In a candid conversation at KubeCon London, industry veteran Kelsey Hightower shared insights on his post-Google journey, the evolution of cloud-native technologies, and his philosophical approach to today’s tech landscape.

Watch the whole interview—it is an enlightening discussion with one of my favorite people in the tech world, alongside icons such as Linus Torvalds.

“Retired, Not Tired”

Since leaving Google, Hightower has embraced what he calls being “retired, not tired.” This doesn’t mean inactivity but rather selective engagement. “Not having a nine-to-five job doesn’t mean you’re not doing anything; it simply means you have the freedom to choose the projects you want to work on,” he explained. His current focus includes advising founders, public speaking, and observing technology trends from a uniquely detached perspective.

Reflecting on the Kubernetes ecosystem, Hightower noted he “stopped thinking about Kubernetes five years ago” to focus on serverless and observability. He observed that while Kubernetes was the initial anchor, the cloud-native movement has expanded far beyond it.

“It’s not just Kubernetes,” he said. “A lot of people are working on everything from security to observability and configuration management. It’s really about platform engineering, which continues to grow—and Kubernetes just happens to be one of the foundational pieces, much like the Linux kernel.”

From Open Source Projects to Sustainable Businesses

One of Hightower’s key advisories to open source startups concerns the crucial transition from project to product. “A lot of people set out to build good technology—maybe even good products. But I’m not sure everyone is truly in the software business,” he emphasized.

He outlined a pragmatic approach to open source business models: “From zero to one, open source makes a lot of sense… very low friction. All the source code is available.” However, he cautioned against unsustainable practices: “If we keep giving this away for free forever, it means you’re going to go out of business.”

Instead, Hightower advocated for a balanced approach where companies evolve their models based on market realities while maintaining their core values.

AI: Tool or Replacement?

On artificial intelligence, Hightower offered a refreshingly nuanced perspective. “If you see yourself as an employee who is measured by metrics… you’ll look at the machine as an equivalent,” he said. However, “if you see yourself as more than what happens on a machine, then your views will be different.”

He encouraged a more intentional approach to AI adoption: “What part of this tool makes sense for me?” Rather than trying every new AI tool, he focuses on understanding the technology and identifying its most beneficial applications.

The Philosophy of Intentionality

Perhaps most compelling was Hightower’s reflection on his minimalist philosophy, which extends far beyond material possessions to how he engages with technology and life itself.

“I don’t want to find a useful tool that I don’t need,” he explained, demonstrating his commitment to intentional living. This approach guides his selective engagement with new technologies and his focus on understanding rather than fear or confusion.

Ultimately, Hightower’s greatest aspiration for the next decade is simply “peace” – a state achieved through reasonable use of technology, consideration for others, and the freedom to observe and enjoy technological evolution without pressure.

As the tech industry continues its relentless pace, Hightower’s thoughtful approach serves as a reminder that sometimes the most valuable contribution is not building more, but thinking more deeply about what we’ve already built.

Transcript

Swapnil Bhartiya: Hi, this is Swapnil Bhartiya, and we are here at KubeCon and CloudNativeCon here in London, and we have with us, once again, my favorite guest, Kelsey Hightower. Kelsey, first of all, great to have you on the show.

Kelsey Hightower: Awesome, happy to be back.

Swapnil Bhartiya: You retired, right? So last time, we also talked about your personal life. How is retirement treating you? How are you treating retirement?

Kelsey Hightower: I’m retired, not tired, right? So not having a nine-to-five doesn’t mean you don’t do anything. You just kind of pick and choose what you want to work on. So for me, it’s advising friends that happen to be founders. People still want to hear my thoughts on things. So still public speaking, but for the most part, I get to just observe what’s happening. I don’t necessarily have to be directly involved as much as I used to.

Swapnil Bhartiya: When you were at Google with Kubernetes, you also played a very big role in getting people comfortable with Kubernetes, because Kubernetes was a complicated, complex thing. When I look at Kubernetes back then, and when I used to talk to Brendan Burns and all those folks, they wanted it to be boring. They don’t want people to talk about it. I thought that one time, KubeCon would become like LinuxCon—only hardcore developers would attend. But look at the crowd. This continues to grow. What is driving this?

Kelsey Hightower: I think—I mean, I stopped thinking about Kubernetes five years ago, right? I focused on things like serverless and observability. When you walk around an event like this, it’s not just Kubernetes, right? A lot of people are doing everything from security to observability to configuration management. So I think Kubernetes may have been the anchor, but most people just kind of re-platformed on top. So I think there are fewer people that are maybe new to Kubernetes, of course, than 10 years ago. But I think now the whole cloud-native movement has found itself in all aspects of platform engineering. So I think it’s more about platform engineering continuing to grow, and Kubernetes just happens to be one of the foundational pieces, just like the Linux kernel. That has become a foundational piece, but people are building things on top of that.

