Most developers think building the product is the hardest part. I did once upon a time.
I spent nights and weekends when my kids were very young cranking away at building my own products – crafting complex applications to manage photos, or aggregate news.
What I learned, and what our guest reinforces in this episode of “The Hallway Track” is that the code or the product itself isn’t the hard part at all.
It is finding customers, earning trust, and knowing what not to build.
Adam McCrea is the founder and developer of Judoscale, a third-party SaaS autoscaler that grew from an internal side project into a sustainable business supporting multiple hosting platforms.
Adam makes the case that AI has made writing code easier without making it any easier to understand, maintain, or sell software: the real advantage still comes from years spent deep in one problem. We also get into how Judoscale grew through the Heroku Marketplace years before Adam knew much about marketing, and why bootstrappers need to protect their intellectual property from day one.
Highlights
(00:00) The platform-risk lesson that opens the show
(01:42) What Judoscale does
(02:14) Running Judoscale as a team of two and a half
(03:48) Adam’s 20-year path from Rails developer to founder
(04:45) Building Judoscale’s first version as a side project
(08:09) The IP ownership scare with TinySeed and his old employer
(11:43) Why launching in the Heroku Marketplace beat any marketing plan
(15:50) Spreading platform risk beyond Heroku
(19:50) Why “AI means everyone builds their own SaaS” is overblown
(25:45) How AI actually speeds up Judoscale’s dev work
(31:00) Talking to anti-AI kids about the future of work
(35:12) Chris’ 3 takeaways, where to find Adam and Judoscale
Transcript
[00:00:00] Adam McCrea: It’s still very risky, to have, all of your eggs in one basket that someone else is in control of.
You know, that platform risk is very real and we still have platform risk. It’s just spread out amongst several platforms but the reality is like we started on Heroku, so Heroku is still the majority of our business.
But if Heroku were to go away tomorrow, at least we wouldn’t be out of business.
[00:00:28] Chris Vannoy: Hi, welcome to The Hallway Track. I’m your host, Chris Vannoy. My guest today is Adam McCrea, founder of Judoscale, a third-party autoscaler that started as an internal tool at his day job and now runs as a two-and-a-half person business across six or seven hosting platforms. Adam built the thing a lot of engineers say they’re going to build, a side project that grew until it could pay him.
But the interesting part of the story isn’t the code, it’s everything around the code. Getting the intellectual property in writing three years after the fact, landing in the Heroku marketplace before he knew what distribution plan meant, and spending 10 years on one problem until nobody can catch up on a weekend.
We get into build versus buy in the AI era and why Adam thinks the everybody’s just going to build their own SaaS now argument is overblown from where he’s sitting. If you’ve got a side project you keep almost launching or you’re the one making build versus buy calls right now, this one’s for you.
[00:01:28] Chris Vannoy: Howdy, folks. my guest today is Adam McCrea, who is the founder of Judoscale. Adam, welcome. Uh, that was a very short intro. You wanna walk through a little bit more about yourself and Judoscale and, like, who you are and what it does?
[00:01:42] Adam McCrea: sure. Thanks for having me. yeah, so I am the founder of Judoscale, which is an autoscaler, a third-party autoscaler, basically a drop-in replacement for the autoscalers that you might find in cloud platforms like Heroku, Render, Amazon ECS for when you need, a little more control, a little more visibility, you just need a little more power in your autoscaler.
So that’s all we do. We started out as a, a Heroku add-on, and now we support a few different platforms. And we’re a small team, so, I kinda do it all. I still do a lot of dev. I do marketing. I do a lot of different things that I’m sure we’ll get into.
[00:02:14] Chris Vannoy: you mentioned it’s a small team. Like how small are we talking
[00:02:17] Adam McCrea: I always say it’s a team of two and a half, which is to say there are two of us full-time and one of us part-time
[00:02:22] Chris Vannoy: Okay, cool. All engineering other than yourself, I suppose, which as you mentioned, you have to wear a lot of hats
[00:02:28] Adam McCrea: Yeah, yeah. All three of us are devs. my, full-time teammate, Carlos, he lives in Brazil. he’s on the Rails Core team, so he is a very good developer, the best developer I’ve ever worked with, which is why I, sought him out to join the team. yeah, he’s awesome. he pretty much does purely dev.
