The Change Resilient Podcast Episode 3: Chris Aniszczyk on the Power of Open Source for our Current AI Times
Chris Aniszczyk is an open source technologist with a passion for building a better world through open collaboration. He's currently a CTO at the Linux Foundation focused on infrastructure and developer experience while running the Cloud Native Computing Foundation (CNCF).
Furthermore, he's a Partner at Capital Factory where he focuses on mentoring, advising and investing in open source and infrastructure focused startups. He has a strong passion for supporting open source focused founders and entrepreneurship.
For disclosure, Chris is an investor Autoptic.
Listen on Spotify Apple, Amazon, YouTube, and wherever you enjoy your podcasts.
Links from show references below:
Cloud Native Computing Foundation
Finite Games, Infinite Games (James Carse)
Show transcript (slightly edited for clarity, machine transcribed, apologies for errors):
[00:00:03] Steve Semelsberger
Hello everyone, welcome to another episode of the Change Resilient Podcast. My guest today is Chris Aniszczyk. Chris is an open-source technologist with a passion for building a better world through open collaboration. He's currently a CTO at the Linux Foundation, focused on infrastructure and developer experience, while he also oversees the Cloud Native Computing Foundation, the CNCF, which we'll have a chance to talk quite a bit about today. Chris is also a partner at Capital Factory here in Austin, Texas, where he and I both lovingly reside. There, he focuses on mentoring, advising, and investing in open source and infrastructure-focused startups. Chris has a strong passion for supporting open source founders and entrepreneurship as a whole. For disclosure, Chris is an investor in my company, Autoptic. Chris, welcome to the Change Resilient Podcast.
[00:00:52] Chris Aniszczyk
Awesome, glad, glad to be here.
[00:00:54] Steve Semelsberger
Yes, I'm happy to have you as well, too, because you are traveling quite a bit. You mentioned that you have a whirlwind series of adventures coming up. Care to share some of the places where we may find you?
[00:01:05] Chris Aniszczyk
Yeah, I mean, totally. We have, KubeCon in Asia next week, in Shanghai, so I'm kind of going to be off to that, where we're hosting an event for, obviously, our Kubernetes community, but also our AI community out there called AgentCon, and also PyTorchCon, so it's a bit of a. busy week there, and then I'll be off to, Europe, kind of a few weeks after that for our AI event out there, so it's, gonna be a crazy, few weeks for me, but I'm always excited to be at home and spend time with. Local, local founders like yourself that are working on interesting problems.
[00:01:42] Steve Semelsberger
We appreciate it, Chris. Austin's gain is the world's loss, and vice versa. You know, as I know you are. but enjoy your time here. So, let's dive in. So, Chris, you've spent a big part of your career inside communities that have experienced a ton of change. You were early on at Twitter, you've been a key part of the Linux Foundation and the CNCF for more than 10 years now. So, when you look back, what originally drew you to open source, and what's kept you working so diligently on open source-oriented initiatives for the better part of a decade?
[00:02:18] Chris Aniszczyk
Yeah, so it's a, it's an interesting question. You know, I'll start with the, what drew me to open source in the beginning is I think I just always had the ability to want to tinker with things. My dad was. a mechanic, exposed me to computers, and, you know, systems at an early, you know, age. Eventually. you know, I found, you know, something called Slackware, you know, at the time, which was one of the early Linux distros, and started to kind of play around. with Slackware, build my own computers from scratch, systems. Mostly at the time was like, hey, how can I kind of, you know, build things to learn from them, and also maybe play video games like, you know, any other kid. you know, wants to do when they're interested in video games, but eventually I got exposed to Slackware, the Linux community there, and eventually a Linux distro called Gentoo at the time. which is all about, kind of building things from scratch, but, you know, I think the memory I have from the early days that early kind of, you know, keeps me involved in open source is kind of the spirit of wanting to help. folks out with, like, no expectation in return, truly for, like, the ability to share knowledge and learn and help each other. And so, in the early days, I had issues where. I think I got, like, a new 486 box, was trying to go set it, set it, you know, up probably, but I couldn't get freaking the sound working at the time on Linux. The ALSA drivers, if I recall correctly, and I jumped on. you know, IRC, which was, like, a very popular chat, you know, network back in the day to kind of talk to folks, went on forums, and all I remember is, like, there was probably, like, hundreds of folks. in, in kind of the, the Gen 2, you know, you know. Channel that basically walked me through and helped me. you know, get sound, you know, working over the course of, you know, whatever, a few hours, and really expected nothing in return but the sharing of knowledge. And I kind of just that, that spirit of learning, sharing is something that kind of always, you know, stuck with me. And, you know, I've been through a variety of different communities, from early days, you know, Linux work, to. you know, Eclipse, to Java, the JVM, like, all sorts of things, and kind of that spirit of sharing, you know, helping folks out has kind of always been there, even if people are working in, you know, overlapping or competing industries, especially today, where open source is so prevalent. So that's kind of the original thing that kind of got me hooked and. You know, got me really deeply, you know, involved in not even, you know, starting my old. company a long time ago that was around, kind of, open source development environments with, productionizing and scaling Eclipse, related things, but, you know. these days, it's like, I just find no better place in the world to kind of learn what is actually happening, kind of, from an innovation point of view. I feel like open source communities are almost always at the forefront of trying to make stuff, you know, work in interesting, you know, ways, and especially with my current job, like, the amount of stuff that, you know, I see across the world in different organizations happening, and open source projects is just. amazing. So I just, just love there for, like, the helpfulness, the, the curiosity that, you know, Curious Height Itch That Scratches for me, and just kind of truly learning what's best and greatest out there is kind of how I. I love it. So I love, love making a career out of it, especially. It's been kind of fun. It's nice to see open source not being something that is, like, stereotypically, that's, you know, someone working on in the basement, where it's, like, literally just powers the whole world, everything that we're, you know, using from, you know, phones to cars, infrastructures, it's all powered by open source projects like Linux, Kubernetes, and countless, countless others.
[00:06:03] Steve Semelsberger
Countless, countless others. We're gonna talk about that in a minute.
[00:06:05] Chris Aniszczyk
Yep.
[00:06:06] Steve Semelsberger
Glad to give a shout out to IRC. My friend Andrew Busey wrote a book way back in the day called The Secret of the Mud Wizards. He started iChat. That was my first Austin company in 97.
[00:06:15] Chris Aniszczyk
I love it.
