Subscribe
Share
Share
Embed
Come join us as we discuss everything open source with guests that are pillars in the industry. Welcome to The Secret Sauce.
Welcome back to the show. We're about to learn the secret sauce.
Kelsey Hightower:I'm happy to be here. You're you're in my neck of woods.
Bdougie:Yes. Yes. We we made the flight up to to hang out with you and a few other folks up here in Portland. We've actually crossed paths a few times. I don't know if you know this, but, I I MC ed a remote version of GitHub Universe, and we cross paths in the green room.
Bdougie:But also I've, like, literally walked down the street at KubeCons and like bumped into you a few times and said hello.
Kelsey Hightower:Did I say hi?
Bdougie:Yes. Yeah. We talked
Kelsey Hightower:for a bit. That was nice.
Bdougie:Yeah. And I I don't think you knew who I was. I don't think I was a big deal back then. I we're working on being a big deal. But you're a big deal.
Bdougie:We're here to talk about the secret sauce, Kubernetes. You have a a eclectic background. It's like a you've done a lot of stuff. So can you explain the the audience who you are and what you do?
Kelsey Hightower:Yeah. I think at the end of the day, you do it long enough, you just become a technologist like most. I started in tech support, system administration, full time software engineer, VP of engineering, product manager. I've done some DevRel. So eventually though, you just understand that if there's a problem to solve, you put on whatever hat makes the most sense and you go execute.
Kelsey Hightower:And then I think a lot of people know me recently, from the last part of my career right before I retire from Big Tech, my work around Kubernetes. Yeah.
Bdougie:And that's honestly where I crossed paths with the original KubeCon. I was working doing DevRel at a company called Netlify, and, I kept getting videos of you speaking because I was doing a lot of live coding. And that was that time when you were just kinda open up a terminal. I don't know if you still do this in your live coding.
Kelsey Hightower:If it makes sense to explain something. Got it. Right? It's like doing the early days of any technology. When you talk about it, it only goes so far.
Kelsey Hightower:I remember, speaking at a conference in Portland, the first HashiCorp, actually. Yeah. And Kubernetes had been out for a couple years, and they were announcing a Kubernetes competitor, Nomad. They had a nice logo, they had a slide up, and it was like, what is this thing? What does it do and how does it work?
Kelsey Hightower:And so the next day, I decided to compare, live, nomad and Kubernetes, and that's where the live terminal helps. Yeah. Right? Because people wanna see what is it like in action. So I just took the app that I know how to run on Kubernetes, and I just remixed it to run on Nomad, so I can explain how they were similar and how they're different.
Kelsey Hightower:So I just kinda use it as a backdrop as a tool to narrate, for the audience.
Bdougie:Yeah. Can you explain what Kubernetes says?
Kelsey Hightower:To me, I used to explain it as the system you would build if you had enough time and expertise. Most people that have to manage servers at the end of the day, their goal is to get software in the hands of your customers, and the way we tend to do that is using servers. Simple as that. If I gave you 2 servers, you do not need Kubernetes Yeah. At all.
Kelsey Hightower:So what do you do when you have a 1,000 servers? And this idea of like, hey, which of the 1,000 should be running the front end? Which of the 1,000 should be running the back end? What happens if any of those servers were to go away? If you've been doing this long enough, you have your scripts, you got tools like Puppet, Chef, and Ansible.
Kelsey Hightower:You have a way of doing this. But I think what happens is, over time, all of those practices can be bundled back up. And when you bundle them back up and you slap a name on it, that is what Kubernetes is to me. This idea of, like, hey, package your app in an abstraction. We call them containers.
Kelsey Hightower:Yeah. If you have to decide what server to run on, maybe the one that has the most memory and CPU free, the active scheduling that's built into Kubernetes. And you add a bit of metadata so you can describe what you wanna do, give it to the system, and it takes the stuff we've been doing for the last 20 years and put some semantics around it.
Bdougie:Yeah. So you mentioned Nomad, and there's a couple other Kubernetes like orchestration platforms. Why did Kubernetes win? Alright. Sorry.
Bdougie:I'll take a step back. Did Kubernetes win and, then why?
Kelsey Hightower:I mean, I think there people talk about winning, like, the English win. Right? It's one of these things out of convenience and what Kubernetes brought to the table. Some would say, Google, Red Hat, cloud providers being behind Kubernetes. How can you compete with that kind of inertia, that kind of force behind any project?
Kelsey Hightower:But what Kubernetes does represent is like, you have Docker that has paved this way and before Docker, all the config management tools paved the way that automation should be a thing. So on top of automation, we get Docker, brings an abstraction. But Docker is missing something. Docker really works great on a single system. Where you install Docker, you can build images, store them and run them, but what happens when you go back to that 1,000 server problem?
Kelsey Hightower:Kubernetes brought in the missing parts of Docker in my opinion. So Docker is still there, but it solves the problem above it. And I think the reason why it won is because the inertia, the expertise. You have a company like Google that has been doing this for 15 years prior that says, you know what? If Docker has won, can we add our expertise on top and turn it into an open source project?
Kelsey Hightower:And if you've ever felt it for the first time, like how straightforward it was, I think, for a lot of people, I think it was just like, yo, this solves actual problems. Yeah. Whereas, Mesos, which came before, you had to explain too much to people. They didn't understand why this complexity was necessary. And also, Mesos, which was the incumbent, predates Docker.
Bdougie:Yeah. Who who made Mesos?
Kelsey Hightower:Mesosphere, but it comes from one of the original papers that Google wrote around this problem. If you have millions of machines and you wanna balance out their usage of compute, right, you don't wanna run apps on servers that go underutilized. That can be tons of cash. Yeah. And so you wanna have these algorithms, these scheduling routines that say, hey.
Kelsey Hightower:We can use our servers 50% efficiency when the world is getting 7. Right? And so then that paper turns into Mesos, the actual implementation, the actual open source project, but it was way too complex. It wasn't clear how to use it for the average person. There's multiple ways to do it if you're doing big data.
Kelsey Hightower:I don't know if you remember the big data era.
Bdougie:Yes.
Kelsey Hightower:We had Hadoop and Spark. Those can run on Mesos. If you just wanna run your web app, you can go with marathon. But you start looking at the plethora of choices, people got real confused. Even experienced people like myself looked at Mesos like, yo, I don't think this is for me.
Kelsey Hightower:Let me just stick to config management. Docker comes out simplifies the problem, gets everyone to buy in on this concept that your app should be packaged. Not as Python, not as Java, but as a container image. And once we get the world there for a couple of years, Kubernetes comes in and say, hey, here's what's missing. And I think that is easier to move our industry forward than the other attempts.
