Subscribe
Share
Share
Embed
Come join us as we discuss everything open source with guests that are pillars in the industry. Welcome to The Secret Sauce.
Welcome back to the show. We're about to learn the secret sauce.
Bdougie:Thanks so much for taking the drive over the bridge.
Solomon Hykes:Thank you for having me.
Bdougie:Yeah. Yeah. And, so folks might know who like, they might know who you are, but they might not know who you are. So do you explain who you are and, what you do?
Solomon Hykes:Sure. So I'm, I'm Solomon. I'm the founder of a startup called Dagger dot io. And before that, I was the founder of Docker. Excellent.
Bdougie:Perfect. Yeah. And, like, so we've been having conversations with folks in open source about their strategy and, like, the wins and also losses. And, we'd love to talk about, first, Dagger. Like, what is Dagger?
Solomon Hykes:So Dagger is it's a platform that reinvents CICD pipelines. So it it takes those big static proprietary pipelines that are kind of a pain to deal with for a bunch of reasons. And it turns them into lightweight, easy to create, easy to run pipelines that are made of code. Yeah. Excellent.
Bdougie:Yeah. And this in the world of CICD pipelines, like, I I feel like it's I don't know if everyone thinks, oh, it's a solved problem. Like, why is Dagger doing this now? Yeah. Well,
Solomon Hykes:oh, you know, honestly, we we the way we started, was as a start up was by bringing the band back together. You know, my cofounders were the earliest employees at Docker. We worked together for years, and we just wanted to build something cool together. So we went out there and, just met with a lot of teams about their problems, you know, asking them questions about their problems, and a lot of them complain about, pipelines, you know, deployment pipelines, build tests, and all of that is, you know, interconnected. And we just realized, the it's just a major source of of pain still for software teams.
Solomon Hykes:You know? So so
Bdougie:who who on the team from the original team at Docker came or the early team at Docker came over to do Dagger?
Solomon Hykes:So my 2 cofounders, Sam and Andrea, were basically the first engineers at Docker before it was even called Docker. It was called dot cloud. You know? And, we went to the same school. And so coming out of Y Combinator in 2010, I, you know, called them and asked them if they wanted to join this new startup I was, I was doing, in in in Silicon Valley, and and they said yes.
Solomon Hykes:They packed their bags and and joined, you know. So Sam and his come from,
Bdougie:or where are you from originally?
Solomon Hykes:I'm French. I don't sound French, but I am French.
Bdougie:Yeah. You told me.
Solomon Hykes:I have an American dad. So Okay. You know, I'm really more French than I sound. And, salmon and dryer are also French. Yeah.
Solomon Hykes:Okay. And then Eric Eric Barton is our CFO and, you know, operational wizard. And, he was the CFO at Docker also.
Bdougie:Okay. Cool. So you I actually what you mentioned, it was called dot cloud. And also you mentioned you did YC as well, which I didn't actually know Docker went through YC. So Yeah.
Solomon Hykes:So I went through YC twice now. The first time, summer 2010, and we were called, yeah, dot cloud at the time. And it was very I mean, it was a really important experience for me. You know, it really kind of beats a sense of product into me. Yeah.
Solomon Hykes:And, yeah, it would definitely had a huge impact on our trajectory. You know?
Bdougie:Yeah. And what problem were you solving the dot cloud, the Docker? Like, did you did you see this work somewhere else that you brought it to dot cloud?
Solomon Hykes:You mean the Yeah.
Bdougie:I, basically, I'm just curious. Like, did you see 2010 is like, okay. We had AWS at that point. Mhmm. We had a couple other new cloud providers come around.
Bdougie:Yeah. So at that point in the industry, it was interesting to start Docker. So, basically, I'm asking, like, a lot of folks will build something somewhere else. Like, we should build this and stuff.
Solomon Hykes:Yeah. Well, you know, actually, there's a straight line that starts in 2008, 2007. Okay. Where I yeah. As, you know, a very, young professional, I graduated in 2006, so I'm a software engineer slash sysadmin.
