Rails Business
Brendan Buckingham and Ryan Frisch talk about developing with Ruby on Rails and how to leverage it to build a business.
Rails Business
Ernesto Tagwerker
Use Left/Right to seek, Home/End to jump to start or end. Hold shift to jump forward or backward.
In this episode, the hosts are joined by Ernesto Tagwerker, founder of Ombuabs and FastRuby.io, a company specializing in Ruby on Rails upgrades as well as AI solutions.
Topics covered include:
- Rails Upgrades
- Technical Debt
- Developer Onboarding and more!
Chapters
00:00 Introduction to Ernesto Tagwerker
00:21 Ernesto's Developer Origin Story
01:10 Discovering and Embracing Rails
02:19 Building a Business Around Rails
02:51 Challenges and Solutions in Rails Upgrades
03:40 Client Needs and Service Offerings
06:18 Content Creation and Misconceptions
07:33 Best Practices for Rails Upgrades
10:34 Open Source Contributions and Tools
13:32 AI and Machine Learning Initiatives
18:27 Developer Experience and Technical Debt
29:46 Advice for Solo Developers
33:11 Conclusion and Final Thoughts
LINKS
- OmbuLabs
All right. Today we're joined by Ernesto Tagwerker founder of OmbuLabs, which includes Fast ruby.io, uh, conference speaker, open source contributor. Did I miss anything in there?
Ernesto:Uh, no. I think also like father of three and overall, uh, old person who likes yelling at the cloud,
Ryan:love it. Yep. That's important. We like to sort of just start off asking everyone sort of what it, what it was their developer origin story, right? Like what brought you to Rails, what gets you excited about Rails, uh, why do you continue to want to contribute in the Rails community?
Ernesto:Yeah that's a good one. I started programming like professionally in Java and that's back in the day when, you know, struts and web works was the way to go and loved it at the time, have a. Pretty boring story in a way. Like I went to engineering school and I just liked working with computers when I was a kid and I thought I could just, play video games and get paid for it. And I thought that's what working in software meant. I know now that I'm wrong, um, it is still like a lot of fun and I'm sometimes surprised like I get paid to work on, on these things. But um, no, back in the day I was, working with Java and one of my old bosses was like, oh, check out this rails thing. And I was like, oh yeah, that's pretty cool. You know, it's like a code generator. And I kind of dismissed it. Right away I was like no. Java's like for professionals rails is, for people who are just like doing hobbies and yeah, I was very dismissive at the beginning and then I got into it and I realized like it was so cool that yeah, it could just turn to be like a one person framework where you could do so much with so little time and. I hated the rails magic at the beginning. And then I was like what's going on? Why are things happening? Why is, how does this know? How does this HTTP request translate to this action method? And I got past that learning curve, I was like, okay, yeah, this is really cool. I like the. The missing XML configuration files, like you don't have to do anything with that. And yeah, I got to basically build my first agent agency around that. At some point, I left that agency and I started freelancing on Rails, and then that turned into a software as a service product built on Rails. And then eventually that turned into an agency that. Was called, it is still called OmbuLabs. And and then eventually that turned into a product I service to upgrade Rails, which is Fast ruby.io. But yeah, that's kind of like long, a really long story to say. That's why I mean, I love Rails because it gets you so much with so little effort. But then at the same time, once you have built like a successful business on top of Rails just start to realize like upgrading a huge monolith can be a huge pain and a huge endeavor, and that's when I saw there was an opportunity there.
Ryan:Yeah, can. Talk a little bit more about, you know, that opportunity. What do you actually do during, you know, the upgrade process for a customer? Like, you know, how do you, how do you, uh, and I'm particularly interested in like how you structure it, like internally, if you can share any of that. Right. Because software. It was like software as a service, or, service as a service really.'cause it require, you know, you're an agency. Like, how, how do you manage that? Because it, every app is different, I imagine.