Swapnil Bhartiya: 100%, yeah. So you’re saying that you advise your friends and you do speaking. What are the core themes of those discussions nowadays?

Kelsey Hightower: I think for all VC-funded startups, they have to come to the realization, especially in the world of open source—I think a lot of people set out to build good technology, maybe even good products. But I don’t know if everyone’s really in the software business. That’s very different. That is enterprise sales motions, procurement, marketing, and I don’t think a lot of technologists realize that. To make the pivot from “this is a really cool open source project with a lot of GitHub stars” to business—these aren’t necessarily the same things at all. So I think a lot of the advisory work is that: how do you go from project to product? And for a lot of people, that’s a big shift, because you actually have to talk to customers. You can’t just put up a pricing page and hopefully they’ll come and figure out what to do next. They need support, probably professional services. These are pretty hefty projects that require integration work, and so a lot of the work is really about maturity—meeting them where they are, and then moving them up the maturity curve so they can actually understand what it’s like to be in the software business and not just a cool project.

Swapnil Bhartiya: Since you mentioned open source, what happened? I’ve seen that a lot of companies—I don’t want to name any company, but you know the founders of those companies. They have that vision. But as the company grows, as more people come in, CEOs are replaced, they sometimes forget the roots of how and why they started the company. How do you advise folks? If you are going the open source route, go this route, or just go proprietary route altogether?

Kelsey Hightower: No, you will forget your roots, because you will have to do business. Giving away software for free is not the root of business. You will be out of business guaranteed. So I think, how do you get started? Zero to one is very different from one to 100. So zero to one, there is an advantage to having something that’s freely accessible in terms of price. Maybe open source is the right way to get started. Like, why would someone want to use your project when there’s a million other projects that already exist? So if you have a price tag for a very immature thing that’s just starting out, I’m not paying a price for an alpha. So I think the challenge is open source is a great way to raise awareness, get mind share, build a community around it. Very low friction. All the source code is available. People can come help out. So in the early stages, zero to one, open source makes a lot of sense. You just have to be clear that that is not sustainable. Going forward, I can’t give the whole thing away for free and hire people to do support and documentation. The math just doesn’t work. If we keep giving this away for free forever, it means you’re going to go out of business. I think over time that business reality sets in. And so now if you don’t start off with a pricing page, you have to add one. If you don’t have a marketing team, you will add one. So I think it’s not really them getting away from their roots, it’s just that their roots only got them so far, and now they have to go to the next stage. And so sometimes you’ll see the CEO get cycled out, maybe for someone that can go from one to 100, but that same CEO probably wasn’t the right person to go from zero to one.

Swapnil Bhartiya: But there are cases where companies totally abandoned their open source community by changing the license. So what you’re saying is the trajectory—the open source and business like Red Hat model is an established model. SUSE’s model, established model. We can say Canonical’s model is also an established model. I don’t want to name any companies, but you know who I’m talking about. How to avoid that pothole?

Kelsey Hightower: To think that no company will ever evolve based on the current situation is kind of insane. So let’s say your software was free and people could pay for enterprise features, and you did that for 10 years. Great. Year 11 comes and a competitor takes your free thing and offers more features than you or a managed service. So what do you got to do? You got to do something different. You can’t say, “Well, we can’t abandon our roots. We’re just going to go out of business so we can remain pure.” It doesn’t work in business. You have to make decisions when you hit that crossroad. And you can imagine the empathy involved if you are an employee, or you’re employing 1,000 people—they rely on you to make good business decisions so that revenue flows in. And so when you get to this crossroad a decade later, you’re going to be forced to figure out, how do I make people pay for something? One avenue is to stop giving away value, and I think that’s really hard to do after 10 years. If I’ve given away 98% of the product, then there’s only 2% enterprise features. Either I have to have a lot more new features that require payment, or I may have to do weird things like redo the license. I think going forward, everyone has kind of a good understanding: don’t give away the whole product ever. And I think what that means for new open source companies is maybe the software or the source code is open, but maybe the release binaries are not open. You got to download those from my website. Maybe you have to give me your email address in the beginning, then maybe new versions are going to require payment, and you separate the product and the project. And I think we just learn now, right? So now that we all know better, maybe we have to do that to be a little bit more clear about the commercial relationship to the open source project.