Bdougie:Yeah. And so, like, Kubernetes started inside of Google, and you eventually got word of Kubernetes. Like, what's your what's your introduction to Kubernetes?
Kelsey Hightower:I'm working at CoreOS. Yeah. We're competing.
Bdougie:Coreos?
Kelsey Hightower:Yeah. Coreos is what the company was. The employees that worked there identified that way, but the initial pitch was Google's infrastructure for everyone else. So when you read all the white papers, it's like, what's necessary to do the things the way Google does? Now some people listening to this be like, no one has Google problems.
Kelsey Hightower:Why are you doing this? Right? But the thing is there was a lot to learn from that way of doing things in terms of compute. So CoreOS came out and said, hey, first of all, there's this white paper called Chubby and it talks about how Google used a key value pair to store configuration data and to do name lookup for all your services. So you start with this key value store, etcd, what people will be familiar with in the open source world.
Kelsey Hightower:But they also had containers internally at Google. Not necessarily like Docker, but that was a component as well. And so now that Docker's gone mainstream, you bring in the key value pair. You bring in this idea of containerization. Then you need an operating system And so we had a Linux distribution that was only designed to run containers.
Kelsey Hightower:All in bet. And we were building the scheduling piece. So if you read the Mesos paper or the Borg paper, we talked about that 1,000 server problem. How do you manage these containers across all these machines? And so we were working on a tool called Fleet.
Kelsey Hightower:We made a big bet on system d. This is the in its system that was growing in capabilities. Instead of just starting and stopping processes, you can actually do dependency management. When s e d starts, then start Postgres. However, the order you need it.
Kelsey Hightower:And so we were like, yo, this is the future, so let's build a Borg like thing using Linux as the backdrop. And so I'm sitting there, like, yo, we're killing it. We're starting to get, you know, buzz as a startup, people are starting to recognize the model that we're going, we're starting to get some adoption, and then Kubernetes gets dropped. You know, we get this noticed a day before, hey, Google is releasing an open source project, it's called Kubernetes, it's gonna be built on top of Docker and it's gonna bring a lot of the things that Google has experience with to the container orchestration market. And so just like everyone else, we're outsiders.
Kelsey Hightower:And we're like, hey, does this disrupt our business? Do we need to pivot? And I remember early discussions were we wanted to build a compatibility layer. So not just adopt Kubernetes or continue working on Fleet, but maybe we do our own implementation so we didn't get necessarily locked in. Now, nights and weekends, I was like, you know what?
Kelsey Hightower:I'm a start contributing to this thing. I gotta walk through the code base. I gotta really feel it for myself to see if this is where we're going or not. And I think I just decided to start contributing even when Cora was wasn't ready. So I kinda hedged a little bit just as a kind of a neutral technologist.
Kelsey Hightower:If this is going to be the future, I need to go and verify. So I became a contributor.
Bdougie:Now today, we're all using Kubernetes. Well, we're not all using good buddies. I have, like, I had this statement that when I started writing code at Netlify, so 2016, they built their entire deployment platform on Kubernetes. I also learned Docker around that same time. I learned enough to get the job, and then immediately, it was like, okay.
Bdougie:Problem solved. I became a React developer, like, immediately. So I went from doing mostly full stack and back end to then doing only front end. And I think that might be an allegory of, like, where we are today with orchestration, where you don't necessarily have to know exactly what's happening because a lot of times those problems are solved. So I guess question today, like, I have a new startup.
Bdougie:Do I go start with Kubernetes, because I know I'm gonna have this problem, or do I wait until the problem occurs?
Kelsey Hightower:If you got no customers, right, you have no you you you really not go use Heroku. Go use Cloud Run. Go use something that just says, listen. Here's my app. I'm gonna just package it up and deploy it, get a URL, do what you gotta do.
Kelsey Hightower:You know when Kubernetes makes sense at some point. If you're scheduling like a batch job and long running, you have a diversity of workloads, of course, you're gonna wanna be able to manage that. But I think most people are kinda jumping the gun. Now, what I would recommend, maybe do package your app in a container because that's gonna give you optionality. I can take a container and run it in Lambda with some work, run it in Heroku, run it on the VM.
Kelsey Hightower:I do think that probably makes sense to normalize the way your code gets shipped around. But do you need Kubernetes out of the gate? Probably not.
Bdougie:Yeah. Yeah. I mean, it's it's tempting because, like, a lot of times I think I I talk to a lot of founders and and start I do a bunch of podcasts. So I talk to a lot of founders, and a lot of times it turns into you're looking for a problem for your solution. And sometimes if you already know how to use Kubernetes or you don't really know how to containerize your applications, you end up reaching for that first.
Bdougie:And it's something that I always try to push back on. Like, if I don't understand it out the gate, then we probably don't need it yet.
Kelsey Hightower:Well, I mean, the truth is sometimes things evolve to just be best practices. Yeah. Do you need a CICD pipeline when working by yourself? Probably not. You can totally build it on your laptop, upload the binaries to GitHub and call it good, but that's not repeatable.
Bdougie:Yeah.
Kelsey Hightower:That's gonna eventually cause some headache down the road. So a lot of times, we just know better. Go ahead and just create the pipeline just for notes for yourself. When the second person joins the team, they can just use the same pipeline. Right?
Kelsey Hightower:A lot of the decisions you've made get captured in an automated way. Linux is the same way. Do you does everyone need Linux? Right? That could be overkill if you're not really doing very much.
Kelsey Hightower:So I think most people decide, like, hey, at some point, like HTTPS, they become industry standards. You adopt them because the cost of doing something else outweighs
Bdougie:Yeah.
Kelsey Hightower:The complexity of just taking the industry standard. There are managed Kubernetes solutions, right, where the cluster's fully managed and you're just working with a kube API. So I do think it's one of these things that if you have no customers, just put your app in some container image, deploy it in the thing that has the least overhead. And then when the time comes, you know, bring it in when it makes sense.
Bdougie:Yeah. Yeah. So I'm gonna restate the the previous question I had, not the last one before that, which is what's the secret sauce for Kubernetes?
Kelsey Hightower:Nobody knows. Kubernetes is caught up in the same thing that culture's caught up in.
Bdougie:Yeah.
Kelsey Hightower:Right? Sometimes the right thing comes out at the wrong time and then you can always do a retrospective hindsight 2020 about why it didn't work. The sentiment was just right for Kubernetes. Right? Kubernetes comes out after Docker is successful.