Solomon Hykes:I'm out there just doing, sysadmin scripting, work, you know, and I'm just excited for personal reasons about containers. This this pretty immature tech in the Linux kernel, it required patching the kernel. It required a lot of hacking and and, you know, kind of, tinkering. But I I thought it was very exciting, and I just kind of pursued it, you know, not knowing where it would lead me at all, not knowing anything about the business of software, at all. And this is in France.
Bdougie:Yeah. Okay.
Solomon Hykes:And so everything else is just iterations and iterations of the same obsession with containers, container tech, and its potential. And, you know, dot cloud was a phase where we bundled that container tech in a platform as a service product, so we competed with Heroku, etcetera. That's what went through YC. And then later, as the PaaS business turned out to be generally a really bad business, We just kind of pivoted and spun out the container engine, and, you know, that was Docker.
Bdougie:Yeah. Well, we we both laugh at it, but I'm curious if folks probably aren't even aware, like, the PaaS business and, like, the Herakus of the world. Like, could you explain that statement of, like, why PAS was bad back then?
Solomon Hykes:Yeah. So, I mean, it's honestly, it's, that well, the reason PAS was exciting was the same reason I'm excited about Dagger today. So it's the problem that's still not solved. Yeah. But PaaS platform as a service was an attempt to solve this problem of application delivery to the cloud.
Solomon Hykes:Right? Yeah. I have codes. I want it to run up there. How hard can it be?
Solomon Hykes:Turns out it's really hard. You gotta, you know, gives, like, a 1000000 steps. And the idea of platform as a service was to abstract away all the steps, all the automation. So you just push your code and then magic, magic, magic, and then it's running, you know. And the so that's the magic, and it worked really well.
Solomon Hykes:The problem is that it bundled 2 fundamentally different parts of, the the application delivery process, the the the pipelines, the delivery pipelines, and the hosting. Yeah. It was all in one. You so we did that. Heroku did that.
Solomon Hykes:You know, Google App Engine did that. A a 1000000 startups did that. You give us the code, and we'll take care of the pipelines and the hosting. And turns out you gotta unbundle it. You you know, you can deliver great pipelines or you can deliver great hosting, but at sufficient scale, you can't deliver both in a world class way.
Solomon Hykes:And, of course, the elephant in the room that made that theory that it was a bad idea, very concrete was AWS. Because AWS did hosting better than Yeah. Than matter how good your your, you know, deployment experience was, your tooling, your pipelines. Amazon was gonna give deliver more scalable, more reliable, and cheaper hosting than you. Yeah.
Solomon Hykes:So that that affected us. You know, we we we were basically reselling AWS compute with with, you know, and the margin was justified by a better deployment experience. But, eventually, if your customer is large enough No. Then they're gonna hire someone, and then they're gonna that someone is gonna say, let's do it. Let's rebuild it on Amazon.
Solomon Hykes:Yeah. So that's That basically killed PaaS as a category.
Bdougie:Yeah. The the classic build versus buy Yeah. Conundrum that everyone hits. And I think once you get to a certain level, you can invest and they're like, okay. We're gonna build and save some money by that engineer is now paid maybe half the cost of what we were paying AWS for the past couple years.
Solomon Hykes:Yep. Yeah. Exactly. Build versus buy. And so that was the lesson from PaaS.
Solomon Hykes:But but Docker came out of PaaS. Yeah. You know? Because Docker was a necessary, building block to solve the problem, and the problem is still not solved. It's still a mess.
Solomon Hykes:You know?
Bdougie:Yeah. So Yeah. And, like, a lot of the newer contenders for that hosting side, like, they allow you to provide the docker image Mhmm. To then get the quick push or get ops going as fast as and then you see, like, the cost has been now trivialized of, like, her the joke is Heroku is $7 a month, but then you have to, like, pause dynos or or whatever dynos are. But a lot of the other services also follow that same pattern of, like, okay.
