Rethinking DevOps and Infrastructure as Code with System Initiative’s Adam Jacob

This is Textron tv. Hey everyone. Welcome back here to techron tv.

You know, unfortunately, my next guest, he comes on here too few and far between, I think the last time he is on, I I bet it's been almost a year since Adam has been here with us. But he, it's always a breath of fresh air and it kind of charges my batteries to, to have him on here. Oh, thank you.

So I'm, I'm happy to, uh, thank you. I'm happy to introduce you to Adam Jacob. Adam is the co-founder and CEO of a company called System Initiative.

But he's done so much more. Adam, welcome back to Text Drug tv. It's great to see you, my friend.

Um, a lot of people out here may not know Adam Jacob. com 10, 11 years ago. I think most of my audience knew, or at least heard of Adam Jacob, but we, it's so much more.

Yeah, I mean, what a weird thing that, uh, that, that what we do is big enough that people knew who I was. But yeah. Uh, I'm Adam.

I, uh, I'm basically a systems administrator back when that was a job title. It's really sort of who I am as a person. I love infrastructure, uh, I love and I love sort of building tools for people who builds things.

Um, and then I started a company called Chef, which did configuration management, which sort of, uh, predates DevOps, but then sort of became a lot of the core automation stuff that was DevOps and then inspired like all the infrastructure as code tools and a bunch of those things along with Puppet and a bunch of other sort of tools of that era. And then System Initiative, uh, is the company that I have now. And, you know, its real goal is to, is to say that, you know, all that work that we did while fantastic and certainly better than it was when we started, uh, in, you know, the late nineties, early two thousands, um, doing sort of ubiquitous automation.

The end results have sort of been kind of mediocre. You know, if you, if you look at what people wanted to get out of it, especially in the large enterprise, but kind of everywhere there's a, what we actually wound up with was a bit of a mess, primarily from the user experience angle. So if you think about what it's like to do this work now, it's kind of sucky.

Like it just is really hard. Um, and, and most of that is because of the way we interact. Um, and so I became convinced from my experience as Chef and my time sort of helping other people build and deploy these large complex DevOps solutions that ultimately we needed to fix the user experience problem if we wanted to have better outcomes.

And so system initiative is our attempt to to, to fix those user experience problems. I love it. Before we go any further, just for anyone who wants to like go check out System initiative, what's the website?

com. Um, and yeah, they can, they can check it out. It's all open source.

You can check out the source code. You can build it yourself. It's, uh, it's about to be, uh, like knock on wood, you'll be able to just go like, sign up and use system initiative as a SaaS here in a hot minute.

So, wow. Not yet, but soon. All right.

So you're gonna come back when that's on and, and Yeah. Kind of announce it here. We'll get, yeah, we'll go deeper kind of there on ramp.

Totally. Very cool. So Adam, you spoke a lot about, I mean, things have changed since Chef and Puppet and Ansible and those kind of tools and Yeah.

Gave rise to all infrastructure as code. Um, you know, today we have things like platform engineering and S res, and we're shifting left. Then we have GI ops and the whole cloud native stack, and you know it.

And then of course we can mention ai, right? Sure. And the effect that that's happening.

Are we even having a conversation if we're not talking about ai? Am I really here or am I an AI Who even knows we have to pay our tax to the bas? Yeah, Absolutely.

So, you know, but the times they are a changing, as someone once wrote a song about, right? Yeah. And so what does that mean?

What does that mean for our core DevOps crowd? What does it mean for, uh, you know, for the way ops is done, for the way software is made for the way we deliver? Right.

Yeah. You know, I, it's interesting. I, I'm 46, um, and I didn't really, you know, tech always moves through cycles.

I sort of had a, I naively obviously, it sort of felt like all the progress that we had made wasn't cyclical, you know what I mean? Like, we were really, we had changed it and now it was going to sort of stay, Right? It's a new paradigm.

Yeah. And I think, I think kind of what we're learning a little is that it is a little cyclical and people do forget why we learned the lessons that we learned. And, you know, the core of what DevOps is and its principles, you know, and, and the things that make us effective.