Swapnil Bhartiya: Yeah, I remember Eric Raymond wrote “The Cathedral and the Bazaar.” We tend to forget those things. Now, I want to change the topic, and I want to talk about something hard and not hard, which is AI. We had this general, brief discussion. What are your thoughts on where we are heading with AI in general?

Kelsey Hightower: I’ve had this discussion a lot in the last six months, and I think it does boil down to how you see yourself. If you see yourself as an employee that is measured on metrics—how many emails you process, how many sales leads you close, how much source code you write—so if that’s how you see yourself in the mirror, you will look at the machine as an equivalent. You will look at this thing as a partner. You will look at this thing as a replacement, because you’re only limiting yourself to what a machine can do, right? So for some people, that’s what they’ve been doing for the last 15 years. They’re just a person operating the computer, and they have never done anything more than the computer will allow them to do. So for me, when I think about software development, it starts before you touch the keyboard: creativity, whiteboard sessions, design. You’re watching how people work. You’re observing the world. Then you build. But some people don’t have that flow. They just sit behind the desk, get assigned a task, and then they become one with the machine. So if that’s your viewpoint, you look at AI as like a replica of yourself. If you see yourself as more than what happens on a machine, then your views will be different.

So for me, I’m like, wow, people are very excited about artificial intelligence. What would it look like if we pour a trillion dollars into real intelligence, real people, right? You see a lot of people saying, “Oh, it’s gonna require all of this power to feed the machines. We gotta give them all this data so the machine models don’t starve.” But you say, “Well, what about like people?” There are people that don’t have enough power. There are people that don’t have enough food to eat. And so it’s kind of weird where artificial intelligence gets the spotlight when the people who create the actual data set that feeds the machine, which is normal people doing creative things, and all their actions become data sets that feed into the models. Sure, these things will create maybe new things, but if, in my mind, anything that a machine could do is because we taught it, and so to me, I see humanity as something more than a person who operates a machine or who wants to live their life behind a screen.

And so there are parts of AI that I think are going to be amazing. For example, when it comes to elderly care, this is like someone who’s maybe later stage of life, they may have dementia, they don’t even recognize their family anymore. So you can imagine how hard it is when a family member tries to assist them 24/7, that means they’re gonna have to give up a portion of their lives to sit with them and be with them around the clock, because they do need someone there. And so when people start talking about humanoid robots in that use case, that makes a lot of sense for me. How do we give people quality care around the clock in a way that allows everyone around them to also live their lives? So that’s one where I would say, look, I step back and say that might be a net benefit that is a good use of technology to fill a very painful gap that we haven’t solved completely. But the idea that we use the same AI to do all the creative stuff, to write the music, to make the movies—maybe we can reserve something for the humans to do as well.

But more practically, when I look at the low-level technologies, things like MCP, where we give our tools semantic meaning, we give them a description, and in some ways, that level of description makes it easier for humans to realize and discover what’s available, but then it allows an LLM to do more than just generate content. Maybe it allows the LLM to start operating our tools for us. So in some regards, it’s a really nice automation framework that gives us a new way of interacting with our existing tools. And as someone who’s seen technology evolve over time, I kind of welcome that next step, and then it’s on us to see what we can do with it.

Swapnil Bhartiya: I mean, if you look at the whole industrialization, the same thing, same fear, was there that it will replace humans. There were a lot of conflicts between machine versus human labor. But then if you just look at it, there are a lot of industries. There are a lot of jobs which are risky for our lives. Why should a person risk their life to do the job where the machine can do it safely? But are you seeing that right now, the way every company, everybody’s obsessed with AI to just get into everything—is it a temporary phase? Over a period of time, will we realize that, no, it is a tool, not a replacement? Or do you see it as an existential threat to humanity in general?

Kelsey Hightower: It’s going to be up to people to decide. Some people, I remember when Apple launched the Screen Time app, what makes someone want to write an app to track how much screen time? Maybe you get to a point where the tool in our pocket became an addiction, and some people don’t even want to talk to anyone else. They just want to—even in the same room, they don’t want to text the person instead of making eye contact. So it’s very clear that technology always has that potential to be just a tool that we reserve for use when we need it, or it can become an addiction. And I think when you look at things like social media, mobile devices, the risk for people to over-consume is really high. And so when you think about AI, in this particular case, we’re literally personifying the technology. We are referring to these things as, like, reasoning, right? What happens when they become emotional, right? Like, when we can’t discern the difference? And then also, why are people pushing for that? So if that’s where some people are pushing it, what happens when they get there? And I think that’s just a thing that humans are just going to have to ask themselves: just because we can, should we?