Bdougie:It's $25 unlimited. And then once you get to a certain bandwidth or a constraint, then it's okay. We're gonna have to talk to you about whatever enterprise offering is. And, that's the world I came up out of. So, like, I shared before we record, like, I learned how to code professionally in 2013, and it's around the same time that Docker, like, came out and, like, was publicized and everyone was using it.
Bdougie:And I was specifically in a meetup in Tampa, Florida, my hometown, in a Ruby meetup. And someone who had contributed into the Docker code base was, like, raving about it and was like, oh, this is the the best thing since sliced bread. We all should just use this stuff. And I honestly could not grok what he was talking about, because I knew what a VM was, but Docker is, like, slightly different than what VMs do. And it's, like, more of the bundling side.
Bdougie:Well, I'm explaining that, but you can probably explain it better. But, yeah, I I wanted that you to explain the concept of containers as well.
Solomon Hykes:Yeah. Well, it was it was a lot of what you described, like, it was just I mean, I was so immersed in it for so long. I'll it I struggle to, you know, explain in simple words what seems so obvious to me. That was and, honestly, I'm having I have kind of have the same problem now with Dagger in the pipelines. You know?
Solomon Hykes:So what saved us with Docker and and replicating that with Dagger is, first, you got it's okay that the whole world doesn't get it yet as long as you have a community of people who have the same gut feeling as you and you kind of finish each other's sentences, and then that community kind of snowballs from there. And then over time, together, you kind of find easier and easier ways to explain it. But yeah. So the the for containers, really, the, the tech existed before to, you know, basically slice up an operating system into isolated sandboxes. Right?
Solomon Hykes:So you can kind of virtualize. Fundamentally, that's what a VM is. That's also what a container, is. The difference is that we figured you could stack them, instead of choosing, a level of virtualization. Do you want, like, a very lightweight virtualization, which is containers or a very heavyweight a more heavyweight one like VMs?
Solomon Hykes:You know, you're choosing between speed and security, basically. We figured containers actually can be used for more than just a virtual machine. They can be a unit of application delivery. You can actually encapsulate an application in there.
Bdougie:Yeah.
Solomon Hykes:And and then it can be a container on top of a VM or a container on top of a physical machine. It doesn't matter. So we we we just figured we can use this tech for something more important than slicing up a machine. It can be for packaging up, you know, a a software component, an application, and then shipping it to another machine, another developer, another cloud provider, and it's gonna work the same way. So we use the shipping container analogy, you know, to explain that.
Solomon Hykes:You know, let's just all standardize on this format to package up an application or a piece of an application that's gonna run-in the cloud, basically. Yeah.
Bdougie:Yeah. And, so where we're recording, like, we're just, like, over our shoulder is the bay. And just around the corner is Alameda where all the shipping containers are literally stacked on park. Yeah. And I I watched a a talk from, like, 10 years ago that you were explaining this where when it did click for me, where you're talking about the shipping container where it could be filled with whatever.
Bdougie:Yeah. It's just a container. And between here and China or here and wherever it's coming from, like, you don't have to understand how it's been delivered. It's just it's delivered. And you open it up, you use the thing, you close it up when you're done, it goes back across the the ocean.
Bdougie:And, like, at the time in 2013 when I was learning how to code, I learned on Ruby on Rails first. And the thing that in the Rails guide they taught you was to use VirtualBox. And VirtualBox made sense for me to be, like, a VM and, like, a situation where Rails works Mhmm. Out of the box, because Rails didn't work it out of the box in every machine or win Windows version and stuff like that. And that comes up being a lot of sense to me because I'm like, oh, I don't know how to install rails and brew homebrew and all this.
Bdougie:Like, now it's a little it's more trivial than it was before.
Solomon Hykes:Yeah.
Bdougie:But just in the last couple weeks, we built this thing using Rust. And, like, I don't know Rust the way that everyone else knows Rust or the person who built the the thing in Rust is. But I can run the thing on my machine because of a Docker image. And that container is just like, I am Yeah. I clone it, the repo from GitHub.
