David Mytton - Arcjet and console.dev

Everyone says you should write a blog post, but the problem is you've got to spend a lot of time because they've got to be technically interesting, and you have to do it for 12 months, 18 months before you even see any results. Most people give up before they start seeing the results. Yeah. Technical founders in particular would love to build the product, but it's the easier bit of building a startup.

Hi, everyone. You're listening to Scaling Dev Tools. I'm joined today by David Mitton from Arcjet and also console.dev. Thanks, David, for joining. No problem. Great to see you. Yeah. Great to see you. And I should say that we were just we bumped into each other, in an office, and we were like, okay. Why don't we just do this? So

Completely unplanned, But I've been listening to the podcast for ages. So, yeah, we're excited to be on.

Yeah. And I've been listening to, to your podcast as well and console dev 1, which yeah. I wanna see more more episodes, actually. It's

Yeah. I know. Getting into running a startup full time, have had less time to do console dot dev and the podcast newsletter's still going. Yeah. But the podcast, like, you know how much work it is.

So yeah. Yeah. So, David, could you tell us a bit about Artjet?

Artjet is a tool to help developers protect their apps once they go into production. So if you think pre prod and prod pre prod is your writing code. There's lots of tooling, whether it's something like Semgrep for static analysis or GitHub's code scanning. There's lots of tools to help you protect your code as you're writing it and writing bad code, essentially. But once you push your code into production, then there's not as much for developers.

Yeah. That's that's really, really powerful. And, like, what do do you find, like, people are coming for, like, specific are there, like, specific cases where people are like, yeah. We need ArcGerry.

Yeah. We've built these core primitives. The the bot detection, if you want that, you can just use it by itself. If you want rate limiting, you can use that. But you can also compose them.

I was speaking to someone that was, like, having loads of, spam like, people were, like, requesting phone, like, verify like, the 2 factor auth 1. Yeah.

And it's expensive. Right? Yeah. Sending SMSes all the time. Yeah. Yeah. So you need to rate limit that. You wanna prevent automated clients from submitting forms. And, really, a human will only submit a form. Well, ideally, they'll submit it once because they'll be successful, but maybe you have a typo in your email address and you can detect that.

Yeah. Yeah. That makes total sense. You've been working on ArcGap for it's, like, a year or 2?

Yeah. Just over a year ago. Yeah. June 2023.

Yeah. And so when I first came across, your work, David, it was actually with, Console before, and which is amazing, and people should subscribe. It's like if you're looking for, like, dev tools newsletter, like console.dev is awesome.

Thanks.

And there's podcast, and it's great. So I kind of was curious about, like, the journey from, like, console to, ARP Chair.

Yeah. Well, if we go back just a little bit further, my the first company that I started, Server Density, in 2009 was a server monitoring product, so, like, Daystar, Conurelyc. And we built that in 2009 and then sold it in 2018. Obviously, a lot happened in in in training years, but sold it in 2018 to StatPath. They're a large security platform, and have, like, a huge had a huge network.

Yeah.

Like, you can go on Hack and Use and there's stuff there, but it's not really for specifically for tools.

You have to look for and filter.

Yeah. And there's no control over what shows up. You'll get really interesting stuff, but it might not necessarily be dev tools related that you wanna try out. And so just in the period of 2009 to when we started console in 2020, the there was a massive explosion in the number of tools and companies and open source, everything just trying to push tools to developers. And so it was very difficult to know not just what you should, where to find things, but also what you should pay attention to.

Yeah.

And it was through that that I started thinking, well, what's the developer tooling like for security? And as I I said earlier, like, you've got this there's a lot of stuff as you're writing code, but that's only half of what you need to do. You gotta actually deploy your application, and then just it wasn't as good. Like, the where's the Vasile of security or the super base of security, like Yeah. In those two examples are just one of many examples of really high quality developer tools Yeah.

Yeah. That's amazing. And and I have questions, on on that journey, but I just wanted to say something that was, like, that I think says a lot about you and about how you run is that one of my friends, is good friends is, like, running a dev tool, and I won't name them, but their dev tool is, like, really good now, and it's amazing. And they're getting, like, tons of customers, but at some point, they it was, like, the early stages. And in their own words, like, it was pretty janky.

And,

yeah, it was just, like, really kind of a cool aside for anyone.

Thanks. That's so that's funny to hear on the other side. I do remember, we'd see idea is this it's things that I like and so I review them and I only put good things in there. I applied the same with the sponsors. Well, it is sponsored. You could pay and it prioritizes me looking at your tool because we have quite a long backlog, but it does it's not pay to pay, kind of pay to be in the newsletter. And I Which

is kind of like one of those things where, like, everyone says, but you actually Okay.

It was good that you got a real example of us actually returning the money to someone. Yeah. And it's not like we'll never include it. Like, products evolve and things Yes. Things progress.