Swapnil Bhartiya: There was some study also a couple of weeks ago—I don’t know if you know that people who rely on these chat GPTs, and they’re getting emotionally attached. There was a movie also “Her,” if you remember, where he was talking—people got attached. They’re like, people who use it for emotional support because people treat them as a friend. You just sit on there and chat. And that can be—and I mean, I have two kids, and I worry, because they’re addicted to their phones. They’re on the Pokémon Go, versus actually interacting with the world. So yes, you’re absolutely right that the fear is there. We don’t know how things will shape up, but we will see. Time will tell. And I can ask—right now, if we just get out of open source technology, what keeps you excited, whether it’s technology or people? If it’s technology, what is the next thing after Kubernetes or platform engineering, as you mentioned, that is keeping Kelsey excited? Hey, this is what I’m looking forward to the next five or 10 years.

Kelsey Hightower: Yeah, I think I’m looking forward to the next five to 10 years is peace. I don’t need to be excited anymore. I do get excited from time to time, but I don’t need to be excited. I’m at the point now where it’s like, if you can tell me that I’m going to have peace in five to 10 years, that’s exciting, because in order to get peace, that means people have to be reasonable. That means we have to use the technology responsibly. We have to consider others. And so in that world, you can have peace, and then if you want to get excited about something, you can always jump into a piece of technology, and build something, add new features, fix bugs—that will hopefully always be an option. But the thing that makes me excited the most is when people can find, like, peace.

Peace for me is like, you know, I retired. Everyone hits their number. You walk away from the daily grind of a nine-to-five, worrying about the stock price, worrying about the feature set, and so to be able to walk away from that and then observe technology without those kind of pressures, that is pretty peaceful. I can look at the technology in a certain way. No one has to win. I don’t care if this one wins or that one wins. I can just enjoy the game. It’s like watching the championship game, and I’m not necessarily rooting for any team. I can just enjoy the match. That means I can watch the game peacefully. If you pick a team and that team isn’t winning, it’s very stressful to watch that game. So when I say peace is one of these things where peace requires balance, and so for me, having the option to choose who I work with, choose not to work at all—that’s a really great place to be, and so the more people who get to experience that, that’s super exciting, because a lot would have to happen for that to exist.

Swapnil Bhartiya: So when people look at Kelsey—I mean, people always read you as a technologist. If you look at today, are you a humanist or technologist? Who are you?

Kelsey Hightower: I don’t know. I just try to be honest. I try to be real. I try to be transparent. The good thing is, I don’t have to be anything, right? And so I think now it’s just more of—most people meet me and I just explain the journey, but they want the nuance. How did you feel in a scenario where you were unsure? How did you feel when you took a two-month vacation and you weren’t sure if you would come back to be irrelevant because you were gone too long? How do you feel when the technology you bet on is no longer the top thing anymore? And so it’s that nuance where people just want guideposts. They just want someone to talk to about those things. So I would say I’m trying to maybe head towards philosopher, if you get good at anything long enough, you can observe it in different ways. You find different ways to articulate it, and then you form your own philosophy. So these days, the technologist evolves into the philosopher, and then you try to get people—maybe, if you have it yourself, you can share that wisdom with other people.

Swapnil Bhartiya: I remember early times when I used to talk to you. You used to call yourself a minimalist. I am a minimalist, so I think that your philosophy toward minimalism is not just about your house, but I think towards life itself. And if you can share that vision—what minimalism means not just, oh, I have less furniture in the house, but what does it mean for life?

Kelsey Hightower: For example, you see the AI wave is very exciting, and you can try to do what everyone else does. You can follow all the new model releases. You can download them all. You can try Cursor. You try this, you try that. You got every—you don’t want to miss anything. So you’re just all over the place, right? You even have alerts whenever something comes out in AI news. You want to subscribe to everything. You watch all the channels. You don’t want to miss nothing. As a minimalist, I’m more intentional. I would be way more intentional, I would say, for my life and my lifestyle.

What part of this tool makes sense for me? There are problems I don’t want to have, right? So if you show me some AI that organizes your email, because you have so much email, I don’t want that problem, so I’m not interested in AI that organizes email better, because I never want to have that problem. So it’s living with intentionality. I don’t want to find a useful tool that I don’t need. So it’s more like that. So I look at it, and so when I say intentionality, I look at all the people building great things, and say, “Wow, what would this be the most useful in society?” I have parents that will be aging. Will this be helpful for them? How? How can I leverage it? How can I influence it? And so then it helps me guide my attention. You know what? Maybe I pay attention to this part. Maybe I pay attention to this part. And I’m also one of those people that says, “Look, I don’t understand it.” I think I maybe understand a lot more than I did last year. And so when I talk to people, I ask them, “Do you know how it works?” And most people in this space, they know how certain layers work, but maybe not the next layer. Then I go search for the next person. “Do you understand how this layer works? Write it down for me,” because I do think what’s going to be very important during this time phase is that, who adds a window to the black box? A lot of the fear, a lot of the analysis, is literally based on lack of understanding. So we’re all just making things up. We’re repeating things we don’t know is true or not, and then that just kind of maybe fills the hype cycle a little bit more. But I’m one of those people like, “Hey, I’d rather think about this stuff from a position of understanding than fear or confusion.”

