Software Misadventures - Kelsey Hightower - On retiring as Distinguished Engineer from Google at 42 (Part 1)

Episode Date: July 24, 2023

Kelsey Hightower was a Distinguished Engineer at Google, where he worked on Google Cloud Platform. In this first part of the conversation, we delve into pivotal moments in Kelsey’s career journey ranging from buying his first car by working at mcdonald’s after school, to starting his own computer store that turned into a music studio after 6pm, to hacking on python infrastructure with the core developers. Through these stories, we learned a ton about how Kelsey thinks about acquiring new skills - getting paid for it, breaking into the world of open source, navigating corporate politics, building trust within a team, and much more.

So again, I was in survival mode even after having some success. So year 10, year 15, I'm still that person that's like, hey, there's not a lot of distance between you and the person living on the street. And so I'm like, I got to save as much money as possible. The tech industry may not be paying these kind of numbers forever. I got to get what I can, save what I can, because I'm still in that mindset. I don't have a college degree. I'm very fortunate to be in this scenario. What if it changes? So you have to unlearn that over time. I have to unlearn the whole, keep getting as much as possible.

Welcome to the Software Misadventures podcast. We are your hosts, Ronak and Guan. As engineers, we are interested in not just the technologies, but the people and the stories behind them. So on this show, we try to scratch our own edge by sitting down with engineers, founders, and investors to chat about their path, lessons they have learned, and of course, the misadventures along the way. Hi everyone, we're thrilled toadventures along the way. community and a distinguished engineer at Google, Kelsey recently announced his retirement at only 42.

We are honored to have caught up with him and enjoyed our conversation so much that it lasted two and a half hours. So to make it easier to digest, we decided to split it into two episodes. In the first part, we delve into the pivotal moments in Kelsey's career journey, ranging from buying his first car by working at McDonald's after school, to starting his own computer store that turned into a music studio after 6pm, to hacking on Python infrastructure with the core developers. Through these stories, we learned a ton about how Kelsey thinks about acquiring new skills and getting paid for it,

breaking into the world of open source, navigating corporate politics, building trust within a team, and much more. Please enjoy this conversation with Kelsey Hightower. Okay, so Kelsey, again, thank you so much for coming back on the show. Super excited to have you with us today. The big thing, the big announcement that you did recently, I think it was sometime this week or last week, you mentioned that you are retiring from Google.

And this is not retirement from tech, but retiring from the regular nine to five. That's a huge milestone, something we definitely want to talk about. And we have so many questions like I'm sure many others. Before we dive into that, let's talk about your career a little bit. I'm sure you have been reflecting on your career. I think you've been working for what, 20, 25 years at this point. So when you look back, what are the major milestones that you see? And how do you think they changed or impacted your perspective on life as well as work? Yeah, man, I think my first job, like many people, you know, you're working at fast food, anything within walking

distance of your house is fair game. And so, you know, 14, got a work permit, and got my first job at McDonald's, right? It's like something that as soon as I get out of school, I could walk there. And I think what you learn from earning your first paycheck that, you know, it's possible. You kind of decide what your work ethic is going to be. And if you work at a place like fast food, you know, you got angry customers, they want food for the cheapest price, and they want it fast. And even in those jobs, though, you can develop a work ethic that's surrounded around, you know, you can be proud of that job. You know, a lot of times I think society looks down on people who work those jobs. But I think I really was proud of being able to earn my own paycheck.

That means I could buy my own school clothes, right, and open my first bank account. I even used that job to buy my first car. You know, back then, this is like 98, 97, I bought my first car, 86 Jeep Cherokee for $1,200 for my McDonald's job. So I think that was my first taste of independence, first ability to earn money. And so that kind of set the tone for like, look, it probably only gets better from here. Right. And so you go through fast food and I've worked at lots of fast food places, but eventually you're going to graduate high school and you're asking yourself,

what do you want to be? And back then I was living in Atlanta, Georgia and Georgia. If you have a B average, you can pretty much go to any public school. That includes Georgia Tech, Georgia State, UGA. These are really good universities inside of Georgia that you can go to. I opted not to go to college. And that kind of creates a scenario of where like, hey, lots of great jobs require college, especially in 1999 when I was graduating. And I decided to go to the A plus certification route. Right. I would just go in bookstore and found a book that was like, hey, you can get an IT certificate.

And a lot of the roles and jobs that I saw that were posted, that's all they really require. If you have A-plus certification, you can get started. And they were paying a lot more than fast food, so I figured that's where I would actually start. And I was a bit too afraid to get my own job at that point because if you look at all the enterprise jobs, they all wanted like five years of experience and all of these things that no one in their right mind could afford, like mainframe, you know, Java.

Like this wasn't a thing that you can easily, you know, do on your own. And so I decided to just open a computer store. So I did some contract work for Bell South. This is during the transition from dial-up internet to high-speed internet. And a lot of the people who used to install phone lines, they didn't necessarily have expertise on dealing with computers, installing network cards, drivers, all that stuff was just kind of out of scope. So they used contractors.

Did that for about a year and a half. And then I decided to open a computer store, number one, to run the kind of like contracting business where you come in, you run Cat5 cables, you set up network printers, that kind of stuff. And then a local computer store so people can come in and like get viruses removed, get a 56K modem installed because there was a lightning storm the day before. And so that's kind of how my journey starts early, early in my tech career. By the way, why not college? You mentioned you decided not to go for college. Well, I mean, look, I got there and then I remember the first set of courses, you know, I went to a small community college and I was like, are we really learning Word and Excel? Is that really the intro into computer stuff is Word, Excel? Like this can't

be it. No way. And I kind of felt like I was going to spend four years and not really have skills. And I decided that I just wanted to do an accelerated path of like, look, a lot of people doing this work, they seem to have skills. And let me try this because worst case, I could just go back to college if I'm wrong. But what I learned once I got that A plus certification is just everything opened up. It was just fast. Like, hey, you have the skills. Come on, learn.