Kelsey Hightower:Kubernetes did not make containers successful. It rides on the back of containers being successful. The thing I think we got right on the community side, there was no foundation in the beginning. There's no level of politics like we have now in the beginning. In the beginning, we have 2 or 3 people that decide that there should be a community based conference around Kubernetes at a time where people didn't think it was going to work.
Kelsey Hightower:They're like, oh, this is Google, they're gonna abandon it in a couple of years. This is no threat to anything. They're just not serious about enterprise. And so it kinda got ignored by the incumbents for a long time. But at that beginning, we decided to have a small community around it.
Kelsey Hightower:Joseph Jacks, Patrick Riley, people like myself, we decided to create Coupcon. And when Coupcon came out, if you notice who was there, it was the practitioners leading the charge. There wasn't a lot of sponsors. There wasn't any of the vendors yet because the vendors didn't even know what this was. So we had a nice organic runway that said Kubernetes must solve actual problems.
Kelsey Hightower:Not hype around platform engineering, not hype around DevOps, but you actually have to solve a problem. And so I think having that 1 year, 2 year head start of solving actual problems, that let the community form around this authenticity layer. And then when we had KubeConna celebrate those milestones and those achievements, then it became sustainable. Once you bring in the VM layers, you bring in the cloud providers, our foundation was so solid that we got the APIs right. So if you were a small startup and you wanted to build your own augmentation to Coupe, you can do that without negotiating Yeah.
Kelsey Hightower:With all the core engineering. So I think all of those ingredients contribute to this outcome where Kubernetes looks more like Linux than it does like Oracle.
Bdougie:Yeah. I I wanna go further down this this path, but I actually wanna take a step back and talk about about your your journey to get to where you are now. I remember watching a talk that I don't know what conference it was, and you talked about there was somebody who weren't at a code in prison and went through, like, your tutorial and, like, it made a huge impact in her life. And that was, like, extremely inspiring to me because I think the the the opportunities you have, like, sort of building a platform like, I've been doing this for 10 years, not this particularly, but writing code for 10 years. Years.
Bdougie:A career switcher did sales, got an MBA, learned how to code because I wanted to build an app. At that point, I told everyone I knew about code. It didn't matter what it was. It was, like, Python, Ruby, do it. It's gonna change your life because it did for me.
Bdougie:So I wanted to talk about your trajectory because you said you did tech support, but also you also mentioned, like, you worked at McDonald's as well prior to that. Your your origin story, can you talk a bit of, like, why you chose computer science as a a pathway?
Kelsey Hightower:I didn't choose computer science Yeah. As a pathway. I didn't go to college, didn't study computer science. 1999, 2000, the north stars were Bill Gates, right, Steve Jobs. These are huge huge companies and every job posting 1 had 2000000 years of experience in proprietary technology that you could not afford.
Kelsey Hightower:Period. So if you're 19, 18 at that time and you look at the world of technology, the gates were really, really high. Ain't no YouTube yet.
Bdougie:Yeah.
Kelsey Hightower:Right? There's no this is all you have to do on your end. It was just this kind of vague industry that looks like it was only for people who had a computer science degree. It felt like you had to have all these years of experience. So where to start?
Kelsey Hightower:And so for me, when I got into the the field, anything was better than fast food.
Bdougie:Yeah.
Kelsey Hightower:Right? And so for me, I got in through a plus certification. Right? I did the certifications and I saw something on social media that all these people are stuck in certification land. The truth is, getting certifications for a lot of people that don't know how this works, it's almost like running on the treadmill.
Kelsey Hightower:You're not running in a marathon, so why would you stay on the treadmill every day? Technically, you're not even making progress. You're running in place. But it is still a workout. It is still a form of progress.
Kelsey Hightower:It beats sitting on the couch. And so for me, getting the a plus certification was like, look, I want some guided learning path. And when I got the a plus certification, my confidence grew a little bit. And I got my network plus and my confidence grew a little more. And luckily, there was a job fair that said, hey, if you wanna learn how to do high speed Internet install, so this is when people are going from dial up to DSL and the existing technicians, you know, I guess they're in a union, they only do phone lines.
Kelsey Hightower:Computers is not part of the contract. You gotta go outsource that. And so people like me came in And that's my entry into the tech sphere. But I also opened a computer store shortly after because I was still afraid to apply for one of those corporate jobs. It was easier to lease a spot in Atlanta and start fixing computers and making service calls.
Kelsey Hightower:Everything after that is really just this entrepreneurial mindset that says, if there's a problem to be solved, aka a job posting, get the skills, get the job and then move on to get my pay to accelerate where it goes. The last 15 years, I think, is me understanding that, hey, you can solve problems at a global and industry level, not just at the job level. And that's the part that people see when it comes to open source and my involvement in the Python community and the various communities.
Bdougie:Yeah. Yeah. And I and and you explaining that to you as well, because I don't think I knew the earlier part of your your journey as well around opening a computer store and even the certifications. I did see your tweet, though, recently. I do agree with that, the the treadmill.
Bdougie:And I talk to a lot of folks who they they always ask, like, what language should I learn first? Or, like, what should I build? Or how do I get more interviews? And I think they're also all valid questions, as long as you have, like, the goal, which is, like, get the interview, get the job, or build the project you're trying to
Kelsey Hightower:Everyone says there's a lot of luck that goes in this, and there's a reason for that. Yeah. I put all this effort into Kubernetes in the early days. I wrote a book on Kubernetes. Kubernetes could have turned out to be a flop.
Kelsey Hightower:Yeah. And then what? What would the narrative be? Right? It's like professional athletes.
Kelsey Hightower:If you lose the Super Bowl by one point, just one point, your team is terrible, they should fire the coach, they should fire everybody. Even though you made it to the championship game, you should reevaluate the whole existence of your team. If you win the same game by one point, the wind blows the other way.
Bdougie:Yeah.
Kelsey Hightower:It's a dynasty in the making.
Bdougie:Yep.
Kelsey Hightower:And so when I think about our careers, you start where you start. If you start with Python, learn the fundamentals. And I and people say, hey, what's where are the fundamentals? And I always just give people a very generic set of advice. Go down until you hit the hardware Until you know how your software runs on servers, CPUs, memories, GPU's.
Kelsey Hightower:And then go the other direction till you get to the people who use your stuff. And you start to understand the UX and all of these things around productivity, user experience, and etcetera.
