The Stack Overflow Podcast

Who is building clouds for the independent developer?

Episode Summary

We chat with Gabe Monroy, the new chief product officer at DigitalOcean, about the customers big cloud companies are missing, how to get ready for the millions of new developers poised to enter the market, and the near-universal relevance of XKCD.

Episode Notes

We kick things off by weighing the merits of two gender-neutral regional pronouns: the familiar y’all and the under appreciated yinz. Now that’s covered...

The global population of developers will hit 45 million by 2030, up from 26.9 million in 2021 (EDC). What platforms will they want to build on?

Did Kubernetes solve all your problems? Did it create new ones?

It seems there’s always an XKCD relevant to our conversation. Today, it’s How standards proliferate.

Episode Transcription

Gabe Monroy In large measure, the small businesses and entrepreneur community is being wildly underserved by, you know, kind of the big cloud folks, because as I said, they're out there chasing enterprise revenue. A lot of what they're doing is really focused on lift and shift of existing workloads and trying to get people kind of into cloud. And that's good. You know, there's a lot of systems that can benefit from that move. And so I think the carving out of the niche is just staying laser focused on what are the needs of that smaller team, small developer community and just making sure you're putting together a product roadmap and a delivery plan that's focused on nailing that crowd.

[intro music]

Ben Popper Here's a question for the audience. Is your code compliance ready? Think about it. Stay ready for SOC 2, ISO or HIPAA compliance audits with Rewind backups for GitHub. Rewind offers automatic daily backups and rapid recovery for GitHub with auto reports, Amazon s3 sync, security assessments and more. Learn more at rewind.com/for/github.

BP Hello, everybody. Welcome back to the Stack Overflow Podcast. Today, we are going to be talking about many things cloud, infrastructure, and container. All the good things that we talked about on the podcast. I am Ben Popper, the Director of Content here at Stack Overflow and I'm joined as I often am by my wonderful co-hosts, Cassidy and Ryan. 

Cassidy Williams Hello!

Ryan Donovan Hey y'all!

BP Y'all is my favorite gender neutral term. But yes, I guess that there's a bit of a southern inflection there.

RD In Pittsburgh, we had yinz.

BP Yinz?

CW Oh, yeah! My husband says that. He's from Philly, too.

RD Oh, yeah. Yinz or yunz. 

BP Hey yinz?

RD Hey yinz. 

BP Oh wow, today I learned. Our guest today is Gabe Monroy, the Chief Product Officer at DigitalOcean. So I want to say Gabe, welcome to the show. 

GM Hey, folks, thanks for having me!

BP So Gabe, this is a podcast all about sort of the art and practice of software. So usually we start out we ask folks, tell us about your first computer, your first programming language, how'd you get started with all this stuff?

GM I actually have a picture of this, believe it or not on my first computer was AT&T 6300. And it's probably about like nine years old. And the first programming language I ever started to dabble in was actually C, because it was just because it was kind of what was available at the time. And so, you know, really had to kind of work through that on my own and started, you know, picking up books and stuff. You know, actually, in my early days, I was doing decent amount of kernel programming and stuff like that, just because the lines were blurrier back then between sort of user space, kernel space, it was kind of all one big mess at the time. So yeah, a lot a lot of money in Linux, and mostly centered around game development. Weirdly, that was like the thing that kind of pulled me in early days.

RD A lot of people get there through game.

BP That's a pretty common refrain. You know, I was a kid, I was just in love with gaming. And then that kind of drew me into the world of programmer. That's what I wanted to do, I wanted to build games for myself.

GM And, you know, most of the games back then, like, the reliability wasn't great. So a lot of it was like, Man, I just got the sense that I'll start downloading this or I figured out how to get this thing on a floppy and it didn't work. Damn it, like how do I go fix it? So a lot of problem solving on that space.

BP And so then, yeah, I guess tell us where it went from there. Did you go through some like formal education in CS? Or did you just end up falling into industry? How did you end up working in the world of software?