And it turns out I was a fast learner. I was super curious. And it only took me a couple of weeks to get the hang of what was going on. And the idea of self-study really resonated with me. So, you know, when you're learning about Windows drivers and why they crash, you know, you can go research like, you know, we didn't really have like the internet we have today. There's no Stack Overflow back then. So you're literally reading through the Windows documentation, the manuals, and you're really learning from

experience like, yo, every time I put in this network card and you start installing the driver halfway, if it freezes, you're basically reformatting the computer. Like, there was no real easy way to recover from some of those errors. And so you kind of understood, like, hey, close everything out before you start installing this driver. Make sure there's no competing processes this thing needs to complete. You restarted at this point, et cetera. So that field knowledge became super important. So anyone that actually had hands-on experience, you really rose to the top really

quickly because there was just no substitute, right? There's no like, oh, just Google it and find out in real time. No, no, no. You had to have the experience. Makes sense. So this idea of self-study, at least for me, I kind of learned that when I was an undergrad, which was probably a little too late to learn that. Or I like to say in school, what I learned was learning how to learn was undergrad. Learning what to learn was my grad school. I'm curious how you got to that realization of self-study being a thing where you can pick something up, have the confidence to try it out and also go out in the field and do real work

and not just something on the side, which is a toy project. I mean, honestly, I kind of felt that was the only option, to be honest. You know what I mean? I didn't know anyone in tech that looked like me at all. I didn't have like a community of people I can just say, oh, if they did it, I can do it.

That wasn't the mindset. You look at those jobs and you say, look, four they did it, I can do it. That wasn't the mindset. You look at those jobs and you say, look, four-year college degree preferred, mainframe experience. I don't have any of those qualifications. So I wasn't quite brave enough just to apply anyway. And so I kind of looked at that and says, you know what? I don't see myself getting that job anytime soon. So I'm going to just go with what I thought was the path of least resistance. You go to the bookstore, they don't ask any questions, you buy the book. You learn everything in the book and you pay your $100 and you take the exam. If you pass the exam,

now you have the certificate in your hand. And then it was easy for me to see the jobs that said, if you have A plus certificate, these are the things you can apply for. And so I had more confidence going that route. Now, once I got into it, and I'm seeing all the people that had like three and four years of experience, I'm like, oh, man, how I'm going to keep up with these folks. And I realized that after like the first two or three on-site service calls, you learn like, okay, I don't know what I'm doing, but I want to pay attention. And the first time you go out by yourself, right, you're sitting there and it's like, okay, I don't know what I'm doing, but I want to pay attention. And the first time you go out by yourself, right, you're sitting there and it's like, yo, I don't think this DSL line is ready because I'm getting no signal. All my tools are telling me the thing isn't working.

I tried everything. And so you call the support line like, yeah, we haven't turned up, you know, the D slam in the street to make that work. Give us a moment. And then all my tools are turning green now. I'm like, okay, I do know how to troubleshoot this thing. And then that first install you get, your confidence level goes to 10. I ran into my first challenge.

I was able to be creative and debug. And it turns out you don't know how creative you are until the next person who rides with you just to check your work is like, oh, you did a really good job terminating that Cat5 cable. We never thought about running it through the wall that way, like your attention to detail, because I was so proud. Like you get your first tool bag, you have all these advanced tools. And I was just like, this is better than fast food. No way am I going to mess this up. And so at that point, people started saying, hey, you're really good at this. And so when people started calling in, they're like, hey,

Kelsey, can you take my jobs for the day? Because I know it's going to get done right. And so now I'm starting to say, oh, yeah, I'm pretty good at this. And so then when I kind of started the business, and I remember there were customers that say, hey, could you hook up this two offices together? And that was the first time I seen like Windows remote services where the app is in one office but you can access it from another office and they're like can you do that I was like yeah the truth is I've never done it before and so you're spending like three days like try this restart this calling the people on the other end all right restart this now and it doesn't work and the business owner's like hey are you sure you know what you're doing

i was like look i won't charge you for all the hours just give me a moment and then once that worked i was like okay the next time i ever need to do anything with like windows domains and multiple sites i definitely can say yes now so that became the habit like just say yes and given enough time if it's supposed to work it eventually will learning on job that was a pretty big jump though right from like starting your business like working from for someone to actually start your own business like in terms of right instead of just having problems where you kind of know like what are the boundaries now you have like all sorts of unknowns right like how do you get financing like you know how do you deal with employees

and things like that how did you that's not how that works when you're that young so when we think about businesses today this idea of getting a loan for 10 million dollars started that's not how that works in the real world that that's venture capitalist that's a new thing that thing is only like 15 years old that's not new all. The way it worked back then was you're doing these service calls and you're only there to hook up the internet and you must leave. That's it. That's the only thing you've been contracted to do. So when that customer says, hey, we have three computers, we would like them all to get online. That's a whole different game, right? And so I was like, you know what? I need to learn how to do that kind of work. So I can't do it during the contract, but I can definitely do it as my own

business. And I decided that, look, there's a market opportunity. So at that point, it's just me going to Office Depot, buying the switch or router, and then spending like $50 to buy it, charging the customer $100 plus your time. And so you're really just kind of being always, you can't afford not to be profitable. Like everything must be a profit at that point. So getting the computer store, there was a decision to make. I can either rent an apartment or sleep on people's couch and pay $1,000 a month for this little office space on a little strip mall. So you can hang your little sign at the top. It wasn't like, I'm going to get a loan. No, it's

like you're 19 years old and you go and call the number that said space for lease. They look at you and it's like, look, do you have the thousand bucks a month? I'm like, I'm pretty sure I can get it. And you sign your name and they'll give you a one month or one year lease and you pay a thousand bucks a month and you have this space and you just got to do the math. Can I make more money than it costs to house this stuff? And the answer was yes for a very long time. So you're really going check to check. Right. This isn't like a big business that's going to go IPO.

You're literally like, hey, I just need a place to do my thing. And then one of my colleagues that was working with me, he decided that, hey, he would help with the business. So he would also go and do some of those service calls. So now we can have two people out there generating profitable income, doing this kind of work. And then the computer store just became a place where, you know, if you were in the neighborhood or you looked in the phone book and you needed, you know, quick fix, I was there. You're a kid and you want a gaming PC, like CompUSA and Dell were not as accessible to folks. And so they would just come in and say, hey, I want this network card. I want this GPU. And I would go to the north side of town, buy it for less than I was selling it, charge a small premium on top. And then that was the profit.

And so it was just more like, hey, as long as I can pay all the bills every month and make more than it costs, we can keep doing this. Wait, so I am realizing that my idea of entrepreneurship is very warped by just sort of starting in Bay Area where like that is the norm exactly to your point. But right there is still the situation of like why, you know, take the risk of having not being able to make more than like you spend. And then why not just like stay at the company and then you kind of, you know, just work up the ladder. Well, at some point you realize though that you can make a lot more money because if someone's paying you $100 to go and install high-speed internet access, right? What are they charging the customer? There's a gap there, right? And so what I realized, I remember the first job I quoted the customer, $15,000, right? It was a huge warehouse that was sorting mail. And I was probably two years into kind of doing this kind of stuff.