Bdougie:It has a Docker image in there, or I pull the image down specifically for this situation, and it just works. Yeah. I could now ping the server and do things I should be doing.
Solomon Hykes:Yeah. That I mean, that it's a it gives you kind of a line of abstraction. Right? The separation of concern between what's inside the box and outside the box, and that's, like, the that's the core of it.
Bdougie:Yeah. Yeah. So your your time at Docker eventually ended. So, like Mhmm. The the past thing didn't work out.
Bdougie:Docker had a lot of success and a lot of adoption. So, like, the assumption is, like, 1,000,000 of dollars, 1,000,000,000 of dollars evaluation. So, like, what was the the thing that never clicked at the time that, your your time at Docker was running?
Solomon Hykes:Yeah. Well, yeah. So I left that for 10 years total, you know, at that on working at the whole adventure, you know, France to the US to to Dockers.
Bdougie:It's a long time. Yeah. It's like a long time. A whole lifetime
Solomon Hykes:in town.
Bdougie:For sure.
Solomon Hykes:So, honestly, the main factor was just that I was tired and, you know, had gone through a lot a lot. You know? And so that was really the the the the main thing. At the time, I mean yeah. I was always as a founder, you're always the worrier, I think.
Solomon Hykes:So I was always worried that we weren't doing enough or we're doing the things wrong or we weren't seizing the opportunity. Because, really, that's what happened. We all of a sudden, like, for 5 years, no one cared about this container thing that we were doing. And then all of a sudden, everyone cared. Yeah.
Solomon Hykes:And so window opened where everyone came rushing in, and demand was almost unlimited. You know? And, but competition also, of course, showed up to mop up that that, soak up that demand. And so we were under enormous pressure to build out a business, build out a complete platform, you know, and so we need a strategy to do that. And I think when that happens, we were just not equipped as a team and as an organization to rise up to the challenge, you know.
Solomon Hykes:And and the core reason for that is for reasons that now I think are were wrong. I decided I was, you know, I I I didn't I wasn't the right CEO for that moment, and I needed to hire an external CEO. That's not necessarily a bad thing to do, but there's a time to do it. Yeah. And I did it at a terrible time, I realize now, because you need to wait for a natural checkpoint Yeah.
Solomon Hykes:Where you kind of you've kind of you took a risk. You're doing, you know, you're kind of going from one checkpoint to the next. And in between checkpoints is a wrong time to bring in someone from the outside because you're in the middle of trying something that no one's tried before, and no one really if it's working, which Docker was, no one really understands why. And you don't you don't have time to explain everything. Like, well, it's it's just, you know, trust me.
Solomon Hykes:We're we're doing this. Are you in or are you out? But when you hire a CEO, they need ownership, you know. Otherwise, why are they here? You know?
Solomon Hykes:And so it's it's an it's a really disruptive thing to do to transfer ownership to an to to an external CEO. And, and and, honestly, I don't anyone would have failed in that situation, but I we failed from there because then you add unlimited money, which we had. I mean, we raised, I think, 300,000,000 total and, you know, huge growth of the team. And then what you have is at the core of this whole organization that's building, you have a lack of clarity on what are we doing, what are we not doing. So what almost killed Docker is a complete inability to say no to anything because no one felt empowered to say no decisively.
Solomon Hykes:And so any middle manager could come up and say, I have this great idea that will get me more head count and, like, you know Yeah. Make me look really smart. And maybe it is a good idea, but you can't do all the good ideas. And we did all the good ideas and the bad. Yeah.
Solomon Hykes:And then we spent $300,000,000 building a $60,000,000 business. You know, that's what we did. Wow. Yeah. And so it almost killed the company, and it definitely killed, you know, my motivation as a founder to keep going.
Solomon Hykes:Because, you know, as a founder, you need ownership. You need to kind of as a CEO founder, which is now I know what that's what I am. You need you need to run the show, you know. Otherwise, you just go build your own show somewhere else. So that's yeah.