And then, John, who lives, locally in Ohio with me, is a friend of mine that I’ve had forever. He is a developer, and he has, you know, a day job where he’s, you know, a, a senior level developer. And then with me, he does a lot of writing on the blog. And we also get together, at a co-working space once a week where, he basically acts as my co-founder, therapist, and everything in between just to bounce ideas off of and, push me out of my comfort zone.
[00:03:14] Chris Vannoy: You essentially built this, as I understand it from previous conversations, like it started off as like a side project essentially, that then grew until at this point you’re supporting, like you said, two and a half people, and all that fun sort of stuff.
Which is kind of like for a lot of entrepreneurial-minded engineers, kind of like living the dream.
Not all of us, but a lot of us have had this idea in the back of their head of like, “I’ll go build some product and then it’ll eventually it’ll be able to support me and then I’m not working for the man anymore.
I’m just doing my own thing.” and so you’ve actually walked that journey. so I’m curious if you could walk us through a little bit of like how you got to this point in Judoscale. Like first off, like how long are we talking for this total journey to get to today?
[00:03:53] Adam McCrea: it was definitely my dream, and I do feel like I’m living the dream right now, and then, and I feel super, super fortunate, to be where I’m at right now. I started my career as a software developer in 2001. I, I got into Rails in 2007, and that’s when I got my first full-time job writing Rails applications.
And Pretty quickly I had that itch where it was like, I’m reading everything that 37signals is writing.
I’m seeing you know, just different software businesses pop up left and right and a lot of them independent and I’m like, “Man, that’s what I wanna do. I wanna be able to run my own product and get to decide the things that I build and not have to answer to anybody but my customers.”
Like you said, a very common dream among software developers. And, along the way I had a few ideas that I built out but never really launched. I just kind of put a half-hearted attempt into it. but in 2016, I was working for, um, a small product company that was, basically like a, sort of a lead generation service for real estate agents and we had a small engineering team. It was just three of us,
We had just moved our hosting from, Engine Yard to Heroku, an effort which I led because I’ve loved Heroku since like 2010. and we were loving Heroku. We were actually saving a lot of money compared to Engine Yard, but we were still burning a lot of money running a bunch of server resources we didn’t need, during our slower times.
So we needed an autoscaler. and there were a couple autoscalers for Heroku at the time. The Heroku native autoscaler did not exist yet, but there were a couple third-party ones. We tried them and they just didn’t work very well. I kind of saw an opportunity. I’m like, “I, I feel like I could build something that’ll do exactly what we need.”
So that’s what I did. I, I built an autoscaler sort of on my own time. you know, I, you know, ran it by my boss first. I was like, “Hey, I think I have an idea where I could build something for us that would really help us out. Do you care if I build it on my own time and we use it?” Like I said, we’re a small team.
We’re super informal about things. He’s like, “Absolutely, I don’t care what you do.”
The first version of it had no UI. our Judoscale app was hard-coded into it. it was as MVP as you can get. No way anybody could sign up for it.
There was no billing, there were no accounts. It was, it was just very much answering the question: Can I auto-scale our specific application in a way that’s better than the other things that we had tried?
And was able to prove out very quickly that, yes, we can. that was, like, in the fall of 2016.
So once I proved out the idea, 2017 was when I brought it into the Heroku marketplace so that anybody could sign up for it. and from there, it was a side project for four more years until 2021 was when I was finally able to go full-time on it. And then I think I brought Carlos on in 2023
[00:06:51] Chris Vannoy: I’ve tried to do that on my own. Like I mentioned, it’s a common dream. It’s one I had, right? So in that same sort of time period, I think I went to a micro conf in Las Vegas in like 2014, something like that, right? Like there was this whole like ecosystem of like small bootstrapped like companies. And actually a lot of the work I did like for day job was oftentimes taking stuff I learned from those communities and then implanting it into clients like big companies, ’cause I was working at an advertising firm at the time. So I’d get on some call with Microsoft and then I’d take some learning I picked up from that whole crew and just start applying it over there instead, like in marketing and stuff like that.