And I looked at it, I was like, look, cable costs this much per foot, Ethernet cable. And, you know, clips cost this much. Faceplates cost this much. I was using computers to build routers. So I was just building routers out of small, cheap PCs. I put BSD on there. You put two network cards. And you had, like, a really high-performance router than the ones you would buy from Cisco.

You would buy neck-ear switches. So all up, parts and labor, I'd probably spend $5,000, $6,000 on everything. And then you would make $8,000 in two weeks. So for my entrepreneur, you're like, do you know how many $100 installs I would have to do to get eight thousand dollars? So even though you weren't going to be a multimillionaire, there was an easier way to sometimes get these jobs that would give you a ten thousand dollar profit for like two weeks of work. So it was one of these things where it was a compliment to like what you were doing. But you found out real quick is like, yo, I can make way more cash per install if I knew how to

do network printers in the whole nine so you're basically just in this hustler's mentality like yes first step was 50 bucks an install then you get 150 dollars to install but mathematically you can only drive to so many offices in a given day like you were breaking the record if you did eight. But I could go do one job for $15,000? You got to throw that into the mix. So it just became something you added to the thing. I was still not in the mindset that I could ever go get one of those other jobs, though. You got to remember, even though I was an entrepreneur, I've had all of this experience,

I still looked at those job postings as too far out of reach. And also, I was having fun. You know, like the computer store, after it closed at 6 p.m., it turned into a music studio. I ended up doing a lot of digital studio conversions where people were using the kind of, you know, analog gear where there's like mixing boards. And then Pro Tools came out. And you can actually start to remove a lot of that stuff and just do it all in the computer. And so I decided in the computer store, I had one whole corner dedicated to like a Pro Tools setup, only had one. And so musicians would come in, I won't name drop them, but a lot of them were pretty famous musicians. They would

come into the studio and say, wow, look at that, two screens, you have this Mac, you have this Pro Tools setup. And the bathroom in the store, I remember cutting a window into it and putting a microphone in there because it had some acoustic value that people looked at. It's like, wow, look at this little studio in the store. And so they would come in, it's like, we want the whole thing. And I would sell them that one, right? And just buy another one later. So I would dismantle it, go on site and sell it. And so I just kind of felt like I could do whatever I want. I was like the computer guy, someone would say, Hey, I know a computer guy. And I was the computer guy. That is incredible. This is what you're describing is having this hustle mindset,

you're looking at inefficiencies in the system or rather how some of the inefficiencies are being leveraged by the business and how you can do it yourself we see today there are a lot of influencers on social media who are trying to just talk about a lot of this stuff and back then there was no social media were there any influencers or not influencers is not the right word here were there any people around you who you would look up to that resulted into you thinking like this or books that you read like what resulted in you having this mindset um you gotta remember at that time people were just starting to stand in line for Windows 98. People had talked about Bill Gates as this person who dropped out of college, started a software business.

I think it was very clear that the Internet had legs at that time. Right. It was like things were finally working. You know, AOL was kind of still the top of the castle. A lot of people were just using AOL for everything. You know, you're playing games on Windows. You know, everyone now had a 56K modem, right? You were just, everyone had dial-up. People were starting to get viruses, right?

And then they needed the viruses to be removed, right? So there was already this huge ecosystem. And when you would read the news or you would look in the papers, it was always these, like, whiz kids who dropped out of college to get into the tech game. And then we were already at the time where the system administrator was the gray beard in the basement, you know, that has been hacking since, you know, the late 80s. And now the person that you probably wouldn't hire before, you're now hiring now. And so even then you started to see the job requirements starting to relax because they knew, hey, open source is coming. If you're a business that was only buying IBM stuff, you were starting to get to the point now where Linux is becoming a thing you could use. I remember buying my first Linux magazine where there was a CD inside where it was like, hey,

you know, here's the distro for the month. And it came with all these free utilities. Like, I don't know if open office was quite a thing yet, but you started to see that there was going to be open source challengers to the things that were very expensive. And so I remember getting my Linux Plus certification by just downloading one of these things, playing with it really deeply. And it's like, yo, and that's when I started using it for routers. Like, yo, this open source thing is a great opportunity. So I kind of was just paying attention because I figured if I'm not going to college, then it's my responsibility to pay attention and learn. I have to be kind of like my own college. And in many ways, I think at that time, Linux Magazine was well ahead of the universities, right? Because they were like,

these people are building this now. You buy the magazine and you're reading about in real time what's going on versus people learning about Microsoft Excel in school. You know what I mean? So I don't know. It just felt like a really good time to be doing it. And I guess when you're busy, you're not really worried about anything because you're just too busy.

How can I fail when I'm this busy? So for many people today who might be in college or maybe choose not to go to college, what advice would you have for them in terms of things to do and also things that they should probably avoid? Because that's the second part of something I don't know if many people think about very openly. I mean, look, we're a product of our environment. So this whole college thing, why are even people care about college, right? We care because your teacher told you, you need to go to college. We care because certain professions require college. You cannot get certain professions without college. So it's a key to many opportunities for a lot of people for a long time. And I think there's been this idea that if you don't go to college,

like we've seen the studies, people who go to college make this much in their lifetime. People who don't go to college make this much. So if you keep pushing this on people, what do you expect to happen? You're going to feel like, if I don't go to college, I've already lost in life. So all throughout your life, you're being told what to do. You take this test, you get this grade. And then someone says, hey, you're going to be a failure if you don't go to college. And I think the bad part about this was we didn't figure out how to make college easy and accessible to everybody. We just started giving people a ton of debt to do it. Even colleges that are not even great.

They're like, hey, you're going to get a bunch of loans. The federal government is giving a bunch of grants. So we'll just raise the tuition And everyone will walk away with $80,000 in debt But did they learn anything? And so when you don't have any requirement On the actual quality and the outcome And people just pay

Then you never know what you're going to get And how does an 18-year-old post it know If your college program is any good or not? We didn't say what colleges were great. We didn't say what the bar should be to make sure you were getting your money's worth because most people have never paid for school before, right? K through 12 is typically, mentally, it's free. And so now you go to this place, you're like,

eh, it's no different than what I was doing before. I guess this must be right. And then you get the bill. But I think there's a lot of professions. If you look at like electricians, for example, a lot of them have a really good apprentice program. They have a way to go from a journeyman to someone that's really skilled at what they do.

And to me, I think a lot of tech is more in that. It's more of a trade than it is about computer science. I think we need to be honest with ourselves in society. Computer science is great, right? The idea of computation as a study, it's great. It's probably more math than it is about what we do as a trade, like writing code. Yes, you could benefit from the theory, but let's stop. Come on, people. You import the library and you call dot sort, right? Whether that's logarithmic import the library and you call dot sort. Yes. Right? Whether that's logarithmic time or whatever, you know dot sort is going to work.

