 Welcome back everyone to the live CUBE coverage here in Chicago for KubeCon, cloud native conference, CNCF's event. We've been to every single KubeCon, it's been affiliated with CNCF, except for the ones that Kelsey and the team put together before the CNCF took over. We've got two great guests here, Kelsey Hightower, famous CUBE alumni, representing himself. He's out, independent, doing his own thing, freelancing, doing some cool stuff, and Rob Whiteley, he's the CEO of Coder. We're here to talk about a concept around how the developer environment is changing. It's all spurted out by Kelsey's tweet. Guys, thanks for coming on the CUBE, Kelsey thanks for talking to Rob, thanks for allocating this time to have a discussion. Looking forward to it. All right, so let's set the table. So, I'll set the table, Kelsey, you wrote a tweet, okay? I'm pretty good at tweeting it, I saved it. That is said, essentially, everything you need to code is on your laptop, the power. So the context, Apple just has their event. They announced the M3. Couple of days before that, Qualcomm announces their entry into the desktop processor game, and you can just see the benchmarks going up and up, more faster, more cores, GPUs locally. And I said, if hard work continues at this trajectory, remote dev environments are going to be less appealing to developers. Even though I was working in cloud, some of the big initiatives were bringing emulators from the cloud to your desktop. I want that fancy database on my desktop, where I'm actually writing my code, so I can have that fast interloop. And latency is a thing. Developers want fast feedback, because the job is already hard. The last thing I want to do is be waiting 10, 15, 30 seconds to do something, and the last thing I'll say here is, I just got a $30,000 laptop. We ain't paying $50, $100 a month for every developer to have a little small footprint in the cloud, just to do some of this feedback loop stuff. All right, so that, Rob, your reaction to that was, well, first of all, I don't want to bet against Apple, so that was not my tweet, that was not my intention. Well, first of all, fully selfish. I represent a company that makes remote development environments, so, but the first thing I would say is, I think you're right, which makes for a terrible debate, but, dot, dot, dot, I do think it only makes sense for small or individual developers. If you have a large team, if you're a bank and you have 10,000 developers or 25,000 developers, never mind the cost of shipping $3,000 MacBooks to all those developers. The chance that they're not going to be able to support and create the kind of environment you're talking about on that machine with VMware Workstation and Docker and all the tooling that's needed to have that kind of robust environment, whereas I can centralize that in the cloud, I can have a small team of five people, 10 people, support a dev environment that scales to 10,000 developers that gives them a good experience. And so I think it is true for a citizen developer or an individual developer, but what I think is challenging for a lot of companies is if they don't have 10x developers, folks that really know how to tinker and build their own environments, I don't know if remote development is the path forward for them. I think it will be the cloud. Well, I want to table that remote hybrid conversation. I want to get back to the tweet because I think what I observed is, well, I appreciate the debate here in discussion, but the traffic on that tweet went supernova. It's like almost 2,000 likes, 691,000 views. So it hit a core and I think where it hits the core it is, we're in an era now, a modern era of high velocity, coding coming in with AI assistance, that's going to be amazing. Collaboration tools getting better and the machines are getting better. The local host or the local, MacBook, it could be a Dell, it could be HP, whatever. More processing power is going to be on the machine at the point of coding. Okay, now that's, I think. Well, we also have another thing that's happening. A lot of the cloud APIs don't require, you're never going to have a dev environment that matches production, it ain't happening. If you require the PayPal API, PayPal is not going to send you their stack so you can run it locally. That ain't never happening, but you can hit those APIs remotely. Those same type of companies are shipping emulators locally and the goal is to write software, not integrate into production. So there's a loop here, right? So if you're a photographer, you're going to walk around with a small camera and you're going to take a photo. If you want to put it on a billboard, of course. You got to go take it to the laboratory and do his thing but I think there's a difference between, I need to write code. And when I say local, there's a lot of things that go into local development. If you write iPhone apps, you have a local emulator that allows you to have an iPad, iPhone, local, your machine. Now you got more GPUs on your laptop than you got in the dev environment. And so as this stuff starts to get faster and faster, I don't see a reason why we shouldn't leverage that thing so we can recreate these things. And yes, I do not want a Kubernetes cluster on my laptop. But if you're doing development correctly, you shouldn't need one. And this is one of the conversations, Rob, that's coming up here. Kelsey, can you get your thoughts? Is that platform engineering has evolved as the buzzword to become the word for setting up for devs to be devs, developers to write code. And some devs don't want to get involved in the infrastructure side, some just want to code, some want to get in there and get and set up their stack or interface. That's what it should be. What's the scenario? What's actually happening? Tell us what do you think's going on? I think a lot of platform engineers don't understand that you have to make the transition from a system administrator writing scripts and automating things. And you got to put on the product hat. If you're a platform engineer, your goal isn't to give people Kubernetes. That's just an ingredient