Anyway, that was fun. But like four years, every timeI did the typical developer thing of like, you actually did the opposite of what I’m gonna say. where like I have some idea, I’ll go build the thing, right? And then I’ll go figure out who actually needs it. What I think is interesting with your story is like you started off with like you already had the need, right?
You were working at a company that like we needed to do this, and then just built it. one of the things I’m curious about is like you were doing that for a company you were working for. How were you able to extract that out into its own thing that was actually yours that you could then spin off as its own thing?
Would that involve any like, negotiation with the employer or anything like that or was it not quite so much?
[00:08:10] Adam McCrea: Well, like I said, initially it was a very informal agreement. There was absolutely nothing written down about it whatsoever ’cause I had no idea what I was doing.
in 2021, as I was getting ready to go full-time, I actually applied to, a program called TinySeed, which, think they bill themselves as like an accelerator for bootstrappers.
So imagine something like Y Combinator, but built to be remote for bootstrappers, a very different sort of mindset. a small amount of investment, a small amount of equity given up, but what you get is a community and mentorship. And it was really appealing to me at the time, one, because the small amount of investment would allow me to go full-time on it with minimal risk.
And as a solo founder, I was definitely feeling sort of alone and ha- and feeling, wanting a community, around me. So anyway, that was really appealing to me. But as, part of the process of joining TinySeed, you know, I applied sort of as a long shot. and when they were interested in, Judoscale joining TinySeed, that’s when they started due diligence.
And they asked the exact question that you just asked, which is, “So you built this thing while you were an employee somewhere. Do you have any sort of documentation that you own this thing?” that was a terrifying moment for me because I was like, “Oh no, I don’t have that at all, and I haven’t talked to these people in like three years.”
Fortunately, they were very cool. I went back to my previous employer and I was like, “Hey, I just need some proof that I own the, intellectual property for this thing.” And I wrote up a little thing. They signed it. It was all fine.
Lesson learned for future bootstrappers out there is, get that in writing from the beginning if you’re building something on the side while you’re working for an employer.
Yeah, it comes in handy later on. Like, not only, like, investment, but if later on you want to flip it and, like, sell it to someone else, that same question’s gonna come up again, right? One of the things I’m interested in generally is, like, this, AI-powered world we’re in now. It’s like that dream of, like, I’m gonna go build something on the side is even easier than it used to be.
Like, when I was doing it, I don’t know about you, but, like, it was a lot of, like I had two small kids in the house at the time, What I was building was, photo gallery management software, like, sort of stuff, right? ‘Cause I was coming out of newspapers and I’d built the same system, like, multiple times. And so I was like, “Oh, I’ll go build that, and then I can go figure out who to sell it to and stuff.” But it was a lot of, like, I had small kids, so I’d wake up early in the morning and go to, like, a coffee shop before I went into work, and then I would work on it, right? And that sort of stuff. And it was months and months and months and months and months of building the thing before I actually went and talked to anybody to actually buy it or use it.
But that’s a separate topic that we’ll get into later, I suspect. but these days it’s so much easier to, like, have an idea and, like, crank something out, right? much more quickly. what I’m interested in lots of times is, like, software engineers that step into this world, like, they don’t really have a lot of good resources for the business half of this, right?
What things when you’re starting something at an employer that you pull out, what steps should you take, right? And you hit on one of them, like making sure to check ahead of time. Like, “Hey, by the way, can I get some paper that says that this is actually mine and not yours?” Because oftentimes when you get employed in the first place, you signed a piece of paper that says the opposite, in most cases.
Any other lessons from early on that, like, somebody starting this journey, like, today would probably want to keep in mind that you had to learn the hard way?
That was the one where I felt like, was scary for a moment, where I felt like I, I didn’t know what I was doing. And, there are definitely things that I feel like I did right along the way that really helped me. And I’d say the biggest one And this is specifically as a solo developer with no marketing experience, no real business experience, just sort of like building a product on a whim because it’s something that I want to exist in the world, is that I put it in a marketplace.
So specifically the Heroku Marketplace ended up being a fantastic way to launch a product. And I tell people about this some people are shocked that, Heroku takes a substantial cut of revenue. You know, it’s, it’s like Apple’s App Store marketplace. They take a significant cut of revenue, and they’re like, “Well, how, how could you agree to that?”