Come on. And I think it's time. And we're starting to see a lot of kids say, hey, if I want to get into tech, at some point, you got to have the skills. So if you want to go to college, I think that's fine. Make sure you become a well-rounded student. But just know it's still going to be your responsibility to learn the actual skill. And even when you graduate, it's still going to be your responsibility to keep up with the skills required to still be someone that people want to work with or choose to hire. And I think that is the new realization that

people are coming at, right? You meet someone and say, hey, I have 20 years of experience. And it's like, hey, man, I don't think you've learned very much after you graduated college. You're still here writing the same bash scripts that you were writing 20 years ago. You're still stuck in this mindset that everything is what you've learned 20 years ago. So that person actually has 20 years of one-year experience. They forgot that it's a continuous education cycle. And so whether you start by self-learning, you're going to do it at some point. So don't think college means that you don't have to do it. So I think hopefully people understand. Go to college if it makes sense for you.

I will still caution people, do not believe that you have to walk away with $100,000 of debt. That is insane. That should be challenged. People should look at it. And I think there's this idea that I can only get a good computer science education from a certain set of schools. That's insane. The quality of education you're getting even at community colleges is fantastic, right? And then you can supplement it, right? Maybe you go to a school you can afford, but then you start getting to the open source community. You start writing code and publishing things on GitHub. And I promise you, when you graduate, and I've been a hiring manager, one person went to, I don't know, Stanford, Harvard, whatever, and they say, hey, I have a CS degree, and then you graduated from a local community college where no one knows its name,

but then you tell me that you were contributing to Terraform and Kubernetes and your code is on GitHub, and you tell me exactly what my team could do to actually leverage these technologies, you're getting hired. You think I'm saying, ah, man, that's great, man, I know you're ready to go, but your community college is no match for Stanford, so you're not hired. That's not going to happen. I promise you that if it's a legit team, a legit hiring manager that's focused on success, your skills will be given the proper credit they deserve. It's just on you to go acquire those skills and be able to articulate and show that you actually have those skills.

So I would advise people to do both. So this happened to you as well at some point, because while you were running this business in Atlanta, I think, or near Atlanta, at one point, I think you got hired at Puppet through, I think... No, no, my first job is at a Google data center, maybe 2003 and 2004. You got to remember, I had a decent amount of money from the business. And I decided that, okay, I don't want to run a business anymore. There's a lot of work that goes into that.

You got to do payroll. I have a couple people now. And there's a lot that goes into it. And I was about to have my first child. I was like, you know what? I'm going to go and just try my first job. And I was about to have my first child. I was like, you know what? I'm going to go and just try my first job. And it was funny. It was like, hey, there's this Google data center and they're hiring for a system administrator. I don't even know what that is, but I'm going to go to the

interview. And I remember being in that interview because I felt confident now that I could pass these type of interviews. I think I have enough skillset. And I remember the job required a lot of Linux and they were asking me these Linux specific questions. I was like, look, I can't, I don't know the Linux answer, but I do know the BSD answer. And there were three kind of more senior system administrators at Google. And they were like, okay, we can swap these questions out to be BSD centric. And it was so funny. I got so lucky. One of the people doing the interview, he had a free BSD logo tattooed on his leg. So he is just like the biggest BSD fan.

So when I told him that all of my Unix knowledge is from free BSD, he was like so happy. Like, oh, you're one of us. And he was like, so how do you do this? And I was like, make world. He was like like so how do you do this i was like make world he was like so how do you do this piece i was like you got to do this with this and then you put this over here and he's like wouldn't have you use it in real life i was like a lot of my network installs i use it as a router you put it in this configuration then you can get xyz he was like okay hi and you hide it on the spot and so I had that job and I realized that this wasn't enough for me.

Someone that was an entrepreneur, it became a very easy job. But it was nice working at Google Data Center, right? Because you have thousands of servers. And people are swapping out RAM and CPU just because a new version came out that had a higher clock cycle. Swap everything.

And then I learned about pixie booting, the burning process, machine databases, how to run Cat5 and fiber optics the right way, ladder racks, cooling, cold side, hot side, power audience. I mean, I just learned everything about what a world-class data center is. Also learned how to debug and troubleshoot at scale. The scripts we used to write to provision top of rack switches.

I mean, it was a really great job, but my mindset was, I'm just collecting skills. So now three months is here. It's time to move on because I look in the thing, I can go from like 30 something thousand to 45,000. So I go to the next small company and they're doing the voiceover IP. And so they're taking Asterix open source and turn it into something that you can power Cisco IP phones from. And so if you were to pay Bell South and Cisco for a PBX, it's expensive. But voice over IP was coming very popular, like calling cards, office that needed 12 lines, but you're kind of paying the price of like three. And so I learned everything about that. And so once I got those skills, I'm like, there's another job that pays 65,000.

So I'm swapping. And so my career, I just bounced around learning everything quickly. And then I will look and say, oh, I can finally check boxes on these other things. And then I got really good at interview. When I go to interview, they say, hey, what would you do for this? And I say, you can run this command, but I'll tell you how it works. And the interviewers are just looking at why are we asking this guy more questions? Let's just hire him now. When can you start? I can start tomorrow. And you go to the other job and say, hey, they pay this much. What are you doing? Not, okay, I got to go. Because there was no loyalty. I can even realize back then there was no loyalty in the job market. No one was working 20, 30 years at the same place anymore. And so

I just started getting money. And then the first time I stayed at a place like three years was, I went to a financial institution. And I remember, I think I was making like, I don't know, 50,000 or something at a job. I was in web hosting. I used to work financial institution. And I remember, I think I was making like, I don't know, $50,000 or something at a job. I was in web hosting. I used to work for a company called Pier One. And they were hosting YouTube before they got bought by Google. So their specialty was like low latency game servers. So if you wanted to have your own game server,

you go online, you fill out this form, and we would provision that server for you automatically and give you IP, and then you could do whatever you want with it. And they were doing that for YouTube as well. And so web hosting, I think, is where I gained the most skills the fastest. Because when I started there, before starting to write code, I was doing tech support. And so people would call and say, hey, I have this web server, and my email isn't working anymore. And they have like 100 domains on there. And you have to log in and try to, you know, is DNS set up right? Is spam assassin doing the thing right?

