Always be dogfooding, with Andrew Lisowski of devtools.fm

One thing I really truly believe in is dogfooding. Using your product, being your own customer, and, like, filling that role makes you better at your job in immeasurable ways. Because I can, like, talk to the CEO and be, like, these are my use cases. These are must haves. The all these things.

Because you're actually talking as a customer.

Oh, yeah.

Hi, everyone. You're listening to Scaling Dev Tools. This is a really cool episode because I'm joined today by Andrew from Dev Tools FM, which is one of my favorite podcasts, and you should absolutely listen to. Andrew, thanks so much for joining.

Thanks for having me.

Yeah. So just for background for people that haven't heard of Dev Tools FM. So Dev Tools FM, do you wanna go in and give an intro to it?

Yeah. Sure. I'll say our tagline. Dev Tools FM is a podcast about developer tools and the people who make them. So that really encompasses all of the stuff that we do is we wanna focus in on a person who is making something that helps developers' lives easier. And then we also try to, like like, go into who they are, where they came from, if if it's an interesting part of their story, and how that kinda, like, led into what they're making now.

Yeah. And I think, like, as one of my favorite episodes is the one that you did with Mitchell from HashiCorp. Mhmm. And from what I remember, it was, like, they're kinda you were talking about, like, the origin story. You were talking about, like, the licenses. You're also talking about Mitchell's flying. Yeah. I feel like that was a lot, actually. That was that was that was awesome to hear. Like, the man loves flying.

Yeah. It's cool it's cool to dig into people's other hobbies, especially if they kinda, like, tangentially connect with, programming. Like, he had built out, like, a flight panel for himself, which is I I love creating software for your own personal needs. Like, I've tried to make startups maybe in the past, but they've all just turned into, like, personal tools that I use to, like, enhance my life, and then I might share them with my friends sometimes.

Yeah. Do you wanna talk about a little bit about, like, some of the projects that you've felt? Because you've maintained and built quite a lot of open source projects yourself. Yeah.

So my first delve into open source was was with a project called Gimp. It's the JavaScript image manipulation program. It was created by a guy in Ireland who is now a politician. So the, the library had basically languished for, like, 2 years because this guy was transitioning from being a developer into being, like, a person in political office and caring about Cork Ireland. So, like, the the project was in need of a of a maintainer.

Yeah. Maybe that's such a, like, a good question. Not everyone would know what it's like to actually, like, maintain a popular open source project.

It it's rough, and that's, like, one of the the main themes of my podcast is, like, maintaining that open source because, like, everybody who, like, likes programming can probably get a library out the door, but, like, maintaining it and, like, taking feedback and knowing when to say no, like, those are all hard things that you learn by doing doing the thing. So it's like, when you first start out, you make a package, you're like, oh, people have come to this. They want me to do things for them. They want me to enable things for them. But as you get further along in your open source career, you go, oh, new additions, I'm, like, more scared of those.

And, like, does it like, on a weekly basis, like, how much time do you think you spend now?

I don't do it as much as I used to. At my old position, I was more, like, officially, like, paid to do open source, so I did a lot of it. I I would just, like, openly just, like, create lots of packages, take on lots of maintenance burden. But then when I got to my new job, I was much more product focused, so I've spent a lot less time on it. One one interesting phenomena I heard about before I experienced it myself was GitHub notification anxiety.

So what like, when you see, like, people creating issues and stuff, like, yeah, like, how like yeah. How PR

is welcome.

Yeah. PR?

I've typed though those two words more times than I can count. And when a PR comes in, I prioritize PRs. So if there's if there is a PR, I will go review it. I if it's a thing I'm okay with, I'll merge it. If there's an issue, I'll breathe through it lightly. Go, yeah. That seems like a problem other people experience too. If you wanna fix it, feel free to.

Yeah. That makes so sense. That makes a lot of sense. Yeah. And do you think that there's something that, like, dev tools, maybe don't typically understand about, like, open source? A

lot of dev tools companies these days are doing it right. Like, they're focusing on DX first. Like, I look back to, like, before the Internet and, like, dev tooling was, like, archaic and, like, hard to use tools. Like, you had to be a genius to, like, even read through the docs and understand what's going on. So I think dev tools companies nowadays really do have that focus on, like, just, like, making the the simple case simple and making it just obvious.

Yeah. I I see what you mean there. That's I didn't realize there's, like, so much of it. I never considered that. There's a big difference between the JavaScript ecosystem.

In my opinion, it it's huge. Like, other ecosystems feel like almost like they're adversarial to the developer. Like, they assume you know a lot, but JavaScript is kind of this, like, lingua fact of, like, programming languages. Yeah. So, like, people when they do their docs will literally, like, point them at, like, complete beginners.