[00:06:16] Steve Semelsberger
So I tracked you there, Chris. Thanks.
[00:06:18] Chris Aniszczyk
Love it.
[00:06:18] Steve Semelsberger
My gosh, now, CNCF is dozens, hundreds of projects? Give us a sense of the magnitude of CNCF today.
[00:06:30] Chris Aniszczyk
Yeah, so CNCF, we're a little over, you know, 10 years, 10 years old now. You know, we started, essentially. you know, with the simple premise of providing a neutral home for Kubernetes and the cloud-native computing, you know, movement, right, which essentially was the way that web-scale, internet-scale companies were running their infrastructures, you know, the folks like Google, Netflix. Facebook, Meta, etc, how they were scaling their systems, and kind of trying to build an ecosystem, and making that accessible, and democratizing that technology for, for everyone. And so, in that short 10 years or so, we've grown to, I think we have a little over 300,000. contributors, globally, representing, I think, almost, like, 10,000 or so organizations worldwide, about 250 open source projects, so not, not only just Kubernetes, we host, essentially containerd, which is a form of the Docker engine, which runs the majority of containers out there. We have things like OpenTelemetry, which you're very familiar, you know, with, which has kind of become the de facto. quote-unquote Kubernetes of the observability, you know, world, just standardize how telemetry collection. you know, metrics and all that stuff is done. So it's a lot of very innovative, you know, projects. So yeah, it's been a crazy journey in kind of taking those principles that a lot of the internet-scale companies pioneered to make it work for them, and democratizing that technology and making it available. You know, for everyone, and we've had a bit of a second wind recently with kind of the AI boom, where, you know, all AI has done is actually made it. made the problem of running distributed systems, you know, properly and in a scalable fashion even more so important. So it's just been a fun journey, to say the least.
[00:08:20] Steve Semelsberger
Has the proliferation of Kubernetes surprised you?
[00:08:24] Chris Aniszczyk
Oh, so in the beginning, as someone who did not work on Kubernetes before in a distributed system sense, I started really cutting my teeth on, well, there's Linux itself, but Mesos was a technology that we built and pioneered at Twitter that we collaborated with the Berkeley AMP Lab, you know, at the time, and that was all about, like, how do you basically go scale distributed systems across multiple. computers and data centers. And so we kind of built our. Approach at that time, versus the rest of the industry wasn't. Google wasn't really sharing much of how they, you know, did stuff outside of maybe some research papers they've written. All the other companies weren't really interested in doing that, so I cut my teeth. Mesos, and I understood the value of. part of the reason we did something like Mesos is because once you're, once you scale to a certain level, things kind of get expensive, whether it's, like, storage or, you know, caching, and if you're not properly utilizing, your, your data center and machines that you cut up, like, if you're like, oh, this service is only using, like. you know, 20% of, you know, the RAM for caching, and it's only using, you know, maybe 30% of the CPU, it leads to an extraordinary amount of waste if you can't actually reuse and slice. You know, workloads to particular aspects of a machine versus, like, partitioning out, like, here are just my data, here are just the machines for databases, here are the machines for caching, here are the machines for, you know, this task. It's a very inefficient model, so it's all about kind of, like, slicing and bin packing for kind of a technical term of workloads. That methodology, I knew, was always going to be a way of how most of the world is going to run workloads as more and more internet. you know, services, you know, come online and have to scale to support all these users. What I was not aware that I, I thought Mesos had architected a little bit better with its, kind of, two-level scheduling approach and so on. I thought it was a little bit more advanced. Kubernetes, had a lot of great ideas, you know, in it, but it was actually not the technology that Google itself used internally. They had something called Borg, which is actually a little bit more similar to Mesos. Kubernetes was a inspiration, based on how Borg was done with some kind of newer ideas baked into it. I think what Google. did and was prescient about is they understood it would take an ecosystem, and it couldn't just be Google only, like, they kind of had to get partners and other organizations involved and truly build it, you know, for something that it's more than just. you know, itself, and that kind of surprised me a little bit of, you know, how quickly Kubernetes, you know, grew and kind of took over, you know, every major. hyperscaler private cloud basically offers Kubernetes as a product, as a service, essentially. Not even counting companies that are, you know, building on top of Kubernetes, so it's hundreds of folks out there, maybe even thousands now. So it's like, the concept of, like, bin packing. workloads in an efficient manner, that growth didn't surprise me, but the fact that Kubernetes won out so handedly against Mesos, Docker Swarm. you know, Slurm, like, all these other kind of. resource managers. That definitely has shocked me quite a bit, even to this day.
[00:11:41] Steve Semelsberger
KubeCon is an amazing experience. Shout out to what you and the Linux Foundation do with that. Encourage anybody who can attend, whatever continent, to go. I had a chance to go last November, it was awesome. And, yeah, during my day job, I talked to engineering leaders, infrastructure platform, DevOps leaders, and Kubernetes is almost expected at this point. It's been really cool to see. So let's continue to play this game of, you know, maybe what's surprised us, and let's invert it, so. look back, if you can, 10 years ago, and think about some of your beliefs from then. Is there something that you thought then that you no longer believe now? Maybe about open source in general, maybe about the world, however you want to answer that question.