Bdougie:Yeah. And so I have a a comment and and thought as well because I know a lot of folks who actually, quite a few folks who learn with React first. They learn how to build a website, how to build a website that interacts with JavaScript, and then straight to React. And they might be 3 or 4 years in their career, and React is it. So for those folks, would the recommendation be okay now?
Bdougie:Learn the next layer.
Kelsey Hightower:I mean, look, if you wanna be a plumber, you don't have to learn anything else but how to go and do plumbing. And there's a lot of people and I tell them it's a very dangerous situation of being in tech. You can have 20 years of 1 year experience. You'll be paid well, you can make 6 figures all day, you can be super comfortable in your job. And then maybe if the landscape changes, you'll find yourself in a weird spot where you got to figure out, are you willing to gear up and do it all over again?
Kelsey Hightower:Learn a new piece of technology, unlearn some things so you can learn some new things. But I do say for people that are really trying to be well rounded and expand the type of problems they can solve. For example, there are a lot of people who are doing front end for a long time. When mobile came out, they didn't know what to do. I I don't know.
Kelsey Hightower:This is not the web. Right? They wanted to have they wanted the new thing to work the old way. Put a browser on the phone and just let me keep with my existing skills. They didn't think about, oh, how do I leverage the fact that I'm running natively on this hardware?
Kelsey Hightower:How do I optimize for this particular screen size versus just trying to remix the web to run better locally? I think that is the thing. So I would just say, look, depending on what you wanna be able to contribute to this game, it's worthy to expand your mindset. So if you're doing front end development, it is a form of development, has a lot more user interaction. There's elements there.
Kelsey Hightower:But it doesn't mean you're restricted to a web browser. You can go into mechanical engineering. Maybe you can redesign the toaster, or you can figure out why native u wis are more acceptable by some other people.
Bdougie:Yeah. I I wanna talk about open source as well. Because open source is the journey I took and the reason why I'm we're building open sauce because my path would not have been my path without open source and folks dropping blog posts and documentation about how to build stuff in Ruby and Python. So that's why I'm here. But I know you have a a journey of, like, why you do open source.
Bdougie:I sat on Twitter space actually on Friday, and it was titled, like, why you do open source. So my question to you is, like, why do you do open source?
Kelsey Hightower:In the beginning, it's access. Yeah. Right? So if you're 18 trying to figure out how to break into this industry where you don't see a lot of clear north stars, but then you go to Cop USA, these are the computer stores back in the day.
Bdougie:Yeah. RRP.
Kelsey Hightower:And they have they have the, kind of computing magazines, and it's like Linux dish on the back, and there's literally an actual CD in the back. And you can install Linux. Right? This form of UNIX that's open. It seems like enterprises are going that way.
Kelsey Hightower:In the year 2000, you're starting to see job posting say, aix or Linux. And you're starting to see that door open up a little bit. And now finally for the first time, the software is accessible. You can install it on the old PC and now you're getting access to the knowledge. And then even in the book, it will tell you with screenshots.
Kelsey Hightower:Here's how you get it installed. Here's how you get this thing that look like Windows. How to recreate the world that you come from and now you feel a little bit more empowered. And so now, I'm on the consuming side, where I'm consuming this knowledge and I just wanna get a job. But eventually you get on the other side, the creation side.
Kelsey Hightower:And again, what tools are the most accessible for creation? Programming languages, most of those tend to be open. And then you also start to understand that you don't have to accept these tools the way they are. You can change them. And so my first contributions to any open source project was, virtual inf for Python.
Kelsey Hightower:Right? So if you're using Python and you have to manage dependencies, it can be a mess. 2 different versions of Python with different dependencies, it's like a shotgun blast to your file system. And so virtual amp gave you a way to create virtual, you know, directories or whatever to manage your dependencies without conflict. But it didn't work the way I wanted it to work.
Kelsey Hightower:Instead of accepting that, you can just raise your hand. Go to the open source community and it's like, hey, I found this bug. This is what I expected and this is what happened. And then you can try your hand at fixing it. And once I got to that point where you're no longer just a consumer and you can actually start to contribute, that's when I kinda understood what this really meant.
Kelsey Hightower:It is a way to shape industry standards.
Bdougie:Yeah. Yeah. I I use the correlation, somewhat we just said to, like, wanting to play in the NBA and or just play basketball in general. So, like, if you wanted to play at the highest level of basketball, you're gonna find pick up games at the YMCA, or maybe you find there's, like, a a camp, or you can, like, run drills or maybe a league. Like, that all gives you an opportunity to level up.
Bdougie:It's not all the NBA, but it's a stepping stone. I find open source is kinda like finding those pickup games and opportunities just to raise your hand and do a thing.
Kelsey Hightower:You know, one thing I think when we say community, a lot of people get confused. They think community is purely the booths you see at the conferences, the free stickers, the free socks, the open funnel to the VC funded startup. Yeah. But when we think about community too, we also think about you literally get to engage with the people who created these things. Right?
Kelsey Hightower:These are the inventors of said technology. This is almost the equivalent of being able to create a light bulb next to Thomas Edison, real time. And this community tends to be very welcoming, maybe for egotistic reasons so that people can tell you everything that they know. But, nonetheless, the community is super inviting. You can come in and say, hey, I wanna understand how operating systems work.
Kelsey Hightower:And someone will be patient with you and it's like, alright, let's walk through the code. Let's walk through your contribution. Let's walk through the potential fix that moves this forward. I think that's the thing that a lot of people that have never contributed to open source don't understand what's so powerful about this particular segment of our industry.
Bdougie:Yeah. And I I feel like the, the line of questions I've been asking, I could kinda summarize your answers, like, it depends with an answer. I'm curious the so open source. Like, oh, actually, let's talk about the cloud in particular. So the CNCF now incubates a bunch of projects.
Bdougie:They graduate a bunch of projects, and there's a lot of success out of CNCF, 100%. Do we need like, CNCF projects seem to be the ones that are blessed. Like, is there a chance for projects that are not CNCF backed in the cloud?
Kelsey Hightower:The the challenge is when we say open source, we're not talking about also the responsibility that comes with it. If you have a million people and one maintainer, all 1,000,000 people go to the site, they download the free software. Yeah. They use it and they expect it to work. These days, people want it to work as good as the software they pay for.
Bdougie:Yeah.
Kelsey Hightower:Right? Even though I paid you $0, my expectation is the same if I would have paid you a $100 a year. And when that is not lined up, what do I do? I can go on Twitter and complain this software is trash because it doesn't do x y z. I can go to GitHub, this open forum, where I can just open a ticket and say, this thing should add this feature so I can be better at my job and make way more money that I'm making now, be more efficient.