Ernesto:Yeah. Yeah. And you know, we've been doing this for more than six years, so we learned from experience that different companies have different needs. And, we work from startups have one person in their, their engineering team to, teams of 50 or 60 engineers adding, code to the single, to one single monolith, one thing we've learned is that, big corporations need to know how long it's gonna take. So that's why we came up with this idea of the roadmap, which is an audit to look at the code base, look at the dependencies and estimate like worst base case scenario, how long it's gonna take to upgrade. And initially when we offered the roadmap, it was free. You know, it was like, we'll do the roadmap for free. And we learned a lot. The first time someone took the roadmap and did the upgrade themselves, we were like, okay, there's enough value here to charge for it. I think we started charging very little $600, like $1,200. And then eventually it got to the point where we charge $12,000 for a roadmap. And that's just, to go from rails four, two to five two. We'll look at the code base, we'll do the roadmap, but that's $12,000. And we know that's also like super expensive for startups and smaller teams. So for those teams, what we offer is more of a. Fixed cost, monthly maintenance service where we just jump into it and we try to get as much done as possible, upgrade a little bit at a time, maybe three months into that engagement we are happy to share like ballpark estimates with our clients. it's not the same as a roadmap, but it does give our clients an idea of how long it's gonna take to upgrade their application. Yeah, depending on the type of needs that you have, we're happy to accommodate. And we did have a couple rough quarters two years ago where we were, we were working too much with corporations and everybody just got spooked by, you know, fears of recession and stuff, and we lost a lot of work then. And then we realized like, okay, we have been working way too much with big corporations and we need to be more accessible and more startup friendly. So that's how we came up with the idea of let's do a fixed cost monthly maintenance service that is, just$4,000 a month, which for us is like on the lower end of what we charge. But yeah, that's basically how we approach it. Depending on the client and their needs, we offer different options.
Ryan:Yeah, that's great. I, I think you, like, I know you. You and your team publish a lot of content, right? And you, I don't, you don't really hide any of your processes, right? You put 'em out there for the world to see. And if people choose to, attack it themselves, that's great. And if they don't want to or don't have the capacity you know, you're there to help. But for those that would maybe want to, DIY, it. What do you, what do you think is the biggest misconception out there in terms of rails upgrades? Like, I'll give you one example is before I started the upgrade process for my app, you know, I read in several places to, lock down the gem file, right? Like specific versions, uh, and, for those listening I work with Ernesto's team to do our upgrades. I'm, I'm one of their clients and it's gone incredibly smoothly. But I did notice that we never once locked down the Gemfile. Right. So, you know that, I don't know what's right or wrong.
Ernesto:wrong,
Ryan:Right. I'm curious what you see out there as like those misconceptions that maybe people go down a wrong path on.
Ernesto:yeah, we don't really advise that you lock your Gemfile unless you see that there's a version of a dependency that's gonna cause problems. But no, I think. of misconception, the one that's like top of mind is, oh, why don't you just upgrade from Rails four, two to Rails eight? Just do it. And it's like, yeah. Yeah. I mean you can, and it's very easy to do it with, you know, the typical Dumb Rails block project that has no logic. Yeah. If you have like a silly, simple. Small application and you wanna do a rails fort two of eight, go for it. But. We don't recommend it for big clients and big rails, monoliths and a lot of the projects that we work with are, have been around for more than 10 years. And you know, that's 10 years of commits and complexity. And when we approach the upgrade, we do one minor version at a time. There's one exception, like rails five, two to six, one. I would say that's a safe one to say, okay, we're gonna do it in one, one jump. But. All the other jumps, you want to go like four, two to five oh and five oh to five one because then it's much easier to know what broke. Ideally you have like a test suite. It covers most of the critical path and critical features that the clients will notice if it, they break. once you get the test suite to run, you deploy it to production, um, after, you know, proper qa if something breaks because. Things will break in production. When you're upgrading like this, even if it's like a minor version, can roll it back quickly and check, uh, the failure. And it's much easier to know what changed between like four, two, and five Oh in Active Record, Active Record callbacks than if you look at four two to eight zero, and you look at all the changes in the callbacks, you're like, it's really hard to find like the, you know, the needle in the haystack when you do so many jumps at a time.