[00:12:27] Chris Aniszczyk
Oh, yeah, interesting. So, maybe there's two things here. One is. I used to, I used to think that. I think most people, most, like, let's call it enterprises, customers out there, want a very strongly opinionated stack. At least they, that's what they claim they want. It's like, they're like, oh, I want to go buy, you know, DevOps, for lack of, like, I want to buy this stack, right? And I think people. crave and want simplicity, and I thought, like, initially, in the beginning, my role and kind of the job was to go provide a lot of that simplicity, but I think what really people want is a base platform that gives them some. level of guarantees that they're not necessarily fully locked into, but the ability to have the flexibility to tweak it maybe to their unique enterprise. And the best way to kind of. you know, coaches, like, there's a reason, like, if you're familiar with a technology called, like, Cloud Foundry, you know, for example, like, had a very strongly opinionated platform, but eventually lost out in the long term because. they basically had just one way of doing things. This is how you do storage, this is how you do networking. So I think my lesson is, like. enterprise vendors always crave simplicity, but I think, but at the end of the day, they want flexibility and optionality to kind of tweak something in the platform, so it's something that I've kind of chained my mind on. Like, I think, like, if you truly want to build a technology that has. global-wide adoption. Simplicity is necessarily the wrong approach. You need to provide enough base. And, you know, maybe good developer experience, but also the flexibility to kind of tweak things, because enterprises environments are not, you know, they're not homogen, like, they're very different at the end of the day. So, that maybe is a bit. you know, a bit controversial, but I think most enterprises crave some flexibility, even though they claim to always want something. you know, simpler, and so that's, like, one thing based on, kind of, my experience. The other thing is, like, I used to always think, like, hey, if open source, open source is enough, if you just have code out there, it's gonna be great, and, you know, we'll kind of live forever, but I think, what I've learned is open source is always not enough to kind of, like. guarantee. true neutrality or understanding that code may be there forever. We've had some cases in the industry where I thought, like, oh, you know, HashiCorp. has some great software, like Vault is amazing, right? Redis had some good stuff, it was open. You know, when you kind of have purely single-vendor open source. there's a risk that, you know, folks may change licenses, and I think a lot of people, a lot of enterprises don't always understand that thing. We've had a couple industries where, there are a couple companies in the industry that has happened, where, like, HashiCorp changed licenses to try to maybe extract. more revenue out of their customer base, and so it upset people. So I think open source itself is not enough to build a truly neutral community. You're gonna have to, like, pair that with, like, open governance or multiple, you know, ownership to keep people fair. So those are maybe two. Things that, you know, my opinions have kind of changed over the last. tenure, and I, I have some, some, some others, but maybe we'll, we'll kind of, we'll kind of go there to.
[00:15:56] Steve Semelsberger
Work them in later. I'm hopeful.
[00:15:57] Chris Aniszczyk
We'll work them in later, and I guess the other, the thing with that latter point is, like, when those types of actions happen, you. once you, once you earn developer trust, it's amazing, right? But if you make a change, and once you lose that developer trust, it is extremely hard to get it back. Like, I always thought it's, you know, developers, you know, they love you if you kind of work with them, but if you break that trust. once, it's extremely hard to get back. Like, HashiCorp used to be one of the most beloved developer companies, probably top 3, you know, they change a license, everyone just. like, threw them out, like, we're forking Vault, we're forking, you know, Terraform, so it's, developer trust is very fragile. Once you earn it, you really have to try to, to, to keep it, in my opinion.
[00:16:51] Steve Semelsberger
So, as you know, the podcast is called The Change Resilient, and we think a lot about resilient systems, systems of a variety of different types, inclusive of software systems, of course, and a few days ago, we're recording, Chris, you and I, today, August 31st, 2026. So, a few days ago, I talked with John Allspaw, who's the founder of Resilience in Software Foundation, and is a deep practitioner of resilience engineering. quite scientific. You know, very precise, loved my conversation, excited for our listeners to explore that one. For you, Chris, given the purview that you have of the industry, what do you think makes a software system resilient?
[00:17:33] Chris Aniszczyk
That's such a, such a broad question. I'm actually excited to listen to that podcast, too, once it's out.
[00:17:41] Steve Semelsberger
We talked as much about skateboarding and music, I think, as they maybe Resilient Software System, so you and I might have to find some other subjects here.
[00:17:49] Chris Aniszczyk
Oh, that's cool.
[00:17:50] Steve Semelsberger
John was wearing a Mike McGill 80s skateboarding t-shirt, so we had to just....
[00:17:53] Chris Aniszczyk
Launch into that.
[00:17:54] Steve Semelsberger
But tell me, Chris, you know, what do you think helps systems? You know, you're talking about flexibility, adaptability...
[00:17:59] Chris Aniszczyk
Yeah...
[00:18:00] Steve Semelsberger
What are attributes of resilience that you look for?
[00:18:01] Chris Aniszczyk
Yeah, I mean, for me, like, inherently, I live in the world of, you know, distributed systems, right? And these things are complex, there's usually multiple failure scenarios that you can't, you know, predict, and I think the. the things that make systems generally resilient are, you know, there has to be some level of transparency and observability built in them so you kind of know what is going on, and more importantly is, you know, no system is perfect, right? And so there will be. mistakes that happen, you know, in systems, usually caused by humans, but these days, they could actually be caused by agents that may not know what they're actually, you know, you know, doing, potentially. And so, when that happens, you know, what do you do, right? And, you know, I think in my. experience, like, where I even did, I did, you know, I did some, you know, made some mistakes in my career multiple times regarding, like, you know, systems that I was responsible earlier in my career, and, you know, I remember there was a time at. you know, Twitter, where, you know, there was a certain incident that happened because of, like, a very, like, simple mistake where there was a variable that, you know, a Boolean, essentially, that was about, like, publishing. you know, binaries to a certain place, and, you know, by default, like, someone, it was never, like, typically in part of our build process, and so the value was, like, always assumed, you know, false, but then eventually, like, someone did a. documentation, you know, change, and, like, added, you know, the, you know, the variable there to, like, publish, you know, everything, you know, everywhere, and that caused, like, a major incident, and so. Very, very, very hilarious one that we'll have to maybe dive into other, but what happened in that case is, like, you know, yes, we all, you know, went and solved the problem at the end of the day, but after that incident. we essentially, you know, even though I was the guilty party for making that change, we went through essentially what is known as, like, blameless post-mortems, right? So, like, here's what happened. you know, why did it, you know, why did it happen? It's not anyone's fault. What can we do? What did we learn from this? What can we do so this doesn't happen the next time? And let's go codify that knowledge into our systems and designs and share it with everyone. So, I think. Systems, you know, especially from, like, the perspective of, like, humans and cultures, is, like, having that culture of, like, learning from. mistakes in kind of this blameless way, where people feel, A, safe to share and fix things in a fast, you know, way, is truly important, in my opinion, in truly building resilient, you know, systems. I think technology, for sure, is important, but, like, setting that. system up, so when things clearly go wrong, they could kind of be self-corrected, and that knowledge is shared and disseminated as much as possible.
[00:21:01] Steve Semelsberger
Transparent, flexible, learning-centric, but blameless. Are there other organizational and even human attributes that help, you know, the people-process-systems around the technology, but build more resilient capabilities? Anything further you want to add to that?