GM Yeah, I did end up kind of jumping straight into just being a practitioner. One of the first things I did was I set up a bulletin board system when I think I was like, 12 or something. And it actually became pretty popular. And I had like a bank of four modems. I remember, it was called gostrey BBs, you still find references to it online. But yeah, it was it was really fun and interesting to see the this idea of a community of folks and you know, all the tech require to kind of get get that all up and running and maintain kind of version compatibility with the latest bulletin board software. There's a lot of that work. And you know, from there, I started going to consulting and just kind of helping folks set up networks, you know, land tastic and just like slang and cable and just you know, this before the days of cable we were doing, I forget what preceded that, but a lot of kind of wiring and stuff like that. And then later on, got a job that was a company that ultimately got acquired by Intuit and I was doing call center work. It was like, sort of level three on-call ultimately, and so it was a lot of AI x and you know, a lot of just like opsi stuff, and I got a real big I don't know I learned a lot about active passive failover systems and just kind of old school IBM, like tech P series, E series, you know, servers and stuff like that. And, you know, kind of from there, I wound up on my way into the world of startups in New York City and Yo ultimately started my own companies started and sold two companies, the last one to Microsoft came to Microsoft and fortunate to pair up with Brendan Burns over there, the co-creator of Kubernetes. And Brandon, I just, you know, did a lot of stuff and open source and Azure services are really driving kind of this idea of openness and open source inside Microsoft. And, you know, I think one of the things I'm more proud of is, you know, we built the fastest growing service in the history of Azure with AKS. I think that was a pretty cool, pretty cool thing. And later on, you know, I ended up leaving, and currently, on week four, of Chief Product Officer at DigitalOcean, where I'm gonna be really laser focused on, you know, a different segment of the market really enabling and empower developers, which I'm super excited about.

CW Dang, week four, that's exciting. And so what was the thing that made you move over to DigitalOcean? Because they are doing a lot of cool things.

GM You know, honestly, for me, it was really about this idea of 44 million new developers are going to rock up into this space by the year 2030. That's like stats from IDC, right. And I'm a very mission driven person. And I really want to make sure I can do my part and empowering you know, who are those people? Right? They're young people, they're students, they're startups, entrepreneurs. And I just kind of felt like the big cloud companies. They're extremely focused on enterprise revenue, right? And it makes sense, right, because it's all about digital transformation. For these enterprises. The stakes are really high. cloud companies are getting paid handsomely for this, right. So it works out for everybody, the world needs it. But the world also needs to be focused on empowering the next 44 million developers who are gonna roll into the space. And that's really where my heart is. I think DigitalOcean is for out of any of the cloud companies, the best, I'd say, to kind of drive that mission forward. That's certainly my point of view.

RD Yeah. I mean, when I was thinking about this, like cloud computing spaces, is one of those spaces where all the big players are in it, right? How do you carve out a niche as a sort of cloud focused company?

GM It's a great question, right? Because the space is crowded, as you say. But I do think that in large measure, the small businesses and entrepreneur community is being wildly underserved, by you know, king of the big cloud, folks. Because as I said, they're they're out there chasing enterprise revenue, a lot of what they're doing is really focused on lift and shift of existing workloads and trying to get people kind of into cloud. And that's good the years right, you know, there's a lot of systems that can benefit from that move. And so I think the carving out of the niches, you're staying laser focused on your water, the needs of that smaller team, small developer community, and just making sure you're putting together a product roadmap and a delivery plan that's focused on nailing that crowd. And it's cool, because like in leadership team meetings already at DigitalOcean, it's like, that's what we talked about, like we talk about, what does the small business community need? Everyone in the entire company is laser focused on that. I think that's different. And I think that leads to a different set of outcomes.

BP Yeah. What you said makes a lot of sense to me is really interesting. Sometimes when I hear right, these stats, like the 44 million, it just seems incomprehensible to me, you know, like--

CW Yeah, it's such a big number.

BP Where are all these people coming from? And like, where are they going to sit? I don't know. [Cassidy laughs] But yeah, like, I think, you know, one of the things that you said that really did, you know, stand out to me was like, at a corporate level, you can get trapped in sort of that SaaS mentality, that enterprise mentality of like, what gets us the best valuation and what's gonna make sure the business thrives. But some of the stuff that I think is most exciting to people who are hands on working in development in open source is like, How can I be impactful, and empowering to the next generation and make sure everybody has access to be able to learn these skills and to come into this industry if they want to? But Cassidy, maybe you could speak to that a little because I know it's something you've been thinking about. 