to a actual platform. And you're constantly dealing with buy versus build. So like to your point, if you struggle figuring out how to give people a consistent experience in terms of a dev stack on their laptop, you might be better off buying a product that may be hosted where people click a button and they get a consistent experience and environment. So I think the role of a platform engineer is to say, how do I take what's available and turn it into a product that my team can use and only a platform engineer can do that. And that's important. So, Rob, let's get back into the coder side of it. You got a product that kind of is a collaborative, I call it the collaborative tool, but it's basically people working together. So now we have the kind of post pandemic view of hybrid work, which now spans the developers. So I'd love to get you guys opinion of how hybrid works now, knowing some of the changes going on in the landscape around platform engineering and some of the opportunities that are emerging. What is the hybrid work environment? Because you have teams all over the world, you have different teams that might have some privileges. So you got security concerns, you got latency concerns, a lot of moving parts to worry about. What's your, how do you guys look at this hybrid environment when some people come back to the office, some people aren't, some want to be remote. What's the state of the art, I guess? Well, what I keep hearing, especially from larger organizations, is we have to provide developers with choice because there is still, even in an economic downturn, even post pandemic, there's still a talent war. And if you don't provide a decent developer experience, then I'm just going to pick up, everything's borderless now, I can just go work for the next company. If I'm helping a bank with their transformation initiative, there's 10 other banks that probably could use that same skill set. So I do think there's a desire to, say, work from wherever you're most productive, work however you're most productive. And so hybrid, for me, is ticks on two factors. Hybrid is, in the working sense, is work from wherever you want, home, in an office, whatever. Then there is hybrid from a technology sense of, and here's where I will concede, there will always be a local component to developing because nobody wants the latency of keystrokes and not supporting all the different themes on their particular IDE of choice or their keyboard maps. So I think there is a local component, but actually where that source code is stored, compiled for some companies will be centralized. And so I almost think we're going to decouple the, for lack of a better term, the presentation layer of the IDE from the horsepower of the IDE that may sit in the cloud. And so you may end up with a hybrid is also an architecture, not just where you're choosing to work from. Casey, what are you seeing out there? What's your vision of how the development environment is going to be easier, simpler, faster, more collaborative, inclusive, safer? I think, look, all these complex enterprises rely on a lot of open source software. They've fought it for a while and now it's the thing that underpins their operations. And if you think about it, software development is a collaborative activity. I think the companies are like, we shall invent all the things in these four walls. That's a losing strategy. The world is collaborating whether you participate or not. And so how does the world globally participate in creating these industry standards? They do it with the best talent, with the best tools anywhere. And so if you don't have a company that can collaborate with the rest of the industry, you will be left behind or your bill will double because you don't participate in this ecosystem. So we know what the blueprint is. Talented people not only choose the best tools, they create them and they share them with each other. And then those go on to become the best standards that they have. So I think the model is developers are very unique and that they also build their own tools. I think that's on point. I'll just reiterate, we've been saying in the queue for over two years now, it's the developers of the new de facto standard bodies. Because remember back in the old days, you had the IETF, IEEE, we build all these standards and then we go salute the standards. But now no vendor can control the standards if the developers reject it. Because one of the standards like a developer goes on stage and say, hey, we all have this problem, right? Here's how I solved it. It's available on GitHub. Three months later, they're getting their A round and they have a booth at KubeCon. That is now the pace of the way this works. There's no more get-to-be. Kind of like Koder. Kind of like Koder, they build their own stuff. All right, so I got to get your perspective. When you wrote that tweet, so I said that Dave Vellante and we're like, hey, look what Kelsey just said. And we're like, we just said that on our podcast because we're talking to Dell guys and the PC guys, they're like, they're not irrelevant anymore. They can have the renaissance is coming back to hardware, right? And so we love speeds and feeds again. So we were riffing because we were just talking to a CIO that said to us, I'll say off the record, he won't say his name. It takes 20 minutes to boot the PC up in this development staff because they have too many agents on loaded on it. So from a security standpoint, you have these PCs that are bloated with like every single observability agent on the planet. And so boot up 20 minutes. Kelsey, that's not productive for a developer. So like, this is a whole nother little side nuance, but back to the point of latency, all this junk is in the machines too. But the funny thing about that is I've worked in those environments and most of those companies have no idea whether that laptop is secure or not. We bought all the things, we checked all the boxes, now no one can get any work done, mission accomplished. Look, that is not the way you do things. Number one, I think there has to be a way to develop safely in a, no developer