[00:21:21] Chris Aniszczyk
I mean, not, like, you know, I think not siloing teams, you know, this whole, like, original how DevOps started, like, you know, essentially. minimizing walls, being able to share, you know, information. I think generally some of the things that the cloud-native movement pioneered along with a lot of the DevOps movement was about, like, hey, you know. you know, codify things into, you know, traceable, auditable, you know, artifacts. There is no, like, silos of, like, taking something and chucking it over the wall to a release team. Like, just low barriers, high fidelity information, across the board, I think, is good. Like, treating folks like they're all on the same. You know, teams, so it's not like, oh, like, you know. we built this, but it's really the release team that screwed up, or the security team that screwed up. I think that's this whole. culture of sharing, learning, and blamelessness, I think, goes along, you know, goes a long way. And of course, like, the technology has to support. observability, traceability, auditability, and all that stuff, because if you don't have that, it's very hard to, you know, even if you have the culture right in your organization, which a lot of times is not there, if you don't have the technology even, it's just impossible. From my experience. And, you know, these days, I come across a lot where folks, like, have Kubernetes rolled out, you know, pretty well, they have pipelines instrumented in a good way, but culturally, their organization is still stuck in kind of a. you know, let's call it a prior generation, almost like, you know, almost waterfall-esque, you know, style from us who have been in the software industry. So, it's, I think it's truly technology paired with good. you know, cultural, you know, practices around transparency, breaking down, you know, walls between teams in terms of, like, information hiding and siloing, and then, you know, of course, paired with, like, blameless things and lessons, lessons learned. So I, one of my. Favorite kind of interactions you know, in kind of the early days of CNCF was. early days Kubernetes is a pain in the butt to, like, roll out. There was all sorts of issues, because early days technology, I had a friend over at Spotify at the time, David, where he, you know, they had this whole, like, blameless, you know, you know. culture, but they kind of codified it into something they called, like, Kate's Fail stories, where they would just, like, share these stories internally across their whole organization, but they also shared them publicly to the world, too, and they weren't embarrassed about that, where some, sometimes the mistakes were truly silly, but other times they're a lot more complex, so just that willingness to. you know, share things without worrying about, you know, you losing your job or looking foolish as an organization, I think goes, goes a long way to kind of build resilient, you know, systems.
[00:24:20] Steve Semelsberger
Take the situation super seriously, don't take ourselves so seriously, it's hard as humans to do.
[00:24:26] Chris Aniszczyk
Good way to summarize it.
[00:24:28] Steve Semelsberger
So, platform engineering. For many of our listeners, this is a subject that they're deep. practitioners of, and they could probably teach me something, maybe teach you something as well, too. But for other members of our audience, platform engineering has been something that's emerged over the last couple of years that they may not know the history of or the purpose of. But I know it's very important to CNCF. So, can you talk a little bit about platform engineering as a discipline? Why are organizations embracing it? What are some of the best practices around it?
[00:24:59] Chris Aniszczyk
Yeah, you know, I really consider platform engineering kind of the implementation of, you know, DevOps, which is more of, like, a cultural movement, where platform engineering is really kind of the implementation of those, you know, principles, and of course, it's important to CNCF, because. A lot of the technologies platform engineering teams, you know. you know, need our CNCF product. Like, alright, need, like, some kind of central orchestration system. You have Kubernetes, need a way to observe things, usually open telemetry is involved in that. Need a service catalog to understand what is actually being used in production, and who owns, you know, who uses it. Well, there's things like Backstage for that. So we're very much at the forefront of providing the tools for. platform engineering teams to be successful, and the reason it's such a trend in industry now is, you know, there's two things going on. One, there's a desire to, like, ship software, you know, faster, to kind of get feedback. You know, and kind of improve your product, and also paired with efficiency. So, if you're a large company, you know, say, like, early platform engineering teams at, like, Spotify or, like, an Uber, for example, where you're providing a service where there's millions upon millions of riders, listeners. And you have a software development pipeline that is, like, building features and responding to issues. Anything you could do to improve that pipeline, whether it's, like, you know, making your developers 10% more efficient to be able to ship faster, features faster, or, you know, do something that, makes things more efficient to run on to save costs on the cloud or something. when your user, when your usage goes up, any of those, like, 10-5% changes could add up to millions upon millions of dollars of savings, right? There's a reason, you know, companies like Google or Meta, Facebook, Spotify will hire even, like, kernel engineers or JVM engineers, because if they could just go, like, tweak. something small enough to make, you know, savings, it benefits the whole organizations, right? And so, that's kind of why, you know, this stuff has kind of really. risen to the forefront, and honestly, like, you know, like I said, it's really about the implementation of kind of these DevOps, you know, principles. It's also really about empowering. developers, too, for a lot of, you know, what I say would say, like, self-service autonomy. So, like, hey, I need a, want to go test out a service on, you know, something, or a new. new hardware, like, basically making that as automated as possible in a good, nice audit trail, versus, like, the old system of, like, oh, I'm gonna go. create a ticket, and someone's gonna go, you know, rack me a server, provision me an account. It's all about automating all that, that toil away, and really empowering developers to, you know, get stuff done. And, and that, and, and the get stuff done. is not just, like, raw output. It's usually, thing with platform engineering, it's usually paired through some, like, success criteria, right? You know, a lot of people are familiar, maybe in your audience, like DORA metrics, which are very classic, or the kind of new variation called, Space Metric, which is really about developer. you know, kind of flow and satisfaction and friction and all that good stuff. So, it's really, really about pairing, like, the work that developers do with some kind of concrete. you know, outcomes and business goals and all that good stuff. So, yeah, it's really, almost every organization that's modernizing out there is, you know, establishing a platform engineering team as part of that. You know, you know, you know, journey.
[00:28:36] Steve Semelsberger
For newer companies, startups that are growing, is there a development team size where it roughly might make sense to create a platform organization?
[00:28:47] Chris Aniszczyk
Oh, I mean, I think that's a great question, and I'm trying to. think about this from maybe some companies. You know, I would say once you start growing to the point where you need to have, more than a few, like, engineering. managers, and have maybe, like, multi, or multiple products that risk. fragmenting your, development stack, that could be, a trigger. But yeah, it's, you know, it's generally, I would say it's more of a large, like, medium to large company thing, because the benefits are significantly. higher for those style organizations based on just either efficiency gains or cost-saving gains that happen. For earlier-sized, you know, smaller startups and so on, you know, I think it's valuable to set yourself up right, but you're generally, as a startup, you're still trying to, like. you know, depending on what stage of the journey on, you're like, you know, like, product-market fit, or, you know, testing, you're, like, there's a lot of, you know, movement that it's hard to, like, standardize, like, a golden path on something if you're just, like, trying to figure out the business, if that makes sense.