You know, teams are more effective when they work closer together. They are better when they have, uh, you know, cross-sectional groups of experts who are working together to solve the same problem. It is better when you build the application and the infrastructure holistically together at the same time.

'cause they're identical, right? Infrastructure without applications is meaningless. Applications without infrastructure don't run.

So, like, you know, it is obviously true that those things are better. And when we design the systems for them to be better, it is, it works. Um, I think, I think what we're learning now is that that user experience of doing the work of what are all the things I have to stitch together?

How do I put it all together? You know, you rattled off a bunch of technologies. You have to know all of those technologies.

You also have to understand the things that they abstract, right? So maybe they abstract the deployment process, or maybe they abstract AWS or maybe they, whatever it is they do. And you have to load all that context into your head, and then you have to be able to orchestrate it all together in a way that makes sense.

And it's brutal. It's brutal. It's hard to understand.

It's, it's difficult to work with. Um, and so I think we're just in a moment where we're recognizing that the results and the experience of doing that work don't match. And we're asking ourselves, what are we gonna do about it?

And so, you know, one answer is sort of go back to some of the old ways. Like maybe we went too far, you know, shifting left. And what we need to do is have a better separation between ops and, and Deb, maybe the problem was that we didn't have a hard enough separation, you know, which is, yeah.

One of the arguments I see getting made, they're not making it that way. 'cause, 'cause whatever they're disrupting, you know, then they're like, right, we've discovered this new thing that your ops team should be behind an API. And I'm like, man, it was always behind an API.

It's just sometimes it was a ticket, you know? But like, but that's 'cause I'm an old person and I'm, you know, whatever, having grumpy old person syndrome, you know, not even that old, but, you know, like there's, You're not that old. Wait, wait, you got time yet to get Grumpy?

It's only gonna ha Yeah. But I'm, I'm kind of into it. Like, I'm happier now than I was in my twenties.

So like, let's go for 60. I'm gonna be stoked. So I think there's like a, Absolutely.

Anyway, I, I think there's a, uh, but I think that's what's happening. I think that we're in this moment of transition where, where as an industry, the recognition that the results tend toward mediocre, they, and, and that mediocrity and I think was a growing recognition that, that mediocrity isn't the fault of an individual tool. It's not like one thing or another sucks.

It's that when you put it all together, the experience of doing it is just not good. So a couple of thoughts on that, Adam. First of all, what you said about, you know, it, it's almost, uh, you know, how things become more complex, less complex over retire rate it, the, the, the fact is, if you don't ride your P's and Q's and constantly weed your guard and do tend to, to drop into a state of mediocrity, I think, right?

Yeah. It, it's hard to keep your edge. It's hard to keep improving.

Um, but I I also think, you know, like one of the knocks you're hearing from some, some of the, you know, peanut gallery is, well, DevOps doesn't scale. Once you get over a couple of hundred developers, it really kind of just bogs down. Sure.

I mean, look, that's the same argument people make about software develop every software develop development methodology that's ever existed. Yep. You know, it, it just turns out that it's hard to, um, to align strategy as you get larger.

So strategy, I, my favorite definition of strategy is John Lewis Gatti's definition, which is how do we align our limited resources with our unlimited aspirations across time, space, and scale? I love It. And, uh, and when you, when you think about that definition of strategy, and then you think about why is it hard to, to scale our DevOps problems, the problem is that it, it's not that it's hard to scale DevOps any more than, it's hard to sale scale anything, any given software development methodology.

What's hard is how do I keep all of those people connected across timespace and scale to the same thread and to the same objective and to the same outcome, and the same, the same focus. And I think, again, when the, when the process we use, or the tools we use get in the way of that communication, they become the strategy, right? Like, like we become our, our, our, our ability to change tactics at the low level of scale.

You know, an individual developer doing work, an individual team trying to ship something breaks down. Because the complexity of the systems that we put in place to try to automate them, constrains them in ways that don't allow them to, to do what they need to do in a way that makes sense to them. Because, and they're the closest to the problem.

And so then, you know, in order to solve their problem, they have to go up the chain, right? They have to figure out, well, I have to break these rules. I have to change these constraints.

I have to do this, I have to do that. And, you know, it breaks down, but it's not necessarily because DevOps is the problem. You know what I mean?