Yeah.

And, like, in other programming paradigms, I just don't see that as much. They kinda, like, assume you're, like, a senior level already and you know how all the things fit together, but that's not really true for everybody.

Yeah. So what made you wanna start a DevTools

class? I didn't at first. Like, by by virtue of myself, I'm a very lazy person and, like, what

Everything you've said so far doesn't point to that.

Yeah. It might not, but, like, something that stuck with me early in my career was, like, the best programmers are the laziest, and those are the ones who are gonna automate things. And Yeah. When it comes to a lot of things, I'm very lazy. And when I was approached by my cohost to to start the podcast, my first thought was, like, I don't wanna do that.

Because you're actually talking as a customer.

Oh, yeah. It was so cool, like, in the first few, like, months of dev tools and d script where I'd be, like, there's this feature, there's this workflow that doesn't exist, and it's making my life terrible. The the one that that it was was we had just integrated video and you could have sequences where it's like, it's your face and my face. And how we wanted to edit it was your face is on screen when you're talking, my face is on screen when I'm talking.

Yes.

Just that alone, doing that for my 1 hour podcast took me an hour and a half of just clicking, and I could feel my wrist starting to hurt more and more. So I was like, one weekend, I was just like, I'm done. I'm coding this.

Really?

It will take me

code.

Yeah. It it is going to take me as much time to code this as it is going for me to do this this time. And, of course, that wasn't true. We all know with automation, you think you're gonna save a lot of time, but you really just invest a whole lot of time upfront. And so that's what it was here too.

Seriously?

Yeah. Wow. But I have, like, laser vision for, like, what the product should be at this point. I also built the design system. So, like, during that, I was at, like, my pixel vision, like, honed in. Like, I could see, like, oh, that's 2 pixels off right there. Like, that doesn't line up how the system should be. So, like, I I can look at the app and probably find a few bugs, like, even right now.

That's so funny. So what was the impact? Like, you just people you were just suddenly did you become, like, the kind of

It's it it has been weird because, like, I have, like, this large body of work that's, like, Descript specific work, basically, and people know about it. So people will just randomly come up to me and be like, hey. Can I use your podcast files for x, y, or z? And I'll just see myself pop up in places. So, like, the CEO is doing some product announcement videos. Oh, there's my face because they needed a multi track sequence project to, like, show off this new capability.

Really? So you're kind of like the darling, the showcase podcast.

I'm probably, like, the the most known podcast internal one. Yeah. But yeah. Like, I'll see product managers use my videos for test stuff. One thing that really, like, tickled me was our AI team. Like, we have a AI clips feature now.

Yeah.

And I was at the off-site. I'm like, I don't use it much yet. And they're like, what are you saying? Like, you're the one person that should use it. We trained on your shit. Like, if it's gonna work well for anybody, it's gonna work well for you. And it just, like, it kinda blew my mind. It's like, wow. My my my dog fooding exercise has become integrated into this company at a very weird scale.

That's, like, seriously cool. This is, like, such a cool, like, career tech or whatever. Like, just like dog food. Life is the same for start ups. Right?

Yeah. It's the same for developer tools. It's like, auto, that automated release tool. The reason we built it is because we needed it. We're Yeah.

Yeah. Do you think do you think dog fooding is like is that has that been one of have you seen that in a lot of the successful dev tools?

I I would say, like, I don't think you can make a good thing if you don't use it. Like, I it it would be hard for me to work in an industry where it's like, oh, you're making, like, accounting software for, like, professional accountants. And it's like, well, I can't dog food that. I can't really understand the needs of the customer at all. So, like yeah. In dev tools, though, yeah, it's it's a great way to go.

Yeah. Kind of like a broad question, but, like, from the guests that you've spoken to, have you spotted any kind of patterns or things that you would do for your own dev tools?

One that one that struck me is just, like, people will have obsessions, and, like, those obsessions might not yield a company or a tool like year 1.

Yeah.

It might take them till, like, the 6th time they've rebuilt it

Yeah.

To, like, find a thing that's, like, oh, I the this is what I wanted. This is, like, a useful thing. It's taken all of your those past failures and kind of, like, rolled them up into, like, the the new great thing.

Yeah.

And I think that is a a great theme across most development. Like, when I'm talking to juniors, like, the basically, my only advice is go build a bunch of side projects. They're all going to fail. They're all going to go nowhere, but you're going to learn a whole lot. And I think that's true of dev tools too.

Yeah. Yeah.