Kelsey Hightower:So then who funds this?
Bdougie:Yeah.
Kelsey Hightower:Who works for free? This is the part where I think a lot of people have gotten confused around the sustain sustainability of Open Source. If you look at the relationship Linux Foundation, the same organization behind the CNCF that they have with Linus Toward, well, he has a pretty nice setup there. He's earned it. Yeah.
Kelsey Hightower:He didn't go and start a company. He didn't try to be the next IBM. He chose to say, listen, you all can have this software, this thing. He stewards the community and he tries to make sure that the code base and the quality of the project stay as high as possible. And in return, there's a foundation around him where the big players, the Samsungs, the Nvidia's, the Google's can contribute some and keep this thing running.
Kelsey Hightower:And so far that has been very sustainable. Linux is over 30 years old. Yeah. Now for these new projects that come out with the same expectation, how many more Linux towers will you think you're ever going to get in this lifetime? Right now, I think, and maybe rightfully so, if you're 25 years old and you create the next big thing, think about how difficult it would be to say, I'm not going to try to become a billionaire from this idea.
Kelsey Hightower:Do you know how many VCs, you know, how many VCs will line up behind you to say you should start a business. We will give you 50,000,000. We'll give you a 100,000,000 if the idea is big enough. And so once that happens, then I think what ends up going is that person, instead of forming the community of different maintainers, that if this person were to stop, the project can continue, that get delayed for a decade. What happens is now they can just hire those maintainers full time.
Kelsey Hightower:They can pay them full time, which I think is a good thing, by the way. Yeah. But then at some point, we need a
Bdougie:profit. Yeah.
Kelsey Hightower:And when we don't get a profit, what happens? Rules start to change a little bit. We go from it's 1 big party to there's an price submission to get in. Yeah. And if you don't pay, and rightfully so, we're not gonna be able to afford to keep paying these people to work full time.
Kelsey Hightower:And any software after a decade of being worked on becomes so complex that to become a maintainer takes so much effort. And again, we go back to who's willing to work for free. So we're in this weird situation right now.
Bdougie:Yeah. I've I've been talking a lot about this to actually with with the team at Open Sauced, but also folks just in around the industry. Because I feel like the and you'd mentioned this before we actually hit record, like, the age of the Rich Stallman's and all the sort of free as in beer open source, I think we're we've a 100% shifted into this new form of open source. And, I'd I'd admit, if I was 25 today and I discovered, like, the next big thing when it comes to the cloud, I probably would be taking the money as well.
Kelsey Hightower:Well, I mean, to be fair, a lot of the earlier stuff, at least in my opinion, was alternatives.
Bdougie:Yeah.
Kelsey Hightower:We have Unix, Solaris, AIX. They have their own command line tools. And so if you think about some of the earlier, more successful project, they're alternatives. Yeah. Grep versus Bash versus Corn Shell.
Kelsey Hightower:And so now that we got all the alternatives, now the alternatives become better Yeah. And they're sustainable. People want products now. Yeah. People want UI.
Kelsey Hightower:They want
Bdougie:Yeah. Documentation.
Kelsey Hightower:They want integration with enterprise software. They want FIPS compliance. They want security patches. They want a full experience on top. And I think open source trying to deliver that under the old model of volunteer development is tough.
Kelsey Hightower:Right? I think it's easy to kinda you built back end APIs. Yeah. I can get you a headless API all day. Check that in.
Kelsey Hightower:It's probably gonna work. But when you start telling me you want a mobile app for iPhone and Android, you want me to test across all browsers, and you want me to keep up with the latest and greatest experiences of the web and mobile, that takes a whole team and a whole different set of skills.
Bdougie:Yeah. In the modern it's like the web in particular. There's, like, no there there's not really a lot of certifications, like, official ones that people are sort of looking at. I know GitHub has had their certs, that are now out and live, not technically web, but more Git. How do we validate your how does an engineer validate themselves with this engineer today?
Kelsey Hightower:Man, it depends. If you're selling software, you validate yourself by revenue.
Bdougie:Yeah.
Kelsey Hightower:And we're being honest. You could be the best engineer you want. Zero revenue, then that's just what it is. But that doesn't mean you're not a good engineer. It just means you're an engineer that can't sell software.
Bdougie:Yeah.
Kelsey Hightower:These are different things. And I think we're trying to measure a lot of these startups based on their ability to sell software against their ability to create great technology. And when it comes to the web, how do you validate a painter? Right? Is it the way they use the brush?
Kelsey Hightower:The brushes they use, the color palette they choose, the camera the photographer chooses to shoot with, we tend to judge them by the output, the work itself. And as we know through history, a lot of times, it's not until a 100 years after you die that people find the ability to appreciate what you've created. So even living artists struggle while they're alive to get appreciation for their work until the market comes around. So I think the thing that makes software hard is number 1, we do wanna solve actual problems. Then number 2, building a business around software.
Kelsey Hightower:To me, these are fundamentally 2 different things. So look, learn the skills, programming languages, design patterns, user interaction and you can go out and solve problems. Building a product, now you need a different set of skills, and I don't know if you can measure the 2 the same.
Bdougie:Yeah. And speaking of problems, like, I don't know if this is a problem coming down the road, but we're seeing a lot of projects that you you talk about the 10 years. It gets complicated. Stuff changes. You gotta make a profit.
Bdougie:HashiCorp recently relicensed TerraForm. We're seeing this, I think, with Red Hat also relicensed Fedora, I think it was recent.
Kelsey Hightower:Years ago. Years ago. I remember. I mean, I remember the Fedora split. You have Red Hat, and for the most part, you have all the clones sent to OS, and you get to a point where Red Hat is like, yo, we're not this is not gonna be as easy to access anymore.
Bdougie:Yeah.
Kelsey Hightower:We would like you to use Fedora. And if you're an enterprise, you're like, no. No. No. No.
Kelsey Hightower:No. Everything now surrounds Red Hat, its certification process. We can't have a new version of our web load balancer. Yeah. We need the version that ships in Red Hat Enterprise.
Kelsey Hightower:They eventually say, no. No. No. You have to pay. If you want Red Hat, that now costs money, and that's the only people that can get it.
Kelsey Hightower:And so then you now are doing good at business. Yeah. But if you were a contributor back then, think about it. Red Hat ships over 10,000 open source projects as part of that distribution. What about all those people?
Kelsey Hightower:Do they get a cut? What's the actual product here? Is it the software or the support? And so I think that's a line that Red Hat probably did the best ever in defining for a real stable open source business model. They've nailed it.