Bdougie:Yeah. Yeah. Appreciate you you sharing that.
Solomon Hykes:Yeah. The Docker survived it. You know? Yeah. Like, eventually, against all odds, Docker survived as a as a business and is now actually doing great.
Solomon Hykes:You know? So the next chapter has yet to be written, but, I think Docker has a shot to continue writing the story.
Bdougie:So Yeah. It it I've it definitely feels like Docker has pretty strong foundation and footing today. And I know, like, DockerCon's coming back, maybe this fall. So super excited to to see that happen. What year was that when you sort of had transfer of the of power in the CEO?
Solomon Hykes:Almost, like, 6 months after the pivot. Okay. And what's crazy with that pivot is every every chart started going up into the right exponentially immediately.
Bdougie:Yeah. Could you explain the, the pivot?
Solomon Hykes:Oh, so the pivot was we were running dot cloud, which was a platform as a service, you know, competing with Heroku. So, again, like I said, deploy delivery pipelines and hosting of your application all in one. Give us the code. We run it. And so, we were just kind of gradually growing that.
Solomon Hykes:It wasn't failing, but it just wasn't blowing up either. And then, we started the side project internally to extract the the container engine that was powering it and spitting it out as a as a, as an open source project that you could build on top of. So it was sort of like we're we were if if PaaS is a car business Yeah. We're selling a complete car and we're competing with other carmakers, we decided, hey. Why don't we sell you know, why don't we, open up the, share with everyone the blueprint to our engine.
Solomon Hykes:Okay. And let's build an a whole ecosystem of people building all sorts of cars and other vehicles on top of this standardized and standardized engine, you know, something like that. And it turns out there was enormous demand for that. Yeah. For fundamental reasons, like, you you need this low level building block to solve the overall problem of application delivery of the cloud.
Solomon Hykes:You need you need an a base ecosystem to build on top of. You know? You can't you can't solve the problem with the big black box. Yeah. You know, you need a lot of tiny little boxes, then you just assemble your own thing.
Solomon Hykes:So there was demand, and so that was the pivot. We opened that project, and it was so popular so quickly. There was so much excitement that we eventually decided to focus on that. We renamed the company from dot cloud to Docker. We eventually sold off the PaaS business to one of our competitors.
Solomon Hykes:And yeah. So pivot.
Bdougie:Okay. Yeah. I didn't realize that the the pivot to Docker was to pivot. But can you explain the the Docker and Mobi connection? I never quite understood that.
Solomon Hykes:Oh, yeah. So that was, yeah. So that like, fast forward a few years, we're now in the middle of the Docker frenzy. You know? And, honestly, if you weren't if if it's hard to picture actually, it's really easy now to describe because because we're in the middle of another frenzy, like the OpenAI frenzy, like the Yeah.
Solomon Hykes:LLM frenzy with at in the middle of it, this one company OpenAI. So that was the container frenzy, and at the middle of all of it was Docker. So we were we're we were in the same situation that OpenAI is in now. And it's, of course, like, who wouldn't wanna be in that position, but also very overwhelming, very stressful, you know, very disorienting, and it it can kill you. And, you know, actually, you know, there's you know, the book Crossing the Chasm?
Solomon Hykes:Yes. You know, about how to get how to get to the product market fit, basically. But, what I discovered at the time is that there's a second book that follows up. It's a sequel. It's called Inside the Tornado.
Solomon Hykes:Oh, okay. Happens once once you have successfully crossed the chasm, and you're now in the tornado of violent product market fit. Yeah. And and what the book basically says is is what I experienced, which is that that can also kill you. You have to actually
Bdougie:Yeah. Tornadoes immediately. Yeah.
Solomon Hykes:It's like you're you're crossing the desert. There's no water. And then all of a sudden, you're swimming in a river or something, and you have to not drown now, you know Yeah.
Bdougie:That's instantly. Yeah. It's it's it's I know someone who works at OpenAI, and I they're definitely in a tornado right now. Yeah. Because they're, like, they're when I talk to them, you could definitely feel it in, like, their mannerisms, but also the energy of, like, we shipped another thing, and also we have a bunch of pushback from this other thing.