Yeah. Any more than agile is the problem. The problem is, well, it's really hard to communicate in that way.

Well, you Hear people saying agile is the problem now too. Yeah, Of course. And like, everything's the problem.

And, and well, look, it is, right? If you look at how we, you know, like if you're doing like the scaled Agile framework or whatever, I don't think that's a framework designed to deliver strategy, you know, outcomes through strategy, under data assist definition to customers. I think it's a way to think about resource optimization.

It's a way to think about, you know, it, it tends to think, one of the things that I think this is a particularly spicy take, but, um, you know, well, a lot of our industry focuses and has learned a lot from, from factory optimization, right? We look at, we, we, we look at those folks for inspiration, and there's a ton of inspiration to be found there. However, the work that we do to produce this stuff is not a factory, right?

Like, these are teams of highly creative people doing highly creative work trying. And we never do the same thing twice. You don't see an engineer stamp out the same line of code twice, obviously, by definition, that's exactly what factories are designed to do, right?

Yeah. And so when we start to design our companies and we start to design the way we interact with each other, it's not that those micro those optimizations for, for throughput or for whatever are bad necessarily, but they can build you something where if your goal is to produce the best outcome they might get in the way because the creativity of the people in the middle gets a little stifled. Yeah.

And I think one of the things we know from DevOps is that the teams who did best, they were highly creative. They did have really wide remit. They could go wherever they needed to go.

And that's very difficult to do in practice, to your point, it's very hard to keep at that Level and very hard to Scale. Very hard to scale. And like, you know, it's not impossible to scale, right?

Yeah. Um, but, but it's hard. And the hard parts about it tend to be, um, they're both systemic.

So how do we build the systems that allow people to do their job? Because that is the, the, the expression of our culture, right? We talk about culture is separate from technology.

That's a lie. Like the way we do the work is the culture in a very meaningful way. 'cause culture is what you do, not what you say.

And, you know, and how we connect that up. So it is a systems problem and it's a, a social problem still. So, yeah.

I think the other, another thing at play there, Adam, is that there's a, there's a double-edged sword to the whole, hard to pin down what DevOps is, right? Hmm. You know, purposely there was, there's no manifesto here.

There's no Yeah. Sort of definition that we all gather around. Everybody has their own kinda little thing.

And so people were like, it just made it, it makes it easy to scapegoat. It does. Is what it comes down to.

It does, yeah. Right. Because it's it is, but it isn't, it's not that.

But it is this, it isn't that, it's this. Sure. Yeah.

It, I mean, this is the same thing with open source. If you ask most software developers, if they understand open source, they'll say yes. And then if you ask them to name the three legal levers that open source uses to, or what its values are, and why they can't answer those questions, no.

But they know how to use GitHub. You know what I mean? So like DevOps is the same, but they weren't free in beer.

Not, you know, they don't care as much about the free and freedom part of it Until they do. I it's, it's an interesting Until They have to, Until they do. Yeah.

Right? Mm-Hmm. It's, I just, I think that you can, it does make it easy to scapegoat.

I think one of the reasons that I, you know, am am very purposefully saying that what system initiative does is, is trying to reinvent how DevOps functions. It's trying to figure out how do we solve this user experience problem is that I really do believe that those core fundamental ideas that were present in DevOps are Right. Like I went through that.

I, I know, I, I believe it to be true. I've seen it happen. I know that when it happens, it's transformative.

I know that it can transform really large organizations and has, and I think the, and all of that's true, and for too many people, it's locked behind the gates of this, like, of this difficult experience of getting to the top of that mountain. And, you know, um, if, if you ever played video games, like there's some video games that are just really hard, dark souls, uh, really those sorts of things. Uh, you know, getting great at DevOps organizationally is a little, like getting great at a very, very difficult video game.

You know, and I love it. It's, it's, once you do it, you love the game. You're like, this is the best game I've ever played because I'm a master at it.

And then some poor person tries it and they're like, ah, you know, I got stabbed by the first guy and I can never make it through this game. You know? Right.

So like, that's a little where we are. And I think it makes sense that people are trying to brand it differently. That they're, you know, talking about some, some other way of doing platform engineering or SRE or whatever we want it to do, because, you know, it's always tempting to, to say, well, let's get rid of this branding because it's got too much baggage.