[00:29:58] Steve Semelsberger
Oh, it makes a lot of sense, yes.
[00:30:00] Chris Aniszczyk
Yep.
[00:30:00] Steve Semelsberger
So, let's talk about AI. Of course, here in summer of 2026, you know, we have to talk about AI. And so, I think the latest data is basically every developer is using AI to help them code, you know, 90% plus. There are a variety of different camps associated with AI as applied to production environments. So, what are some of the things that you're seeing around AI in the world of operations? You know, whether that's new forms of observation, analysis, notification, you know, or fixed remediation recommendations, or anything that you're hearing about or seeing. in AI as applied to production.
00:30:46] Chris Aniszczyk
Yeah, it's, we're definitely, you know, living in a new world, there's no question, you know, there. It's, you know, I think two pivotal moments this year for me, at least, were, I think earlier this year, it was, like, Cloudflare was. Basically stated that they have more traffic now coming from agents than humans, which is, absolutely crazy for one of the kind of larger. you know, CDN, you know, and kind of clouds, out there for infrastructure. And then, a couple weeks ago, you know, GitHub has been a little bit flaky. you know, they've had, they've had some incidents, right? And, you know, developers are getting a little upset at them, back to our point of, developer. love and appreciation how quickly you could lose that, potentially. But, GitHub shared a post-mortem recently, I think a couple weeks ago, on an incident when they were down for a little bit, of basically, you know, the amount of pull requests and commits. that they've had has basically doubled since, since April. I think it went from, like, 1.8 million to 3-something million in, like, or bill, it's got to be, it's billions, like, I have to go pull up the data, but it, the graph they have is, like, it's incredible if you kind of, you know, look how crazy things are happening. So there's been a huge change of, you know. Obviously, developers are using AI tools, because inherently, I think. you know, I don't want to be mean and say developers are lazy. Developers are efficient. If they could go find a tool that makes their life better or easier, they're gonna go do it. And obviously, coding assistants like Claude Codex have passed that bar, you know, significantly, especially for coding. I think for the world that you, you know, tend to operate in terms of, like, operations, change resiliency. I think there's a lot of interest, and, you know, I've definitely seen people use, a variety of, let's call these, you know, either. analysis tools to kind of figure out what people are actually maybe deploying in production, to maybe trying to debug a particular fail, you know, you know, scenario, and trying to collate as many logs as possible. It seems like AI is pretty good at that. that type of work of collating as much information, as possible. But, you know, I think. you know, I think the holy grail a lot of people want is, like, hey, I want, like, something like an AI SRE, where I could just have this magical thing that could go, something happens, they could go figure it out, what's going on, immediately apply a plaque to solve the problem, so I would never have to pay an SRE again in my life again. And so you say, that just is not reality, but even though people. you know, want that type of, you know, functionality, like with, like with anything. It's, you know, it kind of goes back to, you know, the whole early cloud-native movement, where a lot of people thought, like, oh, you know, the fact that we're moving to cloud, we're not going to need to have, like, IT folks that rack servers and so on. Well, it just shifted, like, the job changed, right? A job changed from, like, instead of. that, you now have, like, SREs and folks that work, you know, and maintain and observe the systems. We're probably going to go through a similar little, similar evolution as some of these AI. tools get a little bit better, where I think instead of the humans doing a lot of the. you know, actual, you know, grunt work and remediation work. They'll be maybe overseeing and instructing and setting up guardrails for a lot of these AI power tools. But we're still very, I think, early when it comes. To this, to this world. So, I'm trying to figure, like, you know, answering. your question, other, like, trends, you know, the other big thing I see is, like, sandboxing, which is essentially a fancy word for isolation. You know, when you have agents doing crazy things, you want to be able to put them in a little box in case something goes wrong. You don't want them to take out. You know, the rest of your system. So there's a big. big sets of discussions, like how do you properly isolate agents, which are essentially your non-deterministic systems, right? And there's been a lot of famous cases out there of agents, like, breaking out of environments, doing things that people unexpect, so I think there's a huge. Huge opportunity in trying to figure that out properly, and obviously that's going to be paired with. you know, proper operations, observability, auditing to kind of make sure that everyone is, you know, behaving and then doing what, you know, humans, humans inspect the, the agents to do. So, hopefully that. That answers your question a little bit, like, yeah, isolation, and just, you know, essentially, it's, like, controlling. Controlling the blast radius when something. inevitably goes wrong. And that principle still applies for even the pre-agent world, where it would be like a developer writes a new service and does something crazy. You want to make sure that if. something bad happens, it doesn't take out the rest of the system, so you gotta properly be able to roll things back. Like, these practices don't necessarily change that much, it's just the environment that we operate in now, where agents are. A little bit more nefarious, when it comes to their environments and their ability to tweak. tweak things.
[00:36:01] Steve Semelsberger
The complexity of these systems, and now the velocity of change, the breadth of change, the quantity....
[00:36:07] Chris Aniszczyk
Correct.
[00:36:08] Steve Semelsberger
...frequency of change, those are growing dramatically, dramatically. It's fairly common to hear things like. oh, well, our deployment frequency is up 2X, and the PR per employee are up 2X, so we're in this, like, 4X world, and there seems to be more and more need for people who can set policy, who can understand appropriate risk appetite, you know, who can help recover, who can learn continuously. You know, the human piece of this, you know, in a world of such vibrant. change. It feels, to me, at least from what I can see, you know, just one person at a limited vantage point, like, people who can figure this stuff out and be orchestrated, I know we use the word orchestration in a lot of different ways, but, you know, when I think of a conductor, you know, who can have the orchestra actually play well and know when it's out of tune, like, man, that is an important skill set. And I know it's kind of a wild time where there is some fear in the. market. we interviewed a person named Brett Hurt, who you know as well here as well, too, has a new book called Love Conquers Fear, and Brett sees AI as really bringing in a period of abundance for all of humanity, and yet, you know, he's an entrepreneur and a technologist. He sees that, you know, things will be bumpy, and it's easy for us to become fearful, you know, especially in the face of such change, and so, you know, I'm betting on the need for lots of SRE-like talent on the human side, you know, even as the AI agents assisting people get better.