I've ever met wants production data on their laptop. I don't want it. What am I going to do with it? I don't want it in the first place. So give me data that represents production and I won't go crawling around for Oracle passwords. So I think there's been a bit of a part where we've been locking down laptops but not providing any experience. So I think where we got to get to is like, listen, are you here to help or harm? And right now I think what developers are saying is we know what we need to move fast. I want the ability to install Postgres or Oracle on my laptop. I want the ability to create the same schema on my laptop. And can I not get a snapshot of scrub data that I can now test locally on my laptop? How hard is that? And no one's really built those tools. So now you see developers building it themselves. Tools like Dagger from Solomon and the Docker team. You've got things like test containers because developers are stepping up and saying, listen, this is chaotic, let's solve this problem. And hopefully the security people can join in and make that safe as well. Yeah, and we were saying too, on the AI side, the big trend also is to have local horsepower for model training on the smaller open source models that are getting traction. They may not be the largest, they're more specialty models are emerging. Why would I want to put that in the cloud if I can do it here? Then push it to the cloud and then do a little alchemy around model merging. So you got the AI phenomenon coming right around the corner as well. How do you guys see the AI changing the developer experience, productivity and just environment? Yeah, so I agree AI fueled a lot of this because we had a supply chain constraint. We could not get GPUs out to laptops. And so if you were trying to train a large language model, which was already a large data set, then centralizing that data set and renting GPUs in the cloud became a more productive path for a developer. I see the pendulum swinging. Apple and others are getting better at the supply chain. We're kind of through that. GPUs will get more powerful out at the laptop. I still think there is a, how do I replicate the data set to train the model? So I've still had customers express concern, but I think what you're bringing up, which I totally agree with is there is a short-sightedness to all of this. I think we are where the cloud was 15 years ago, where going to the cloud, man, I picked up my workload, lifted it, shifted it, and I ran it exactly as I used to, but it's now in the cloud. And that had an advantage, but it wasn't until we re-imagined things in the cloud that there was actual longer-term sustainable benefit. I think the same thing's happened to the laptop. Generation one, I picked up the development environment and I made it a VDI or I made it an instance in the cloud and I just shifted it. Same experience with just latency now. I think once companies start re-imagining what it will actually be like to be in the cloud and change development, then we get to a point where I don't think developers will need all of that infrastructure on their laptop in order to have a good experience, which I do admit is needed today. Or to Kelsey's point, maybe someone comes with a Novell idea to make it seamlessly seem like they're their local or leverage both the power and the cloud under the covers. I'm also saying that we've all been a benefit of the cloud figuring out how to scale down. So you get S3. This is a huge scalable thing. Only Amazon can run. And now you can go to MinIO and you can run an S3 binary on your laptop. And as you develop against the S3 protocol on your laptop. So I think what we're saying is like number one, cloud has created these wonderful experiences. You're in the SaaS business. You're in the experience business. So if I come to your cloud and I want to develop an environment, I'm clicking one button and my ID pops up. I'm writing my code. But here's the thing. Docker runs on my laptop. What's stopping you from allowing me to click one button on my laptop and a consistent development environment comes up. So I think what people are asking is like I've tasted the future. Yeah, yeah, exactly. And my laptop is capable of running it. And if you won't do it, the next vendor is like, hey, we have local support. And I think this is the tug of war. Because now we're using the same tools. Like if you look at a lot of these hosted environments, they're using containers under the covers. I can run containers locally. Why are you not giving me this experience locally? I was just talking to the CEO of Chroma. They do the vector database. They had the local version. Now they're working on distributed version. So again, the same kind of movie. You know what, honestly, what I hope is that cloud should mature to the point where it can run anywhere. It can run in AWS, GCP, Equinex. And it's good for everybody. Building managed services. And those that can't necessarily leverage those for whatever reason, do we have to trade off the experience just because it's local? I hope the answer is no. I mean, the key word is environment, right? It's all one cloud operation. And I think what I like what you're saying is I do believe we've hard-coded the development environment to the laptop. And if you can decouple it and then run it anywhere, maybe centrally in the cloud for security, maybe locally for performance and latency, there is no choice now. You can replicate a Docker and get there, but there is no push-button easy developer experience to do it. It's very tinkerer. My first time I saw things like the web or virtualization and even 3D printing. Oh, that's freaking amazing. I can't wait to see the, hello, provision me 10 clusters of Kubernetes and running a low-code, no-code environment with full data pipelines. Go. Just gets built. But to be clear, I don't think having Kubernetes on the laptop is what I'm saying. Yeah, no, I know what I'm saying. What I'm saying is I depend on these five APIs and this code base to do my job, make it appear on my laptop. And if that's Docker or Kubernetes or some other thing, we now have the