And like, if it weren’t for the Heroku Marketplace, I would not be doing what I’m doing today. There’s just no way.
Because I launched it in the Heroku Marketplace, and I didn’t do any marketing at first. I didn’t know how to do any marketing at first. And it’s not like it took off. It definitely didn’t take off.
It was four years before I was able to go full-time on it. But slowly, over time, people on Heroku would find it in the marketplace. They would try it, they would use it, and it just grew sort of on its own without me doing anything other than continuing to improve the product, supporting it, making sure people liked it.
I don’t recommend launching a product and not marketing it at all. That’s not what I’m saying.
Part of it was timing. You know, it was a great time launching it into a marketplace was super helpful.
I think, like, at the end of the day, you just need to have a distribution plan when you launch something. for me, I didn’t really think through the distribution plan other than the fact that, like, the Heroku Marketplace is how I find things to install on my Heroku app, so I’m gonna put something in there too.
And that worked out for me. But if not a marketplace like that, if you’re gonna launch something, you have to have some sort of plan for how you’re going to distribute it, how you’re going to get people to be aware of the thing and want to try it out. I’m no expert on that. Uh, it just worked for me this one way.
[00:13:56] Chris Vannoy: That’s kind of the key point and like, uh, some of the work I do these days is also teaching like young, folks coming right out of CS degrees, right? Where in the old days you would just go get a job, and a lot of them are finding that a little bit difficult. And so some of what I’m coaching them through a little bit is like, maybe just go build a business of your own, right?
Like go build an app or something like that. And one of the things I mention to them is like the hard part of this is not the code, right? even an autoscaler, which is a fairly technical product, I imagine that was not the hard part of building this business, right? The hard part was finding customers, keeping customers happy, making sure you’re building something that people actually want, right?
And so oftentimes working with like engineering or like software developer folks, they wanna put the bar pretty high before they release something. Like I did this too. Like I said, I worked for months on something and like time away from my family and waking up super early and going to go work on something. When it’s like oftentimes you can early on, the simplest, smallest, stupidest thing you can do to figure out is this something someone will pay me for often helps a lot. And the other bit you mentioned there is that having an idea of who the buyer is and where they hang out, right, so that you can find them also helps a lot.
And you stumbled into one of those where those marketplaces are a great place to do that sort of thing, right? The other thing that I think you probably benefited from is like it was a vendor-specific marketplace. So early on in those first four years, I imagine everything you built was for Heroku, right?
You didn’t have to worry about any other hosting platform or anything like that for quite some time. I know you support more of them now than you did then. how much is that related to, as you mentioned, sort of the waning influence of Heroku versus seeing other market opportunities other places?
[00:15:50] Adam McCrea: Yeah, I wouldn’t call it market opportunities. It’s more about risk mitigation. whether Heroku was booming right now, whether everybody loved Heroku or, whether everybody loved to hate on Heroku as, as is often the case, it’s still very risky, to have, all of your eggs in one basket that someone else is in control of.
You know, that platform risk is very real and we still have platform risk. It’s just spread out amongst several platforms and that was, that was the goal, going beyond Heroku. I think we support like six or seven different hosting platforms at this point. but the reality is like we started on Heroku, so Heroku is still the majority of our business.
But if Heroku were to go away tomorrow, at least we wouldn’t be out of business. We, we aren’t completely dependent on Heroku like we once were
[00:16:42] Chris Vannoy: Yeah. And it is one of those concerns as the business grows. Essentially, when you’re first starting off, you don’t have that risk yet, right? It’s like the risk you have is that the whole thing doesn’t work and you’re creating something no one actually wants, right? So until you get to a point where the business has grown enough that it’s like, “Oh, this is actually a real thing,” and like I’m full-time on it or I’ve got multiple people and that sort of thing, is when you have that risk of like, “Oh, wait. I probably don’t wanna be locked into one vendor.” but early on, I think that vendor lock-in is probably fine, right? Because it helps constrain what you’re working on and helps get it out the door faster and start the learning process a lot quicker.