[00:37:39] Chris Aniszczyk
I, I very much, agree with you, and I think that. you know, in the world of agent speed change, like, I think resilience is really about, like, kind of understanding why an agent maybe had made a decision, right? And, you know, being able to have access to all the. decisions, you know, logs and traces, everything you kind of, need to understand, and so I think the role of maybe a next-generation SRE or human responsible for this is kind of adjusting, maybe tuning how the agent works, adjusting, maybe, the guardrails of a, you know, policy. Framework and system, and then kind of maybe even evaluating. remediation changes that are suggested by, you know, other agents. So I think it's really more necessarily about how to build. And kind of, design maybe, like, the, the, the operational envelope in which, like, an agent. is kind of, I think, maybe the role of, you know, future. future SRE work for a, for maybe a way to maybe codify what I think will happen. We're going to become conductors, I like the word conductor, I think that's maybe a good kind of thing. It's kind of like, you know, you don't necessarily control. the individual, you know, sax player or anything, but you're kind of are responsible for, like, conducting the, you know, where folks are going, even though you don't individually control each independent actor in the orchestra. So I do like that, do like that analogy.
[00:39:13] Steve Semelsberger
I like that a lot as well, too. And, you know, as you can see behind me, there's a couple guitars and a drumset.
[00:39:20] Chris Aniszczyk
Love it.
[00:39:21] Steve Semelsberger
...that I play not very well, of course.
[00:39:24] Chris Aniszczyk
We all have our passions (laughs).
[00:39:24] Steve Semelsberger
Oh. this, like, human spirit of resiliency, you know, you've seen it in open source and through your career and in life as a human being. I think sometimes for some of the open source initiatives that power so much of the technology of the world. Like, probably easy to lose sight of the fact that there are a few people who will slog it out. and will keep contributing, and really, really care about these systems, and you have a chance to get to know some of the key contributors across, yeah, I imagine dozens of projects. What keeps people going? Is it altruism, Chris? Is it, sometimes ego? What are some of the things that when push comes to shove, you know, we'll, we'll equip someone to continue to update...
[00:40:17] Chris Aniszczyk
It's interesting, I feel like, obviously, the answer, I think it depends on the type of open source project and person you're dealing with. Like, let's use a quick example of, Linux just turned 35.
[00:40:29] Steve Semelsberger
Wow.
[00:40:29] Chris Aniszczyk
Right? So that's a. That's a very long time for an open source project, and then Linus, the creator of Linux, I think, just, he's, like, in his mid-50s, I think he's 56, right? So, he's been doing this for a very long time, and part of me, you know, thinks that it's a lot of, there is an altruistic element there, where truly he does it because. You know, he loves building a system that many people depend on and use in the world throughout, even though it could be a little bit stressful, because any. you know, break in Linux just definitely breaks, breaks for your role, but he does love sharing knowledge, you know, he's from Finland, it's kind of the Finnish-Nordic. you know, spirit, I think, applies there. And I think, you know, there are folks like that, and they exist in many ways, not, sometimes their project doesn't rise to the success. you know, of Linux, but I think Linux had done a good job of reinventing itself over the years to be successful in new environments. Like, originally, Linux was kind of more of a hobbyist thing. It wasn't really necessarily, built for. you know, essentially the mass server market, which eventually came to dominate, where almost every online thing we use, or touch, or order, order Lyft, Uber, all that stuff is, you know, even order something from Apple.com, running, you know, Linux servers back on the end, so it truly, like, has evolved in many things, and most of phones, like, all Android phones run Linux, right? So....
[00:42:01] Steve Semelsberger
...it took over about 30 years ago. I grew up in the era of AIX and HP-UX and Solaris, and a variety of different flavors of Unix. It wasn't clear that one would rise.
[00:42:13] Chris Aniszczyk
And to Linus' credit, he was like, if someone came and is like, hey, we want to go stuff Linux on an embedded device or a phone, he could have clearly told them, like, not interested in making that work and whatever, people could go fork it, just like, you know, any other open source project, but to his credit, he was like. sure, why not, right? And, you know, eventually it evolved where, like. Linux is on phones and cars and satellites. So, like, the, when open source projects are, like, healthy to accept, you know, feedback and evolve with where maybe their community or customers or end users take them leads, I think, to higher chances of success. And, like, Kubernetes, I think, is a little bit similar in those veins, where, you know, when CNCF first started, it was all about containers and Microsoft. services, right, and orchestrating those things. Eventually, Kubernetes evolved to. run VMs also. People use projects called, like, Kubernetes from Red Hat to run VMs side-by-side to containers. People also run databases on Kubernetes, so Kubernetes itself has kind of evolved to not only just orchestrate containers, but other types of deployment artifacts, especially with the rise of AI, Kubernetes has kind of improved itself to really work well with GPUs, which historically, it was really not. built for that use case, and so that community has a good job to kind of, you know, evolve. But yeah, it's interesting, like, I've been. around quite a bit. I've seen a lot of projects that. you know, some get a lot of attention. I feel like Linux and Kubernetes have done a great job in building truly nice ecosystems, and then, you know, sometimes you're running the cases like OpenSSL, which a lot of people remember Heartbleed, where, you know, a funny little bit of story at the Linux Foundation is we kind of tried to help that project, and if you kind of, like, dived into OpenSSL, it was really, like. it was, like, two guys named Steve, basically, working out of Europe that, that were really responsible for something that probably derives trillions of dollars of value through it on, you know, a weekly basis from a transaction perspective. But they really. they were not the best at wanting to accept help and build an ecosystem, they just kind of had a very technical passion of maintaining that particular. projects, so it's truly, you know, I think a lot of open source developers are there to, like, kind of learn and share what they learn, and that generally is what kind of works and things, you know, grow. There's obviously folks that do it for more selfish reasons, and I don't think necessarily starting a. business is a selfish thing, it's just, like, that's what people do sometimes to try to monetize their passions or hobbies. It's truly variable all around. I think open source is big enough to support all sorts of, you know, cases to build and sustain projects. I think, you know, we have a weird problem where. I think a lot of times right now, we're. We're stuck in a bit of a, like, plethora of riches problem, where we have. It's like, it's like, there's so many different databases and caching systems you could use, and, like, it's almost like a new one is created, and so when you have major security issues, and the industry is using all sorts of everything, sometimes it's hard to rally folks to, like, hey, let's go all, like, fix. this thing, right? This database, because there's just so many of them out there, so that's kind of a harder problem, I think, to kind of, you know, deal with. We have no, like. we have, we have no, like, Rainforest alliance program or whatever, like, trade association seal that, like, this is, this is good quality open source that you should consider using versus just, like, a random repo on GitHub. But, like, we truly have so many options out there, which is. blessed for innovation and so on, but it makes maintenance and security a little bit of a trickier, you know, problem, because developers will just, like. I'm just gonna go use whatever I find on GitHub that helps me get my job done, you know, faster, and that problem's actually been more exacerbated with AI, where it's like. Hey, I got a problem then. This is you gonna write something which AI probably learned and copied off an open source. Project. So, truly crazy, you know. problems that we have to deal with, but I think everyone, we're doing our best to kind of rise, you know, rise to the child's.