And that's been done through side project after side project just, like, making things that nobody really cares about.

Interesting. So making the things that nobody cares about.

But you care about them.

Care about them. And eventually

Eventually, it might become a company. Yeah. Probably care about them. Or a dev tool. Like, my current iteration of the static site, storybook generator is called Floosh.

Yeah. And are you gonna try to get people to

This time around? Yeah. Maybe. The the things I've built it on are still experimental themselves. Yeah.

Yeah. Yeah. Yeah. That makes sense. And so if you were gonna put it out there, like, how would you do that now?

I wouldn't put it out there until I'd fully integrated it into our design system and then ran with it for a while.

Side dog tested it.

Yeah. Yeah. We we would dog food it for for quite a while before, I I would release it. And then once I release it, I probably just, like, tweet it and then just be good about saying no. Because, like, a lot of people are gonna wanna add a lot of stuff, and I'm gonna have to be the arbiter arbiter, the benevolent dictator for life saying yes and no to what what I think spins in the project.

What what was the other one?

Just like, there's, the pattern where you have a bunch of resizable panels in, like, a page. There's a nice React component, for that from Brian k or I don't know if it's k. Brian Vaughn, from the he used to be on the React team, works at Remix now. It's really good, but it doesn't support pixel values and all of my designs come in pixel values. So it just doesn't work and I kind of like I forked it for Descript and added all that stuff that I wanted, but it it got again to that point where it's like, yes, it works.

That's really cool.

But I'm easily distracted.

Yeah. No. No.

No. That's

that's super, super cool. Sometimes people ask me about, like, podcasting. Like, dev dev tools say, I'm interested in starting a podcast. Like, should I start a podcast? How should I start a podcast? I'm sure you have a lot of thoughts on that.

Yeah. It's a lot more work than you think it will be, and it's always more work. Like, it's very rarely are you gonna go, okay. I got bigger, now I do less. Like, even if you even if you hire an editor, that just frees you up to do more different things.

Yeah. That that makes sense. And do you think that there's been other benefits around? Like, you were mentioning, like, all the people that you've met and, like, the conferences and

Yeah. I it's it's crazy the people I've met at this point. Like, when when we started, like, our first three episodes was just like us, and then we interviewed one of Justin's coworkers, who made Patch Package, and then we interviewed my co one of my coworkers who were just, like, doing random stuff. And I would have never thought it would get to the point where I'm, like, interviewing Guillermo Rauch of Vercel. Yeah.

I'm talking to DHH for an hour today. Yeah.

Wow. Yeah. How did you prepare for the DHH call?

For him,

it's basically the same process as everybody. I I I start by going to your Twitter.

Yeah.

I open up every link I can find on you, and I just start reading. So, like, with him, a a lot of that episode focused on his recent, like, spicy programming takes. So I just, like, went into those, read the blog post about them. Like, I I I was there for the typescript dramas. So, like, I had my questions to ask.

Yeah.

So it's it's it's very interesting to see how people react to the different creators.

Yeah. Was that was that by far the kind of, like, most

that was that was probably our biggest episode as of late just because he has such a huge following. But, like, we've had a few ones that pop off, but they're always just, like, big people. So, like, we had Evan Yu on, like, episode 13.

That

was a good episode, but it had really bad audio. So I was, like, damn. Oh, my

yeah.

Wish it could have been better. We we had Rich Harris. He also had bad audio. It's, like, the higher profile of the person, like, it seems like the worst setup they have. Like Guillermo Rausch, he was in, like, a skyscraper. You can see all of San Francisco behind him. Terrible mic. Potato quality, camera.

Yeah. Someone someone needs to get the DevTools founders, Successful successful guys, better mics.

Yeah. Please don't use your AirPods on my podcast.

They're the worst. Yeah. So what what do you think makes a good, if someone's gonna start a podcast for their developer tool, and they're like, you know, Andrew calling, like, what what what should I do? Like, give me what

Like, if you were start you had a developer tools company and you want to start a podcast for that. Yeah. I think the Jam people are good people to look at for community building. I think a lot of dev tools and probably especially dev tools startups are all about community and building a bunch of people that are excited about the thing that you're making, and they're doing really good at that. Like, they have their own podcast.

Yeah. True. That makes subtle sense. Yeah. And, like, what what do you think makes, like, a good podcast episode?

That's that's a hard thing to answer because, like, during the episodes, as I mentioned, I have kinda, like, producer anxiety. So sometimes I'll finish recording an episode and be like, that was terrible. Like, I said one thing wrong and the whole episode was off. Then I'll go back and, like, watch it during editing. I'm like, okay.