My SQL DB is corrupted. You got to restore it. And then you get like 15 minutes to troubleshoot a server that you didn't provision. And so I learned that, man, I'm fast. I can be really fast. And I started writing code for that provisioning system. And I remember going to this financial institution. These are the jobs that I was the most afraid of.

These are the jobs that needed all those years of experience. I'm thinking to myself, like, these are the experts. These are the people. This is the reason why you have to go to college. If you want to be on their level, you need all of this stuff. And so I'm very intimidated. Like, look, I went the kind of grassroots, underground, hardcore, you got to be fast.

They have no time for delay. There is no excuse that you can't do a thing. You better learn it because you got to go. So I get there, and I realize how slow the enterprise is. Oh, my God. This thing was, I mean, the questions was too easy. I was like, is this a trick or something? This is not even.

I saw this coming. I saw this coming. This is what? Five years of experience. This is the one? I was like, no way. And so I'm just showing off in the interview. I said, listen, you can definitely do that.

But here's the problem. If your MTU setting is this, you can never get this much throughput. Let me whiteboard for you how this works, right? Like this is how many bytes you're moving. You got to do this thing. If MySQL ends up doing that, you to make sure you the the eye notes on this and like you got to compact that or it's going to get slow right the the kernel has a ring but oh yeah i don't know what y'all do is when you hit this problem you call the vendor you call

someone like me to fix it and then you just close the ticket and you're the professional. Now I understand. I'm not saying everyone in enterprise is like that, but the people I met were like that. And I was like, whoa. So I was like, look, I think I did a good job. So remember, I went to this interview on my lunch break. And that's the kind of place where you have to wear a shirt and tie. So I bought my first shirt and tie to make sure I was looking appropriate for the interview. And so on my lunch break, my manager at the time, Joe Rodriguez, we're still friends today. And I had just got into the engineering side at Pier 1. And I remember just going to this interview, getting my suit on during lunch. And it's like, I don't know, 40 minute drive north. I do the interview. I'm talking to my wife on the way back. It's like, I don't know, 40 minute drive north. I do the interview.

I'm talking to my wife on the way back. I was like, babe, I think I did a really good job. I think my skills are really at that level now. And I remember the recruiter called, right? Because you have a recruiter. They call, they say, hey, they want to offer you a job. And I was like, well, how much does it pay? And they were like 90 something thousand.

I almost had to pull over, right? You go from making 50 to 90? Your life is different. Your life is very different from 50 to 90, especially in 2007, 2008. And so now I'm like, I'm ready. And when I got to that job, I didn't understand corporate politics. I didn't understand corporate politics. I didn't understand the hierarchy.

I didn't understand that some people, their boss, not because they're the best, is because everybody else has quit. And they're the last person standing with the most time. And so that took a long time for me to realize how to navigate enterprise. It's not even about having great technical skills. You have to make people comfortable. Some people are playing defense and you don't even know it. And so it took me about a good six to eight months to really understand how to deal with that. And that's when I understood

that my personality had to come out. The thing where you make people feel comfortable, the thing where you make someone smile. I remember when I brought in some automation tools like Puppet, my team at the time was so scared because he was like, we used to run these scripts and if they didn't work, they just brought things down. And so they would type out these scripts and ask people to review what they've typed. Like, hey, before I hit enter, make sure I have everything right. I'm like, dude dude you're so paranoid and i remember when i started bringing in a puppet the the chance of success was so high now like no

more of this random stuff okay and it was his turn to run it and he probably waited three or four months to get in on the action and so we finally taught him how to like use the new stuff and he was so afraid and i remember standing over his keyboard, I just hit enter for him. Like, stop being so scary. I hit enter. And it just works. And he's sitting there. It's like, why are you so confident? I was like, because, you know, we know how it worked before. We know why it's broken. But we've also been running this so much. Jira now drives the automation. I don't even know why you're typing this manually. You could have done this so many other ways. But I made him feel confident though. I made him feel like this new way could allow him to heal from the trauma of the old way. And when you start to build confidence in the enterprise, people start to trust you. You become the go-to guy when there's

an outage, people call you first. But I had to learn how to make sure that everyone felt like if they got called, they could do what I do. And so I started focusing on writing more documentation, having those engineering lunch and learns where you show everyone how the system works. You let them try it and ask questions. Never make anyone feel stupid. Never make someone feel like they're less than you. And then you try to make sure that it felt like a team win. And I think that's when I finally became like, not just an engineer anymore,

but like a technologist that can actually show people how they fit into the future. And it's also right around the first time that I gave a community talk, right? I had brought in Python because I wanted to learn Python because at this company, everything was like mainframe and written in COBOL. community talk, right? I had brought in Python because I wanted to learn Python because at this company, everything was like mainframe and written in COBOL. And so the way it worked in financial services probably still does today in 2023. You get these large files from like,

you know, people who process credit card transactions. And they're in like these fixed length fields. So you can imagine like the first 16 bytes is the customer name. Next 16 bytes might be the name of the thing they bought, etc. And you have to walk these files. And there's another interesting thing about these files is that instead of representing numbers in bytes, like a whole byte, it would be a nibble, so half a byte. So you didn't want to waste space because, you know, network links aren't as fast. And also they were destined to go to the mainframe, which was optimized for processing these things quickly. So instead of wasting a whole byte of memory, you can do half a byte, a nibble, and you can represent these numbers. So you can get a lot more data in a file and it's less to unpack and the mainframe can move real fast. And so I was like, oh, this is why everyone's sticking to COBOL, because COBOL had the libraries that let you create nibbles and read them back. So I was like, all right, I got to learn Python. But there were no libraries with things like EPSIDIC and PACDECIMAL. And so I was like, wow, I got to learn this stuff. So I just started learning Python, started speaking at meetups just so I can build a community. And I remember meeting maybe someone

from like the NumPy community. And they're like, look, it's just an array of bytes. And so what you have to learn is like, what are the byte layout? And you just have to do the same thing. So I remember then looking up like Wikipedia and seeing what a nibble is and then the basic algorithm to take a number and compress it to represent it in half a byte. And I remember making this Python library that did that. And I was able now to just use Python to process these jobs and format them in the way that the mainframe can process. And when people saw that, they were just like, oh, we never thought that that was a possible thing.

You don't have to replace the mainframe. We can use Python. We just need to understand. And that's the moment I understood the fundamentals. Like COBOL is good at a thing, but really just look at the bytes. If you just look at how the bytes are laid out, any language can arrange the bytes in any way you want. And so I'm like, wow, this is what happens when you understand the low level. You understand the low level, you can take any tool and do whatever you want. And my guess is going back to the college thing, maybe that's the thing you learned in college, like, hey, what bytes are, byte arrays the details now. Why does this stuff work this way?