Yeah. Yeah. It was really cool. And but so so then going back to what you're saying about, the origins of ArcGex, could you tell us a bit about, like, yeah, the initial like, how did you get it out there, like, initially and

yeah. Well, this is the hardest bit, and it's it's something I've seen. So I've made a few DevTools Angels investments, and it's something I've seen through console as technical founders in particular would love to build the product, spend all the technical founders in particular would love to build the product, spend all the time on the technology. And it's not that it's easy, but it's the easier bit of building a startup because 50% of it is a product, 50% of it is just the go to market and getting it in the hands of developers. And it's just really, really difficult.

Yeah. Yeah. And so you in terms of, like, if you're thinking about, like, blog like, are you writing a lot of content, currently?

Yeah. So I'm my goal is to do a blog post a week. We've got DevRel on the team. He's writing a blog post a week as well, and I try to encourage the team to write blog posts. Yeah. It's hard to get engineers to do it just because it's not their job, really. Yeah. And they're obviously building the products and they're writing code, And it's quite different discipline writing Yeah. Prose than writing code. Yeah.

have, like, a process, like,

to do it? It depends on what I'm writing about. Often, I'll have no idea at all, and then I'll come up with an idea and I'll write the blog post in 3 or 4 hours,

and then

it will go.

Like, it just kind of has to come out of your filler out sometimes. But

when I look back on the blog posts that I've done really well, it's writing about something that has taken me a long time to solve, technically. Mhmm. So a recent one was we wanted to put, reduction into our Go logging on the server because we'd log a lot of data panel requests, but we don't want to put information into the logs. So personal information, IP addresses, all that kind of stuff because it goes into we use Datadog, goes into Datadog for a bit, and then it goes into long term storage and just GDPR compliance. If someone wants to remove part of their personal information that you put in there Yeah.

Yeah. That's really that's really interesting. I feel like it's like it's just this process of, like, every time you do something, reflects, like

Yep. Every time you do anything, I think, unless it's just, like, saying that next year, I see this template, probably take some notes and then write a quick blog post. Yeah. And that's probably the easiest way to get started because you've already done most of the work, which is the technical bit. Yeah.

Yeah. Yeah. I I feel like sometimes as well, like, I don't know if it's maybe this is slightly different, but this is more for, like, the onboarding stuff or, like, guides and things of, like, how to set things up is like it I feel like it also forces you to reflect on, like, every step of, like Yeah. Oh, this is we don't make this that clear. I actually Right.

Did you have someone on your podcast recently talking about this, like, the onboarding side of things? Yeah. I remember listening to that episode. It's like you build it once, and then no one ever uses it from your team again. Yeah.

I I I feel like there's, like, so many, like, things that are just, like, so basic that it's, like, you might go back to, like, this big article and, like Yep. I we had where, like, the Docker image was out of date for them, like just stuff like that where, like and and things have changed and

it's Onboarding, quick starts, like, the first bits in the in the docs, things change all the time. And you want your developers to write documentation at the same time, but maybe it's in a different repository and they just didn't they forgot to do the PR correctly or they just didn't think about it or they haven't had time to do it or there's a rush to get something out. And there's all these reasons why you can't always sync the docs with a release. So you need someone whose job it is or at least they have the responsibility to to test that on a regular basis.

And that's the CEO.

I think it should be the CEO. Yeah. Yeah. Even up to hundreds of people. Yeah. I think the CEO should sign up to their product. Like mystery shopper type approach. Like, you actually use your products. Did emails make sense? Did they come out?

Yeah.

It's the same kind of thing, isn't it? You're a developer. You haven't got much time. You wanna solve a problem That's

a good way to think about it.

And to sign up for the product.

Yeah. That's such a

good way

to think about it. Yeah. That's that's amazing. And then so you are someone who's like if you think about, like, DevTools founders, like, you understand content, I think, more than more than, like, the top top naught point naught naught 1 percent of people, right, probably because because of console dev. Do you think so you've got the blog posts. Are you also doing, like, newsletters and stuff?

Yep. So we do a a monthly newsletter. So you sign up to the product Yeah. And you you get subscribed to that. And that started out just sending interesting links about security.

And I think like I get is that like a kind of do you consciously think about what you want people to feel about Arcjet? Or

We have done we we spent a bit of time thinking about it because the word Arcjet, when I was coming up with the name for their company, literally had list of Wikipedia entries of, like, all the engineering terms, all the science, all the physics terms, and went through every single one. Yeah. I was like, does the domain exist? Is there a startup? Is there a trademark?

That's quite a good connotation for

So that's what we wanted to use. We wanted to say kind of sci fi and space and futuristic. And so we a couple of weeks ago, we started discussing this properly to, like, make sure our branding and how we talk about the company and how we do our design is has that feel without it being cliche. Yeah. Because everyone well, all a lot of dev tools have this have a very similar feel, and so we're conscious of not seeing the same

Yeah.

But also taking this kind of futuristic space type feel, which is linked to the name.