[00:46:39] Steve Semelsberger
Those humans, and even as you described that, that circular loop that.
[00:46:42] Chris Aniszczyk
Yeah, totally.
[00:46:44] Steve Semelsberger
You know, I'm a Steve, I was born in 1970. Steves...there's a lot of Steves from the 70s, not so many Steves in the. 20s, we can talk about that.
[00:46:53] Chris Aniszczyk
Oh.
[00:46:53] Steve Semelsberger
And as you know, you were describing some of the motivations for people.
[00:46:57] Chris Aniszczyk
Yeah.
[00:46:57] Steve Semelsberger
And how Linus said, sure, you know, the things like Linux for mobile. You know, it made me think a little bit of James Carse's work. Have you ever read Finite Games, Infinite Games?
[00:47:08] Chris Aniszczyk
I have not, but I've heard of it. It's been on, it's been on the list for me to definitely look at.
[00:47:13] Steve Semelsberger
Yeah, it's really cool, and Simon Sinek wrote a derivative book around it as well, too, and...
[00:47:19] Chris Aniszczyk
Is it, like, console, like, game theory, or, like, what's, what's the, what's the kind of.
[00:47:22] Steve Semelsberger
The core premise is, you know, to decide what game you're playing. In a finite game, by definition, you try to win.
[00:47:29] Chris Aniszczyk
Yes.
[00:47:29] Steve Semelsberger
infinite game, the point is to keep playing the game. And it feels like Linux is playing an infinite game.
[00:47:35] Chris Aniszczyk
It is, and it's like a, to me, I consider, like, a positive sum game, where, like, it is the, the pie of value continues to grow for, as more participants kind of enter. It's not just a zero. something where the pie is limited, right? And that's sometimes hard for people to. to understand, right? Where sometimes vendors, especially in open source land, will. Be worried, like, oh, like, we gotta be the only. we have to, like, hijack this space, and if you kind of go back to our earlier discussion around Mesos and Kubernetes, Mesos has really had one company, really, behind it, one vendor, right? They kind of tried to control everything and be, even though it's an open source project, you kind of go through. But with Kubernetes, they wanted to build a positive sum game where there would be multiple people making money over time, so that the pie grew with the more participants, where I think Mesos, the pie was limited because they didn't allow for. prefer that, but I love, I love those types of games, yeah, yeah, yeah.
[00:48:41] Steve Semelsberger
Yeah, that's a really good case study. So, we'll wrap up here in a few minutes. I had a couple of other topics that I was hoping to get to. So, open source as applied to language models. So we're in a moment where there's incredible innovation with commercial models, and that's, you know, a lot of what's being deployed and utilized, and yet, I know a number of platform, infrastructure and, you know, even CTO leaders who are excited about open source. Some of this is token optimization now from, you know, the whole token maxing movement that lasted a couple weeks, it felt like, in early 2026, you know? But I think that the bigger concept is, like, how do we use resources? People, you know, process economic resources. to right-size how inference can perform. And so, can you talk a bit about what you're seeing with open source as applied to language models (LLMs), and where organizations and even maybe open source movements are experimenting today?
[00:49:43] Chris Aniszczyk
Yeah, I'll, this is a good, timely topic, right? You know, it's, if you kind of look at the data, if you're familiar with a service. open router. You know, out there, which kind of, you know, tracks. let's call it model usage across, you know, the board. I actually got, I think Stripe is in the process of purchasing them for some multi-billion dollar amount, but as of about 2 months ago. open models, such as, you know, Quinn, Kimi, you know, Minimax, all those, overtook proprietary models and overall usage on, through the open router, platform. And, that's just one, you know. Popular, you know, tracker. Essentially, of model usage, but directionally, that's going to an area of what I'm actually witnessing when talking to enterprises. Like, what people are basically doing is, you know, proprietary models are, you know, clearly. expensive right now, right? We're kind of in a big build-out, that's the way things are priced, you know, and what I'm, what a lot of people are doing is, like, you know what? I could go take an open source, you know, model, let's not, you know, discount where the model, you know, comes from. I could go. run this, you know, in my cloud or, you know, in some other, you know, environment, it's going to be significantly cheaper per token, and more importantly is I could go fine-tune that model with data that is very unique. to my organization that I'm actually not willing to give out to, maybe a frontier model or a cloud company, and there are a lot of, cases where that is essentially happening, where it's getting easier to essentially fine-tune these models, because there's more. you know, software out there to do that through startups, and eventually, you know, it will become easier. And to kind of think about it this way, it's a little bit similar to what kind of happened with the cloud and running stuff on-premises and the cloud. Like, sometimes it's cheaper to, like, run on cloud, sometimes maybe it's cheaper to run on-premises, depending on your scale and what your goals are, and being able to. Write and tweak software in a way where it makes that whole process easier. is great. Like, that has gotten significantly easier over time with the advent of things like Kubernetes and the whole cloud native set of projects around that. For AI, it's still early, it's not there yet for, like, mass consumption. It takes a long time. kind of do it, but, there was actually an article, I think I was read, I think it was last week, where. Thomson Reuters, which, I'm sure you're familiar with, basically launched its own. kind of, they call it their own frontier model, but all it was, like, Quinn fine-tuned on proprietary, you know, data that Thomson Reuters has. And if you're familiar with them, like, I don't consider them a cutting-edge technology company, but they are multiple billions of dollars of revenue, and, you know, they were already spending probably, you know. I think in the article, it's, like, $100 million on AI. you know, things, right? They've made the decision, or they've experimented with the decision, maybe it's cheaper if we could kind of fine-tune our own, you know, models. And I think that path, that decision-making path, is going to go through almost every, you know, enterprise, you know, out there. Like, the margins are too crazy. Right now, for people just purely to use a proprietary model. I think we're going to be in a position where there will be some innovative stuff that happens in proprietary model land. For sure. But the majority of the world enterprises that need to take advantage of AI in their systems are going to use open models and have those fine-tuned for their specific, needs, because it makes A business sense because of cost, and then B, really the most valuable thing that. companies have, is they're. proprietary data in how their business runs, or whatever specific asset they have. Like, that makes very little sense to give away. You'd rather just use that to your advantage by fine-tuning and tweaking models on that. So yeah, the rise of open models will continue to accelerate faster than, I think, anything we've seen, and as, like, an old-school Linux enthusiast. I think I learned the most when I was building my own computers, compiling my own Linux kernel from scratch, and doing all that stuff. When we make this as easy as possible for, you know, kids and adults alike to go. build their own little model or fine-tune it on their own little local Mac Mini or whatever computer, it will truly unlock a crazy amount of, you know, innovation and usage in new areas we don't understand. So I don't think there's any going back, it's just gonna get easier and easier, you know, to do it.