[00:17:23] Adam McCrea: I think that framing of you don’t have that risk yet, like applies to so many other things. It’s like, when you’re getting ready to launch like a brand new idea or you’re just building a brand new idea, it’s so easy to get caught up in like, well, you know, I got to think about scalability and I got to think about like, what’s the legal structure around our business entity and like, there’s just so many different things you can get caught up in.
And the reality is if you don’t have a single paying customer yet, none of that stuff matters
[00:17:54] Chris Vannoy: you, if you don’t have someone’s credit card to charge, you don’t need the ability to charge a credit card. Like, you don’t need to wire up billing until somebody’s willing to pay you. software engineers are particularly bad at sort of raising that bar. The other one that are particularly bad are like the, entrepreneurs, like very early, like they don’t have a thing yet.
‘Cause I end up doing a lot of that, like it’s free consulting basically. Like, you know, they buy me a beer, like something like that, and I’ll help you out. And the advice I constantly give them is like, “Do less.” What can you do tomorrow to try and figure out if somebody will pay you for this,” right?
And the interesting thing is like that was the same advice I was giving like 10 years ago, right? But these days, the amount you can do and like what could you do tomorrow is significantly higher than it used to be, thanks to AI tools and things like that. which I think really both helps and hurts.
I think it helps bootstrappers quite a bit. it hurts those who want to go seek funding, right? Because the bar for funding, venture funding specifically, has gotten so much higher because they’re expecting you to have a working demo, to probably already have some revenue coming in, to already have essentially a business built up before they can invest in it. But if you’re bootstrapping, you can get something fairly impressive out the door pretty quickly, right? That then you can go start the process of discovery, right? To actually figure out who will pay for it and that sort of thing. that’s also interesting for your business, which is a little bit further along. Do you see any, proliferation of more competitors or something? Like how has the space you’re in changed in the course of AI? Let me get specific for a minute. You’re also selling a fairly technical product, right? Your buyer is probably software engineers. do you get any sense of more people, thinking like, “Oh, I could probably build an autoscaler myself,” right?
Using AI tools or something like that. Have you seen that sort of effect in your business? Or what effects are you seeing as far as that goes?
[00:19:50] Adam McCrea: It’s really hard to tell. the general like industry sentiment among software developers is that, so many of these SaaS businesses can just be built in-house now super easily. I haven’t gotten that direct feedback from anybody that’s like– Nobody has left Judoscale and told me, “Yeah, we, we built our own autoscaler in-house.”
Nor has, nor have I heard from anybody that’s like, “Oh, we were gonna look into Judoscale, but we built our own autoscaler in-house.” I’d have no way of getting that feedback, right? Like I’d have no way of knowing if somebody might have looked into Judoscale, but they decided to build their own autoscaler in-house.
So it is so hard to tell. all I can say is that as a very small software company who pays for, probably over 30 different SaaS products that we use to run our business, so far we’ve replaced one of them, with no plans to replace any others. the reality is that’s not how I wanna spend my time.
That’s not what I wanna focus on. I wanna focus on our core competency, which is autoscaling. I don’t wanna do any of the other things we pay these SaaS products for, because at the end of the day, it’s worth it. It’s worth it to pay these companies to run these products that I don’t need to think about.
My take is that the idea that everybody’s just gonna build their own tools and SaaS is dead because everybody’s just gonna build their own tools, I think that’s way overblown. I don’t really believe it, and I’m not, super scared of it. You also asked about competitors.
I haven’t really seen that change, since the like big AI, you know, boom of the past couple years.
Uh, we’ve definitely gotten some competitors, prior to that, some other like entrances into the Heroku marketplace in particular that basically just cloned what Judoscale does. And, that’s just the way it is. There’s not much that we can do about that. I still think, you know, we’re gonna do a better job, we’re gonna provide better service, and people are gonna be willing to pay for that even if we have some competitors that try to do the same thing and undercut us.