Swapnil Bhartiya: Well said, excellent. I will drag you back to the tech world, because when we talk about all this, AI, Gen AI, I mean, the fact is that it’s very resource-hungry. I mean, I can run it on my laptop also, but not at that level. I can write a chapter on it. And when we look at hyperscaler, when we talk about open source, are you seeing—I mean, does that get back to open source, where these things will become more democratized, not too much tied to resource-hungry, power-hungry technologies?

Kelsey Hightower: They’ll get better. People figure out better ways of doing it. Maybe they become the baseline—maybe they become like a little library you use to give your app natural language processing. Maybe it’s something you use to generate support commentary on a ticket. I think it’s—I mean, technology tends to do that. When we figure out how to do something, people will move around and share that knowledge and create competition. That’s just the normal flywheel of innovation. There’s no way one company is just going to dominate forever. You may enjoy a three to six month lead, but forever, I don’t know about that. I think now it comes down to what role do you want to play in all of this? As an observer, when the tools get mature, I’ll use them as necessary, but I think as someone that’s standing back and just watching other people surf the wave, I’m happy to sit this one out for now. Yeah, I don’t feel like I need to be fully ingrained in this, like I was with Kubernetes. I need to pick and choose the waves that I want to surf. And when I tell people this, they get a little amazed, like, “Hey, this is the biggest wave ever.” It’s like, “Are you sure about that?” There’s waves in medical science right now. We’re curing some diseases that people thought they would never have a cure for. There’s breakthroughs in energy right now. There’s all kinds of things happening around the world. So most people don’t surf all the waves, either. They ignore most of them. And so for me, I’m—this is part about being intentional. My intentionality looks at this AI wave and say, “You know what? I’m gonna let this one settle down a little bit, see what comes out of it, and I just stick to the fundamentals.” Is the thing underneath the AI changing? For example, if someone will say, “Hey, Kelsey, this is cool AI thing we’re using to drive Kubernetes.” I was like, “Could you show it to me?” It’s like, “Great, you can now chat and ask Kubernetes to deploy an app for you.” I said, “Okay, that’s, I guess that’s pretty cool, but it’s still Kubernetes.” Like, “Yes.” “Can I see what it generates?” “Yeah, it looks like the thing that I still understand.” So the fundamentals aren’t changing here. It’s like, “No, we’re using a new tool to drive the fundamentals.” I feel safe in that. So the way I look at it is the way people get excited about the new models. I get excited about my own model. When I’m learning new things, I’m improving my own mental model. If I learn how to cook a new dish instead of buying it from a restaurant, I’ve improved my own capabilities. When I learned a piece of new technology, I improve my own capability. So imagine that people walked around and thought about their own unique one-of-one model. You have a context of the world. You have your own worldview. At this very moment, we’re pulling in new training data, and we’re deciding, do we retrain ourselves and become a different person over time? So I’m looking at this time period as like, “Man, maybe I should approach the same intentionality that maybe a company like OpenAI is doing for their model to my own.” So now I’m like, being more aware of the things that I’m putting into my mindset, and then I’m taking notes a little bit more, and I’m asking myself, like, “Did I improve?” So that’s the way I look at it. Now it’s like, “Hey, artificial intelligence, I don’t think there’s anything wrong with real intelligence, and I can improve my model, and I can’t remember everything, so there’s nothing wrong with reaching for a calculator from time to time.”

Swapnil Bhartiya: And, yeah, I mean, since birth—not just humans, every animal—we have our own training data. I mean, some are hardwired right from the very beginning, and we evolve. And that’s why we read, that’s why we watch. Now we are more in the consumption phase, not in the learning phase. We don’t read books anymore that we used to earlier. Anyway, so when I look at it, you look like I think you are becoming less often engaged, more of an observer.

Kelsey Hightower: Yeah. I mean, I think I was always an observer. When I used to build things, it comes after the observation. I’m not just building, I’m observing the world and deciding to create based on that inspiration, or I see a need, then I build. And so whenever I got involved in any piece of technology, it was based on how I observed the world. Some people say they’ll ask me a question. “I want to get involved in open source project.” I’m like, “Why?” “So I can make a name for myself