Solomon Hykes:Yeah. No. It's it's it's really insane. It's it gets a really unique experience, and it really a it ages people. I mean, I didn't have white hair when I before the pivot.
Solomon Hykes:And yeah. So to answer your question, somewhere in that tornado phase, so, you know, 2013 to 2018, I guess, 2017, There was a lot of, like, push you know, a lot of pushing and shoving, in the in the broader ecosystem around containers because Docker unlocked something enormously valuable. And, that created a lot of opportunities and also threats for a lot of very valuable companies, including Red Hat. And, Red Hat saw Docker as a mortal threat and acted accordingly. You know, first, they decide, oh, this is our Linux.
Solomon Hykes:So they latched onto Docker, and we were very close partners initially. And then they realized this is not our Linux because Docker wants to be Linux and Red Hat. So there's no room for us to do our thing. Yeah. And so then they pivoted to Kubernetes, which was, you know, the perfect vehicle for them.
Solomon Hykes:And they've done that very successfully, by the way. But in the middle of all that is just a lot of competitive dynamics. And so, that resulted in a lot of open source drama. You know, a lot of just a lot of criticism, a lot of debates, a lot of flaming, a lot of trolling. Kind of, like, again, the LLM stuff now.
Solomon Hykes:OpenAI is not open, you know, blah blah blah blah. So all of that stuff. So we were experiencing that. And so the problem when you're in that situation is that you're you're getting a fire hose of negative feedback. And, you know, a lot of it is important, valid, well intentioned negative feedback.
Solomon Hykes:Like, hey, I'm a user. I'm rooting for you, but you got to fix this thing. Mixed with a lot of bad faith bullshit. But it's all mixed together.
Bdougie:Yeah.
Solomon Hykes:And it's really hard to tell it apart. And so we we just kind of that really can that was the most tiring part for me. And so, eventually, you know, a feedback that rose around, above was the the noise was you need to draw clearly more clear lines between what's open and can be shared with the whole ecosystem and what's actually yours. You know? And so and it's true.
Solomon Hykes:We'd blur the lines because we were naive. We just figured we were going to reinvent how open source and business combine. You know, in our in our minds, we invented the thing, you know, all of it. The the the product experience, the open source community, the brand. I mean, no one cared about containers before.
Solomon Hykes:So we just had this and this is me, you know, like, the designer, you know, the dictator mindset. You know? It's about control of the experience end to end. So just let us do our thing. But eventually, you realize, okay.
Solomon Hykes:You need to open up parts of this to the ecosystem if you want them to follow along. It's not all about the perfect design. It's also also about, you know Yeah.
Bdougie:Collaboration. Yeah. Community.
Solomon Hykes:Yeah. I was gonna say business, but, yeah, collaboration with people who also want their own thing to succeed. Yes. So it's really about the distinction between community and ecosystem, I think. You know, you have to be able to collaborate with people who don't who are partially aligned with you, but not completely aligned with you, even with competitors.
Solomon Hykes:So, like, that's what open source is about. Anyway, so Mobi was our attempt to do that, and, I think it failed. But it was basically if you picture what CNCF is, that's what Mobi aims to be, you know, a a collection of individual projects that forms the components of the Docker platform, and you were free to go and use the pieces you wanted. And we did spin out a lot of components. We spun out run c, container d, build kit, notary, a whole bunch of them.
Solomon Hykes:But I think CNCF ended up being the home. You know. Yeah. And and and Moby was, all of it was copied after the Red Hat model, actually. You know, they have Fedora Yep.
Solomon Hykes:Which is, you know, they have you have Linux and all the upstreams. And then downstream from that, you have Fedora, which is open source. It's community driven, but it's Red Hat's community, You know? And then downstream from that are the commercial products, Rell. So Mobi was our fedora, but turns out it wasn't really needed, I think.