Kelsey Hightower:I think a lot of people wish to replicate. The case of HashiCorp I think is the way they did it. HashiCorp did a good job. I applaud them. For over a decade, they gave us Terraform, Console, Vault.
Kelsey Hightower:All of these tool that a lot of people use every day in their job and the majority of people probably pay $0. They IPO. They're a public traded company now. And I don't know if anyone's ever worked for a public traded company, but you gotta grow revenue 20% every year forever. And if you don't have a plan for growth, then people stop investing in your stock and then funding goes the other way.
Kelsey Hightower:And so in the case of HashiCorp, if you communicate to people like me, I've contributed to HashiCorp projects before. I've contributed code not because I was tricked, not because I thought the license is gonna stay the same way forever, just because I wanted to be more than just a consumer of the projects. I want to scratch my own itch so I contribute the code. And who cares if they're gonna make a profit on top? But when you swap the license and the way you communicate it, I think people should have just been told we need to make cash.
Bdougie:Yeah.
Kelsey Hightower:And the only way you we found to do that is to force you to pay for the software at a certain level and certain degree. Yeah.
Bdougie:Sorry. Do you watch TikTok at all?
Kelsey Hightower:Not really, but people send me videos.
Bdougie:Alright. Well, there's a video I watched last night about Pyrex. So Pyrex plates, you probably got it at your wedding. I got some at my wedding as well. Turns out capital Pyrex, so if it's spelled all caps
Kelsey Hightower:Mhmm.
Bdougie:That was from beginning of time until, like, the nineties, quality product. And then it was sold to, to another company, which I forgot the parent company, lowercase Pyrex, different product. And, I asked my question well, basically, what the deal is, if you heat it up and cool it down too quick, it's gonna shatter. Like, glass shatters in your oven. And turns out I had lowercase Pyrex during the pandemic when I was making a sourdough bread bread.
Kelsey Hightower:It exploded.
Bdougie:It exploded in the oven. And I was like, oh, this is interesting. I thought this was supposed to be high quality. I bring this up because the relicensing, the difference of, like, profit, like, we expect that now the open source that has to make a profit, are we getting lowercase versions? No.
Kelsey Hightower:No. No. So I I actually think if I had to play both sides of this, if I'm at HashiCorp, this is a net win. Yeah. We are finally being very clear about our intentions.
Kelsey Hightower:We are building a software business, not a charity. We don't hand out free software and expect only the most willingness to pay. Now we're saying there's gonna be a split between the effort it takes to solve these enterprise class problems. And so, now we're gonna have a way to say, no. So when you say, hey, at this feature, at this thing that doesn't make financial sense for us, we can say no, and now you have somewhere to go.
Kelsey Hightower:There's forks. There's a fork of Terraform. There's a fork of Vault, and the community can choose to go that way. But I think the quality will stay high on both sides because they've been doing that for a decade. I don't expect that to change, but I do expect them to be much more focused, much more clear.
Kelsey Hightower:And now you gotta pay full price for the value that they add to the puzzle. That's business. I'm okay with business. Yeah. The other part though is when we say open source, that little fork button that's at the top of GitHub, that's what a community's gotta put their mouth money where their mouth is.
Kelsey Hightower:So if you're complaining about the license change, that means you're willing to step up. Right? Fork the project, rename it, and keep contributing in a way that adds value for 10 years. And if you're not gonna be doing that, then the price and the license change is justified.
Bdougie:Cool. I wanna switch gears. I wanna ask you the question which is, like, what tech are you excited about today?
Kelsey Hightower:I'm I'm 40 I'll be 43 tomorrow. Oh, happy birthday. Alright. 43 tomorrow, and I don't think there's nothing that excites me anymore because after 20 plus year career, you just see the cycle. Right?
Kelsey Hightower:And so the cycle continues. The industry is very excited about AI right now.
Bdougie:They are.
Kelsey Hightower:That's an abstract thing. And it's very easy to be excited about a thing you don't understand. Right? Because there's no one can answer the questions clearly. And then we get groups of technologists, they split typically.
Kelsey Hightower:Some just wanna be magicians and they just wave the trick in front of everyone. You can raise a lot of money, asset bubbles form. People just like, hey, help us do a thing we don't understand. And you can ride that wave. And then we get the educators.
Kelsey Hightower:These are the people that counterbalance a little bit and they say, yo, listen. Let's be practical. For example, people are very excited that AI will write all the software. Meaning, humans won't have to type into our text editors anymore to produce software. Now you could just use some tool and it would do it for you.
Kelsey Hightower:And that's very naive because what is software engineering? It's making some decisions based on a design. Humans have a problem. We design what we think is the right solution to said problem, and the very last thing we do is tell the computer what we want to do. That's it.
Kelsey Hightower:That's the last thing we do. So at best, maybe we get more time to think. Maybe we get more time to design and document. And then we'll have some help writing the code. But right now, people are overly excited about this idea that robots will be able to dictate the day to day.
Kelsey Hightower:What I am excited about though is people understanding how we've gotten to such boring work throughout our day that we can take a generative AI model and use it to respond to people via email. Why are we sending people artificial responses when we now need to reconsider? Did we even mean any of that? Was half of that email just a waste of time and platitudes that were unnecessary? I hope people go back to real intelligence.
Kelsey Hightower:When I grew up, the goal was to get the real thing if you could afford it.
Bdougie:Yeah.
Kelsey Hightower:And now people seem to be very happy about being fed the artificial thing. This is insane. This is backwards. So I'm hoping that if AI has all of this attention, then real intelligence gets even more. So go outside, interact with a real person, be awkward, let it be spontaneous.
Kelsey Hightower:I hope that is the thing that we all return back to once we get bored of the artificial.
Bdougie:Yeah. Yeah. I it it's interesting because we started this conversation with Kubernetes and how it's like now automating the things that it was trivial to not trivial to do, but things that got basic enough to then automate. I think AI is also covering a bit of that path. And I heard a conversation recently about, like, AI.
Bdougie:If it was named something different, probably wouldn't be as exciting. But if I like, for example, if, like, Kubernetes search. So, like, you're doing a LLM and you called it Kubernetes search, it probably would still be as as exciting. I'm not sure if that's the right way to apply that.
Kelsey Hightower:Think about it. These large language models, let's be very specific. We're not talking about all of AI. That's been out for a long time. If you put into Google Maps, point a to point b, there's a lot of ML, AI at work to try to find the shortest path for you based on tons of heuristics for the models that they're using.