Brendan:Yeah. I think to me that also stands out as like a, a good practice in general because if you do a huge upgrade like that, it's just gonna be a huge pr, right. I. And like we've talked on this podcast before as well, just about incremental development and that's like another form of it. Only you're doing upgrades, so you're incrementally going step by step, keeps things small, easier to review. So yeah, that's a really good advice. I.
Ernesto:Yeah, and you can find a setup that. Where you can also do, do like gradual deployments. Um, you can set up Kubernetes run, I dunno, 5% of the traffic with Rails five oh and you know the rest with Rails four, two. And that's a way to de-risk the process as well. We did this with SoundCloud back in the day when we did their upgrade they work very well because then only 5% of your traffic might see an error in production then you can quickly go and, and fix it.
Brendan:Yeah, that's also really good advice too. We, we do something similar, 'cause we've got multiple regions so you can kinda deploy to one of those regions and test it out and yeah, it, it works pretty well. That, that also makes me think I know I've, I've read some of your documentation back to what Ryan was saying. You guys have a lot of good material out there, blog posts and I forget the name of your PDF book. That's kind of like a guide, I think, do you guys still use, I know in there you at least talked about using like the gem file.next, I forget the name of it and everything You can, yeah. Yeah. Next Rails. Yeah. Is that something you guys still do, I assume?
Ernesto:Yeah. Yeah. And it's super helpful for debugging. You know, we use, uh, we maintain this ruby gem called Next Rails. It's very similar to like, boot and appraisals, but it does give us a couple more tools that we find super helpful, like tracking deprecation warnings and keeping them all in one file. But yeah, we're, we're also happy to share all this information because we know our prices are pretty high. Like when you hire experts, you kind of pay the price, right? So know that, people from Argentina, I'm originally from Argentina, right? I live in Philadelphia, know people in Argentina would love to hire us, but we just can't. Work with them because our prices are too expensive for them. So our way to, to get around it is to publish as much as we can in our blog and to share our recipes so other teams can, who don't have the budget to hire us, can just take those recipes and apply them and basically do it themselves. So we're always happy to, know, maintain tools that are. We're gonna help teams all over the world to do an upgrade. We recently took over the rails pump.org project, which is, you know, open source and this guy in Germany was doing it manual and he did a great job and helped me like take it over. And that's like entirely open source. Anybody can go to rails pump.org and submit their gem file dot lock and look at the. Dependencies that are gonna be a problem. but any sort of tooling that we can open source, we'll, we're gonna try to do it 'cause that's kind of like our policy is like open source by default. And there are only a few things that we keep closed source and that's more related to our secret sauce because there's still like a secret sauce that we don't want to share. But I would say it's maybe like 20% of the work we do is close. Close source today.
Brendan:Yeah. That's really cool. That's a, a great approach to hear that you, you know, like to, obviously you're, you're the experts, so having you do it is gonna be the best way. But to be able to give back to the community in a way that, you know, others can benefit without having to pay for it. That's really cool to hear. this is a good place to kind of switch gears a little bit to, well, I guess one question I have before we even do that is we've talked a lot about Rails upgrades and you've, you're, you and your team focusing on those. How much of your work is not upgrades though?
Ernesto:Oh yeah, that's a good question. I think these
Brendan:I.
Ernesto:about like 30% of our work is more on the AI front, so we build, custom machine learning models for other clients. And do that with Humble Labs, which is we're in the process of like pivoting from humble labs, just being a consultancy to an AI consultancy where. We basically analyze, uh, clients', data sets, and we try to figure out whether there is a potential AI solution that could help them solve like a very specific problem. So yeah, 30% is more about AI and 70% is more about tech debt remediation, Ruby and Rails upgrades. I think AI is like we all know, right? Like AI is changing the world and. Everything is gonna look different in like two years, right? So we
Brendan:Totally.