resources locally for that other thing just to work. That is what developers are clamoring for. Not, how do I run Docker or Kubernetes on my laptop? You want a declarative environment where you just say, this is what I want. Today we have an imperative environment where you have to build every single ingredient yourself and get to it and I think that's not scaling. Guys, great discussion. I think we settled in. We're all in alignment. Developers first, developers are excellent, developers drive standards and go faster, make it easier, and let's get more horsepower on every machine and server. Okay, that being said, final track of conversation here since you're both legends, you guys built a great product over there doing great work. Kelsey, you're well-documented in history and Kubernetes of which goes back to the formation, the stories there. Look at this place. I mean, you got to be pretty pumped about how growing up this is. This is a big event again. Just more kind of a pinch me moment of like, look how big it is. Look how significant Kubernetes has become. Eat containers around it, cloud native companies. The young crowds come in and I saw a bunch of folks first night, just graduated from Cornell, a bunch of schools. They leave in their jobs, come in into tech. Just more diverse minds, younger, old, hanging around together, building stuff. I mean, what's going on with you in your mind? How's it going? Yeah, I mean, I think it's easy to come up with new ideas but they take decades to propagate. And when you look at the badges, you see the visas, you see the college students, and they're all coming and they're talking about the same ideas. And that shared vocabulary that has now made it to the shared lexicon, it's like here now. So that part is not being debated. Now what's being debated is like, what do you do with it? Which is a great place to be. So we went from what it is, what is it? So what do we do with it? And now this year I'm seeing people like, here's what we've been doing with it. And now people are like, what should we do next? And so I'm really happy that we've gotten to the point now where this is mature enough that people are asking what's next. Yeah, one of the things I love about open source is it was kind of like the underdogs and then it became the standard, it is the software industry now. And so as people come in, they sometimes don't know the old expression of standing on the shoulders of giants before you. So I have to ask you guys, final question is as you look at the industry now and people watching that might be younger or coming to the industry, how are we going to, how do you see this refactoring of the developer experience going next level from here, knowing what went on before and what should people know about going forward? What are some of the ethos, the standards, share your vision of, because it's going to go next level, but it's not this, it may not go the way we think it is because it's now standard. It's the software industry. It's not like open source already won. It's not like... Yeah, it's a good question. Look, here's what I would say. So Coder was started by a couple of very young gentlemen and they built a product the way they thought it should be built, not the way they were told it should be built because they didn't know. They just thought, this is the way we should develop software and it resonates with people in enterprise because it's a cleaner, purer way to build software. And I think that's the joy we're getting back to with open source is people are kind of reimagining things, a younger generation is coming in that doesn't have the same biases and questions and I'm like, well, I would never do it that way because my IT department would never allow it. My dad did it that way. Yeah, you know, and so I do see there's a kind of a, I used to work tinkerer, but there's a sort of an innovator mindset that's back and I'm seeing that enterprises are willing to, because they can wrap guardrails around it, enable that innovation without having to stifle it, which they were five years ago. Because of what's your vision of how things might get re-backed? My biggest saying is like, it's easy to predict the future when you're working on it. And now we understand how to work on it. And so now that if you have an idea, you can go open source without permission and even the incumbents, the big proprietary vendors, they're no longer afraid of like the wall garden. Hey, that's a good idea. In many ways, they're like the first ones rolling it into the enterprise product. So that distribution model now is like, seeing that outlet, I think people get to decide and it doesn't have to become a big megacorp to do it. That's what I think is so exciting right now. And it's inclusive, you got end user customer and user customers, entrepreneurs, open source groups. Half of this stuff comes from companies doing it in production. Argo CD, one of the most popular CI CD systems in the cloud native space, comes from the tax people into it. We've got the people that- Backstages. Yeah, backstage coming from Spotify. Spotify, yeah. And so you're starting to see these companies like, hey, not only are we buying things, we're creating them. So I think that is like the end game of this whole open source movement. Yeah, I mean it's a beautiful thing. It's a whole other level, another generation. Guys, thanks so much for coming on theCUBE and sharing the discussion and sharing your vision. Rob, I appreciate you taking the time on this segment. Thanks for calling me a legend. I've never gotten that before. You guys are legends. I'm legend to Jason, but I'll take the lead. And legend is in our own mind. You're on theCUBE, we're gonna get a legend. I'm a CUBE alumni now, you're good. Thanks so much, appreciate it. Yeah, thank you, John. Thank you, John. Thank you so much. Okay, live CUBECon here in Chicago, theCUBE. We'll be back with more coverage. Day two, got a full day tomorrow. We got to have experts come in. We're going to have old school OGs coming on. We're going to have the young guns here. Stay with us for more coverage after this short break.