Kelsey Hightower:And you've been using it so much, you don't even need to say AI anymore. You just say I'm using the map. Easy. Straightforward. That's how technology should be.
Kelsey Hightower:It should get buried behind the solution to the problem, not to be the thing. So LLMs are it's impressive to a person that doesn't understand. If you don't know how to write, which is dangerous now because so many more people are illiterate, They can't express themselves. So you go to this thing and say, write me an email to follow-up with a customer about this thing. And it writes an email that you believe that you're either incapable of doing or you don't have the time to do.
Kelsey Hightower:If that's impressive to you, then to me, something is wrong. I can see how, using Photoshop for the first time, you just get one of those brushes and you just move it across. You're like, oh, I'm a artist now. And the truth is, it will enable more people to become an artist. And nothing's wrong with that.
Kelsey Hightower:I'm all for that. But I hope we don't just disrespect what it means to be an artist with perspective, what it means to observe the world and then go draw it versus having no experience with the world and saying, I just wanna artificially generate a thing that I don't understand and just put my name at the bottom. Yeah. That's that's what I I don't want to cheapen the human experience Yeah. Because we have these tools.
Bdougie:Yeah. I was laughing because I was thinking about the my brother I have a twin brother, in out in Nashville, and he's been working in kitchens for years. And I've been cooking for the past couple years, and every time I get a new gadget, I'm like, hey, you ever heard of mandolin? And I'm like, oh, and I could slice lemons like this. And he's like, yeah, Dude, I've had this for years.
Bdougie:Like, this has been around, but it's exciting to me. So I think with LMS or chat GPT or whatever, that's that's my dad texting me. Like, hey. Look what I found or check out this research I did.
Kelsey Hightower:I do. I do. So at a higher level, I do wish, If I think about I saw some LLMs. They're trying to give people better workflows like book a flight, and once I get there, build me an itinerary so I can go watch a concert.
Bdougie:Yeah.
Kelsey Hightower:And you take all the context that this thing knows about me, and it should give me something that's personalized. This is a pretty decent experience.
Bdougie:Yeah.
Kelsey Hightower:But when I think about how apps are built today, they're all little walled gardens. None of them export any data about themselves. If I wanna go to my bank, I gotta log into my bank, click around just to see my account balance. Wouldn't it be nice if that app, with my permission, was constantly exposing a common data model? Then it'd be very easy to build these workflows.
Kelsey Hightower:Right? What music do I like exported by Spotify? Then anything can pick up that preference and use it as input to their decision making engine. I think people are excited now about all this stuff being aggregated and now being able to have a better API. So when I think about a lot of the AI we're seeing today, it tells me that the software we've written in the past has a terrible API.
Kelsey Hightower:Yeah. Nobody wants to click through 50,000 different apps to try to put together a vacation. People just wanna be able to string through this is what I want and have the right apps do what they need to do in response.
Bdougie:Yeah. It it's the, so I know there's a concern around data and who owns the data when it comes to now where AI is taking us, especially with LMs. So Reddit just announced their IPO, but also announced their deal with Google Search for 60,000,000, which I think they probably could have asked for a little bit more than that. But so at this point, as me, a user, a citizen of the Internet, do am I concerned about if I'm setting up my vacation and what's happening with the data and how it's like gonna discover about me later?
Kelsey Hightower:The truth is people don't care. Yeah. We can educate people all day. Like for example, look at these houses. They're not secure.
Kelsey Hightower:You have glass windows. I can throw a rock and break the glass window and enter the door. So we rely on people just being good citizens not to come into your house uninvited. Yeah. We're not building secure homes.
Kelsey Hightower:Right? We're just building enough guardrails to say, please don't come in here unless I open the door for you. So when I think about where we are with data, people don't care about their data anymore. Should we educate them? Yes.
Kelsey Hightower:When you put on the smartwatches, when you use the app and you click around, all your data is being shipped to the other side. And so everyone has this innate fear of Big Brother, but as long as you make it convenient, they will pay you to take their data and then buy back the product that's built on top. So the reality is that's where we are today. Now, what does a shared responsibility model look like? Even though I'm not really excited about the web 3 people, you know, the crypto plus, you know, thinking this through.
Kelsey Hightower:But I do think the idea that what if we were intentional about our own data? What if I own my own data? What if I was the person that when I go to the hospital, they have to give me my data that's about me and I decide who gets to use it. If nothing more, I think what we end up having is the ability to create our own work flows. Like, I buy financial instruments, but I gotta log in to get the data out in order to put into another tool.
Kelsey Hightower:This is this is extremely silly, but it because that data is not mine. It's actually their data and it's all loan to me. So I would love it to be the other way around where it's like, hey, these services that you use, it's your data. You always have access to it. Maybe we encourage people to store their data on their own end and then push it the other way, but then there's a lot of responsibility come with it.
Kelsey Hightower:Who's gonna back it up? Who encrypts it? Who make sure that it doesn't get stolen? But I do think it's time for that conversation. At least there's move the needle towards being intentional about your data and where you give it.
Bdougie:Yeah. That that's actually a good segue to talk about open sauce data, which we're gonna break for a demo showing open sauce. Yeah. So Open Sauced, I mentioned, like, in passing, we're basically building the inside tab with GitHub. Think of, like, a full on product, though.
Bdougie:So, we took 260,000 repositories, gathered that data because it's open source and public. So we're not looking at code. We're looking at specifically Git commits and PRs. So we can actually infer a lot of information just based on that. That that's basically the product.
Bdougie:So, like, the thing that I I recently was talking through, and actually came from my conversation with Brendan Burns about, like, Kubernetes and success on sing adoption. What he saw was issues was a good showing of interest. PRs was a good showing of adoption. And then when you look at folks who are doing issues and PRs outside of Google or the Kubernetes team, then that's a signal that, okay, there's probably somebody outside of Google we can talk to. So we're building that pipeline, which I'm calling stars to PRs.
Bdougie:So you see a star is a great indication of someone interacted. They're saving it for later. But what's the what's the the conversion of them coming back and opening issue and saying, hey. I started this thing, or I'm trying to use this thing. Here's my problem.
Bdougie:So, like, that's hopefully what the demo will show.
Kelsey Hightower:Yeah. I mean, I think I think some of those measures are really good early indicators if you think about the first couple of years of Kubernetes where we didn't have all the features we have today. Yeah. Yeah. So seeing people filing issues saying, hey.
Kelsey Hightower:This thing is great, but if you had this Yeah. I could really use it. That flywheel. You're right. It's it's beautiful.
Kelsey Hightower:But then what happens when you become VIM?