Ernesto:of that and we want to make sure like we're building our AI tooling to power some of the upgrade work that we do so we can focus more on strategy. And the AI tooling that we build focuses more on like shipping the implementation of, of that strategy.
Brendan:Well, that's cool. And a, a quick question there. You don't need to get into specifics about what you're building, but my question is basically are you primarily building internal tools? Are you primarily building for clients or both? I guess with ai, I.
Ernesto:Yeah, both. We. Love to, you know, work with our clients. And also we learn a lot from working with our clients. What we want to do is like take those learnings and implement our own internal tooling to make upgrades faster and maybe more cost effective for our clients so we can deliver more value without them having to pay more. One of the tools that we're gonna be shipping soon is, I don't think it's gonna be public, but at least we're gonna start using it internally, like an AI tool that will help us like address deprecation warnings. So we have all the tooling to basically track any deprecation warning that is present in our client's code base. And what we can do is basically feed it into one of our machine learning models to say, Hey, these are all the deprecation that are present now. Submit a pull request that addresses it in, all the files. Now, of course that's gonna be like, supervised and monitored by our senior engineers, and it's not gonna be like live and submitting poor requests without, adult supervision. But, um, that's one of the first things we're, we're working on. another one that we're working on is, like I said, like the roadmap used to be free. And there is a way to offer a free roadmap for clients. We're working on something like that's gonna be like AI powered and LLM powered based on our knowledge base on official guides, on our client's code base. And that's not. Live yet, but probably in the next like two months we're gonna publish that and that's gonna be available to anybody to use for free. might, add a way for you to like feed your email so we can contact you later if you opt in. That's one of the things that I love about running like a consultancy is that we can use engineering as marketing and I've been always. Like excited about that idea of like, oh, I can build this tool that's useful, that people can just use, and then I can, you know, reach out if they want me to reach out and be like, Hey, did you find the tool useful? want us to tell you like how long it's gonna take to upgrade? Because. Our version of the roadmap is gonna be, or the free version is not gonna have estimates. So that's kind of how we say, okay, if you wanna pay for the roadmap, we're gonna have like two senior engineers give you set of like worst best case scenarios to know how much it's gonna take to upgrade.
Ryan:Yeah, I think that's a great sort of, so it's like a hybrid model, right? You're giving just enough away to peak, uh, people's interests and then, you know, when you want to take it to the next level, you know, that's when you guys can step in and provide value, right?'cause the estimate's, the hard part.
Ernesto:Yeah, and maybe, you know, like the free roadmap will give you an idea of like, oh, these are like kind of like low hanging fruit tasks that you can do by yourself that are very non-disruptive and. get it done. And then once things get thorny or you try to do a deployment and it didn't work, or you run into blockers that, you can't really dedicate the time because if your product roadmap gets in the way, you can reach out and we can help. Yeah, we offer a lot of information, but there's always like the other side is if you don't have the time to do it, us to do it for you.
Ryan:Yep. Yeah, that makes total sense. All right, so you, you mentioned, uh, a couple different open source gems and, um, you know, I know you. You speak at conferences and you typically speak about dev metrics and code base analysis, which ties back to the, the work you do with fast ruby in a lot of cases. But you know, how do you see that, you know, being part of the development process, right? You, you talked about tech debt, um, in pre-call we talked about, uh, developer, uh, experience and all of those things to me are sort of interconnected. I'd love to hear your. Sort of high level take on it, and then maybe we can drill down a little bit.
Ernesto:Yeah, and I'll just say, I really enjoyed your episode on Tech Dead. Anybody who's listening. Please go and, and listen to that.'cause think there is a lot of, aspects to tech debt. the problem usually is related to culture and, culture is defined not just by the engineers, but also by the non-technical folks in a company. When I think about tech debt, I think about the developer experience. You know, there's a paper that was published by Google about, technical debt and they basically run surveys, with a ton of their engineering teams, and they were asking about their perception of technical debt, right? So when I think about technical debt, I think perception is very important to, to take into account. Because what, what you, Brendan might think it's tech debt. I might think not tech debt and I'm okay with it. Or there's always like a threshold where we're okay with,
Brendan:yep.
