundamentally different, and I will pick up an AI coding tool, and then you just need to tell me what the delta is between that year that has occurred and now. Yeah. I'm curious. While you were at Amazon, maybe you can't say, but I'm sure you guys had some kind of like copilot tool, right? Like some kind of internal tool working. How...

Like, was it any good? And like, how much of your job did it automate? Because when I was at Google, we just started getting something like that. And it would have probably been pretty helpful for like config changes and things like that. For anything more complicated than that, I don't think it would have done the job. I got out before anything was officially allowed to be used. I think they used a tool called Q.

Which when I talk to developers, they kind of go, it's not good. So I do think that with, I think it's called Cairo, their new tool, it's powered by Claude. And so I have high hopes for that. But I wasn't around for that. I'm kind of glad that I wasn't. I do think, though, that they had spent a lot of time and effort on static analysis. And so I think there's just something to scanning code.

and looking for things without humans having to be in the loop and so i think that ultimately like static analysis with direct prompts right or looking for a specific thing ultimately i think is going to be more powerful than these current set of llms because what i've noticed is I thought that part of like this AI revolution was going to be that LLMs and the AI were going to find these unseen connections between things. That there were going to be like, oh, that looks like a branch of math.

uh that's related but you haven't been looking at it like maybe there's a connection to like this biology research that you're using and then it would just be like oh and then we would make we would start to make like really big leaps And so I think about the millions, probably billions of lines of code at Amazon, which is probably not in the training set of any LLM.

And trying to fit that into a context window doesn't make a ton of sense to me. And so I'm like, well, what is the thing that is going to... to really unleash these insights that AI was supposed to do. That's what we got sold on. Instead, we have these less than a million token context windows that...

that really are trying to like condense and compress all of like the prior contacts that are there. So essentially you can think about an LLM as being stateless. And with the way that attention works and runtime compute works, it's... It's just more computationally expensive. It doesn't scale linearly. It scales exponentially. So I have really...

deep concerns about whether AI agents are truly going to be able to replace senior devs and above. I think right now you are seeing a paradigm where people are talking about you know, the like cloud code being an intern. And so then you're now you're managing a team of interns or entry level folks. They can go and write the unit test. They can go and write the code. Your job is to pull them back in.

But is it actually going to be the developer or some sort of agent that's reviewing that code and this being like, that's a good idea or not a good idea? I don't think we're there yet, and I don't see line of sight to that. Okay, so the compliment to that is what's undervalued right now? What's something that you can't believe more people aren't looking at? I guess in the learning space, for example, like if you're a student or something.

I think now that AI has made everything easier to learn, you don't really need to open up Stack Overflow. You can just have your editor, maybe cursor, maybe GPT or cloud code or whatever.

and you can get so far now like at least for me i i could make a project so much quicker than i could before because i don't have to like go back like all the tedious things i would need to like open up stack overflow for i don't need to do that anymore especially if i'm in like a brand new programming environment so

I think people are underestimating the interesting projects that they could create, especially if you're a student. Everybody does front-end stuff. Try doing some infrastructure stuff. Try doing something interesting that maybe you had heard about in one of your college classes.

like try to create a clone of some kind of infrastructure service like AWS Lambda or something. You'll run into some really interesting problems. Like I've explored doing something like that. And I think if you had that on your resume, You'd stand out so easily. You could post about it on Twitter. Somebody would notice. Somebody would be able to understand that this is genuinely impressive. So I think in the learning space, people aren't really...

people aren't getting creative. I've seen so many like full stack clones on YouTube, but nobody's doing infrastructure stuff. Like there's AWS services that don't have a single YouTube video around them right now. Like there's just so much opportunity, but nobody wants to do the things that are hard. Or I won't say nobody wants to do them. I'll just say...

The reason people haven't been doing these things is because they're hard. Just like nobody was making DSA videos in 2020 because nobody wanted to read the leak code for them and try to understand that crap because it's hard. But if you find it enjoyable... then it'll help you stand out. The world does not need another to-do app. No. It really doesn't. And so, yeah, that would turn heads. It's like thinking about a video.

idea i tried to make aws lambda and this is how i failed or succeeded i would like i'd watch the crap out of that video yeah so what content are you currently binging right now doesn't have to be tech can be anything

Um, it's mostly tech Twitter right now. I never understood Twitter for like the first few years that I was using it, but just recently... I realized why people use Twitter because it's just like a raw place to... I mean, once you get past all the random spam and like... if your feed is anything like mine you might have like random porn just felt like riddled in there for whatever reason but once you can get past that it's really interesting just seeing like the like just drama and just like tech

I feel like news hits Twitter before it hits most other apps. Like, if some major tech thing is happening, you'll hear about it on Twitter sometimes before, like, the Wall Street Journal or New York Times will report it. So, I think, um... Yeah, I've just been banking a lot of Twitter. Yeah, I do think that it's at the boundary of what's going on.

I also think there's a lot of noise there. So it's basically, you know, do you want to be the filter? Because if you do, you can find some really great stuff there. Or it could be that Grok is, you know, sort of like hallucinating. Oh, yeah. And this is a question I'm trying to ask everybody that shows up for the podcast, which is, what would you have been if you weren't in tech? Probably doing math or physics.

I always wanted to make money, but I don't know if I'd be able to make a ton of money doing those. So I'd probably just become a professor or a teacher or maybe go into finance if I was lucky. My uncle is a math professor, and so I think about this parallel life I could have had if I... He also works at JPL and NASA. He's the actual academic all-star in the family. I just make YouTube videos. Seems like he has a pretty decent life.

Well, Navi, thanks so much for joining me on the podcast. I really appreciated the conversation. Of course. Appreciate it.