And then I became the person who could actually start contributing to open source. So I started, I contributed to Puppet, I contributed to Cobbler for like provisioning stuff. And after a while of contributing to Puppet, I remember my manager was like, hey, you know, he wasn't really cool with all of this ops people doing dev work. And so there was a bit of friction there. And again, that's me learning how corporate politics works. You can be punished for going above and beyond. And I got it at the time because if you're an ops and you're delivering things that devs should be doing, you start setting the wrong expectation for all the other ops people. People start thinking they should be writing code. They should be willing to learn on their

own time. So I got it. I was putting a lot of pressure on my teammates because selfishly, I wanted to move far and beyond. And I remember he came back from a conference like, Kelsey, I know what you're doing now. I was like, what am I doing? And he's like, you're doing DevOps. And I was like, I don't know, man. I don't think you get to name what I'm doing. But I think he kind of seen the light. And when he saw the light, I remember he took a real interest into what we were doing with Puppet and why we were writing code to do things like adjust the logs to make sure they fit how we were doing logging on the back end. And I remember he invited someone from Puppet Labs to come check out our setup, you know, so that we could learn how to do it better. His name was

James Turnball. He wrote like the book on Puppet. And I remember my manager was so hyped up. He was like, yo, James Turnball is coming to our office, right? He was in Atlanta meeting customers. So he decided to come meet us. And I think my boss, we got in the elevator and he was introducing me. He's like, Kelsey, this is James Turnball. And James was like, oh, we know who Kelsey is. Like he's the guy contributing to our project, open source. My manager looked like, what the hell? He knows you already? I was like, yeah, you know, I'd be contributing to open source on my own time. And I remember we went upstairs and we're in this like room. And the first step was like, show James how we're using Puppet.

I was like, all right, the way we use Puppet here is I've turned everything to RPMs to make the Puppet code easy. File, package, service. We don't want a bunch of noise. And given that we have people in QA, business system analysts, like've got to make it easy for them to use this. So we have a Jira front end. If you go to Jira in the dropdown, you can pick the RPMs that have been built that are ready to go, and you hit save. And then you pick the environment you want the RPM to be in. And then what we would do behind the scenes is we take that

and we call Puppet under the covers. And Puppet, as it's running, doing the deploy, doing the setup, doing the configuration, we take the output and we put into the ticket. So it looks like someone's working on it, making progress. And when it's done, we move the ticket to verification.

And then the person who asked for the deploy could be a PM or QA. And then they hit approve and it's done. And so no one ever runs Puppet around here. And he saw this. He like oh my god like i i have no advice for you like this is great i think my manager finally realized that what we were doing was like above and beyond and i went to go give a talk at puppet comp the very first one in portland oregon and i remember they gave me a job offer and I went to go give a talk at Puppet Conf, the very first one in Portland, Oregon. And I remember they gave me a job offer and I went back to work

and I told my manager I'm quitting. And I remember he was like, not surprised. We're surprised you stuck around this long. I was there for like three years, right? We knew that you had kind of gone above and beyond what we could provide for you. And he was like, congratulations. And I moved on.

So that was kind of like the mid career. That's when I get to Puppet Labs. So there's a lot of grinding, working in the trenches, on call, you know, fast companies, slow companies. And so when people hear me talk about these technologies, when people look at the contributions I make to these technologies, they're rooted in their practitioner's experience. You know, it's not so much as theory, it's more like, I know the job because I probably had it. I installed and managed VMware. I installed and managed the NetApp. I know how these things work and how these things break. And so, yeah, some people only know me from Puppet, but there is a whole storyline that starts before that. Coming to the enterprise from what you were doing before,

it was very different. You hadn't been in an enterprise world before. And like you said, in an enterprise world, there is more of a structure, also a hierarchy. So at times when you can move much faster, probably much wider than others, you stand out. and there is a skill in for the lack of a better word not pissing people off along the way where like you said your your teammates should still feel that you're part of the team your manager shouldn't be afraid of their job because you're doing so much better and you're probably doing things that they're not even thinking about how did you go about dealing with that like was there friction that you experienced at any point and decided to kind of pivot how you were working in the team i mean you gotta remember coming from being an entrepreneur this idea of having a boss

never resonated the same anyway it's like dude you're a manager i mean you know what i mean like you can't i can quit tomorrow and go somewhere else easily. So I was never feeling that pressure. I was confident in my skillset that I can just go get another job when necessary. But I also felt that it was on me to continue to learn and grow. And so I think I probably was in the position, remember, when you own your business, you're in the customer service business too. And I've had jobs that were tech support related. So I get that you need to listen, see why people are having this friction. And my mindset is I got to solve the problem.

I'm not trying to be right and have you be wrong. To me, that's kind of petty. That wasn't my outcome. I don't want your job. I just want to solve problem. And so if I hear two teams bickering about something, I listen to them like, okay, what are they arguing about? And I just get on the whiteboard. So let me understand this needs to happen. Okay. Let's just talk tomorrow. And so I got a whole day now

to just like prototype the fix. I heard this side, I heard this other side. And now before we get to the second meeting, it's like, hey, before we dive in, I think I heard everyone correctly. Can I show you something? And then you would show a thing and they'd be like, what's that? I was like, I'm just saying, is it right? Is this the thing?

Like, yeah, I mean, what is that? They're like, how much does it cost? Do you buy from IBM or something? No, no, no, it's like a thing we built. It just uses rsync to do this thing and guarantee that things go on the other side. And you can run it twice. So if you're not sure, just run it again. It's going to tell you there's nothing that needs to change.

So we don't have to trust each other. We can just use a tool like this and get trust but verify. And the team was like, okay. I said, so you guys want to keep talking about it? It's like, I mean, maybe we should just use that. I have docs. Here's how you install it. I will circle things in red to make sure it was really clear to see.

And so for the people that had less confidence, they can just step through the guide and have confidence. And I said, if there's any problems, I will totally support it. And if it's not right, we should just remove it. And I think I was showing people how to behave. Like instead of all of this arguing and trying to be right and throwing this guy under the bus, you can kind of just step up. And I didn't do a thing where I said, well, that's not my job. Hey, that looks like Dev.

That Dev needs to do that. I was like, I'm going to try to write this code. And I had a few people in Dev. I said, hey, am I doing this right? It's like, yeah, but, you know, do this instead. And I would fix it. So I was always ready for that meeting. So I think I just chose to be that person. I was like the glue person. Like, hey, I found an opportunity. I think it's very clear what needs to happen. And what I realized is that I just started building trust.