Ernesto:I know Brendan in that episode, he was talking about, hey, you came up with this Active Record query. It worked very well for like five years and now that table up. So now that query is just not working anymore. And. You know that software was good enough for five years and now it became a problem because it kinda went above the threshold that it made it, it's now technical debt and performance is like a really easy one.'cause it's is the app slow? Are people noticing? Yeah.
Brendan:Yep.
Ernesto:technical debt. But another one that we run into that is kind of like a silent killer of productivity is onboarding. You know, like every now and then, like every month we get onboarded to our clients' code bases, and I. Sometimes it takes like a week or two to try to set up your environment with someone's code base. And for consultants, it's just like super pricey to say, oh, you don't have the onboarding process set up. Developer experience sucks for anybody who's joining because they have to talk to two of your principal engineers to get onboarded. To me is like unacceptable. But there are a ton of companies out there that don't test their onboarding enough to say, oh, okay, yeah, this is, this sucks and we need to improve it so we don't waste so much time. Every time someone needs to get added to the team. I.
Ryan:A huge unlock for me personally. And I think you, you both have had the opportunity and may, maybe not you directly Ernesto, but someone on your team had the opportunity to onboard to. To my code base. And it used to be, we used to be in that one to two week window, right? And it would be, I'd have to pair right every time. I'd have to get on the phone and we'd have to just work through all there. I, I think some of that was like older versions of Ruby that, you know, aren't supported anymore on the new, you know, Mac, uh, apple chips and those things are sort of working themselves out. But I had, at some point I just set up Docker. You know, I bit the bullet and got it working and now it went from two weeks to usually an hour, maybe two. And that just makes a huge, huge difference when you're working with someone new 'cause, uh. It used to be we'd, we'd hire, you know, if we were hiring a freelancer, they'd get frustrated and just wander off and never come back like, eh, this isn't worth it. But now I can get them into the code base. They can get working right away. Uh, that's just a, a huge win for both sides. So I, I love that one. To think about,
Ernesto:yeah. And you know, I, I get it. Like we all have priorities and, of us know that security is number one, right? First you wanna upgrade to the, you know, most stable, most secure version of Ruby and Rails. You wanna make sure, like in production, you don't have any CVEs present that could be exploited. Yeah, that's technical. That, that's also like very easy to detect and be like, okay. This is important because we could, lose a ton of data and we could, they could bring our system down if someone exploited this the right way. So I get it, like onboarding can be tech debt, but security issues can also be tech debt. When I think about the tooling that I've built over the years I maintain Ruby critic, which is really cool because it uses this idea of like, you know, churn versus complexity for every file in your system. Then I wrote Skunk. That kind of extends Ruby critic and it adds code coverage to that formula. And it says, okay, which files are like changing all the time, which files are super complex and lack proper tests, to cover it. And then that usually gives an idea of like, okay, some of these files could be a problem in the upgrade project. But then at the same time, you know, skunk and Ruby critic use. Flog use re and for instance, like re is like a sub, it's basically a crowdsource database of code smells that someone. In the internet, things is a bad thing for any code base. I usually recommend anybody is to take a moment with their team and look at those code smells and look at the defaults and try to figure out whether those defaults make sense in your code base. And it's very easy, you know, you do gem install wre app, and then you look at all the code smells and you see a ton of information. Then you're like. I'm okay with those coat smells. You know, some of these would actually make the coat harder to understand and to maintain. going back to the idea of like culture, I think the engineering culture needs to have a conversation and an ongoing conversation about what is tech debt or what's the threshold you're okay with and what the threshold that you're not okay with, and you need to. Make better. Is that how you guys approach it in, in your companies?