Solomon Hykes:So that was a long answer.
Bdougie:Yeah. I mean, I feel sufficiently educated on on that part of the story of the journey. So around this, like, a little shortly after pivot, I joined GitHub, and we built this tool called GitHub Actions. And the thing that I saw really quickly, because it wasn't CICD at first. It was just, like, Docker images that could run flavors of your code as scripts.
Bdougie:And post the Microsoft acquisition, we shipped CICD, like, right after that in 2019. So what we predicted happened where CICD became commoditized, where there were tools people like, like Jenkins and CircleCI. But because you could just add a YAML file in a repo, it was really easy for folks to do GitHub actions. Yeah. So you've you've pivoted your career into well, it's the same dev tools arena, but now you have Dagger.
Bdougie:Yeah. And, you're building a flavor on top of CICD. So do you wanna explain, like like, how this is different than the assumption, which is GitHub Actions or Jenkins and all the other sort of incumbents?
Solomon Hykes:Yeah. So so I guess our our curse as a as a team and as a, you know, as a as a group of, yeah, as a crew, our curse is that we seem to be incapable of just building a better version of something that's already well defined. Yeah. It looks like our curse is to just go and fix the underlying category, you know. And so and and or at least try, you know.
Solomon Hykes:Yeah. And maybe we fail, but the same thing that happened with Docker. We were kind of looking at the landscape. You know, we had these application platforms, very language specific, and then you had infrastructure, which was dominated by VMs. And everyone and he had a 1,000,000 improvements on that.
Solomon Hykes:And our arc was basically saying, this whole thing seems wrong. Let's let's actually change how the industry is structured, which is very arrogant, but it turned out to be correct. It was the air the arrogance of youth kind of, but we're kind of doing the same thing now, and we're looking at these categories called CICD, configuration management, PaaS, ops automation, and dev environments now. And we're thinking that seems wrong. You know?
Solomon Hykes:Like, that's there's a common there's a common thread across all these things. Yeah. Like, the the pipelines don't actually end at what we call CICD. You know? Like, what about in development?
Solomon Hykes:You know, when I'm running a make file and I'm building and I'm testing, that's considered outside of CICD. But shouldn't it be the same pipeline that's then building and testing and deploying it on this on the CI server? It seems like there are artificially artificial boundaries, you know. And so we're our approach is to just kind of rethink the architecture here, you know, and and and say that this could be better. And then our product is a reference implementation of that.
Solomon Hykes:So it specifically worked we're we're saying CICD should cross pre push and post push. It should be the same pipelines. You write them once, and then that you run them during development, and then you run them, during integration after pushing. It should be the same thing. Yeah.
Solomon Hykes:And it should be code. You know? It shouldn't be proprietary YAML. It should be code that that you write in the same language, ideally, that the application is written in. So a Python team, if you're writing Python, a Python application, you should the pipeline to build, test, and deploy that application should also be Python, etcetera.
Solomon Hykes:So that's our approach. Does that answer your question?
Bdougie:It it answered my question. I think where I've I've come to the point where I think I'm I'm sold on what your what your problem you're trying to solve in in Dagger because I I live with this. Mhmm. Before we hit record, I told you that we spend a lot of time in Terraform trying to build what we call the pizza oven. And the pizza oven is our proprietary Yeah.
Bdougie:Situation. Or maybe maybe you call it our Mobi. But it's it's underlying technology that we serve to our our customers. Mhmm. Not today.
Bdougie:Hopefully, in the next couple weeks, we'll we'll serve with our customers. But by the time this comes out, it'll be live. But I'm what I'm getting at is, like, you're back in the position of building the thing that builds a thing. And I think a lot of these conversations that I actually for folks who are building cloud offerings or things in open source, it I was talking to Eric from Stackblitz. They were building a whole startup, and then they end up building the thing that you build your start up with.
Solomon Hykes:Right.
Bdougie:Yeah. And, it sounds like with with Dagger, now you you can write the code. It's actually what GitHub actions originally was gonna be.