I think my concern is that we throw the baby out with the bath water, right. That there is this like, or truth of what it is and does that I think is really meaningful and very transformational. And we also have to recognize that the outcomes aren't what we want.

And so, you know, we need to usher in a new era of more creative solutions because our implementation is letting us down. I want to come back to that in a second, but first I wanna circle back. You're right.

In the beginning you mentioned you're 46 years old and you think you're old. I'm, what am I supposed to say? I, I, you know, I got 15 years on you, but I, I think what happens more than Technology terms, Right?

Well, because technology, because it moves so fast, five to 10 year lifespan, usually. Yeah. Yeah.

But I, I think what also happens, and I've seen this, you know, I've been in tech 30 years. Yeah. I've seen that every new generation comes in, and it's not that they wanna reinvent the wheel, but they have, they think they have a new, unique way of looking at things and what's gone on before it's very easy to point fingers and show its warts.

Yeah. But they, thanks, they lose sight of what it was before. If you think this is bad with DevOps, you should have been here before Agile and DevOps were Around you.

It was worse. And Then you would've had something to really b***h about. Yeah.

But, you know, and, and it's just the way, it's, it's how way in the world. It's why kids always think their parents are right f*****g nuts until happens. It happens in music.

It's just, it's, it's just how it is. It's fine. That's The way deal of the World deal.

It's, it's, it's, you know, I think Yeah. And I think you just have to accept it. It's normal.

It's fine. I think the, you don't have to like fight about it. You don't have to, there's nothing to save.

There's nothing worth saving, right? Mm-Hmm. What all that really matters is, you know, what's the, how's it feel for the people doing the work?

And is it better for those people today compared to how it was before? And, you know, I, I love building, I love building tools for people who do operations work and do, you know, deploy applications like those are, because that's what I love. And so I love building those tools for those people, because that's what I love to do.

And so, like, I kind of understand them and, you know, uh, and I, I think it's fun. I Yep. You know, that it also is, can be transformative in large organizations is even better because that's also super fun.

Like, it's super fun to go into a large organization and figure out how to optimize it or figure out how to like, make it work better. And I think, well, one of the things we've lost a little is that this is a joyful thing, right. That that's what, that's the fun part of it.

Right? And, and so, you know, work, people ask me about tech Or can be what, What, what, why do I love tech? Why do most tech people, it's the problem solving aspect.

Security. Yeah. I've been in security 25 years.

The fact that I can break something and then make it not breakable or not so easily breakable. That's the fun part. It's Fun.

Yeah. I know. And, and people lose sight.

And A lot of DevOps at this point has become really unfun. And so like, you kind of have to recognize that that's true. You know, really.

Like, in the same way that I, you know, like I love music and I grew up with like, hair metal, lots of hair metal bands, and like Mm-Hmm. That music was really fun. And then it died because like Guns n Roses and grunge happened.

Yeah. And next thing you know, like, it was like a completely different era. And it was because it stopped being fun and it started to be a parody of itself.

It started to be like, oh, this isn't fun anymore. Like the party's over tap. And what we need now is like, we need more authenticity.

We need more, you know, I could smell the mold coming off the amplifiers or whatever, you know? And like, it's cool. It's the same kind of vibe that's happening now.

It's just, you know, as we, as we look at what we produce and we look at what the art is and we look at how we ask people to do that work, part of what makes it fun is that you find a way to reinvent it. You find a way to change it. You find a way to, to, to take what was good, to make you good.

Yeah. To take what was good from what came before, and then push it into a new area. You know, push it into a new way of being, push it into a new expression.

And I think we're in a moment where that's happening. For DevOps, we should embrace that truth. That's fantastic.

What a lovely, those are lovely moments, uh, to be in. Um, they're a little tumultuous. They're a little uncomfortable, but like in the end, it tends toward progress.

So I think we're in violent agreement here. Yeah. Give me, gimme some easy to sink your teeth in examples, into examples of how what we do to reinvent DevOps, Adam.

Yeah. Okay, so here's an example. So if you think about what's required to get, um, let's just talk about infrastructure.