Brendan:Yeah. Oh yeah. I mean, it's always a trade off for sure. It's like there's the things you've already talked about. On top of that, there's the selling it to the executives to say, Hey, it's worth spending the money on, right? You've already talked about how much it costs to hire you guys to do just the rails upgrade, then when you're talking about all the other things, it's like there's the time cost, there's the development cost. There's so many different factors you gotta figure out, and it's like onboarding is a really good example of one that's kind of a hard sell until you tell. Your boss that hey, it's gonna cost us OmbuLabs $4,000 or whatever it is to spend two weeks to do it, versus, you know, we do it once and then they can get on board really quickly. But yeah, that, that's the thing I always find, and I know in, in your Sin City Ruby talk, uh, last year did a really good job of talking about the numbers. I think the dollars are always the thing that, the CEO is gonna say, oh yeah we need to address that versus. A more technical person just coming to someone that's not technical and saying, Hey, we really need to do a rails upgrade, or we really need to fix our developer onboarding. So those are another, just something I wanted to point out that I think while you were talking there, I was thinking about a lot I.
Ernesto:Yeah. And that's, when communication comes, comes into play, right? I'm still working on it. I'm not great at telling you like, okay, if you don't upgrade Rails four, two to five, two, or. A seven you could get exploited or, there could be other consequences. Like anybody who is on Heroku today and is running like a really old version of Rails knows it now, but if they, you had told them a year ago, Hey, you need to really need to upgrade because in a year you're just not gonna be able to deploy because Heroku 20 is gonna. you. It's going, you know, end of life. You can't do it anymore. So yeah, today there are like a lot of applications out there that can't deploy because they forgot to upgrade. And maybe they were like, oh, we'll upgrade later. We'll upgrade later. And then. know, later never came and now they're like, oh shoot. Well I guess we need to migrate El Heroku and that's gonna cost you a ton of money if you want to do it quickly.'cause it's now a Rails rescue project, which usually is, is very expensive.
Brendan:That's another really good example of why, to keep up with it. Yeah, I hadn't even thought about that one.
Ryan:You asked, you know, is that. How we do it you know, how we approach things and, and I wanted to touch on, there's always competing priorities, right? Security, performance, just developer experience, culture, you know, all these words are so.
Ernesto:So
Ryan:Tangled, right? Like, it's hard to separate. So I think the way I like to ideally think about it in day to day, this, you know, I, I bounce back and forth between the priorities, but at a high level, I, I wanna think about it as a continual process. I always gotta be moving upgrades ahead. I always gotta be worried about security. I always need to think about what I want my onboarding experience to be. And you know, as the priority shifts, that's okay. I can bounce around a little bit as long as I'm dedicating some time in my schedule. Ideally, at least once a month on these things to move them forward, right? And sometimes, with an upgrade. There's a deadline, so, you know, maybe I'm spending a lot more than once a month on it. You know, security is a good one where you can just sort of run bundle audit and bundle outdated and sort of get a sense of, you know, what the vulnerabilities are in your gems and in your app, you know, and if you get in the habit of doing that, then it becomes easier to maintain. It's not quite as scary. It becomes part of your process. Free call. You mentioned the, this idea of. How do you talk about developer experience to a team of one? Right. Rails is, is known for being a, you know, a single developer platform. Um. I think, you know, my opinion is developer experience is something you need to think about from day one. What do you want it to look like? And if you're not doing it yet, then I'd recommend or strongly urge you to, to spend a little bit of time thinking about it, like what are the pain points for you? And start putting together a plan to, you know, address those. But would you have any advice for the, the single developer in, in terms of like. How to, you know, we've sort of talked about developer experience a little bit, but like, what about the culture side of things? Like how do you define that?