CW I love the idea of empowering individual developers as much as possible, indie developers, indie hackers, whatever you want to call them, those small teams, because I feel like they're very often underserved because companies focus so much on the enterprise side of things, and they want to get the big money from the big tech. And it makes sense, you do get a lot of money from the larger companies. But I feel like the most creativity and a lot of the innovation comes from those smaller teams and those individual developers getting the education they need to build things that are truly innovative, that another company might take a lot more time to build because they have so many processes in place.

GM Totally guys. And I think one of the big surprises for me, you're having spent so much time in big cloud is that you're that phenomenon of small teams and small village. Like, I think when people think about cloud, I think the promise of cloud was that like developers are going to be empowered, right? Like they were all going to be able to just kind of take their credit card and roll up and get the services they need and start innovating. And actually, the way this is this works in most cases is it's actually through central IT and these enterprise infrastructure and operations teams, that a lot of the kind of cloud mechanics works behind the scenes. And so a lot of the products that end up getting built that are ultimately, you know, intended for developers, there's kind of this like self through central IT, motion. And there's a lot of features and a lot of work being put in to kind of empower those folks. So those are controls and enterprise policy and all the stuff, which is important. I'm not saying it's not important, but that is ultimately overpowering the needs of like the self serve, and I still don't think we've seen the unleashing of developers, you know, who can just kind of roll up the cloud and get what they need done. I think it's companies like DigitalOcean, like, frankly, Netlify that are actually focused on that part of the space and likely to deliver for developers in the end.

RD So you've been in the the cloud space for a while now, almost five years or so. Is that right? Have you seen the sort of needs change based on some of the machine learning the big, like, massively parallel applications that they use, like GPUs and such?

GM Absolutely, and I think, you know, my view on this is, there's this concept of, you know, data gravity in kind of, like, the enterprise and cloud space, it's like, where your data, you know, lives is, you know, it's gonna keep you happy on a given platform. But I think when you really start to unpack that, that what the future holds here is it's really about data to enable machine learning into enable AI and those insights for folks. And so I've seen a ton of that, you know, but I think at this point, I would still classify the world of AI and ML as something reserved for folks who have a specialized set of skills, right, we're not yet seeing this really democratize broadly. And I think you're one of the things I'm really interested in is working on the inferencing side of the house, right, I think if you can get more developers using, you know, facial recognition, or sentiment analysis or speech to text through off the shelf models, right, not custom train models, but basic models, and understanding how to build AI enabled applications, that's going to start to create more demand for the ML side of the equation, where you can start to use the data that you've stored with your cloud partner, to kind of enhance and enrich those models and really make the inferencing more applicable to your business. We're the early days of that stuff. And when I look at the work I've been exposed to around GPT, three, and the rest of it, you know, because I have worked closely with open AI in the past, it's still reserved for the technical elite. And I think that's a big opportunity.

RD We get a lot of pitches from Nigeria, from data science people. So there's a huge contingent of people up and coming into that.

GM Yeah, and that's an interesting, you know, thing that I've been wrestling with is like this concept of data scientists. And yeah, I do wonder if like, if there's an opportunity to bring some of that data, science expertise and kind of that discipline to the professional developer audience, right? Because, you know, right now, sometimes to be it feels like a lot of the AI, ML nirvana is being held back by a lack of data science capability. And I think it's my sense is that, like, professional developers are eager to learn the space, but it's not super straightforward. I mean, it's very different than, you know, writing back end software or front end software. And so I really wonder, you know, is there a bridge we can build to help empower professional developers to help out in that space? Or is it something where like, now you're gonna have to go to school to really understand the space, and you're kind of bound by the skilling and talent development there? I don't know, something I wonder about.

BP That's interesting. I mean, like, when you talk about open AI, and needing a certain level of technical sophistication, I think one of the things that I always find really interesting about like, GPT 2 and 3 was that they would sort of throw it out there. And you would see people messing around with these, you know, fun little experiments, kind of showing off what it could do. And then that kind of almost led back like those some of those ideas seem like they inspired, you know, GitHub Copilot, and things like that. Do you mean just that like to be able to work with it, you need to have a certain level of technical sophistication, or do you mean, like to be inside of a company getting to do that kind of work? Because like, one of the things I guess that was sort of inspiring to me about that was the way Open AI will throw it out there. People will experiment and That almost seems to have a feedback loop with some of you know, what they productize