Bdougie:Yeah.
Kelsey Hightower:Where you feature a complete, there's not a lot that needs to change. Some people don't want it to change. It's kinda like now we're in maintenance mode. And in that case, commits slow way down. The number of issues goes way down.
Kelsey Hightower:Is that a sign of a failed project or the sign of a healthy project? Yeah. And I think that is what's not super clear from those metrics alone.
Bdougie:Yeah. Yeah. And, like, what we've been looking at is the closed issues, not all issues, just closed issues. Like, what's the timing of that closing? And I think you could also I think the the thing that Brendan had said in the podcast was I just wanna know if this is a good neighborhood or a bad neighborhood.
Bdougie:And, like, if that if Vem is a neighborhood you're looking for, you'll know really quickly. Oh, cool. Takes 3 years for a issue to close. Totally fine. I know I can put this on my my laptop, and this is gonna be fine for years.
Bdougie:But then the alternative is like, okay. And we've we've touched this as well as, like, Neovim. If you want an alternative, go use Neovim. Like, the vibrant community, new plugins all the time, that's probably where you wanna hang out if you wanna see constant growth and acceleration.
Kelsey Hightower:I mean, and Vim is also one good example where we're all getting old enough for the maintainer of Vim, he died. Yeah. This is the ultimate bust factor that we all talked about when it comes to sustainability. And so in that scenario, imagine where Neo Vim doesn't start rewriting Vim, what, 5, 6 years prior. Yeah.
Kelsey Hightower:What if it was all of a sudden, people look up and say, woah, woah, woah. What's the game plan now? Should we rewrite this thing? Where should it go next? So I think that happening in parallel was really good for the whole ecosystem.
Bdougie:Yeah. This is the thing that we've been we've been tracking as well. So, like, being able to go back and see who the maintainer was in 2017 or the the team of core contributors versus today and how that's changed over time. I used to make this joke about, when I was at GitHub about people would disappear in the cloud in open source, which is they got hired by Google. So, like, you depend on this project.
Bdougie:Maintainer now has a better job. A job take takes up way more time, so now who who replaces them? And I think, like, CNCF does a great job with their projects. They have a sort of like a SIG group. Kubernetes has SIG, but other projects have something different.
Bdougie:Like, you can help and replace that, but some projects outside that
Kelsey Hightower:I've abandoned pretty much all of my open source projects. I'm at a point in my life where I don't use those tools anymore and I got other stuff to do with my time. And so what do you do with a project like that? In many cases, I tried to find a maintainer. And look, a lot of people have stepped up over the years for things that are libraries or, you know, tooling on their own.
Kelsey Hightower:And they do a good job for a couple of years, but they also move on. Yeah. And then as a maintainer, do you come back and try to fill in the gap? And I think the thing now is, if I release an open source project, is it a lifetime contract? And if it isn't a lifetime contract, how long before people get upset?
Kelsey Hightower:Will they take 10 years? Do they expect 25? And so this is why I think the red hats of the world that will step up and say, hey. We'll take these things if they're important to our customers and we'll maintain them effectively forever if necessary. You asked a question earlier about the CNCF and foundations like this.
Kelsey Hightower:I think what they do is say, hey, if you are a large industry player, we're talking about the NVIDIA's, the Samsung's, the Google's of the world. You know at some point, maybe dollars were enough for now. Maybe we'll be investments in these companies or 2 or 3 people from each organization's gonna have to step up and contribute. I think that's what the foundations bring to the table. Yeah.
Kelsey Hightower:They bring a sustainability path that can outlive a single maintainer or a single company that's funding these projects.
Bdougie:Yeah. 100%. Yeah. I I I think it's foundations like that. CNCF, Apache's got a foundation, ton of foundations, all in the Linux Foundation umbrella, help sustainability of open source and help us be able to depend on building our products in the future.
Bdougie:So, like, if I choose to use a Docker or a containerd or whatever it is, I know that's gonna stick be around
Kelsey Hightower:despite waking up towards is it doesn't guarantee a business. Yeah. It doesn't guarantee this is the part that I think a lot of people are starting to struggle with. This attention, it's great. It brings a lot of people to the top of the funnel.
Kelsey Hightower:But realizing over 2 years, if that doesn't convert to dollars
Bdougie:Yeah.
Kelsey Hightower:That BC backed startup model is incompatible with the sustainable open source model, especially if it doesn't lead to revenue.
Bdougie:Yeah. Well, I know we're, now at top of the hour. So anything else you wanted to share with the audience, folks who
Kelsey Hightower:are curious about open source, wanna get involved? No. I think it's I think open source is being challenged now. There's a lot of, you know, AI models that are coming out that say they're open source, but they come with a set of new restrictions that we haven't seen in a long time. And maybe rightfully so for example, Facebook has an open source model, llama.
Kelsey Hightower:And it says pretty much anyone can use this unless you have 100 of millions of customers. Right? So we know who they're targeting. Right? This is gonna be companies as big as they are.
Kelsey Hightower:And so that kinda limits what you can do commercially. Like, you're probably not gonna go find, without Facebook's permission, a hosted Llama model from some cloud provider without a commercial relationship in place. And so now we start to think, is that open source? Is that really open source? Right?
Kelsey Hightower:And I think when people feel that you can't just contribute and then do what you want with the final product, does that count too? So I do think it's one of these things that maybe we need a reminder. Maybe we need to go remind ourselves of the history and why open source exists. And then we gotta decide, do we step up and protect it?
Bdougie:We honestly, we did we didn't touch on your retirement at all, this whole conversation. But how's, how's retired life going to you? You got a boat yet?
Kelsey Hightower:No. That's how you stay retired. Yeah. Boats are liabilities. But I think one thing I've learned is there's no retirement.
Kelsey Hightower:And when you just have this much time, my goal is to buy my time back. And so when you buy your time back, you got 2 options. You work for yourself or you work on yourself. And I've been trying to just do a bit of both. So still advise startups, still speak at the conferences, but also just trying to make sure I figure, like, man, I'm 43.
Kelsey Hightower:I would like to enjoy life and focus on doing that just as hard as I went on learning how to work and all the skills that led to a successful career.
Bdougie:Excellent. Well, wishing you best of luck in this this new journey. And listeners, stay safe. The secret sauce of the podcast produced in house by Open Sauce, the open source intelligence platform that provides insight by the slice.
Kelsey Hightower:If you're
Bdougie:in San Francisco and interested in being a guest on the show, find us on Twitter at saucedopen, and don't forget to check out Open Sauced at opensaucedot pizza.