Ernesto:Yeah, and I think that's very important to talk about because it's so different between, a team of 40 engineers, like adding changes and all, and just one engineer who's working on the code base, who doesn't even have someone to review their changes, you know, so you're just like submitting a pull request or sometimes not even just committing to Maine, deploying to production, and. you are doing what, what you're saying, Ryan, basically keeping everything in mind at all times. So let's say you don't you're not an expert in security. That's the nice thing about Rails. You know, you're, one person framework. You don't have to be an expert in security, but you know, you could. Use tools like Brakeman Bundle audit and basically set up your own to stop you from shipping something to production that could cost you money because someone exploits it, right? So when I think about developer experience for one person, I think about. I think about that, it's like how much time are you investing in the product roadmap versus in maintenance? Like much needed maintenance, right? Not just like, oh, I'm just gonna make this code more maintainable. when you're a one man band or one woman band, you have to prioritize and. You have to think about, okay, I need to add features because I have, I have clients, I have to, you know, ship features for, then at the end of the day, I also have to upgrade. I have to make sure things are performant. I'm gonna use all these. to feed me information, and then that's gonna help me prioritize which maintenance tasks I have to prioritize as well. and I have, I have friends who, you know, are also like in this boat and to build features and I love to work on product, but they also need to do maintenance. And sometimes they forget about it, and then it does become a problem. So, yeah, I think it's a, a bunch of things. One of them is just it all in your head or in your tool, you know, whether it's Trello or Jira. And the other one is I. Yeah, may maybe getting help from someone who is an expert. Sometimes one, one person teams just want to hire someone that is right? And I'm much more inclined to say no. I think you should hire someone who's very senior. Just get fractional access to that person and maybe have them coach you into not just doing product roadmap work, but also like maintenance work to make sure, like the minute you have to add someone to the workflow, someone to cover for you while you're on PTO. That they can get set up quickly. Like you said, you took it from two weeks of onboarding to one day. great because then that means like when that happens, someone can quickly pick it up and you, you can go on vacation and enjoy your time off while someone kind of covers for you.
Ryan:Yeah. I have like a bazillion more questions floating around my, my head, but I, I, I think we're just about outta time. Is there anything, uh, you wanna add Ernesto, or anything you wanna wrap up with, uh, Brendan before we call it a day? I.
Brendan:No, I, I feel the same way. I've got lots more questions but yeah. Ernesto, is there anything else you wanted to add?
Ernesto:No, I'd be happy to, you know, talk about it some more in another episode. I think, I might have not answered all, all the questions and I'm sorry about that. What I wanna leave the audience with is that technical debt and developer experience is not gonna be solved with ai. There's not gonna be a magical tool that's gonna come and solve it for you. Having. Deep conversations with your engineering team or if you're just one person, like deep conversations with you know yourself to try to define what is technical debt and what's acceptable on that front. then tooling. There's just so much tooling. Yeah, there's Cursor, Windsurf, you know, they're gonna be AI agents that will do a lot of these things for you. You just have to make sure you're setting the right strategy. To make the code base slightly better every month, you know, and don't just do product work because it's fun. It is fun. I agree to build something new, but you can't just ignore maintenance because it's gonna come haunt you you know, in the future.
Brendan:Yep. Yeah, that's great advice. Well, cool. It's been great in having you on and uh, yeah, we'll, we'll definitely have to schedule another time 'cause there's, like we said, a lot to talk about. But before we sign off, is there anywhere on the internet you wanna point people to?
Ernesto:Yeah, sure. Well thanks a lot for having me and if anybody's interested in, you know, technical debt and Ruby, just go to Fast ruby.io. if you're interested in some of the AI work we do, just go to ble apps.com and I. Yeah, I'm always, you know, happy to answer questions on, you know, if you just send us a, a message on, on our contact form. I'm happy to connect and talk some more about this.
Ryan:Thanks so much for joining us.
Brendan:Thanks Ernesto.
Ernesto:Thank you.
Brendan:See.
Ryan:I.
People on this episode
Podcasts we love
Check out these other fine podcasts recommended by us, not an algorithm.
Code with Jason
Jason Swett
Standing in the Fire
Very Good Software
Remote Ruby
Chris Oliver, Andrew Mason, David Hill
IndieRails
Jess Brown & Jeremy Smith
Startups For the Rest of Us
Rob Walling
Vibing with AI Code
Brendan Buckingham