GM Yeah, I guess the point I'm trying to make is that, you know, GitHub Copilot, I think is a really great example, right? Where someone ran an end to end product experience. And you know, you read a comment inside of your IDE, and boom, you get your code written for you and have a huge debate the quality of the code, but it's certainly impressive, and I think everyone, everyone will agree that it's a pretty big step up in terms of innovation. But I'm not seeing yet the average company with the average--and I don't mean average in terms of like mediocre engineers, I mean, like average, you have non GitHub company, being able to put that same sort of product and end to end AI, ML experience baked into their applications. And I do believe that that's the future. Right, I think that's where we're heading. And I think there's a lot of work that we have to do to sort of unlock that future vision. And that's definitely something I've got my eyes on.

BP So you're saying it's gonna be exciting when the technologies that now are mostly being used by the trillion dollar companies trickle down and become accessible to everyone to build cool new tools and products and experiences?

GM Absolutely. Right. Because GitHub, Microsoft, Google, Amazon, they don't have a monopoly on wisdom. There's a lot of companies out there, you know, smaller companies in the space, you know, innovative entrepreneurs who are gonna put their own spin on something like a GitHub Copilot, or something like a GPT 3. And that is super exciting to me, like, like, when you unleash all that innovation upon this technology foundation, that's rather impressive. Who knows what we're going to find, right. But my guess is that that kind of thing is going to change the world.

BP So I wanted to just go back for a second, you mentioned something early on about working with Kubernetes at Azure. I know, those are two things that are both popular with our community and on our podcast, and also things that we're, you know, thinking about using internally. So can you tell us a little bit about that project? What you worked on? Sort of like soup to nuts, like how did it come to be had you end up working on it? You know, as you went in? What were you hoping? And then you know, how did you work through obstacles and things along the way to get to that sort of final result?

GM Yeah, so this is a pretty fascinating story. And it's funny, because I'm involved in Kubernetes, everybody's community pretty deeply a founding member of CN CF, etc, etc. But I actually pretty mixed, you know, relationship with with the project, you know, particularly around developer experience.

BP You've been through the drama.

GM I've got drama, for sure. You know, I was building pas systems, you have to empower your development teams inside of companies. And that was what we were doing over days and days was effectively like Heroku, like pas experience, built on top of ultimately, Kubernetes. But in the early days, like we were using Chef, we were using whatever tech we could, and you're close with Solomon hikes in the Docker project, I was actually the largest external contributor to Docker myself for a few months, back in the early days of the project that wrote the volumes of systems. Have you ever done bind mounts or volumes and Docker, I wrote much of that code, fortunately, all gone now, because it was garbage code. Later, you know, I ended up getting in touch with Brendan, because we were looking for better orchestration stuff. And so Brendan Burns and I started working on that. But like, my focus was always in building platforms for developers and I was using Kubernetes as just kind of underpinnings to to enable that stuff. You know, certainly Brendan and I had this belief that Kubernetes was never intended to be like the final layer, where your developers were interacting with it directly. It was intended to be a principle, that abstraction that would sit on top of the operating system, on top of compute network storage, and allow us to build up you know, principally towards, you know, towards that. Because I think one of the things that people have seen over and over time and time again, in the past face is people build a great pass. Like, I'll take Heroku as an example. People love Heroku. I love Heroku. Right. It's fantastic. Until it's not, right? And all the reasons that you love the product experience, and the kind of gated barriers that that provided at certain points, like wait, I need to put a JMX monitoring port on this Java app. And the past doesn't let me do it. Whoops, I have to like, drop off the attraction, you know, and just kind of bail on the past. Well, with Kubernetes. If you're building the platform on top of Kubernetes, when you drop down, it's not super painful. And you can kind of see the other workloads that are running. So that I think was one of the promises for me. But as I got further into it, I started realizing that there was a lot of demand for Kubernetes raw, right, like the platforms that we're building were interesting, for sure, but like people really want other Kubernetes. But as I dug into it more, I really realized it wasn't in developers. It was operations teams. It was people who were previously slanging Puppet and Chef, you know, now they're looking at, you know, containers and stuff like that. And I don't think Kubernetes has really served the needs of the developer community. I don't think it's going to beyond being a set of plumbing that I think is something we all can agree upon, or most of us can agree upon. And yeah, that puts me a bit at odds with some folks who are rock Kubernetes API. But yeah, more I can say on that. But that's that's definitely Kubernetes is not the end all be all for developers.