I haven’t seen any real like new competitors pop up in the past year or two
[00:21:55] Chris Vannoy: That’s interesting. I would have thought like, the thought if this was happening, the everybody’s building clones of like SaaS products and that sort of thing, you would see this wave of competitors, right? Again, autoscaling is difficult, but it is fairly well defined, right? so you would think if that were happening, you would start seeing it, right? also, like I mentioned, you’re a technical product, so the people who are users are also probably, A, familiar with the problem, and B, familiar with generally speaking how to fix it. not that it’s simple, ’cause it’s not. that’s why you exist is it’s a pain in the ass and no one wants to do it. which maybe it’s that. but the other piece I wanna pull at is that sort of like AI, everybody building their own clones in-house sort of thing. And this has always been a problem, I imagine, for folks that are selling technical products. I’ve been on the other side of buying products lots of times, and it’s like I can look at it and go like, “Oh, we could build that ourselves,” or like, “It doesn’t quite do what we want, and so we’ll go build it ourself,” or like that sort of thing.
Occasionally it has been, “This thing is too expensive.” your Judoscale’s price is very reasonable, like beyond reasonable. You should probably charge more, to quote Patrick McKenzie, for what y’all do and what it enables. But, like that sort of build versus buy decision, like has always been around, right?
It’s always happened. I think the only thing that might have changed is the amount it costs to build might have gotten lower. But I think those who have actually built software before realize it hasn’t lowered that much. AI makes code a lot easier to write. It doesn’t really make it a lot easier to, understand, maintain, support, like market, any of that sort of stuff.
It hasn’t really helped all that much.
[00:23:35] Adam McCrea: it doesn’t make it easier to think through all the different edge cases
I think about like an error tracker, right? I don’t want to take the time to think through all the different things that an error tracker needs to do. Like Honeybadger, Sentry, like these products, this is what they’ve been focused on for years.
I know what it’s like to be focused on a specific problem for years. You figure out a lot of different things and like,
There’s no way, AI is gonna like come up with all those learnings just immediately.
That’s just not gonna happen
[00:24:04] Chris Vannoy: Yeah, which has always been the thing in this build versus buy thing is like whatever company I’m building, right? Like auto-scaling our stuff is not our core competency, right? It’s not what we’re gonna think about all day, every day. It’s not what we’re concentrated on, right? We’re concentrated on making widgets or whatever, right? Whereas for your company, that’s what you’re concentrated on is like in and out of how do you auto-scale across all these different platforms to like save customers money, or make them more scalable generally so they can handle a big traffic spike or something like that. You’ve been thinking about it day in, day out for, I’m doing the math, like 10 years almost, you have a 10-year head start from someone sitting down and thinking about it all the time before they could figure all this stuff out, right?
That’s the part that is actually difficult to just throw AI at and have it replace. that’s the other like metaphor I’ve been toying with of like what is good for humans versus good for AI is like humans don’t have context window problems, right? you’re holding inside of your head 10 years of learning about this problem, right? And it’s like stuff is gonna drop out of our memory sometimes. Like, you know, sometimes I forget to take out the garbage, right? Occasionally I’ll forget that. but like we don’t have that same sort of problem of like if I go too many words into a prompt or something, it’s gonna forget what’s at the beginning and then try and redo it. the other thing,
I imagine AI has also helped your business. Like we’re talking about all the headwinds it might cause, which it sounds like it’s not very many. But like you also mentioned like you’ve got across multiple frameworks and multiple platforms at this point.
How much has AI helped you sort of do that risk mitigation and be able to spread a small team across that much additional surface area?
[00:25:45] Adam McCrea: we’re definitely building faster than we ever have, which is great. in particular, like new platform integrations are, much faster for us than they used to be. Like we’ve, we’ve built a couple new platform in terms of integrating with new hosting providers.
Just in the past few months we’ve added a couple. and it’s definitely helpful, when I need to use a language that I’m not an expert in. So we started out as a Ruby on Rails specific autoscaler, and we’ve since expanded to Python and to Node and even Java, ’cause we had some folks using Java Spring Boot that wanted to use Judoscale.
And I haven’t touched Java since 2001. So, it’s changed a bit. AI is, is super helpful when I need to dive into that stuff. So we built our Python and Node, packages before we were using, AI-enabled development at all. But now when I need to go in and make changes to those packages, it’s really helpful to have, you know, AI helping explain to me how everything works.
‘Cause day-to-day we’re all working in Ruby. That’s the language that, that we all know and we all love. but we’ve got to support our Python customers. We’ve got to support our Node customers. We’ve got to support the packages, that go along with those. AI is incredibly helpful, for that because, It’s able to translate the, the concepts that we know in Ruby to these other languages really easily.