Yeah.

I I rely a lot on research and, like, looking into their passwords. So a lot sometimes I'll ask questions and it's like, I I know the answer answer, or I know this is, like, stupid in x, y, and z ways, and you're gonna react to it in these ways and go on the rant I know you're gonna go on. So, like, it it's good questions are almost like manipulation.

Interesting. Yeah. Manipulation. Do you think when when developers are, like, listening, like, why do you think developers listen to podcasts, actually? Do you think do you have a goal of, like

that's a very good question because I, myself, do not listen to many developer podcasts.

Do you

listen to them?

No. So I Do you listen to any podcast?

Not developer ones really. True. Of course. Like, maybe some comedy ones. Yeah. But even those I don't listen to too much. Like, my podcast was started as a dog food exercise. Someone on the Internet said, hey, you wanna make a podcast? I wasn't at all really, like I I can see the value in my own. Like, I try to just make good content, like, have good conversations.

Sense because some people are, like, creators, but they don't really like consumers, I think.

Yeah. And I I especially with the guests that I have on my podcast, I try not to consume other podcasts from Really? Yeah. Like, I don't wanna be influenced by

someone else.

In the questions that I ask this person. In the past too. I want I want it to come from, like, my interest. So a lot of the episodes are, like, fueled by what I'm actually interested in at

the time.

Like Yeah. We had the creator of Stylex on, and I'm interested in CSS styling. So, like, I could go very deep and, like, we often get, like when we prepare for an episode, I create a GitHub discussion. I put, like, 15 questions on it, and the reaction we always get is, wow, this is so thorough. And I'm like, okay. I I did my job this time.

Yeah. Yeah. That's amazing.

Yeah. It sucks when I get to the end of the doc and I don't have enough questions. I'm like, what's your favorite plug in to your thingy?

Right. Yeah.

Yeah. But sometimes that has a good answer too.

Yeah. Yeah. Do you have a goal of, like, like

With money?

Kind of effort. I don't know. Like, just in general because you've got, like it feels like it's kind of like your the work you're doing is the is the goal and the career is like you're kind of doing the podcast to help in the in the job and stuff.

Yeah. We've we've made money off the podcast. Like, another thing to consider when you're starting a podcast is it's not free.

Yeah.

Like, you're gonna you're gonna end up paying for a lot of things. Yeah. Whether it's equipment, whether it's software, whether it's people, you're gonna pay for something. Yeah. So, like, with that in mind, if you're like, oh, I want a podcast indefinitely, it's like, oh, do you wanna pay $2,000 a year for the rest of your life to do something?

Yeah.

In itself, and it feels like you're, like, you're hurting your own content. Like Yeah. The way we structured it was the free episode was, like, around 45 minutes. And then how I would get it down to 45 minutes is that I'd try to go in there and, like, cut out large chunks that, like, could still flow, but still kinda leave you hanging. And it felt icky.

Yeah.

Like, why should we take our our, like, fans' money?

Yeah.

It's it's a it's an odd relationship. So we tried that for, like, a year and then ended up going, okay. Sponsors is probably the way to do it. Yeah. We've been doing Patreon for a year. We're getting a $140 a month. Yeah. That's, like, okay. That kinda pays for an editor for maybe one episode. Yeah.

Yeah. Yeah. But you did well on that.

Yeah.

Right? So

Yeah. When we sign people up, we usually sign them up for like 10 weeks.

Yeah.

And so we had Raycast for 20 weeks. We had Codecrafters for 10 weeks. We just had Clerk for 10 weeks.

Oh, did you? Oh, nice. Yeah.

And then the next 10 episodes are sponsored by Mux.

Oh, nice. Yeah. Yeah. Yeah. Very cool.

Very cool. Yeah. And it's very hardening when we have somebody on, and they're like, oh, I saw you have sponsors. We want a sponsor. And I'm like, yes. Thank you.

In 10 weeks.

Yeah. See you in 10 weeks.

Yeah. Yeah. That's so good. That's great. Well, it's a great show, and I think it's very unique. And, I'm not surprised everyone wants to sponsor it.

Well, I think you were our first Patreon subscriber, if I'm not if I'm not wrong.

And I've got the for anyone watching, I have got the DevTools FM shop.

Yep. Shop.devtools.fm. Shop.devtools.fm

for your developer tools merchandise.

Yep.

Yeah. Andrew, I think we're probably out of time, so I think it's probably time to get some pizza.

Yeah. I'm down.

Thanks so much for joining. It's so cool to kinda have you on. Yeah. Thanks for making a great great show.

Thanks for having me on. Yeah.