RD Do you think those ops folks kind of miss the very low level of contact with server operations?

GM You know, you go to KubCon as I've done many years. They certainly seem excited about it. Certainly, you know, a bunch of them seem super excited. But I think the idea of immutable infrastructure is really compelling and for good reason. I think there's a setup problems as an industry we have to solve. One of the big elephants in the room is this idea of like patching. It used to be that with non immutable infrastructure and all these servers running, you can run patch management software of zero day comes out, and it's like, alright, well, we're gonna patch the zero day. And you know, today with immutable infrastructure, it's like, Oh, we got to run all these CI pipelines again, and you know, fingers crossed, where we've got them all. And so there's a bunch of problems with kind of the current model where you're proceeding around security, where I think a lot of those ops teams are really struggling today with, you know, how to provide end to end security on that. So in some ways, I worry that Kubernetes has created as many problems as it solves.

BP That's always the way it is. You get things that come into play, solve the problems, and suddenly you're like, this is too much conflict. Can we unbuild this now?

CW It's like that XKCD comic was just like, wait a minute, there's five different standards? We should make a new standard to just fix all of this. And it's like, now there's six different standards.

RD Wanted to get back to the developer focus, like I've heard of a bunch of people just spinning up a little AWS space and running easy to run spaces kind of get out of hand and cost people an arm and a leg. How do you find the middle ground?

GM So you're talking about sort of this idea of costs, and like how to do cost management for some of these ideas? 

RD Yeah. I mean, I've had people assigned to just pare down the the cloud compute space.

GM It's a really fascinating question. And I'm not sure there's an easy answer. You're beyond. It's just like making sure you're managing costs. Like you know what one interesting story I like to tell you back when I used to work at data centers and was you know, slinging infra in New York City. It used to cost like, I was paying like 250k a month for a SAVVIS facility. And so the cost was, right. And this is like for startups, right? It's like you have any validated a product yet, right? And you're the capital outlays is massive, right? And so today, it's like, well, you can get started in one of these spaces, DigitalOcean, $5 droplet, right, get started on an idea. And you can kind of scale your costs up as you get traction on your product. And I think that really unlocks innovation. The trick, of course, for these businesses is managing when the thing scales up, you know, when that idea and when you reach that product market fit with that idea and your bill starts to explode. Like, are you prepared for that? The only thing I think the good news is that, you know, most providers have good answers around cost management. And, you know, I would argue that not having to do the capital outlay that I was talking about before, and then having the issue on the other side of your bills going out of control. That's probably a better space to be in than blocking innovation for folks due to the fact that they can't reach, you know, cloud or can't reach IT infrastructure due to the cost of it.

[music]

BP Alright, everybody, thank you so much for listening. It is that time of the show, I'm going to shout out the winner of a lifeboat badge, someone who came on Stack Overflow and help save a question from the dustbin of history. Awarded 18 hours ago to Prafulla. 'How to import an Excel file into a SQL Server.' Alright, if you're having trouble with this, we got an answer for you. This one's five years old, it's been viewed 312,000 times. 

CW Dang.

BP Wow, that might be the most number of views I've ever seen on a question. So clearly something a few people have struggled with. I am Ben Popper. I'm the Director of Content here at Stack Overflow. You can always find me on Twitter @BenPopper, email us podcast@stackoverflow.com with questions or suggestions, and if you'd like to show leave us a rating and a review. It really helps.

RD I'm Ryan Donovan. I edit the blog and the newsletter here at Stack Overflow. I am on Twitter @RThorDonovan. And if you have a great idea for a blog article, please email me at pitches@stackoverflow.com.

CW I'm Cassidy Williams, Director of Developer experience at Netlify. You can find me @cassidoo on most things

GM And I'm Gabe Monroy, Chief Product Officer at DigitalOcean. You can find me on Twitter @Gabe_Monroy and yeah, definitely check out my blog post on why I joined DO, it talks a lot about what I think what the opportunities for DigitalOcean and for the professional developer community going forward.

BP Alright everybody, thanks for listening and we'll talk to you soon.

[outro music]