So if you look at infrastructure as code, uh, so you look at like Terraform, Lummi, the CDK, um, sort of that group of tooling. Um, and then you think about how we build and manage those things and how we attach them to their deployment process. And then how you attach them to like Kubernetes, um, the list of tooling that goes into that, let's just call it long, right?

So there's like, oh, I'm using Terraform, but then I need a way to run Terraform on demand. So am I gonna use Terraform Cloud? Am I do something else?

Then you gotta think about like, well, how's that interact with gi? What am I using for gi? How does that get triggered?

How do I, how do I then think about like, well, I wanna do GI ops for all my Kubernetes stuff. Does that overlap? How do they get triggered?

When do they get triggered? Do we do infrastructure first? All these decisions you have to make that sort of pull those things together, all of which assume you understand how the underlying technology you're trying to automate functions.

So if you don't know, you know how I am works and you try to use Terraform, you're learning both I am and Terraform at the same time. And that's harder than than you'd think, right? Right.

And, and so when we throw that at developers and we say, we shift it left, and we're like, Hey, don't worry, you have full control over your own stack. Here's a present of this big grab bag of technology that you have to stitch together. Understandably, they don't love it.

Yeah. Because it's sort of brought with peril. So one example, um, there's system initiative.

We think about doing that by taking, basically taking all that data and instead of making it code, we turn it into data. We say, what if this was information that you could operate on as opposed to code that you write and then refactor? And how would that change the shape of what we do?

And so it enables us to build new user experiences, new ways of inferring data, new ways of having more intelligent automation, new user interfaces become possible. It can become multiplayer. It can be kind of like work and Figma, which is cool, right?

Um, it has this very fundamentally different underpinning, which is this big hypergraph of functions. And so, you know, how you program it is very different, right? You program it in the, in the application itself.

Like you log into system initiative, you open an editor if you want to, and you start changing JavaScript functions, which change how the graph behaves super different from writing infrastructures code. You can see a similar kind of thing with, um, like Wing, you know, eLog from Wing has this vision that says, well, what we need is a new programming paradigm. That, that if we want the infrastructure and the application to be more seamless, then what we need to do is build a programming language that makes it seamless.

And then we can use, through this programming language, we can change the user experience. So, for example, he builds a simulator inside the language for all the infrastructure, the language supports. So as an application developer, you can write your cloud native application, you can then run that application in a simulator, which gives you this fast local dev loop independent of the technologies that are implementing on the other side.

Super interesting, very cool, very different point of view about how to solve what is fundamentally the same problem, right? And those are just two examples that could come up with more. Um, but you know, they're weird.

They're not just like, they're not like a another layer on top of the existing substrate. They're fundamental reimaginings of, of what the foundational technology of this could be. And to me, that's where all the excitement is right now, right?

It's like, let's, let's break, let's break the ground. Like we can't just build on the house we already have. 'cause it's not delivering what we want because it's wrong by design.

So, you know, then You gotta redesign, make new designs. You gotta, you gotta, you gotta, you gotta, you gotta take what was good about what you had before and then you gotta rip it up a little bit. You gotta get a little punk rocky on it.

Yeah. I love it. Adam, we're at the top of the hour here.

Well, I don't know when people will be watching this up. It'll be the top of the hour, but at the top of the hour we're, when you and I are recording it. Yeah.

And, um, we gotta run. It's been a pleasure having you on edge. Always, Always great to talk to you, Alan, So don't be away.

Don't be away so long. Come back. Tell us more.

Well, I already got you. I'll done whenever you want. Nailed down when this SaaS thing out, right?

Yeah, yeah. As soon as that's happening, of course I'll be in, I'll, I'll talk to everyone. I mean, I love talking to anyone who will talk to me anyway, but, uh, all right.

Yeah, certainly when what I'm doing is launching a company more fully and way be book can you like, you won't be able to shut me up. So it's all good. I'll be here.

I'm gonna hold you to that, man. Good luck with this launch. We'll be looking thank you for it.

System initiative. com, right? INI.

You got it? Yeah. Alright.

Adam, Jacob is here. Uh, Adam Jacob here on Techstrong tv. We'll be back in a.