Solomon Hykes:Right. I heard that several times. Yeah. Yeah. I've heard from several GitHub people that the, you know, they they there was a first version of GitHub actions that never was.
Solomon Hykes:And when they look at Dagger, they they they it's kind of their outlets. You know, they're excited about it. Like, okay. It may actually happen. But, you know, it's hard for GitHub to do something like Dagger because the whole point of Dagger is that it's platform agnostic.
Solomon Hykes:Yeah. And the whole I mean, the logical thing for GetUpToDo is to lock everyone into their platform. Yeah. You know? And so there's kind of their, you know, opposing forces here.
Bdougie:Yeah. It's it's interesting. The thing to look at us as you see some of these tools grow and, like, we we do see, like, this new wave of CICD. So I don't think I it truly is commoditized in the way that GitHub predicted. I think we could go to that wave for a couple years, 2019, 2020.
Bdougie:But now we have I would love to write Go code to deploy my application and not try to figure out how to write bash again or
Solomon Hykes:do YAML config. Yeah. I mean, our competition is mostly bash scripts, make files, and YAML kind of all mashed together. You know? That's that's what we replace usually.
Solomon Hykes:But, yeah, I I think the next phase for CIC, I think, is an unbundling phase where right now it ties hardware and software together, really. You know, the the compute infrastructure to run the pipelines with the software platform to program the pipelines. They're all you know, it's like these French fries, you know. And we're basically going for a hamburger model Yeah. Where you've got you get you get the best computes to run your pipelines.
Solomon Hykes:Yeah. And maybe commodity compute. It may be from Amazon or Google Cloud or whatever. And then you get the best software platform to develop and collaborate on the pipelines, you know. And and I think there's gonna be a best of breed a set of best of breed compute providers, and I think they're gonna be the same for the to run the pipelines and for the rest.
Solomon Hykes:So, again, it's the hyperscalers. Yeah. And then there's gonna be best of breed software platforms for creating these pipelines. And, you know, we wanna be on that list, you know. But I everyone in the CIC space, I think, is gonna have to adapt to that because they're straddling 2 very different worlds that are actually drifting apart right now.
Solomon Hykes:Yeah. So at some point, you gotta pick which side you're gonna stay on.
Bdougie:Okay. There's a war happening. So I I wanted to ask a real quick question. Folks have probably listened to the entire journey of Dagger. How did they get started, as we wind down this conversation?
Bdougie:Mhmm.
Solomon Hykes:How did Dagger get started?
Bdougie:No. How do you how does how do folks get involved in Dagger?
Solomon Hykes:Oh, sorry. Oh, how
Bdougie:do you get started? This is something I wanna use.
Solomon Hykes:Yeah. Yeah. So, I mean, at Dagger dot io, the the the ideal moment to try Dagger is when part of your CICD pipelines are causing problems. Like, it's too slow or you need to test it and and collaborate on it and it's too painful or it doesn't fit the development workflow, like there's drift between dev and CI environments. And so we you know, Tiger sold all of that.
Solomon Hykes:And, also, if you're generally interested in DevOps, application delivery, and where that space is going, you know, we just, like, nerding out on these topics. We have a really fun community of people exactly like that on our Discord. So I I encourage you to join that Discord and just hang out and see what's happening even if you don't have a specific use case for a dagger because usually yeah. That that's the most fun part, I think. Yeah.
Solomon Hykes:There's just a lot of exciting experiments going on, you know, with with Dagger as a toolkit to to create those experiments. Yeah. Cool.
Bdougie:Well, I'm looking forward to joining the Discord, checking out, and maybe, popping in some of these community calls. Oh, yeah. Yeah. Hope folks who are watching, jump in and stay saucy. Secret Sauce of the podcast produced in house by Open Sauced, the open source intelligence platform providing insight by the slice.
Bdougie:If you're in San Francisco and interested in being a guest on the show, find us on Twitter at saucedopen. And don't forget to check out Open Sauced at opensaucedot pizza.