[00:54:39] Steve Semelsberger
For U.S. corporations that are considering running models that have origination from China, but running those models in their own, say, AWS environment, what would you advise?
[00:54:51] Chris Aniszczyk
Yeah, I mean, hey, if you look at the model itself, if it's a truly an open source, you know, project where you both have the weights and the training code and all that stuff, and you could go look at it yourself and inspect it, you know, it's just like. finding an open source project on GitHub, you know, that happens to be written by someone in another, you know, country. The stuff is there, available, and so on. I think, like, open models are good for you to kind of run, you know, on your own in your environment, but also sanity check things, I think. The geo, the current geo, geopolitical climate makes things a little bit. complicated, for sure, but hey, I think the value of open models have been shown, and there are actually open models being produced in Europe. in Japan, even in the US, we have companies that are producing them. There's a company called Reflection AI, which is doing that. NVIDIA just spent a healthy amount of money to purchase a company called Poolside, which is producing open, models. You have Thinking, or thinking machines, you know, that's doing, so, so there are options out there. You just don't have to, you know, go use Quinn. If you go to Open Route. water. and look at, like, with all the open models, there's a variety out there, and they get better and better, every day. So, use one that makes sense for your sovereign needs or geopolitical considerations. I think there's no, there's, these things are only going to get better faster at this point.
[00:56:20] Steve Semelsberger
Great guidance. So, wrapping up, Chris, we've talked a lot about learning, we've talked about games, we've talked about building, tinkering, the work that you did with your dad, and, you know, I've known you for a bit and have thought of you as somebody who pronates on learning and the experience of the human journey. So, what are you excited about learning as you look forward, as you go to the end of 26 and even in 2027, what are some of the things that you would like to learn?
[00:56:48] Chris Aniszczyk
Like, I have a, there's two things, you know, that, you know, I've been, you know, starting recently. Like, one. As someone who travels on an airplane for, you know, way too much of his life, I think I track everything via TripIt and Flighty, which is a great app if you haven't used it before. I've flown probably 3.5 million miles, over my career, which is a lot. I would love to, you know, fly a plane myself, so I've been, you know, taking flight lessons recently, so it was like a personal thing of just, like. you know, appreciating, you know, the thing that's taken me so many interesting places in the world, and kind of being able to learn, you know, how to operate it myself, and so on. So I've kind of started that journey.
[00:57:28] Steve Semelsberger
I have a friend who is similarly learning how to fly a plane midlife. He described it as learning a whole new language.
[00:57:37] Chris Aniszczyk
Absolutely, they have their own kind of terminology. Some things seem a little bit archaic, but I truly respect the amount of, like. like, the engineering and, you know, even resilience they put into these things to ensure safety and so on. Like, the aircraft industry is a good example of a resilient system where lessons learned is truly important, right? They have a whole. regulated framework where anytime there's a accident or issue, it's all written down and shared, so it's truly a beautiful, I think, like, industry and something amazing that we built as humans. So I'm spending, like, learning a little bit more. you know, about, that these days, and, you know, I also. have a bit of a passion of learning a little bit more about, the intersection between open source and robotics. I'm, you know, colleagues with some friends that have been working on open, call it, let's call it open source drone platforms and some other things, and that's just something that, like. very, know very little, you know, about, and something I want to go learn about more and spend some time. So I think, you know, I just love. you know, learning kind of new industries and new things, and Shingcia could take knowledge from another. industry that probably ran into crazy problems and see how I could apply that in my world. I think there's a lot to learn in many different, you know, industries and fields that we could apply to our kind of day-to-day life, and. you know, I think I'm trying to think, what else am I doing? I'm, you know, I'm always trying to stay, you know, healthy. You know, as we all get older, we're all trying to, you know. make sure that we were in tip-top shape, so I'm trying to get back into marathon running and so on, so I'm trying to get a little bit, you know, better, better there, and being a little bit more analytical on my times, and improving that, and trying kind of different techniques to. you know, improve my, improve my speed. So, yeah, it's a mix of. Both, you know, let's call it flight, robotics, and a bit of a personal health. journey, and I know it's something that, it's, you get lost sometimes so much in what you do and your passion that you have to forget that, like, health is truly wealth, and it's kind of the most important thing, to take care of. I actually admire, your other. interview with, like, Brett, which I'm kind of, you know, interested to hear. Like, Brett is very much into, you know, wellness and health, and, you know, sometimes it's just good to be around folks that. you know, care about that problem. It's really, we couldn't do other stuff without, without that, so.
[01:00:09] Steve Semelsberger
Anything important.
[01:00:09] Chris Aniszczyk
To prove that as something.
[01:00:10] Steve Semelsberger
Not easy.
[01:00:12] Chris Aniszczyk
Yeah.
[01:00:13] Steve Semelsberger
Flying 3.5 million miles. That is a lot, Chris. Yeah. Well, you know, I look up here west of downtown Austin, and there frequently are commercial flights, and I've thought of you on commercial flights. Now, you know, when I see a smaller plane up there, I will wave as well, too. It's great you're learning that.
You've been enormously generous with your perspectives, with your stories, with your insights, with your challenges, with your provocations. The work that you do is so important to technology all around the world. It's been an honor for me to have a chance to spend the last hour or so with you. Chris, thank you so much for joining us today on the podcast.
[01:00:51] Chris Aniszczyk
Yeah, I appreciate it, and excited to kind of be on this resilience journey with you. It's been, it's been fun.
[01:00:57] Steve Semelsberger
Thanks again, Chris, and thanks, everybody, for tuning in. We really appreciate your time as well. Have a great rest of your day.