And so it, it explains things and helps us figure out the bugs that would’ve taken us, you know, hours,
It’s not really different than, than I guess any other software team at this point. It’s helping us build faster. It’s helping us, be more competent in, languages and, domains that we weren’t previously competent in
[00:27:29] Chris Vannoy: when it’s helping you do these, does it also help you maintain it afterwards, right? Like, not only do you now support Java, but we were talking about earlier, like you might have customers on it. They might come across bugs or things that are unusual or have feature requests, right? Is it still able to help with all that sort of thing afterwards, and how’s that working out so far?
[00:27:49] Adam McCrea: that’s the best part is that it’s able to, take a customer customer request and be like, walk me through our Java package and where does this happen?” I think that’s even more beneficial than the initial build is diving back into things after maybe it’s been months since I’ve touched the Python package or the Java package.
The ability for AI to help me just like reorient myself with that is really nice.
[00:28:14] Chris Vannoy: I’ve noticed that in Ruby projects specifically, because I work in consulting a little bit, I occasionally get Ruby projects, right? And, the standard sort of code archeology used to be like just open up the model folder and see how deep you got into something, in a lot of cases, right?
And then you open up the gem file and see what it’s all in there and that sort of stuff. But now it’s like I just throw it into AI and say like, “Hey, do all that essentially. Like, tell me what the guts of this thing are. What are the God objects inside of this? Like, where’s the hairy parts of this?” Right?
And it helps accelerate that sort of code archeology step, in Ruby apps. And also I have a client app that’s in Python, that was put together by a non-programmer, not using AI tools, which is fascinating. but that’s sort of like, “Hey, give me heads or tails of this thing. Like, what’s it do?
How’s it doing it? Like, what are the parts of this?” it makes it really kind of easy to get A lot of what is relatively straightforward to do once you’ve done this for a long time, as both of us have, but it’s also kind of nice to have someone, like some bot go in and like do it all real quick, right?
So you just get a summary to go off of, which is also often all you need in a lot of cases, to do that sort of stuff. the other thing we talked about right before this is like both of us are in AI stuff kind of all day, right? we both have kids, that are in high school, in my case college in yours, that are relatively anti-AI,which I find myself having to be very nuanced with when I talk to my kids lots of times. where it is like for our field in particular, AI works really well for this, right? the way I put it is it helps with the what was always the easiest stuff, it does really well, it can write code really well. It can do what I just described, sort of explain it, like that sort of thing. It changes how we work in a lot of ways. But like the hard parts of being a software engineer are still there, right? Which are like, is this the right approach to begin with? is this serving a customer? is this a good decision later on, that sort of like taste and judgment sort of thing. and trying to figure out how to talk to my kids, who my, daughter wants to be a writer, and that sort of stuff, these sort of creative fields. And it’s like thinking through how do I coach them through how does AI work and how to think about it in the world they’re walking into, Like when you take it out of the work, it also has so many like societal problem or issues or concerns, environmental, business, like all sorts of other stuff sort of wrapped around it. whereas we’re able to use it day-to-day. I don’t know. It’s something I struggle with, and I was curious if you’ve run into that as well, or how you’ve been able to approach that.
[00:30:56] Adam McCrea: Or is it just like, da, da, we won’t talk about it right now”? If I’m being honest, right now it’s pretty much a subject that we just don’t talk about. Which,
I have two daughters, both in college, both extremely anti-AI, mainly because of the environmental impact. but also because they’re artists and, They have no interest in AI art whatsoever, and, and I get that.
But I think it’s really hard because I don’t think there’s a single industry Where AI can’t be helpful in some way. you know, even as an artist, it doesn’t mean you need to use AI to make your art, but, you know, maybe there’s a way you can use AI to help you plan your art, be more efficient with your materials.
I don’t know. Whatever it is, it’s just a tool. if you just completely dismiss the tool, you’re gonna miss out on some significant benefits. This is gonna sound terrible. I know what it’s like to be young and idealistic. And if I’m the one to push back on that, I’m gonna become the enemy.