And that's it. And then people start trusting you. So they know when you walk in a room, it's like, you know what? Kelsey's here. And I remember trying to make people laugh. I was like, we lost the file? Like, it's gone. This is people's money in the file, so it's gone. Someone's going to wake up in the morning and be like, my money's gone.

And people start laughing. I said, so we got to find the file, right? They're like, yeah, we got to find the file. I said, what happens if we don't find the file? I said, you know, Chris doesn't work here anymore. I think to find the file, right? They're like, yeah, we got to find the file. I said, what happens if we don't have find the file? I said, you know, Chris doesn't work here anymore. I think he lost the file too. And people would laugh. And I was like, so we got to find the file.

Because I like working with you all. You're actually pretty cool. And so you just show them the find command and some regular expression and you find the file. I said, I found the file. Hey, you should bookmark that command. You can use it later. So you don't have to feel like this again. And they'd be like, yo, thank you, man, because after today, we would have to escalate that we can't find the file. I said, I know, but just tell them that you know how to find files now. And then you leave the credit to them. And so now they email their boss, we found

the file. But if they say Kelsey found the file, then they can still be looking like they're not competent. It says we found the file. So run the command again to make sure you understand how to find the file, move the file into the right place. And if this ever happens again, now you understand how to use the find command. And it was like, hey, thank you so much. And then next time they know how to do it and they say, hey, Kelsey, thank you for making us not look silly right now. And I think some managers caught on that I was doing things like that. It's like, I swear to God, Kelsey was in there.

Now, all of a sudden you can find a file. I was like, yeah, why can't they do that? We were just brainstorming together. It rings true on the quotes I really liked. It's like people don't remember what you said to them. They don't remember what you did to them, but they remember how you made them feel.

And, you know. So you get to Puppet in this case. So the Puppet conference that you talked about. Now, Puppet Conf would have a lot of companies

giving talks. This was your first time there. This is probably where you would meet a lot of other technologists too who are not just thinking about here's how you run a command or solve a problem, but they're thinking about where technology can go and how it can enable them. Can you walk us through what did it look like for you to get that job offer?

Like, what was the conversation like? Well, you got to remember, before Puppet, I had already started contributing to open source. I was working on Python back then. So I used to contribute to VirtualInf and PyPy and PyUtils, like all the packaging tools. I was just contributing back then. And I remember going when PyCon came to Atlanta, I remember me and my buddy were just so excited.

That was my first real tech conference. And I'm looking at the names of the people that were there, like Guido van Rossum, all these people working on stuff. And I'm watching them give talks. And I was just speaking at the local meetups. But I started to get good at the local meetup. But they were on the big, big stage. And I'm just sitting in the audience.

It's like, man, this is. And I now knew the difference between people who had the theory and the people who had the skill. Because it was very clear. Because through work experience now, me learning things that, hey, fundamentals matter. And then now I start to learn the names behind the people who were working on the tools I was using. And so when I started to contribute to open source and started getting my stuff merged in, like I added support for configuration management to Cobbler. And so Michael DeHaan ran that project. And I realized how close the community was. Go to Cobbler, you see Michael DeHaan, he had his email. I want to add this feature. I get it working locally. We're using it at that company. And I email him like, hey, would you take this? He's like, yeah,

I'll review it. And then it gets merged in. Virtual Nth and PyPy and all these tools. I remember working on the mailing list and we were just adding all these bugs. I was just fixing bugs, getting it merged, reviewing other people's code, getting it merged. And then we would cut a release. And then at work we would import the new version. And I knew that some of the features I had worked on. So I had kind of felt like this is how you make technology work. So when we went to PyCon for the first time, I met the people that I'd never met before because we were just working over the internet.

And so getting to see the people that, hey, okay, this is what you look like, right? Like we were just working in our free time because I just have to do this after work. The company wasn't really on board with contributing to open source. And so they had this after work. The company wasn't really on board with contributing to open source. And so they had this hack day. So at the end of PyCon, there's like five days where everyone sticks around, like the core contributors. And we were working on Python infrastructure.

I got access to like some of the core stuff in Python. And you're sitting next to all of these people you look up to. And they're just like, I'm like, oh, man, I'm not using this comprehension correctly. And this guy's like, oh, I did the list of comprehension implementation. And you're asking him, he's like, yo, I learned a little bit about Haskell. Did you get it from there? He's like, 100%. We got it from Haskell. We put it this.

It's slightly different. And so I'm sitting there like, OK, this is what it's like. So after that five days, because it started like Friday and it went over the weekend. So I was just there all weekend, Saturday, Sunday, you're just there all day trying to figure out what to work on. And I was like, you know what? This is the tech game. The job is one place, but outside the job, this is happening. So when I get to Puppet, I had already had contributions merged into Puppet. And so when I went there, I kind of understood, but now I was on the stage and I was talking about one of Puppet's most advanced features. And my talk was really good. People liked the talk. So I was like, wow, I can now talk on the big stage. And so when I got off the stage

and I was meeting a lot of the people that worked at Puppet, we knew each other from IRC already and so it kind of felt like I think the honorable thing was when I got the job offer, it's like, okay, not only can I use the tools, I felt like at that time that was the highest bar to be making the tools and I felt like, okay, I'm at the part now where I get to make the tools. And so that felt really, really good because I felt like that was the milestone. And I felt like that's beyond the path that tells us to go on. Get a good job. Don't make anyone upset. You might get a promotion too to you can just change it without permission so that felt like that trajectory it feels it feels like a really nice juxtaposition with the

earlier when you had this image of right like with the fancy job requirements and showing up being like super disappointed versus now like discovering like this is the real game and to close the loop on the pycon thing what was so dope, working at Puppet as an engineer, your shipping features, I remember when he added a package manager to the Puppet command line tool because we wanted to have official modules.

And I learned about this dependency management. And that was probably the first time that I felt like not going to college hurt me. Because when you think about walking trees and creating graphs of dependencies and the dependency diamond problem, that was like the most CS I have ever had to deal with, right? Because when you get into like recursion, when you actually need it, when you have to walk a graph and know when to stop, I didn't know anything about those

algorithms. So my brain was just exploding like, oh my God, like we got these dependencies, but I need to walk this tree in a way that doesn't take hours to tell a user that this will never resolve and then to exit. And so I remember one of the persons who did go to college, his name was Peter, and he was able to help me understand the graph theory, all of these algorithms. And so I would look at the algorithm and say, okay, now it makes sense. It's all about the data structure. And once you get the data structure right, then the computation becomes much easier. Because I was just like brute force and everything,

like bash scripts. And I was like, I got to pause. Is the data structure even correct? Like, oh, now I understand why people sometimes create their own linked list because you may have to build one yourself or you need to know that you need a list structure to go along with this map structure so you can walk the tree in a more efficient way, or that you need to create a graph here. And so I was like, okay, I got it. And so being next to

someone who was far above my level of expertise, but I didn't feel inferior because my creativity allowed me to even think of merging this thing into the tool, how it should feel when you used it. And then going full circle, I remember after I started giving lots of talks, I remember there was a time in Portland that PyCon came to town. And the first person who gave me my meetup talk, I went to a meetup in Georgia Tech and I saw him. His name was Brandon Rhodes. And I was like, yo, can I speak at the next one? He was like 100%. And he was so cool, so down to earth.