Yeah. And because I because I always felt like console.dev has, like, a really strong brand.

Yeah. It's, like, very that was luck. Like, the domain was available. We just read the console.dev.

I meant I actually meant more, like, in the terms of, like, just everything is very, like, kind of congruent

and Or with the lines

and everything. And also, like, just the way you, like, write about things matches, like, the brand, and it just feels very, like, I don't know, like, well done. I don't know.

That's good that's good to hear. The console branding, we put a lot of effort into the design of that, like the website and, the custom email template as well, make sure it works on dark mode, make sure it works on light mode, and just having kind of this kind of a neon feel. Yeah. And we did that in 2020 before neon became cool. It was cool in, like, the seventies and the eighties and then it wasn't cool and now it is cool again. And I feel like we're a little bit ahead of that. Yeah.

You made it cool again.

I'll claim that. Yeah. But it's kind of how I write about things and then Art Jet is a bit more we put a bit more time into figuring that out because it's a bigger team. Console is just me doing the writing and then my cofounder, who does kind of the the back office side of things and Yeah. Have a limited amount of marketing that we do as well.

Yeah. And I guess, like, the question I guess I have as well is, like, were there any things that you learned from console or dev that you, like, have been applying directly into ArcGit?

I think we we talked a little bit about the go to market. That was a big learning and just how important that is. Like, I spend a 100% of my time on that now and then 100% on Pretty much. Percent on go to market. Yeah. It's 100% on go to market and then another 20% on running everything else.

Yeah. Yeah. Yeah.

It's like the CEO bits. Sort of like, oh, I've gotta review the accounts for our UK subsidiary today. And then Oh, no. The next day I'm reviewing a PR. Yeah.

Yeah.

And just to focus on on the DX is super important, and that means everything from the onboarding through to, the docs are organized and all the code samples. I wrote all of our documentation at the very beginning. Now our our team helped as well. But, every single code example, we've probably got 200 code samples in the docs that I've created for TypeScript and JavaScript. Yes.

Yeah. Yeah. That's that makes so much sense. It's like you almost become like the I guess you're the one that's actually just, like, yeah, poking it and just, like, constantly making sure that everything is, like,

Yeah. Consistent and it works. Yeah. And not just, like, the quick start works. Yeah. That is also important. It has to work, but everything that branches out from that. Because developers start with a quick start, but then they go Yeah. They go a lot deeper. And this is often where it's called a graduation problem Yeah.

Yeah. I think when I, like, started the podcast and before I'd done, like, tried to build my own stuff, like, I feel like it's it's really easy to, like, underappreciate, like, how much those kind of basics are, like, not you know, are, like, really hard to do.

Yep. And it's the stuff behind the scenes.

It's rare.

Like and if you don't do it, you notice. People will notice. But if you do it right, then no one notices Yeah. Because it's what they're expecting. Yeah.

And it's like, oh, I don't really know how we grew. How did we grow? Like, just I know. Whoever just used it, but maybe it's all the work that you do where it's just making it amazing.

It's all the work, and then you also don't know which thing you did actually had the impact. And then 5 years later when someone asks you, you create a narrative around what it was. And maybe that was it, maybe it wasn't. Yeah. But also the 5 years is a long time. Things have changed. Yeah. So there's a lot of, I think, common things that you need to do. Like, we're talking about with the writing just writing blog posts Yeah. On a regular basis.

Yeah. And have you tried working with influencers yet with UpchUp?

It's very difficult to get in, and I think there's a it's, like, kinda like selection bias. Once you're in with one one well known influencer, then the others will see and you get this credibility.

Mhmm.

And so I think it's kind of like the cold start problem. Interesting. And So

it's actually hard to get It's hard to get hard to get to kinda take your money in a sense.

Yeah. They just they just don't reply to emails or they're just not interested or they think the product's too early, and these are all legitimate things and it's good that you can't just pay to be mentioned by an influencer. So otherwise, they wouldn't have the audience because people don't trust them. And so we've been spending a lot of time working on it and I'm trying to understand how to do it and got some really great angel investors on the architect cap table that have been helpful with that, but I think it's just okay it's just continuous. It's all part of the how the product is, like Yeah.

Yeah. That's super, super interesting. Okay. David, that was, amazing. Like, we've wanted to chat for ages, I think, and it's kind of so cool that we just bumped into each other.

Yeah. Funny to be in San Francisco rather than

Where can people learn more about Arcjet and about you?

Arcjet.com, that's the website, and blog. Arcjet dotcom. And then for me personally, it's davidmitten.blog. Don't write as much as I used to on my personal blog. It's mostly about sustainable computing now. And my research on that has kind of dialed down as I've focused on other things. So console dot dev will be the other thing as well. Subscribe to the dev tools newsletter.

Okay. Amazing. Amazing. Thanks everyone for listening. Thank you, David, for joining.

Thanks a lot.