So my goal right now is to be the supporter When my daughters enter the workforce and need to get jobs, they will find out whether or not they need to accept AI as a current reality or whether they can continue to, not adopt it. I don’t know what’s gonna happen. I know there are software developers right now who don’t want to use AI in their daily work.
I imagine it’s very hard for them to get jobs right now.
If I were hiring right now, would I consider, a software developer who, who refused to use AI? I would not. There’s no way. It’s too powerful of a tool, and I think it’s gonna become that way in almost every industry.
That’s my stance right now, is that I’m trying to just be supportive, not talk about it, sometimes I feel like, I’m just burying my head in the sand and I’m not super proud about that. you just had John Nunemaker on and he’s talking about how he’s like working with his kids to like build AI worlds and games and I don’t know, and all this stuff, and I’m just like, “Yeah, I’m not gonna do that.”
[00:32:48] Chris Vannoy: No, I’m not either. well, but it’s at the same time, like, like I mentioned, like I’m giving lots of talks, especially right now to like recent CS grads and like business grads and that sort of stuff, like college kids in a lot of cases. That there is a, like 20, 30% of this audience that is of that. And like a lot of what I’m telling them is a little bit of my own story, right?
Where like, I think we’re similar ages, I was an English major, right? And I picked up how to code by like viewing webpages and like following lots of blogs and like all these tools were like free and available, and you could just like, there was a learning environment at the time where everybody was figuring stuff out. And so you could just like hack away at stuff and like build stuff, and then, 20 years later it’s like, “Man, I’ve got a career in this now.” … And I think we’re in a similar age now where a lot of these tools are subsidized and they’re relatively inexpensive. And especially younger folks even the naysayers sometimes know these tools better than folks our age that aren’t software engineers all the time, right? And so being able to take advantage of free tools or subsidized tools and like go figure it out, right? And then you can probably generate a career off of it on the back end of it, right? I can look at my own life and that’s kind of how I ended up here in the first place. it’s difficult though, when they’re in, like I mentioned, my daughter wants to be a writer, your son, like they’re in art, sort of things.
It’s like that’s hard.
[00:34:17] Adam McCrea: ultimately, if, if your mind isn’t open to it, no amount of, somebody telling you, “No, this is an opportunity. You need to embrace it,” like that’s just, that’s just not gonna work. Like it has to happen naturally in some way.
[00:34:29] Chris Vannoy: especially dad, that doesn’t work. Like, maybe we could, you know, trade kids for a little bit or something. I don’t know. But, like no, I don’t wanna do that. That’s fine. all right. Adam, thank you so much for joining me today. Uh, where can people keep up with you online and, like, uh, learn about Judoscale a little bit more even?
[00:34:47] Adam McCrea: we’re always posting interesting things on the Judoscale blog, Judoscale.com. I’m not super active on social media anywhere. I occasionally post to, to LinkedIn or to, to Twitter as Adam Logic. The best thing you can do is sign up for the Judoscale newsletter.
It’s just once a month. it’s whatever we’ve been working on, whatever we’ve been talking about, and it’s usually interesting stuff even if you don’t use Judoscale
[00:35:09] Chris Vannoy: All right. Thanks again, man.
[00:35:11] Adam McCrea: Thanks, Chris.
[00:35:12] Chris Vannoy: Three things I want you to take out of that conversation. One, if you’re building something on the side while you’re employed, get it in writing that you own it. Adam went three years without that paper and only found out it mattered when an investor started due diligence. It worked out. It won’t always.
Two, distribution is the plan, not the afterthought. Adam didn’t market Judoscale at all for years. He put it where his buyers were already looking, and that was enough to carry it until he could go full-time. Three, AI has lowered the cost of writing code. It did not lower the cost of 10 years of knowing where the edge cases are.
Adam pays for more than 30 SaaS products to run his business, and I run almost as many, and has replaced exactly one of them. That’s the actual build versus buy math, and it hasn’t moved as much as the timeline says it has. If you want more of this, the newsletter is at getaxiomatic.com. That’s where the stuff between episodes lives.
You can go find Adam at Judoscale.com. His blog is really good, and the monthly newsletter is worth your inbox, even if you never touch the product. He’s occasionally on LinkedIn and Twitter as Adam Logic. Thanks for listening. I’m Chris.