I mean, one of the most professional moderators I've ever seen at the time. He ran that meetup perfectly. It almost felt like a little bit conference. And then he became the chair of PyCon. And I remember he's like, Kelsey, can you come give the closing keynote to PyCon? And Kubernetes had just come out two years prior. And so I remember I was like, yo, I got to make this the best keynote ever. And it's the first time I started using the Google Assistant to drive my demo.

And I was talking to Kubernetes on stage. And people were like, is this real? I was like, oh, I spent so much time on that code. I wanted to feel like magic because it felt like my whole career had come full circle doing that keynote for PyCon. You mentioned while at the previous job at the enterprise, at the financial institution, you were contributing to open source on your own time because the company wasn't necessarily supporting it. Now, that requires working sometimes beyond 9 to 5, sometimes in the evenings, nights, and weekends. And also, after you got the job at Puppet,

you, I'm assuming, moved from Atlanta to Portland, I suppose? I was still working remote in Atlanta, but I did eventually move to Puppet, moved to Portland, just because it was a change of pace. You know, my wife was cool with it so we all moved to to the west coast just to be closer to the action in this case like working nights and weekends are deciding to make this move requires also a lot of family support and from what you've described you obviously have an amazing work ethic and seeing that balance like this is something my wife and I

often talk about is when you look at successful people. I mean, not joking, but this is an actual conversation amongst us is that when you look at a lot of successful people, you look at one side of them and you look at how successful they are. And obviously they've put in a lot of hard work behind it. And of course, they're also smart people, but you're not looking at the other aspects of their lives to see what they had to give up. And you've many times talked about balance in life, prioritizing family or just saying yes to opportunities. So while you were kind of doing all this, how did you put that balance in place or how did you think about that? I don't think I was thinking about a lot of balance. You know, I remember taking a job overnight when our daughter was born so I could be there during the day because we couldn't really afford, like, daycare like that.

And so you kind of do what's necessary. But also watching the income go up based on new skills, you're sitting there asking, like, I mean, how else am I going to close this gap? You know what I mean? Like, there's no other way. And so I looked at kind of like just being patient and just going and maybe work at the same place 20 years and, you know, maybe you'll get some small raises because the raises people were getting, like 5%, 6%.

I was like, I don't know if that's going to be the right pace for me. And looking at just how easy it was to acquire skill, right? From that A-plus certification to that part of my career, I was grinding. Now, a lot of times I was doing that work at home. So at least you're there, you're at home. You know, people can see you, but they see you putting in the work, but then they're seeing the results. Okay, now we can go buy a house. We can buy a nicer car. We're starting to feel a little bit more security. The savings is growing. There's this vision where you can see that, okay, we're going from surviving to thriving to saving to investing. And now you're starting to feel like, hey, I remember I haven't

used a resume in a very long time. So a lot of this work became the resume. So it's like you get to call the shots. I want to work at this company. And then you go work at that company. And so I think seeing all of that, that your family extended and internal family, you're seeing results. It's hard to not be supportive when you see the results and the results are actually being a part of your life. Everyone's kind of enjoying those success pieces. Now, it is hard. Like if you're on an airplane, you're traveling a lot, it's hard.

But the price of that is, the output is me and my wife are retired at 42. Not just me. We're both retired at 42. Not just me. We're both retired at 42. And so that's the cost of doing that. And honestly, just to be very transparent, I don't think people realize you can grind that hard. And I probably made more money in the last four years than the 21 years combined prior. So there's not this idea that it's just like,

you just keep growing and saving because you can get lucky. You can go and get involved in something where the equity goes up by 6x. Does that mean you worked harder that year than all the other years? It doesn't mean that at all. So I think a lot of times is I probably worked less in the final years than I worked in all the other years. So it's not really this balance thing. I think it's more of you got to do whatever it takes to get the skill set you need to get yourself in position to get lucky. There's no universal formula for success, like work really hard and success will come. No, that's not how it works. Work really hard, be prepared, and you might get lucky because a lot of times you have to be in position to take the opportunity that feels like, look, you could have took this job or this job. You have no idea which one is going to have that 6x multiplier on the equity that you're granted. That's luck. People think, oh, I went to the, no, you didn't. If you would have took the other offer, we wouldn't be having this conversation. So I think there is a

factor of luck, but I do think you can influence it because you can buy more than one lotto ticket. So I think that grind mode allows you to play multiple lotto tickets. I can be on the speaker circuit. I can invest in startups while working at a large company like Google who's seen their stock multiply multiple times. And so I think that's what goes into the equation. So again, I was in survival mode even after having some success. So year 10, year 15, I'm still that person that's like, hey, there's not a lot of distance between you and the person living on the street. And so I'm like, I got to save as much money as possible. The tech industry may not be paying these kind of numbers forever.

I got to get what I can, save what I can, because I'm still in that mindset. I don't have a college degree. I'm very fortunate to be in this scenario. What if it changes? So you have to unlearn that over time. I have to unlearn the whole keep getting as much as possible. And so it takes a lot to walk away and leave money on the table because when you get really good,

someone somewhere is going to always be offering you money for your skill and your time. And you have to get to a point where you say, no, no more. I have more than I need and I'm out. And so this is an extreme pivot now. Some people will look at this and say, based on your trajectory, you could have been grinding for at least another 10-plus years, but I decided to make a really stop, just cut it off.

It's a hard stop. And so now maybe things will even up over time. So everyone's different. You've got to do what's right for you. I did what was right for me. And luckily for me, I get to cash in on it. Thanks so much for tuning into the show. In the next episode, we'll continue our chat with Kelsey about his retirement, the financial planning that enabled him to retire at 42, how he got started advising startups, and

his perspectives on compensation, turning down a substantial offer from Microsoft, and meeting Satya Nadella in person. And of course, we'll hear about his plans for the future. See you in the next one!